For some workloads such as arm bootup, tb_phys_hash is performance-critical.
The is due to the high frequency of accesses to the hash table, originated
by (frequent) TLB flushes that wipe out the cpu-private tb_jmp_cache's.
More info:
https://lists.nongnu.org/archive/html/qemu-devel/2016-03/msg05098.html
To dig further into this I modified an arm image booting debian jessie to
immediately shut down after boot. Analysis revealed that quite a bit of time
is unnecessarily spent in tb_phys_hash: the cause is poor hashing that
results in very uneven loading of chains in the hash table's buckets;
the longest observed chain had ~550 elements.
The appended addresses this with two changes:
1) Use xxhash as the hash table's hash function. xxhash is a fast,
high-quality hashing function.
2) Feed the hashing function with not just tb_phys, but also pc and flags.
This improves performance over using just tb_phys for hashing, since that
resulted in some hash buckets having many TB's, while others getting very few;
with these changes, the longest observed chain on a single hash bucket is
brought down from ~550 to ~40.
Tests show that the other element checked for in tb_find_physical,
cs_base, is always a match when tb_phys+pc+flags are a match,
so hashing cs_base is wasteful. It could be that this is an ARM-only
thing, though. UPDATE:
On Tue, Apr 05, 2016 at 08:41:43 -0700, Richard Henderson wrote:
> The cs_base field is only used by i386 (in 16-bit modes), and sparc (for a TB
> consisting of only a delay slot).
> It may well still turn out to be reasonable to ignore cs_base for hashing.
BTW, after this change the hash table should not be called "tb_hash_phys"
anymore; this is addressed later in this series.
This change gives consistent bootup time improvements. I tested two
host machines:
- Intel Xeon E5-2690: 11.6% less time
- Intel i7-4790K: 19.2% less time
Increasing the number of hash buckets yields further improvements. However,
using a larger, fixed number of buckets can degrade performance for other
workloads that do not translate as many blocks (600K+ for debian-jessie arm
bootup). This is dealt with later in this series.
Backports commit 42bd32287f3a18d823f2258b813824a39ed7c6d9 from qemu
This will be used by upcoming changes for hashing the tb hash.
Add this into a separate file to include the copyright notice from
xxhash.
Backports commit dc8b295d05ec35a8c032f9abca421772347ba5d4 from qemu
The exception_action() function in user-exec.c is just a call to
cpu_loop_exit() for every target CPU except i386. Since this
function is only called if the target's handle_mmu_fault() hook has
indicated an MMU fault, and that hook is only called from the
handle_cpu_signal() code path, we can simply move the x86-specific
setup into that hook, which allows us to remove the TARGET_I386
ifdef from user-exec.c.
Of the actions that were done by the call to raise_interrupt_err():
* cpu_svm_check_intercept_param() is a no-op in user mode
* check_exception() is a no-op since double faults are impossible
for user-mode
* assignments to cs->exception_index and env->error_code are no-ops
* assigning to env->exception_next_eip is unnecessary because it
is not used unless env->exception_is_int is true
* cpu_loop_exit_restore() is equivalent to cpu_loop_exit() since
pc is 0
which leaves just setting env_>exception_is_int as the action that
needs to be added to x86_cpu_handle_mmu_fault().
Backports commit 0c33682d5f29b0a4ae53bdec4c8e52e4fae37b34 from qemu
Add a comment to do_interrupt_user() along the same lines as the
existing one for do_interrupt_all() noting that the next_eip
argument is not used unless is_int is true or intno is EXCP_SYSCALL.
Backports commit 33271823323483b4ede1ae99de83d33b25875402 from qemu
The function cpu_resume_from_signal() is now always called with a
NULL puc argument, and is rather misnamed since it is never called
from a signal handler. It is essentially forcing an exit to the
top level cpu loop but without raising any exception, so rename
it to cpu_loop_exit_noexc() and drop the useless unused argument.
Backports commit 6886b98036a8f8f5bce8b10756ce080084cef11b from qemu
Since the only caller of page_unprotect() which might cause it to
need to call cpu_resume_from_signal() is handle_cpu_signal() in
the user-mode code, push the longjump handling out to that function.
Since this is the only caller of cpu_resume_from_signal() which
passes a non-NULL puc argument, split the non-NULL handling into
a new cpu_exit_tb_from_sighandler() function. This allows us
to merge the softmmu and usermode implementations of the
cpu_resume_from_signal() function, which are now identical.
Backports commit f213e72f2356b77768b9cb73814a3b26ad5a0099 from qemu
The user-mode-only function tb_invalidate_phys_page() is only
called from two places:
* page_unprotect(), which passes in a non-zero pc, a puc pointer
and the value 'true' for the locked argument
* page_set_flags(), which passes in a zero pc, a NULL puc pointer
and a 'false' locked argument
If the pc is non-zero then we may call cpu_resume_from_signal(),
which does a longjmp out of the calling code (and out of the
signal handler); this is to cover the case of a target CPU with
"precise self-modifying code" (currently only x86) executing
a store instruction which modifies code in the same TB as the
store itself. Rather than doing the longjump directly here,
return a flag to the caller which indicates whether the current
TB was modified, and move the longjump to page_unprotect.
Backports commit 75809229bbf28b371afce14921ff5be98ddc5faa from qemu
Make sure that config-host.h and config-target.h are rebuilt whenever
there is a change in the scripts that generates them; add the dependency
to the pattern rule as suggested by Peter.
Backports commit 553350156d80c18d0127c742f47b7adbd642f3ef from qemu
The WORDS_ALIGNED #define is not used anywhere, and hasn't been since
2013 when commit 612d590ebc6cef rewrote the various ld<type>_<endian>_p
functions to not use it. Remove the #define and the comment describing it.
Also remove the line in the comment about TARGET_WORDS_ALIGNED, since
it has never actually existed.
Backports commit 0d5c21f2b3bf1e0b562a2c74e353d2e03f2f50ef from qemu
This fixes these warnings from shellcheck:
^-- SC2006: Use $(..) instead of deprecated `..`
Backports commit 89138857619b2a023c32200e9af780792ccaa8c3 from qemu
Address size is 40-bit for the AArch32 stage 2 translation,
and t0sz can be negative (from -8 to 7),
so we need to adjust it to use the existing TTBR selecting logic.
Backports commit 6e99f762612827afeff54add2e4fc2c3b2657fed from qemu
Remove some incorrect code from arm_cpu_do_interrupt_aarch64()
which attempts to set the IL bit in the syndrome register based
on the value of env->thumb. This is wrong in several ways:
* IL doesn't indicate Thumb-vs-ARM, it indicates instruction
length (which may be 16 or 32 for Thumb and is always 32 for ARM)
* not every syndrome format uses IL like this -- for some IL is
always set, and for some it is always clear
* the code is changing esr_el[new_el] even for interrupt entry,
which is not supposed to modify ESR_ELx at all
Delete the code, and instead rely on the syndrome value in
env->exception.syndrome having already been set up with the
correct value of IL.
Backports commit 78f1edb19fe11fa0c5d0bf484db59a384f455d3c from qemu
For some exception syndrome types, the IL bit should always be set.
This includes the instruction abort, watchpoint and software step
syndrome types; add the missing ARM_EL_IL bit to the syndrome
values returned by syn_insn_abort(), syn_swstep() and syn_watchpoint().
Backports commit 04ce861ea545477425ad9e045eec3f61c8a27df9 from qemu
Add support for generating the ISS (Instruction Specific Syndrome) for
Data Abort exceptions taken from AArch64.
These syndromes are used by hypervisors for example to trap and emulate
memory accesses.
We save the decoded data out-of-band with the TBs at translation time.
When exceptions hit, the extra data attached to the TB is used to
recreate the state needed to encode instruction syndromes.
This avoids the need to emit moves with every load/store.
Based on a suggestion from Peter Maydell.
Backports commit aaa1f954d4cab243e3d5337a72bc6d104e1c4808 from qemu
Add the Hypervisor System Trap Register for EL2.
This register is used early in the Linux boot and without it the kernel
aborts with a "Synchronous Abort" error.
Backports commit 2a5a9abd4bc45e2f4c62c77e07aebe53608c6915 from qemu
Using "," literal in $(call quiet-command, ...) arguments is awkward.
Add this constant to make it at least doable.
Backports commit 2f4e4dc237261c76734d8ae1d8e09d2983d2f1ca from qemu
Let users of qemu_get_ram_ptr and qemu_ram_ptr_length pass in an
address that is relative to the MemoryRegion. This basically means
what address_space_translate returns.
Because the semantics of the second parameter change, rename the
function to qemu_map_ram_ptr.
Backports commit 0878d0e11ba8013dd759c6921cbf05ba6a41bd71 from qemu
Move the old qemu_ram_addr_from_host to memory_region_from_host and
make it return an offset within the region. For qemu_ram_addr_from_host
return the ram_addr_t directly, similar to what it was before
commit 1b5ec23 ("memory: return MemoryRegion from qemu_ram_addr_from_host",
2013-07-04).
Backports commit 07bdaa4196b51bc7ffa7c3f74e9e4a9dc8a7966a from qemu
Of the two callers, one does not use it, and the other can compute
it itself based on the other output argument (offset) and the RAMBlock.
Backports commit f615f39616c4fd1a3a3b078af8d75bb4be6390de from qemu
Remove direct uses of ram_addr_t and optimize memory_region_{get,set}_fd
now that a MemoryRegion knows its RAMBlock directly.
Backports commit 4ff87573df3606856a92c14eef3393a63d736d11 from qemu
Currently we emit a consume-load in atomic_rcu_read. Because of
limitations in current compilers, this is overkill for non-Alpha hosts
and it is only useful to make Thread Sanitizer work.
This patch leaves the consume-load in atomic_rcu_read when
compiling with Thread Sanitizer enabled, and resorts to a
relaxed load + smp_read_barrier_depends otherwise.
On an RMO host architecture, such as aarch64, the performance
improvement of this change is easily measurable. For instance,
qht-bench performs an atomic_rcu_read on every lookup. Performance
before and after applying this patch:
$ tests/qht-bench -d 5 -n 1
Before: 9.78 MT/s
After: 10.96 MT/s
Backports commit 15487aa132109891482f79d78a30d6cfd465a391 from qemu
For correctness, smp_read_barrier_depends() is only required to
emit a barrier on Alpha hosts. However, we are currently emitting
a consume fence unconditionally, and most compilers currently treat
consume and acquire fences as equivalent.
Fix it by keeping the consume fence if we're compiling with Thread
Sanitizer, since this might help prevent false warnings. Otherwise,
only emit the barrier for Alpha hosts. Note that we still guarantee
that smp_read_barrier_depends() is a compiler barrier.
Backports commit c983895258a771f8a5e4a53950bfb7fd2216651c from qemu
It is not safe to make a direct jump to a TB spanning two pages in
system emulation because the mapping for the second page can get changed
but we don't take care of direct jumps in this case.
However in user mode emulation, this is not the case because there's
only static address translation and TBs are always invalidated properly.
Backports commit c88c67e58b61618a904d2333ceebefc3c852d32e from qemu
QOM instance_init functions are not supposed to have any side-effects,
as new objects may be created at any moment for querying property
information (see qmp_device_list_properties()).
Move TCG initialization to realize time so it won't be called when just
doing object_new() on a X86CPU subclass.
Backports commit 57f2453ab48a771b30aeced01b329ee85853bb7b from qemu
x86_cpudef_init() doesn't do anything anymore, cpudef_init(),
cpudef_setup(), and x86_cpudef_init() can be finally removed.
Backports commit 3e2c0e062f0963a6b73b0cd1990fad79495463d9 from qemu
Newer PC machines don't set hw_version, and older machines set
model-id on compat_props explicitly, so we don't need the
x86_cpudef_setup() code that sets model_id using
qemu_hw_version() anymore.
Backports commit 9cf2cc3d8237732946720d78bf9aec0064026ed8 from qemu
The macro will be used by code that will stop calling
qemu_hw_version() at runtime and just need a constant value.
Backports commit d494352c2f7818aeba184a8ef757569083740bb2 from qemu
This doesn't introduce any change in the code, as the offsets and
struct sizes match what was present in the table. This can be
validated by the QEMU_BUILD_BUG_ON lines on target-i386/cpu.h,
which ensures the struct sizes and offsets match the existing
values in ext_save_area.
Backports commit ee1b09f695dcd8532f470e53297473bd3bc88718 from qemu
Add structs that define the layout of the xsave areas used by
Intel processors. Add some QEMU_BUILD_BUG_ON lines to ensure the
structs match the XSAVE_* macros in target-i386/kvm.c and the
offsets and sizes at target-i386/cpu.c:ext_save_areas.
Backports commit b503717d28e8f7eff39bf38624e6cf42687d951a from qemu
mr->ram_block->offset is already aligned to both host and target size
(see qemu_ram_alloc_internal). Remove further masking as it is
unnecessary.
Backports commit e4e697940dff612b789b0858270c20a8b680f78d from qemu
Its value is alway set to mr->romd_mode, so the removed comparisons are
fully superseded by "a->mr == b->mr".
Backports commit 5b5660adf1fdb61db14ec681b10463b8cba633f1 from qemu
The collision check does nothing and hasn't been used. Remove the
variable together with related code.
Backports commit b61359781958759317ee6fd1a45b59be0b7dbbe1 from qemu
On the one hand, we have already qemu_get_ram_block() whose function
is similar. On the other hand, we can directly use mr->ram_block but
searching RAMblock by ram_addr which is a kind of waste.
Backports commit fa53a0e53efdc7002497ea4a76aacf6cceb170ef from qemu
Include qom/object.h and exec/memory.h instead of exec/ioport.h;
exec/ioport.h was almost everywhere required only for those two
includes, not for the content of the header itself.
Remove block/aio.h, everybody is already including it through
another path.
With this change, include/hw/hw.h is freed from qemu-common.h.
Backports commit df43d49cb8708b9c88a20afe0d1a3089b550a5b8 from qemu
pio_addr_t is almost unused, because these days I/O ports are simply
accessed through the address space. cpu_{in,out}[bwl] themselves are
almost unused; monitor.c and xen-hvm.c could use address_space_read/write
directly, since they have an integer size at hand. This leaves qtest as
the only user of those functions.
On the other hand even portio_* functions use this type; the only
interesting use of pio_addr_t thus is include/hw/sysbus.h. I guess I
could move it there, but I don't see much benefit in that either. Using
uint32_t is enough and avoids the need to include ioport.h everywhere.
Backports commit 89a80e7400f7225d9401b35ef32454b4ab29dc67 from qemu
exec-all.h contains TCG-specific definitions. It is not needed outside
TCG-specific files such as translate.c, exec.c or *helper.c.
One generic function had snuck into include/exec/exec-all.h; move it to
include/qom/cpu.h.
Backports commit 63c915526d6a54a95919ebece83fa9ca631b2508 from qemu
TCG backends do not need most of exec-all.h; extract what they actually
need to a separate file or move it directly to tcg.h. The next patch
will stop including exec-all.h from everywhere.
Backports commit 00f6da6a1a5d1ce085334eccbb50ec899ceed513 from qemu
These are here for historical reasons: they are needed from both gdbstub.c
and op_helper.c, and the latter was compiled with fixed AREG0. It is
not needed anymore, so uninline them.
Backports commit e6623d88f44aae9e9c78276c0cb7bd352283d50a from qemu
Move it to the actual users. There are still a few includes of
qemu/bswap.h in headers; removing them is left for future work.
Backports commit 58369e22cf971448411bfbc8c894b2addebe2111 from qemu
Disentangle cpu-common.h and memory.h from NEED_CPU_H. Prototypes are
not defined for !NEED_CPU_H, so remove them from poison.h too. Only
macros need poisoning.
Backports commit a7d6039cb35592683ecc56d2b37817da2d2f8b00 from qemu
Make SPARCCPU an opaque type within cpu-qom.h, and move all definitions
of private methods, as well as all type definitions that require knowledge
of the layout to cpu.h. This helps making files independent of NEED_CPU_H
if they only need to pass around CPU pointers.
Backports commit d61d1b20610e4655d7846e4cb43d22188e935f5f from qemu
Make MIPSCPU an opaque type within cpu-qom.h, and move all definitions of
private methods, as well as all type definitions that require knowledge
of the layout to cpu.h. This helps making files independent of NEED_CPU_H
if they only need to pass around CPU pointers.
Backports commit 416bf936864f16caad6993b9ebd452fb34f801bd from qemu
Make M68KCPU an opaque type within cpu-qom.h, and move all definitions of
private methods, as well as all type definitions that require knowledge
of the layout to cpu.h. This helps making files independent of NEED_CPU_H
if they only need to pass around CPU pointers.
Backports commit a836b8fa00fa1032ccd234a71b33943627d211ea from qemu
Make X86CPU an opaque type within cpu-qom.h, and move all definitions of
private methods, as well as all type definitions that require knowledge
of the layout to cpu.h. This helps making files independent of NEED_CPU_H
if they only need to pass around CPU pointers.
Backports commit 4da6f8d954429c0cd1471d25cb9dbe909607374e from qemu
Make ARMCPU an opaque type within cpu-qom.h, and move all definitions of
private methods, as well as all type definitions that require knowledge
of the layout to cpu.h. This helps making files independent of NEED_CPU_H
if they only need to pass around CPU pointers.
Backports commit 74e755647c1598a6845df1ee4f8b96d01afd96e7 from qemu
Return the negated value of accel_initialised is meaningless,
and the caller vl doesn't check it.
Backports commit bdc3f61dec2f9c227235bb5f677a0272e1184c82 from qemu
Simplify cpu_exec() by extracting TB execution code outside of
cpu_exec() into a new static inline function cpu_loop_exec_tb().
Backports commit 928de9ee14b0b63ee9f9275732ed3e1c8b5f4790 from qemu
Simplify cpu_exec() by extracting interrupt handling code outside of
cpu_exec() into a new static inline function cpu_handle_interrupt().
Backports commit c385e6e49763c6dd5dbbd90fadde95d986f8bd38 from qemu
Simplify cpu_exec() by extracting exception handling code out of
cpu_exec() into a new static inline function cpu_handle_exception().
Also make cpu_handle_debug_exception() inline as it is used only once.
Backports commit ea284766ec6b9f1712369249566b4c372f3cec8b from qemu
Simplify cpu_exec() by extracting CPU halt state handling code out of
cpu_exec() into a new static inline function cpu_handle_halt().
Backports commit 8b2d34e997371c9729a0f41e3cc624d4300bbe78 from qemu
This comment should have been deleted by commit 0ac087f1f3ae ("removed
unused code") but somehow it is still here. There's no point to keep it.
Backports commit c6f0d9f84c43ae973270df1a77482466558ee487 from qemu
This field was used for telling cpu_interrupt() to unlink a chain of TBs
being executed when it worked that way. Now, cpu_interrupt() don't do
this anymore. So we don't need this field anymore.
Backports commit 3213525f8ab48742db09dab18cb9ae6f36a6c921 from qemu
Move tb_add_jump() call and surrounding code from cpu_exec() into
tb_find_fast(). That simplifies cpu_exec() a little by hiding the direct
chaining optimization details into tb_find_fast(). It also allows to
move tb_lock()/tb_unlock() pair into tb_find_fast(), putting it closer
to tb_find_slow() which also manipulates the lock.
Backports commit a0522c7a55cc8ac76d82884cf8e52f76daa664cc from qemu
'tb_invalidated_flag' was meant to catch two events:
* some TB has been invalidated by tb_phys_invalidate();
* the whole translation buffer has been flushed by tb_flush().
Then it was checked:
* in cpu_exec() to ensure that the last executed TB can be safely
linked to directly call the next one;
* in cpu_exec_nocache() to decide if the original TB should be provided
for further possible invalidation along with the temporarily
generated TB.
It is always safe to patch an invalidated TB since it is not going to be
used anyway. It is also safe to call tb_phys_invalidate() for an already
invalidated TB. Thus, setting this flag in tb_phys_invalidate() is
simply unnecessary. Moreover, it can prevent from pretty proper linking
of TBs, if any arbitrary TB has been invalidated. So just don't touch it
in tb_phys_invalidate().
If this flag is only used to catch whether tb_flush() has been called
then rename it to 'tb_flushed'. Declare it as 'bool' and stick to using
only 'true' and 'false' to set its value. Also, instead of setting it in
tb_gen_code(), just after tb_flush() has been called, do it right inside
of tb_flush().
In cpu_exec(), this flag is used to track if tb_flush() has been called
and have made 'next_tb' (a reference to the last executed TB) invalid
for linking it to directly call the next TB. tb_flush() can be called
during the CPU execution loop from tb_gen_code(), during TB execution or
by another thread while 'tb_lock' is released. Catch for translation
buffer flush reliably by resetting this flag once before first TB lookup
and each time we find it set before trying to add a direct jump. Don't
touch in in tb_find_physical().
Each vCPU has its own execution loop in multithreaded mode and thus
should have its own copy of the flag to be able to reset it with its own
'next_tb' and don't affect any other vCPU execution thread. So make this
flag per-vCPU and move it to CPUState.
In cpu_exec_nocache(), we only need to check if tb_flush() has been
called from tb_gen_code() called by cpu_exec_nocache() itself. To do
this reliably, preserve the old value of the flag, reset it before
calling tb_gen_code(), check afterwards, and combine the saved value
back to the flag.
This patch is based on the patch "tcg: move tb_invalidated_flag to
CPUState" from Paolo Bonzini <pbonzini@redhat.com>.
Backports commit 6f789be56d3f38e9214dafcfab3bf9be7191f370 from qemu
The value returned from tcg_qemu_tb_exec() is the value passed to the
corresponding tcg_gen_exit_tb() at translation time of the last TB
attempted to execute. It is a little confusing to store it in a variable
named 'next_tb'. In fact, it is a combination of 4-byte aligned pointer
and additional information in its two least significant bits. Break it
down right away into two variables named 'last_tb' and 'tb_exit' which
are a pointer to the last TB attempted to execute and the TB exit
reason, correspondingly. This simplifies the code and improves its
readability.
Correct a misleading documentation comment for tcg_qemu_tb_exec() and
fix logging in cpu_tb_exec(). Also rename a misleading 'next_tb' in
another couple of places.
Backports commit 819af24b9c1e95e6576f1cefd32f4d6bf56dfa56 from qemu
In user mode, there's only a static address translation, TBs are always
invalidated properly and direct jumps are reset when mapping change.
Thus the destination address is always valid for direct jumps and
there's no need to restrict it to the pages the TB resides in.
Backports commit 90aa39a1cc4837360889f0e033ca25cc82100308 from qemu
We don't take care of direct jumps when address mapping changes. Thus we
must be sure to generate direct jumps so that they always keep valid
even if address mapping changes. Luckily, we can only allow to execute a
TB if it was generated from the pages which match with current mapping.
Document tcg_gen_goto_tb() declaration and note the reason for
destination PC limitations.
Some targets with variable length instructions allow TB to straddle a
page boundary. However, we make sure that both of TB pages match the
current address mapping when looking up TBs. So it is safe to do direct
jumps into the both pages. Correct the checks for some of those targets.
Given that, we can safely patch a TB which spans two pages. Remove the
unnecessary check in cpu_exec() and allow such TBs to be patched.
Backports commit 5b053a4a28278bca606eeff7d1c0730df1b047e9 from qemu
Unify the code of this function with tb_jmp_remove_from_list(). Making
these functions similar improves their readability. Also this could be a
step towards making this function thread-safe.
Backports commit f9c5b66f487a04d3747dc6997b1503f9258df945 from qemu
Move the code for removing jumps to a TB out of tb_phys_invalidate() to
a separate static inline function tb_jmp_unlink(). This simplifies
tb_phys_invalidate() and improves code structure.
Backports commit 89bba496322d4cf996d42cdd4bb0912231656c3d from qemu
tb_jmp_remove() was only used to remove the TB from a list of all TBs
jumping to the same TB which is n-th jump destination of the given TB.
Put a comment briefly describing the function behavior and rename it to
better reflect its purpose.
Backports commit 133626783aa5a1bf86332fa3e6f7b8efe005f924 from qemu
The check is to make sure that another thread hasn't already done the
same while we were outside of tb_lock. Mention this in a comment.
Backports commit 9962c478b153a18fe88a6509fe58cd178aff8abc from qemu
Initialize TB's direct jump list data fields and reset the jumps before
tb_link_page() puts it into the physical hash table and the physical
page list. So TB is completely initialized before it becomes visible.
This is pure rearrangement of code to a more suitable place, though it
could be a preparation for relaxing the locking scheme in future.
Backports commit 901bc3deb43bf37c85e43955905d003be7ae5fa5 from qemu
These fields do not contain pure pointers to a TranslationBlock
structure. So uintptr_t is the most appropriate type for them.
Also put some asserts to assure that the two least significant bits of
the pointer are always zero before assigning it to jmp_list_first.
Backports commit c37e6d7e3589ecb96914faa21025ad7ba6654aea from qemu
Briefly describe in a comment how direct block chaining is done. It
should help in understanding of the following data fields.
Rename some fields in TranslationBlock and TCGContext structures to
better reflect their purpose (dropping excessive 'tb_' prefix in
TranslationBlock but keeping it in TCGContext):
tb_next_offset => jmp_reset_offset
tb_jmp_offset => jmp_insn_offset
tb_next => jmp_target_addr
jmp_next => jmp_list_next
jmp_first => jmp_list_first
Avoid using a magic constant as an invalid offset which is used to
indicate that there's no n-th jump generated.
Backports commit f309101c26b59641fc1aa8fb2a98a5441cdaea03 from qemu
The setting of tcg_ctx.code_gen_buffer_size is done by the only caller of
size_code_gen_buffer(), which is code_gen_alloc():
$ git grep size_code_gen_buffer
translate-all.c:static inline size_t size_code_gen_buffer(size_t tb_size)
translate-all.c: tcg_ctx.code_gen_buffer_size = size_code_gen_buffer(tb_size);
Backports commit 835154b6e2200460f04719d0028716a37c178368 from qemu
Ensure direct jump patching in MIPS is atomic by using
atomic_read()/atomic_set() for code patching.
Backports commit c82460a560176ef69c2f0662bd280612e274db96 from qemu
Ensure direct jump patching in SPARC is atomic by using
atomic_read()/atomic_set() for code patching.
Backports commit 84f79fb7c6e857edc807e4a251338243ce0cbac3 from qemu
Ensure direct jump patching in AArch64 is atomic by using
atomic_read()/atomic_set() for code patching.
Backports commit 9e269112953be4d670cb0d25042bd6546fcf3e45 from qemu
Ensure direct jump patching in ARM is atomic by using
atomic_read()/atomic_set() for code patching.
Backports commit 7d14e0e2d661479985197203589c38840e1066df from qemu
Ensure direct jump patching in s390 is atomic by:
* naturally aligning a location of direct jump address;
* using atomic_read()/atomic_set() for code patching.
Backports commit ed3d51ecd7fe248d3959e469d53890ac9ffe0cd2 from qemu
Ensure direct jump patching in i386 is atomic by:
* naturally aligning a location of direct jump address;
* using atomic_read()/atomic_set() for code patching.
Backports commit 0d07abf05e98903c7faf204a9a90f7d45b7554dc from qemu
These macros provide a convenient way to n-byte align pointers up and
down and check if a pointer is n-byte aligned.
Backports commit 6b587d3cda48e7ba26de8d30bf0d8a7063970715 from qemu
We are inconsistent with the type of tb->flags: usage varies loosely
between int and uint64_t. Settle to uint32_t everywhere, which is
superior to both: at least one target (aarch64) uses the most significant
bit in the u32, and uint64_t is wasteful.
Compile-tested for all targets.
Backports commit 89fee74a0f066dfd73830a7b5fa137e87888c870 from qemu
The TCR_EL2 and TCR_EL3 regdefs were incorrectly using the
vmsa_tcr_el1_write function for writes. Since these registers don't
have the A1 bit that TCR_EL1 does, we don't need to do a tlb_flush()
when they are written. Remove the unnecessary .writefn and also the
harmless but unneeded .raw_writefn and .resetfn definitions.
Backports commit 6459b94c26dd666badb3547fef1456992a08e60b from qemu
The various load/store variants under disas_ldst_reg can all reuse the
same decoding for opc, size, rt and is_vector.
This patch unifies the decoding in preparation for generating
instruction syndromes for data aborts.
This will allow us to reduce the number of places to hook in updates
to the load/store state needed to generate the insn syndromes.
No functional change.
Backports commit cd694521ca061a5d0436d5df4ec8c17c8f4dfcdb from qemu
Use extract32 instead of open coding the bit masking when decoding
is_signed and is_extended. This streamlines the decoding with some
of the other ldst variants.
No functional change.
Backports commit 026a19c3128678d4fe301fc36e8ffacdc9ecccb8 from qemu
Split the data abort syndrome generator into two versions:
One with a valid Instruction Specific Syndrome (ISS) and another without.
The following new flags are supported by the syndrome generator
with ISS:
* isv - Instruction syndrome valid
* sas - Syndrome access size
* sse - Syndrome sign extend
* srt - Syndrome register transfer
* sf - Sixty-Four bit register width
* ar - Acquire/Release
These flags are not yet used, so this patch has no functional change
except that we will now correctly set the IL bit in data abort
syndromes without ISS information.
Backports commit 094d028a7968236cd2b7f7b96394f7a3b8ad97c8 from qemu
Use tcg_set_insn_param() instead of directly accessing internal
tcg data structures to update an insn param.
Backports commit 25caa94c4a26daaab1e65c6d887e2972aeb5749e from qemu
Add tcg_set_insn_param as a mechanism to modify an insn
parameter after emiting the insn. This is useful for icount
and also for embedding fault information for a specific insn.
Backports commit 1d41478fd428e01f057d3248292e4cdcdb048523 from qemu
There is a bug in ARM address translation regime with a long-descriptor
format. On the descriptor reading its address is formed from an index
which is a part of the input address. And on the first iteration this index
is incorrectly masked with 'grainsize' mask. But it can be wider according
to pseudo-code.
On the other hand on the iterations other than first the descriptor address
is formed from the previous level descriptor by masking with 'descaddrmask'
value. It always clears just 12 lower bits, but it must clear 'grainsize'
lower bits instead according to pseudo-code.
The patch fixes both cases.
Backports commit dddb5223413c5425ae6eaeb3b967627efc9675f7 from qemu
As described in AArch32.CheckS2Permission an instruction fetch fails if
XN bit is set or there is no read permission for the address.
Backports commit dfda68377e20943f474505e75238cb96bc6874bf from qemu
Returning a partial object on error is an invitation for a careless
caller to leak memory. We already fixed things in an earlier
patch to guarantee NULL if visit_start fails ("qapi: Guarantee
NULL obj on input visitor callback error"), but that does not
help the case where visit_start succeeds but some other failure
happens before visit_end, such that we leak a partially constructed
object outside visit_type_FOO(). As no one outside the testsuite
was actually relying on these semantics, it is cleaner to just
document and guarantee that ALL pointer-based visit_type_FOO()
functions always leave a safe value in *obj during an input visitor
(either the new object on success, or NULL if an error is
encountered), so callers can now unconditionally use
qapi_free_FOO() to clean up regardless of whether an error occurred.
The decision is done by adding visit_is_input(), then updating the
generated code to check if additional cleanup is needed based on
the type of visitor in use.
Note that we still leave *obj unchanged after a scalar-based
visit_type_FOO(); I did not feel like auditing all uses of
visit_type_Enum() to see if the callers would tolerate a specific
sentinel value (not to mention having to decide whether it would
be better to use 0 or ENUM__MAX as that sentinel).
Backports commit 68ab47e4b4ecc1c4649362b8cc1e49794d1a6537 from qemu
The semantics of the list visit are somewhat baroque, with the
following pseudocode when FooList is used:
start()
for (prev = head; cur = next(prev); prev = &cur) {
visit(&cur->value)
}
Note that these semantics (advance before visit) requires that
the first call to next() return the list head, while all other
calls return the next element of the list; that is, every visitor
implementation is required to track extra state to decide whether
to return the input as-is, or to advance. It also requires an
argument of 'GenericList **' to next(), solely because the first
iteration might need to modify the caller's GenericList head, so
that all other calls have to do a layer of dereferencing.
Thankfully, we only have two uses of list visits in the entire
code base: one in spapr_drc (which completely avoids
visit_next_list(), feeding in integers from a different source
than uint8List), and one in qapi-visit.py. That is, all other
list visitors are generated in qapi-visit.c, and share the same
paradigm based on a qapi FooList type, so we can refactor how
lists are laid out with minimal churn among clients.
We can greatly simplify things by hoisting the special case
into the start() routine, and flipping the order in the loop
to visit before advance:
start(head)
for (tail = *head; tail; tail = next(tail)) {
visit(&tail->value)
}
With the simpler semantics, visitors have less state to track,
the argument to next() is reduced to 'GenericList *', and it
also becomes obvious whether an input visitor is allocating a
FooList during visit_start_list() (rather than the old way of
not knowing if an allocation happened until the first
visit_next_list()). As a minor drawback, we now allocate in
two functions instead of one, and have to pass the size to
both functions (unless we were to tweak the input visitors to
cache the size to start_list for reuse during next_list, but
that defeats the goal of less visitor state).
The signature of visit_start_list() is chosen to match
visit_start_struct(), with the new parameters after 'name'.
The spapr_drc case is a virtual visit, done by passing NULL for
list, similarly to how NULL is passed to visit_start_struct()
when a qapi type is not used in those visits. It was easy to
provide these semantics for qmp-output and dealloc visitors,
and a bit harder for qmp-input (several prerequisite patches
refactored things to make this patch straightforward). But it
turned out that the string and opts visitors munge enough other
state during visit_next_list() to make it easier to just
document and require a GenericList visit for now; an assertion
will remind us to adjust things if we need the semantics in the
future.
Several pre-requisite cleanup patches made the reshuffling of
the various visitors easier; particularly the qmp input visitor.
Backports commit d9f62dde1303286b24ac8ce88be27e2b9b9c5f46 from qemu
As shown in the previous commit, the string input visitor was
treating bogus input as an empty list rather than an error.
Fix parse_str() to set errp, then the callers to exit early if
an error was reported.
Meanwhile, fix the testsuite to use the generated
qapi_free_int16List() instead of rolling our own, and to
validate the fixed behavior, while at the same time documenting
one more change that we'd like to make in a later patch (a
failed visit_start_list should guarantee a NULL pointer,
regardless of what things were on input).
Backports commit 74f24cb6306d065045d0e2215a7d10533fa59c57 from qemu
As mentioned in previous patches, we want to call visit_end_struct()
functions unconditionally, so that visitors can release resources
tied up since the matching visit_start_struct() without also having
to worry about error priority if more than one error occurs.
Even though error_propagate() can be safely used to ignore a second
error during cleanup caused by a first error, it is simpler if the
cleanup cannot set an error. So, split out the error checking
portion (basically, input visitors checking for unvisited keys) into
a new function visit_check_struct(), which can be safely skipped if
any earlier errors are encountered, and leave the cleanup portion
(which never fails, but must be called unconditionally if
visit_start_struct() succeeded) in visit_end_struct().
Generated code in qapi-visit.c has diffs resembling:
|@@ -59,10 +59,12 @@ void visit_type_ACPIOSTInfo(Visitor *v,
| goto out_obj;
| }
| visit_type_ACPIOSTInfo_members(v, obj, &err);
|- error_propagate(errp, err);
|- err = NULL;
|+ if (err) {
|+ goto out_obj;
|+ }
|+ visit_check_struct(v, &err);
| out_obj:
|- visit_end_struct(v, &err);
|+ visit_end_struct(v);
| out:
and in qapi-event.c:
@@ -47,7 +47,10 @@ void qapi_event_send_acpi_device_ost(ACP
| goto out;
| }
| visit_type_q_obj_ACPI_DEVICE_OST_arg_members(v, ¶m, &err);
|- visit_end_struct(v, err ? NULL : &err);
|+ if (!err) {
|+ visit_check_struct(v, &err);
|+ }
|+ visit_end_struct(v);
| if (err) {
| goto out;
Backports commit 15c2f669e3fb2bc97f7b42d1871f595c0ac24af8 from qemu
Tighten assertions in the QMP output visitor, so that:
- qmp_output_get_qobject() can only be called after pairing a
visit_end_* for every visit_start_* (rather than allowing it on
a partially built object)
- qmp_output_get_qobject() cannot be called unless at least one
visit_type_* or visit_start/visit_end pair has occurred since
creation/reset (the accidental return of NULL fixed by commit
ab8bf1d7 would have been much easier to diagnose)
- ensure that we are encountering the expected object or list
type, to provide protection against mismatched push(struct)/
pop(list) or push(list)/pop(struct), similar to the qmp-input
protection added in commit bdd8e6b5.
- ensure that except for the root, 'name' is non-null inside a
dict, and NULL inside a list (this may need changing later if
we add "name.0" support for better error messages for a list,
but for now it makes sure all users are at least consistent)
Backports commit 56a6f02b8ce1fe41a2a9077593e46eca7d98267d from qemu
Implement the new type_null() callback for the qmp input and
output visitors. While we don't yet have a use for this in QAPI
input (the generator will need some tweaks first), some
potential usages have already been discussed on the list.
Meanwhile, the output visitor could already output explicit null
via type_any, but this gives us finer control.
At any rate, it's easy to test that we can round-trip an explicit
null through manual use of visit_type_null() wrapped by a virtual
visit_start_struct() walk, even if we can't do the visit in a
QAPI type. Repurpose the test_visitor_out_empty test,
particularly since a future patch will tighten semantics to
forbid use of qmp_output_get_qobject() without at least one
intervening visit_type_*.
Backports commit 3df016f185521f8dfa5bd89168722887156405c7 from qemu
Right now, qmp-output-visitor happens to produce a QNull result
if nothing is actually visited between the creation of the visitor
and the request for the resulting QObject. A stronger protocol
would require that a QMP output visit MUST visit something. But
to still be able to produce a JSON 'null' output, we need a new
visitor function that states our intentions. Yes, we could say
that such a visit must go through visit_type_any(), but that
feels clunky.
So this patch introduces the new visit_type_null() interface and
its no-op interface in the dealloc visitor, and stubs in the
qmp visitors (the next patch will finish the implementation).
For the visitors that will not implement the callback, document
the situation. The code in qapi-visit-core unconditionally
dereferences the callback pointer, so that a segfault will inform
a developer if they need to implement the callback for their
choice of visitor.
Note that JSON has a primitive null type, with the single value
null; likewise with the QNull type for QObject; but for QAPI,
we just have the 'null' value without a null type. We may
eventually want to add more support in QAPI for null (most likely,
we'd use it via an alternate type that permits 'null' or an
object); but we'll create that usage when we need it.
Backports commit 3bc97fd5924561d92f32758c67eaffd2e4e25038 from qemu
The visitor interface for mapping between QObject/QemuOpts/string
and QAPI is scandalously under-documented, making changes to visitor
core, individual visitors, and users of visitors difficult to
coordinate. Among other questions: when is it safe to pass NULL,
vs. when a string must be provided; which visitors implement which
callbacks; the difference between concrete and virtual visits.
Correct this by retrofitting proper contracts, and document where some
of the interface warts remain (for example, we may want to modify
visit_end_* to require the same 'obj' as the visit_start counterpart,
so the dealloc visitor can be simplified). Later patches in this
series will tackle some, but not all, of these warts.
Add assertions to (partially) enforce the contract. Some of these
were only made possible by recent cleanup commits.
Backports commit adfb264c9ed04bfc694921b72173be8e29e90024 from qemu
After recent changes, the only remaining use of
visit_start_implicit_struct() is for allocating the space needed
when visiting an alternate. Since the term 'implicit struct' is
hard to explain, rename the function to its current usage. While
at it, we can merge the functionality of visit_get_next_type()
into the same function, making it more like visit_start_struct().
Generated code is now slightly smaller:
| {
| Error *err = NULL;
|
|- visit_start_implicit_struct(v, (void**) obj, sizeof(BlockdevRef), &err);
|+ visit_start_alternate(v, name, (GenericAlternate **)obj, sizeof(**obj),
|+ true, &err);
| if (err) {
| goto out;
| }
|- visit_get_next_type(v, name, &(*obj)->type, true, &err);
|- if (err) {
|- goto out_obj;
|- }
| switch ((*obj)->type) {
| case QTYPE_QDICT:
| visit_start_struct(v, name, NULL, 0, &err);
...
| }
|-out_obj:
|- visit_end_implicit_struct(v);
|+ visit_end_alternate(v);
| out:
| error_propagate(errp, err);
| }
Backports commit dbf11922622685934bfb41e7cf2be9bd4a0405c0 from qemu
In the QMP input visitor, visiting a list traverses two objects:
the QAPI GenericList of the caller (which gets advanced in
visit_next_list() regardless of this patch), and the QList input
that we are converting to QAPI. For consistency with QDict
visits, we want to consume elements from the input QList during
the visit_type_FOO() for the list element; that is, we want ALL
the code for consuming an input to live in qmp_input_get_object(),
rather than having it split according to whether we are visiting
a dict or a list. Making qmp_input_get_object() the common point
of consumption will make it easier for a later patch to refactor
visit_start_list() to cover the GenericList * head of a QAPI list,
and in turn will get rid of the 'first' flag (which lived in
qmp_input_next_list() pre-patch, and is hoisted to StackObject
by this patch).
This patch is therefore altering the post-condition use of 'entry',
while keeping what gets visited unchanged, from:
start_list next_list type_ELT ... next_list type_ELT next_list end_list
visits 1st elt last elt
entry NULL 1st elt 1st elt last elt last elt NULL gone
where type_ELT() returns (entry ? entry : 1st elt) and next_list() steps
entry
to this usage:
start_list next_list type_ELT ... next_list type_ELT next_list end_list
visits 1st elt last elt
entry 1st elt 1nd elt 2nd elt last elt NULL NULL gone
where type_ELT() steps entry and returns the old entry, and next_list()
leaves entry alone.
Backports commit fcf3cb21783b2dae3358fdbe7001cb2f74e0cedf from qemu
Don't embed the root of the visit into the stack of current
containers being visited. That way, we no longer get confused
on whether the first visit of a dictionary is to the dictionary
itself or to one of the members of the dictionary, based on
whether the caller passed name=NULL; and makes the QMP Input
visitor like other visitors where the value of 'name' is now
ignored on the root visit. (We may someday want to revisit
the rules on what 'name' should be on a top-level visit,
rather than just ignoring it; but that would be the topic of
another patch).
An audit of all qmp_input_visitor_new() call sites shows that
there were only two places where callers had previously been
visiting to a QDict with a non-NULL name to bypass a call to
visit_start_struct(), and those were fixed in prior patches.
Backports commit ce140b176920b5b65184020735a3c65ed3e9aeda from qemu
Commit e8316d7 mistakenly passed consume=true within
qmp_input_optional() when checking if an optional member was
present, but the mistake was silently ignored since the code
happily let us extract a member more than once. Fix
qmp_input_optional() to not consume anything, then tighten up
the input visitor to ensure that a member is consumed exactly
once (all generated code follows this pattern; and the new
assert will catch any hand-written code that tries to visit
the same key more than once).
Backports commit e5826a2fd727f0be54a81083f31fe02a275465cd from qemu
The following uses of a QMP input visitor should be strict
(that is, excess keys in QDict input should be flagged if not
converted to QAPI):
- Testsuite code unrelated to explicitly testing non-strict
mode (test-qmp-commands, test-visitor-serialization); since
we want more code to be strict by default, having more tests
of strict mode doesn't hurt
- Code used for cloning QAPI objects (replay-input.c,
qemu-sockets.c); we are reparsing a QObject just barely
produced by the qmp output visitor and which therefore should
not have any garbage, so while it is extra work to be strict,
it validates that our clone is correct [note that a later patch
series will simplify these two uses by creating an actual
clone visitor that is much more efficient than a
generate/reparse cycle]
- qmp_object_add(), which calls into user_creatable_add_type().
Since command line parsing for '-object' uses the same
user_creatable_add_type() through the OptsVisitor, and that is
always strict, we want to ensure that any nested dictionaries
would be treated the same in QMP and from the command line (I
don't actually know if such nested dictionaries exist). Note
that on this code change, strictness only matters for nested
dictionaries (if even possible), since we already flag excess
input at the top level during an earlier object_property_set()
on an unknown key, whether from QemuOpts:
$ ./x86_64-softmmu/qemu-system-x86_64 -nographic -nodefaults -qmp stdio -object secret,id=sec0,data=letmein,format=raw,foo=bar
qemu-system-x86_64: -object secret,id=sec0,data=letmein,format=raw,foo=bar: Property '.foo' not found
or from QMP:
$ ./x86_64-softmmu/qemu-system-x86_64 -nographic -nodefaults -qmp stdio
{"QMP": {"version": {"qemu": {"micro": 93, "minor": 5, "major": 2}, "package": ""}, "capabilities": []}}
{"execute":"qmp_capabilities"}
{"return": {}}
{"execute":"object-add","arguments":{"qom-type":"secret","id":"sec0","props":{"format":"raw","data":"letmein","foo":"bar"}}}
{"error": {"class": "GenericError", "desc": "Property '.foo' not found"}}
The only remaining uses of non-strict input visits are:
- QMP 'qom-set' (which eventually executes
object_property_set_qobject()) - mark it as something to revisit
in the future (I didn't want to spend any more time on this patch
auditing if we have any QOM dictionary properties that might be
impacted, and couldn't easily prove whether this code path is
shared with anything else).
- test-qmp-input-visitor: explicit tests of non-strict mode. If
we later get rid of users that don't need strictness, then this
test should be merged with test-qmp-input-strict
Backports relevant parts of commit 240f64b6dc3346d044d7beb7cc3a53668ce47384 from qemu
Rather than having two separate ways to create a QMP input
visitor, where the safer approach has the more verbose name,
it is better to consolidate things into a single function
where the caller must explicitly choose whether to be strict
or to ignore excess input. This patch is the strictly
mechanical conversion; the next patch will then audit which
uses can be made stricter.
Backports commit fc471c18d5d2ec713d5a019f9530398675494bc8 from qemu
Management of the top of stack was a bit verbose; creating a
temporary variable and adding some comments makes the existing
code more legible before the next few patches improve things.
No semantic changes other than asserting that we are always
visiting a QObject, and not a NULL value. In particular, the
check for 'name && qobject_type(qobj) == QTYPE_QDICT)' is a
bit overkill (a dict visit should always have a name); a later
patch revisits that, while this patch is only changing one
layer of indentation due to dropping 'if (qobj)'.
Backports commit b471d012e5d7bec1d2272738141e121b5581fcdf from qemu
Our existing input visitors were not very consistent on errors in a
function taking 'TYPE **obj'. These are start_struct(),
start_alternate(), type_str(), and type_any(). next_list() is
similar, but can't fail (see commit 08f9541). While all of them set
'*obj' to allocated storage on success, it was not obvious whether
'*obj' was guaranteed safe on failure, or whether it was left
uninitialized. But a future patch wants to guarantee that
visit_type_FOO() does not leak a partially-constructed obj back to
the caller; it is easier to implement this if we can reliably state
that input visitors assign '*obj' regardless of success or failure,
and that on failure *obj is NULL. Add assertions to enforce
consistency in the final setting of err vs. *obj.
The opts-visitor start_struct() doesn't set an error, but it
also was doing a weird check for 0 size; all callers pass in
non-zero size if obj is non-NULL.
The testsuite has at least one spot where we no longer need
to pre-initialize a variable prior to a visit; valgrind confirms
that the test is still fine with the cleanup.
A later patch will document the design constraint implemented
here.
Backports commit e58d695e6c3a5cfa0aa2fc91b87ade017ef28b05 from qemu
By sticking the next pointer first, we don't need a union with
64-bit padding for smaller types. On 32-bit platforms, this
can reduce the size of uint8List from 16 bytes (or 12, depending
on whether 64-bit ints can tolerate 4-byte alignment) down to 8.
It has no effect on 64-bit platforms (where alignment still
dictates a 16-byte struct); but fewer anonymous unions is still
a win in my book.
It requires visit_next_list() to gain a size parameter, to know
what size element to allocate; comparable to the size parameter
of visit_start_struct().
I debated about going one step further, to allow for fewer casts,
by doing:
typedef GenericList GenericList;
struct GenericList {
GenericList *next;
};
struct FooList {
GenericList base;
Foo *value;
};
so that you convert to 'GenericList *' by '&foolist->base', and
back by 'container_of(generic, GenericList, base)' (as opposed to
the existing '(GenericList *)foolist' and '(FooList *)generic').
But doing that would require hoisting the declaration of
GenericList prior to inclusion of qapi-types.h, rather than its
current spot in visitor.h; it also makes iteration a bit more
verbose through 'foolist->base.next' instead of 'foolist->next'.
Note that for lists of objects, the 'value' payload is still
hidden behind a boxed pointer. Someday, it would be nice to do:
struct FooList {
FooList *next;
Foo value;
};
for one less level of malloc for each list element. This patch
is a step in that direction (now that 'next' is no longer at a
fixed non-zero offset within the struct, we can store more than
just a pointer's-worth of data as the value payload), but the
actual conversion would be a task for another series, as it will
touch a lot of code.
Backports commit e65d89bf1a4484e0db0f3dc820a8b209f2fb1e8b from qemu
We have three classes of QAPI visitors: input, output, and dealloc.
Currently, all implementations of these visitors have one thing in
common based on their visitor type: the implementation used for the
visit_type_enum() callback. But since we plan to add more such
common behavior, in relation to documenting and further refining
the semantics, it makes more sense to have the visitor
implementations advertise which class they belong to, so the common
qapi-visit-core code can use that information in multiple places.
A later patch will better document the types of visitors directly
in visitor.h.
For this patch, knowing the class of a visitor implementation lets
us make input_type_enum() and output_type_enum() become static
functions, by replacing the callback function Visitor.type_enum()
with the simpler enum member Visitor.type. Share a common
assertion in qapi-visit-core as part of the refactoring.
Move comments in opts-visitor.c to match the refactored layout.
Backports commit 983f52d4b3f86fb9dc9f8b142132feb5a8723016 from qemu
QEMU 2.6 added support for the XSAVE family of instructions, which
includes the XSETBV instruction which allows setting the XCR0
register.
But, when booting Linux kernels with XSAVE support enabled, I was
getting very early crashes where the instruction pointer was set
to 0x3. I tracked it down to a jump instruction generated by this:
gen_jmp_im(s->pc - pc_start);
where s->pc is pointing to the instruction after XSETBV and pc_start
is pointing _at_ XSETBV. Subtract the two and you get 0x3. Whoops.
The fix is to replace this typo with the pattern found everywhere
else in the file when folks want to end the translation buffer.
Richard Henderson confirmed that this is a bug and that this is the
correct fix.
Backports commit 502c8e86ea07294067578292c6d402601c196019 from qemu
Recent versions of GCC report the following error when compiling
target-mips/helper.c:
qemu/target-mips/helper.c:542:9: warning: ‘memset’ used with length
equal to number of elements without multiplication by element size
[-Wmemset-elt-size]
Backports commit a525decfaa3449f1458ea2d7a06320cf46aebf3f from qemu
Commit b00c72180c36 ("target-mips: add PC, XNP reg numbers to RDHWR")
changed the rdhwr helpers to use check_hwrena() to check the register
being accessed is enabled in CP0_HWREna when used from user mode. If
that check fails an EXCP_RI exception is raised at the host PC
calculated with GETPC().
However check_hwrena() may not be fully inlined as the
do_raise_exception() part of it is common regardless of the arguments.
This causes GETPC() to calculate the address in the call in the helper
instead of the generated code calling the helper. No TB will be found
and the EPC reported with the resulting guest RI exception points to the
beginning of the TB instead of the RDHWR instruction.
We can't reliably force check_hwrena() to be inlined, and converting it
to a macro would be ugly, so instead pass the host PC in as an argument,
with each rdhwr helper passing GETPC(). This should avoid any dependence
on compiler behaviour, and in practice seems to ensure the full inlining
of check_hwrena() on x86_64.
This issue causes failures when running a MIPS KVM (trap & emulate)
guest in a MIPS QEMU TCG guest, as the inner guest kernel will do a
RDHWR of counter, which is disabled in the outer guest's CP0_HWREna by
KVM so it can emulate the inner guest's counter. The emulation fails and
the RI exception is passed to the inner guest.
Backports commit d96391c1ffeb30a0afa695c86579517c69d9a889 from qemu
For KVM to use Transparent Huge Pages (THP) we have to ensure that the
alignment of the userspace address of the KVM memory slot and the IPA
that the guest sees for a memory region have the same offset from the 2M
huge page size boundary.
One way to achieve this is to always align the IPA region at a 2M
boundary and ensure that the mmap alignment is also at 2M.
Unfortunately, we were only doing this for __arm__, not for __aarch64__,
so add this simple condition.
This fixes a performance regression using KVM/ARM on AArch64 platforms
that showed a performance penalty of more than 50%, introduced by the
following commit:
9fac18f (oslib: allocate PROT_NONE pages on top of RAM, 2015-09-10)
We were only lucky before the above commit, because we were allocating
large regions and naturally getting a 2M alignment on those allocations
then.
Backports commit ee1e0f8e5d3682c561edcdceccff72b9d9b16d8b from qemu
The TCG code is quite performance sensitive, but at the same time can
also be quite tricky. That is why asserts that can be enabled with the
--enable-debug-tcg configure option.
This used to work the following way:
| #include "config.h"
|
| ...
|
| #if !defined(CONFIG_DEBUG_TCG) && !defined(NDEBUG)
| /* define it to suppress various consistency checks (faster) */
| #define NDEBUG
| #endif
|
| ...
|
| #include <assert.h>
Since commit 757e725b (tcg: Clean up includes) "config.h" as been
replaced by "qemu/osdep.h" which itself includes <assert.h>. As a
consequence the assertions are always enabled, even when using
--disable-debug-tcg, causing a performance regression, especially on
targets with many registers. For instance on qemu-system-ppc the
speed difference is about 15%.
tcg_debug_assert is controlled directly by CONFIG_DEBUG_TCG and already
uses in some places. This patch replaces all the calls to assert into
calss to tcg_debug_assert.
Backports commit eabb7b91b36b202b4dac2df2d59d698e3aff197a from qemu
Accoding the chapter 7.6 Trap Processing of the SPARC Architecture Manual v9,
the Trap Based Address Register is not modified as a trap is taken.
This fix allows booting FreeBSD-10.3-RELEASE-sparc64.
Backports commit de5f1077446ca455342db149737bdc395a7b9882 from qemu
ldstub [addr], reg incorrectly reads a signed byte from memory which causes
problems in the 32-bit Solaris mutex code. Here the byte value being read is
0xff which is incorrectly sign-extended to 0xffffffff before being written back
to the target register causing lock detection to behave incorrectly.
This fixes the intermittent hangs and MUTEX_HELD warnings issued to the
console when running 32-bit Solaris images under qemu-system-sparc.
With thanks to Joseph Dery for providing a condensed test image to consistently
reproduce the problem on demand, and Martin Husemann for allowing me access to
real hardware for comparison.
Backports commit 4553e10360a0713e31647220ed396942f9a6fca0 from qemu
Since 5e5f07e08 "TCG: Move translation block variables
to new context inside tcg_ctx: tb_ctx" on Feb 1 2013, compilation
of usermode + TB_DEBUG_CHECK has been broken. Fix it.
Backports commit 7e6bd36d61129feb7f667cb09ffec1b7b54b971c from qemu
Xiao Guangrong ran kvm-unit-tests on an actual machine with PKU and
found that it fails:
test pte.p pte.user pde.p pde.user pde.a pde.pse pkru.wd pkey=1 user write efer.nx cr4.pke: FAIL: error code 27 expected 7
Dump mapping: address: 0x123400000000
------L4: 2ebe007
------L3: 2ebf007
------L2: 8000000020000a5
(All failures are combinations of "pde.user pde.p pkru.wd pkey=1",
plus either "pde.pse" or "pte.p pte.user", plus one of "user cr0.wp",
"cr0.wp" or "user", plus unimportant bits such as accessed/dirty or
efer.nx).
So PFEC.PKEY is set even if the ordinary check failed (which it did
because pde.w is zero). Adjust QEMU to match behavior of silicon.
Backports commit 44d066a2f770ee9d61fd1c2a609bdf2a994dfdf7 from qemu
The MIPS TCG backend is the only one to have
tcg_target_reg_alloc_order[] elements of type TCGReg rather than int.
This resulted in commit 91478cefaaf2 ("tcg: Allocate indirect_base
temporaries in a different order") breaking the build on MIPS since the
type differed from indirect_reg_alloc_order[]:
tcg/tcg.c:1725:44: error: pointer type mismatch in conditional expression [-Werror]
order = rev ? indirect_reg_alloc_order : tcg_target_reg_alloc_order;
^
Make it an array of ints to fix the build and match other architectures.
Backports commit 2dc7553d0c0a3915c649e1a91b0f0be70b4674b3 from qemu
In order to simplify arguments of function, introduce a new struct
named X86CPUTopoInfo.
Backports commit ed256144cd6f0ca2ff59fc3fc8dca547506f433b from qemu
Move the architecture agnostic function prototypes for exec.c out of
cputlb.h to exec-all.h. This allows hiding of the arch specific
cputlb.h from exec.c which should be getting close to having no
architecture specifics. Prepares support for multi-arch, which will have
a minimal cpu.h that services exec.c but not cputlb.h.
Backports commit dfccc7602374c9fd3b083208b552d62daa244811 from qemu
To prepare for multi-arch, cputlb.c should only have awareness of one
single architecture. This means it should not have access to the full
CPU lists which may be heterogeneous. Instead, push the CPU_LOOP() up
to the one and only caller in exec.c.
Backports commit 9a13565d52bfd321934fb44ee004bbaf5f5913a8 from qemu
The last two arguments to these functions are the last and first bit to
check relative to the base. The code was using incorrectly the first
bit and the number of bits. Fix this in cpu_physical_memory_get_dirty
and cpu_physical_memory_all_dirty. This requires a few changes in the
iteration; change the code in cpu_physical_memory_set_dirty_range to
match.
Backports commit 88c73d16ad1b6c22a2ab082064d0d521f756296a from qemu
The __atomic primitives have been available since GCC 4.7 and provide
a richer interface for describing memory ordering requirements. As a
bonus by using the primitives instead of hand-rolled functions we can
use tools such as the ThreadSanitizer which need the use of well
defined APIs for its analysis.
If we have __ATOMIC defines we exclusively use the __atomic primitives
for all our atomic access. Otherwise we fall back to the mixture of
__sync and hand-rolled barrier cases.
Backports commit a0aa44b488b3601415d55041e4619aef5f3a4ba8 from qemu
__atomic_thread_fence does not include a compiler barrier; in the
C++11 memory model, fences take effect in combination with other
atomic operations. GCC implements this by making __atomic_load and
__atomic_store access memory as if the pointer was volatile, and
leaves no trace whatsoever of acquire and release fences in the
compiler's intermediate representation.
In QEMU, we want memory barriers to act on all memory, but at the same
time we would like to use __atomic_thread_fence for portability reasons.
Add compiler barriers manually around the __atomic_thread_fence.
Backports commit 3bbf572345c65813f86a8fc434ea1b23beb08e16 from qemu
Although accesses to ram_list.dirty_memory[] use atomics so multiple
threads can safely dirty the bitmap, the data structure is not fully
thread-safe yet.
This patch handles the RAM hotplug case where ram_list.dirty_memory[] is
grown. ram_list.dirty_memory[] is change from a regular bitmap to an
RCU array of pointers to fixed-size bitmap blocks. Threads can continue
accessing bitmap blocks while the array is being extended. See the
comments in the code for an in-depth explanation of struct
DirtyMemoryBlocks.
I have tested that live migration with virtio-blk dataplane works.
Backports commit 5b82b703b69acc67b78b98a5efc897a3912719eb from qemu
Move the ALIAS tag from VTCR_EL2 to VTCR so that we migrate the
64-bit version, as is usual. (This has no particular effect now
unless the guest wrote to the high RES0 bits of VTCR_EL2.)
Add a comment about why it's OK that we don't have the various
accessor functions that the EL1 TCR regdefs do.
Backports commit bf06c1123a427fefc2cf9cf8019578eafc19eb6f from qemu
The regdefs for the ESR_EL2 and ESR_EL3 system registers should not
be marked as ARM_CP_ALIAS, because these are the master copies; the
DFSR regdef in vmsa_pmsa_cp_reginfo[] is marked as an alias.
Remove the ALIAS tags so that these registers are correctly migrated.
Backports commit 094a7d0b9d10812d06be2c5c19288cee4603c693 from qemu
The regdef for SCTRL_EL3 was incorrectly marked as being an
ARM_CP_ALIAS, with the remark that this was because the 32-bit
definition would take care of reset and migration. However the
intention for banked registers as documented in the comment in
add_cpreg_to_hashtable() is:
* 2) If ARMv8 is enabled then we can count on a 64-bit version
* taking care of the secure bank. This requires that separate
* 32 and 64-bit definitions are provided.
and so it marks the 32-bit secure banked version as an alias.
This results in the sctlr_s/sctlr_el[3] field never being reset
or migrated for a 64-bit CPU with EL3 enabled.
Fix this by removing the ARM_CP_ALIAS annotation from SCTLR_EL3.
Since this means it now needs a real reset value, move the regdef
into the same place that we define the 32-bit SCTLR.
Backports commit e24fdd238a159d830a9a65dd9b08f80fba9b9e06 from qemu
MIPS Release 6 and MIPS SIMD Architecture make it mandatory to have IEEE
754-2008 FPU which is indicated by CP1 FIR.HAS2008, FCSR.ABS2008 and
FCSR.NAN2008 bits set to 1.
In QEMU we still keep these bits cleared as there is no 2008-NaN support.
However, this now causes problems preventing from running R6 Linux with
the v4.5 kernel. Kernel refuses to execute 2008-NaN ELFs on a CPU
whose FPU does not support 2008-NaN encoding:
(...)
VFS: Mounted root (ext4 filesystem) readonly on device 8:0.
devtmpfs: mounted
Freeing unused kernel memory: 256K (ffffffff806f0000 - ffffffff80730000)
request_module: runaway loop modprobe binfmt-464c
Starting init: /sbin/init exists but couldn't execute it (error -8)
request_module: runaway loop modprobe binfmt-464c
Starting init: /bin/sh exists but couldn't execute it (error -8)
Kernel panic - not syncing: No working init found. Try passing init= option to kernel. See Linux Documentation/init.txt for guidance.
Therefore always indicate presence of 2008-NaN support in R6 as well as in
R5+MSA CPUs, even though this feature is not yet supported by MIPS in QEMU.
Backports commit ba5c79f26221c0fd7139c883a34a4e75d993f732 from qemu
There is no particular reason to keep these functions in the header.
Suggested by Paolo.
Backports commit 99affd1d5bd4e396ecda50e53dfbc5147fa1313d from qemu
The MAAR register is a read/write register included in Release 5
of the architecture that defines the accessibility attributes of
physical address regions. In particular, MAAR defines whether an
instruction fetch or data load can speculatively access a memory
region within the physical address bounds specified by MAAR.
As QEMU doesn't do speculative access, hence this patch only
provides ability to access the registers.
Backports commit f6d4dd810983fdf3d1c9fb81838167efef63d1c8 from qemu
Indicate that in the MIPS64R6-generic CPU the memory-mapped
Global Configuration Register Space is implemented.
Backports commit a9a95061715ca09abff56a3f239f704c410912c2 from qemu
Physical base address for the memory-mapped Coherency Manager Global
Configuration Register space.
The MIPS default location for the GCR_BASE address is 0x1FBF_8.
This register only exists if Config3 CMGCR is set to one.
Backports commit c870e3f52cac0c8a4a1377398327c4ff20d49d41 from qemu
To avoid cluttering the code with #ifdef legs we wrap up the print
statements into a tlb_debug() macro. As access to the virtual TLB can
get quite heavy defining DEBUG_TLB_LOG will ensure all the logs go to
the qemu_log target of CPU_LOG_MMU instead of stderr. This remains
compile time optional as these debug statements haven't been considered
for usefulness for user visible logging.
I've also removed DEBUG_TLB_CHECK which wasn't used.
Backports commit 8526e1f4e418443a4d6ed0714487e47d45ef9c98 from qemu
qemu-log: dfilter-ise exec, out_asm, op and opt_op
This ensures the code generation debug code will honour -dfilter if set.
For the "exec" tracing I've added a new inline macro for efficiency's
sake.
Backports commit d977e1c2dbc9e63454b2000f91954d02543bf43b from qemu
When debugging big programs or system emulation sometimes you want both
the verbosity of cpu,exec et all but don't want to generate lots of logs
for unneeded stuff. This patch adds a new option -dfilter which allows
you to specify interesting address ranges in the form:
-dfilter 0x8000..0x8fff,0xffffffc000080000+0x200,...
Then logging code can use the new qemu_log_in_addr_range() function to
decide if it will output logging information for the given range.
Backports commit 3514552e04388d8e7686bcf89efd022e892acb5b from qemu
Improve the TB execution logging so that it is easier to identify
what is happening from trace logs:
* move the "Trace" logging of executed TBs into cpu_tb_exec()
so that it is emitted if and only if we actually execute a TB,
and for consistency for the CPU state logging
* log when we link two TBs together via tb_add_jump()
* log when cpu_tb_exec() returns early from a chain of TBs
The new style logging looks like this:
Trace 0x7fb7cc822ca0 [ffffffc0000dce00]
Linking TBs 0x7fb7cc822ca0 [ffffffc0000dce00] index 0 -> 0x7fb7cc823110 [ffffffc0000dce10]
Trace 0x7fb7cc823110 [ffffffc0000dce10]
Trace 0x7fb7cc823420 [ffffffc000302688]
Trace 0x7fb7cc8234a0 [ffffffc000302698]
Trace 0x7fb7cc823520 [ffffffc0003026a4]
Trace 0x7fb7cc823560 [ffffffc0000dce44]
Linking TBs 0x7fb7cc823560 [ffffffc0000dce44] index 1 -> 0x7fb7cc8235d0 [ffffffc0000dce70]
Trace 0x7fb7cc8235d0 [ffffffc0000dce70]
Stopped execution of TB chain before 0x7fb7cc8235d0 [ffffffc0000dce70]
Trace 0x7fb7cc8235d0 [ffffffc0000dce70]
Trace 0x7fb7cc822fd0 [ffffffc0000dd52c]
Backports commit 1a830635229e14c403600167823ea6b3b79d3097 from qemu
Make qemu_log_mask() a macro which only calls the function to
do the actual work if the logging is enabled. This avoids making
a function call in possible fast paths where logging is disabled.
Backports commit 7ee606230e6b7645d92365d9b39179368e83ac54 from qemu
My later debugging patches need access to the origin PC which is held in
the TranslationBlock structure. Pass down the whole structure as it also
holds the information about the code start point.
Backports commit 5bd2ec3d7b47b2252745882795d79aef36380fb7 from qemu
Move declarations out of qemu-common.h for functions declared in
utils/ files: e.g. include/qemu/path.h for utils/path.c.
Move inline functions out of qemu-common.h and into new files (e.g.
include/qemu/bcd.h)
Backports commit f348b6d1a53e5271cf1c9f9acc4646b4b98c1771 from qemu
Not only it makes sense, but it gets rid of checkpatch warning:
WARNING: consider using qemu_strtosz in preference to strtosz
Also remove get rid of tabs to please checkpatch.
Backports commit 4677bb40f809394bef5fa07329dea855c0371697 from qemu
This patch replaces get_ticks_per_sec() calls with the macro
NANOSECONDS_PER_SECOND. Also, as there are no callers, get_ticks_per_sec()
is then removed. This replacement improves the readability and
understandability of code.
For example,
timer_mod(fdctrl->result_timer,
qemu_clock_get_ns(QEMU_CLOCK_VIRTUAL) + (get_ticks_per_sec() / 50));
NANOSECONDS_PER_SECOND makes it obvious that qemu_clock_get_ns
matches the unit of the expression on the right side of the plus.
Backports commit 73bcb24d932912f8e75e1d88da0fc0ac6d4bce78 from qemu
qemu-common.h should only be included by .c files. Its file comment
explains why: "No header file should depend on qemu-common.h, as this
would easily lead to circular header dependencies."
One of the reasons for headers to include it is QEMU_ALIGN_UP() and
QEMU_ALIGN_DOWN(). Move them next to ROUND_UP() in qemu/osdep.h, to
facilitate removing these ill-advised includes later on.
Backports commit e07e540aaa08718c9ff8213067a3dcef31b3e313 from qemu
qemu-common.h should only be included by .c files. Its file comment
explains why: "No header file should depend on qemu-common.h, as this
would easily lead to circular header dependencies."
One of the reasons for headers to include it is HOST_LONG_BITS. Move
that to its more natural home qemu/osdep.h, to facilitate removing
these ill-advised includes later on.
This also lets us use HOST_LONG_BITS in bswap.h instead of duplicating
its definition there to avoid cyclic inclusion.
Backports commit a8139632161d7546218b696cada0a4f64cc78fb7 from qemu
Commit 57cb38b included qapi/error.h into qemu/osdep.h to get the
Error typedef. Since then, we've moved to include qemu/osdep.h
everywhere. Its file comment explains: "To avoid getting into
possible circular include dependencies, this file should not include
any other QEMU headers, with the exceptions of config-host.h,
compiler.h, os-posix.h and os-win32.h, all of which are doing a
similar job to this file and are under similar constraints."
qapi/error.h doesn't do a similar job, and it doesn't adhere to
similar constraints: it includes qapi-types.h. That's in excess of
100KiB of crap most .c files don't actually need.
Add the typedef to qemu/typedefs.h, and include that instead of
qapi/error.h. Include qapi/error.h in .c files that need it and don't
get it now. Include qapi-types.h in qom/object.h for uint16List.
Update scripts/clean-includes accordingly. Update it further to match
reality: replace config.h by config-target.h, add sysemu/os-posix.h,
sysemu/os-win32.h. Update the list of includes in the qemu/osdep.h
comment quoted above similarly.
This reduces the number of objects depending on qapi/error.h from "all
of them" to less than a third. Unfortunately, the number depending on
qapi-types.h shrinks only a little. More work is needed for that one.
Backports commit da34e65cb4025728566d6504a99916f6e7e1dd6a from qemu
As soon as setjmp.h is included from qemu/osdep.h, those old include
statements are no longer needed.
Add also setjmp.h to the list in scripts/clean-includes.
Backports commit 8ff98f1ed2f50cd05c3c5027c7efdf69859ec664 from qemu
setjmp must be declared before sysemu/os-win32.h
because it is redefined there for 64 bit Windows.
Backports commit e89fdafb58038038e3ccb860c5e1068ba063bac8 from qemu
Now that the generator supports it, we might as well use an
anonymous base rather than breaking out a single-use Base
structure, for all three of our current QMP flat unions.
Oddly enough, this change does not affect the resulting
introspection output (because we already inline the members of
a base type into an object, and had no independent use of the
base type reachable from a command).
The case_whitelist now has to list the name of an implicit
type; which is not too bad (consider it a feature if it makes
it harder for developers to make the whitelist grow :)
Backports commit 3666a97f78704b941c360dc917acb14c8774eca7 from qemu
Rather than requiring all flat unions to explicitly create
a separate base struct, we can allow the qapi schema to specify
the common members via an inline dictionary. This is similar to
how commands can specify an inline anonymous type for its 'data'.
We already have several struct types that only exist to serve as
a single flat union's base; the next commit will clean them up.
In particular, this patch's change to the BlockdevOptions example
in qapi-code-gen.txt will actually be done in the real QAPI schema.
Now that anonymous bases are legal, we need to rework the
flat-union-bad-base negative test (as previously written, it
forms what is now valid QAPI; tweak it to now provide coverage
of a new error message path), and add a positive test in
qapi-schema-test to use an anonymous base (making the integer
argument optional, for even more coverage).
Note that this patch only allows anonymous bases for flat unions;
simple unions are already enough syntactic sugar that we do not
want to burden them further. Meanwhile, while it would be easy
to also allow an anonymous base for structs, that would be quite
redundant, as the members can be put right into the struct
instead.
Backports commit ac4338f8eb783fd421aae492ca262a586918471e from qemu
Simple unions were carrying a special case that hid their 'data'
QMP member from the resulting C struct, via the hack method
QAPISchemaObjectTypeVariant.simple_union_type(). But by using
the work we started by unboxing flat union and alternate
branches, coupled with the ability to visit the members of an
implicit type, we can now expose the simple union's implicit
type in qapi-types.h:
| struct q_obj_ImageInfoSpecificQCow2_wrapper {
| ImageInfoSpecificQCow2 *data;
| };
|
| struct q_obj_ImageInfoSpecificVmdk_wrapper {
| ImageInfoSpecificVmdk *data;
| };
...
| struct ImageInfoSpecific {
| ImageInfoSpecificKind type;
| union { /* union tag is @type */
| void *data;
|- ImageInfoSpecificQCow2 *qcow2;
|- ImageInfoSpecificVmdk *vmdk;
|+ q_obj_ImageInfoSpecificQCow2_wrapper qcow2;
|+ q_obj_ImageInfoSpecificVmdk_wrapper vmdk;
| } u;
| };
Doing this removes asymmetry between QAPI's QMP side and its
C side (both sides now expose 'data'), and means that the
treatment of a simple union as sugar for a flat union is now
equivalent in both languages (previously the two approaches used
a different layer of dereferencing, where the simple union could
be converted to a flat union with equivalent C layout but
different {} on the wire, or to an equivalent QMP wire form
but with different C representation). Using the implicit type
also lets us get rid of the simple_union_type() hack.
Of course, now all clients of simple unions have to adjust from
using su->u.member to using su->u.member.data; while this touches
a number of files in the tree, some earlier cleanup patches
helped minimize the change to the initialization of a temporary
variable rather than every single member access. The generated
qapi-visit.c code is also affected by the layout change:
|@@ -7393,10 +7393,10 @@ void visit_type_ImageInfoSpecific_member
| }
| switch (obj->type) {
| case IMAGE_INFO_SPECIFIC_KIND_QCOW2:
|- visit_type_ImageInfoSpecificQCow2(v, "data", &obj->u.qcow2, &err);
|+ visit_type_q_obj_ImageInfoSpecificQCow2_wrapper_members(v, &obj->u.qcow2, &err);
| break;
| case IMAGE_INFO_SPECIFIC_KIND_VMDK:
|- visit_type_ImageInfoSpecificVmdk(v, "data", &obj->u.vmdk, &err);
|+ visit_type_q_obj_ImageInfoSpecificVmdk_wrapper_members(v, &obj->u.vmdk, &err);
| break;
| default:
| abort();
Backports commit 32bafa8fdd098d52fbf1102d5a5e48d29398c0aa from qemu
Now that we are always bulk-initializing a QAPI C struct to 0
(whether by g_malloc0() or by 'Type arg = {0};'), we no longer
have any clients of c_null() in the generator for per-element
initialization. This patch is easy enough to revert if we find
a use in the future, but in the present, get rid of the dead code.
Backports commit 861877a0dd0a8e1bdbcc9743530f4dc9745a736a from qemu
Commit 82ca8e46 noticed that we had multiple implementations of
visiting every member of a struct, and consolidated it into
gen_visit_fields() (now gen_visit_members()) with enough
parameters to cater to slight differences between the clients.
But recent exposure of implicit types has meant that we are now
down to a single use of that method, so we can clean up the
unused conditionals and just inline it into the remaining
caller: gen_visit_object_members().
Likewise, gen_err_check() no longer needs optional parameters,
as the lone use of non-defaults was via gen_visit_members().
No change to generated code.
Backports commit 12f254fd5f98717d17f079c73500123303b232da from qemu
Rather than generate inline per-member visits, take advantage
of the 'visit_type_FOO_members()' function for emitting events.
This is possible now that implicit structs can be visited like
any other. Generated code shrinks accordingly; by initializing
a struct based on parameters, through a new gen_param_var()
helper, like:
|@@ -338,6 +250,9 @@ void qapi_event_send_block_job_error(con
| QMPEventFuncEmit emit = qmp_event_get_func_emit();
| QmpOutputVisitor *qov;
| Visitor *v;
|+ q_obj_BLOCK_JOB_ERROR_arg param = {
|+ (char *)device, operation, action
|+ };
|
| if (!emit) {
| return;
@@ -351,19 +266,7 @@ void qapi_event_send_block_job_error(con
| if (err) {
| goto out;
| }
|- visit_type_str(v, "device", (char **)&device, &err);
|- if (err) {
|- goto out_obj;
|- }
|- visit_type_IoOperationType(v, "operation", &operation, &err);
|- if (err) {
|- goto out_obj;
|- }
|- visit_type_BlockErrorAction(v, "action", &action, &err);
|- if (err) {
|- goto out_obj;
|- }
|-out_obj:
|+ visit_type_q_obj_BLOCK_JOB_ERROR_arg_members(v, ¶m, &err);
| visit_end_struct(v, err ? NULL : &err);
Notice that the initialization of 'param' has to cast away const
(just as the old gen_visit_members() had to do): we can't change
the signature of the user function (which uses 'const char *'), but
have to assign it to a non-const QAPI object (which requires
'char *').
While touching this, document with a FIXME comment that there is
still a potential collision between QMP members and our choice of
local variable names within qapi_event_send_FOO().
This patch also paves the way for some followup simplifications
in the generator, in subsequent patches.
Backports commit 0949e95b48e30715e157cabbc59dcb0ed912d3ff from qemu
The original choice of ':obj-' as the prefix for implicit types
made it obvious that we weren't going to clash with any user-defined
names, which cannot contain ':'. But now we want to create structs
for implicit types, to get rid of special cases in the generators,
and our use of ':' in implicit names needs a tweak to produce valid
C code.
We could transliterate ':' to '_', except that C99 mandates that
"identifiers that begin with an underscore are always reserved for
use as identifiers with file scope in both the ordinary and tag name
spaces". So it's time to change our naming convention: we can
instead use the 'q_' prefix that we reserved for ourselves back in
commit 9fb081e0. Technically, since we aren't planning on exposing
the empty type in generated code, we could keep the name ':empty',
but renaming it to 'q_empty' makes the check for startswith('q_')
cover all implicit types, whether or not code is generated for them.
As long as we don't declare 'empty' or 'obj' ticklish, it shouldn't
clash with c_name() prepending 'q_' to the user's ticklish names.
Backports commit 7599697c66d22ff4c859ba6ccea30e6a9aae6b9b from qemu
We already have several places that want to visit all the members
of an implicit object within a larger context (simple union variant,
event with anonymous data, command with anonymous arguments struct);
and will be adding another one soon (the ability to declare an
anonymous base for a flat union). Having a C struct declared for
these implicit types, along with a visit_type_FOO_members() helper
function, will make for fewer special cases in our generator.
We do not, however, need qapi_free_FOO() or visit_type_FOO()
functions for implicit types, because they should not be used
directly outside of the generated code. This is done by adding a
conditional in visit_object_type() for both qapi-types.py and
qapi-visit.py based on the object name. The comparison of
"name.startswith('q_')" is a bit hacky (it's basically duplicating
what .is_implicit() already uses), but beats changing the signature
of the visit_object_type() callback to pass a new 'implicit' flag.
The hack should be temporary: we are considering adding a future
patch that consolidates the narrow visit_object_type(..., base,
local_members, variants) and visit_object_type_flat(...,
all_members, variants) [where different sets of information are
already broken out, and the QAPISchemaObjectType is no longer
available] into a broader visit_object_type(obj_type) [where the
visitor can query the needed fields from obj_type directly].
Also, now that we WANT to output C code for implicits, we no longer
need the visit_needed() filter, leaving 'q_empty' as the only object
still needing a special case. Remember, 'q_empty' is the only
built-in generated object, which means that without a special case
it would be emitted in multiple files (the main qapi-types.h and in
qga-qapi-types.h) causing compilation failure due to redefinition.
But since it has no members, it's easier to just avoid an attempt to
visit that particular type; since gen_object() is called recursively,
we also prime the objects_seen set to cover any recursion into the
empty type.
The patch relies on the changed naming of implicit types in the
previous patch. It is a bit unfortunate that the generated struct
names and visit_type_FOO_members() don't match normal naming
conventions, but it's not too bad, since they will only be used in
generated code.
The generated code grows substantially in size: the implicit
'-wrapper' types must be emitted in qapi-types.h before any union
can include an unboxed member of that type. Arguably, the '-args'
types could be emitted in a private header for just qapi-visit.c
and qmp-marshal.c, rather than polluting qapi-types.h; but adding
complexity to the generator to split the output location according
to role doesn't seem worth the maintenance costs.
Backports commit 7ce106a96feee4d46bfcdb47127b0935804c9357 from qemu
We started moving away from the use of the 'void *data' member
in the C union corresponding to a QAPI union back in commit
544a373; recent commits have gotten rid of other uses. Now
that it is completely unused, we can remove the member itself
as well as the FIXME comment. Update the testsuite to drop the
negative test union-clash-data.
Backports commit 48eb62a74fc2d6b0ae9e5f414304a85cfbf33066 from qemu
Dan Berrange reported a case where he needs to work with a
QCryptoBlockOptions union type using the OptsVisitor, but only
visit one of the branches of that type (the discriminator is not
visited directly, but learned externally). When things were
boxed, it was easy: just visit the variant directly, which took
care of both allocating the variant and visiting its members, then
store that pointer in the union type. But now that things are
unboxed, we need a way to visit the members without allocation,
done by exposing visit_type_FOO_members() to the user.
Before the patch, we had quite a bit of code associated with
object_members_seen to make sure that a declaration of the helper
was in scope before any use of the function. But now that the
helper is public and declared in the header, the .c file no
longer needs to worry about topological sorting (the helper is
always in scope), which leads to some nice cleanups.
Backports commit 4d91e9115cc6700113e772b19d1f39bbcf345977 from qemu
C types and JSON objects don't have fields, but members. We
shouldn't gratuitously invent terminology. This patch is a
strict renaming of static genarated functions, plus the naming
of the dummy filler member for empty structs, before the next
patch exposes some of that naming to the rest of the code base.
Backports commit c81200b01422783cd29796ef4ccc275d05f9ce67 from qemu
C types and JSON objects don't have fields, but members. We
shouldn't gratuitously invent terminology. This patch is a
strict renaming of generator code internals (including testsuite
comments), before later patches rename C interfaces.
No change to generated code with this patch.
Backports commit 14f00c6c492488381a513c3816b15794446231a0 from qemu
QAPISchemaType.c_type() is a bit awkward: it takes two optional
boolean flags is_param and is_unboxed, and they should never both
be True.
Add a new method for each of the flags, and drop the flags from
c_type().
Most callers pass no flags; they remain unchanged.
One caller passes is_param=True; call the new .c_param_type()
instead.
One caller passes is_unboxed=True, except for simple union types.
This is actually an ugly special case that will go away soon, so
until then, we now have to call either .c_type() or the new
.c_unboxed_type(). Tolerable in the interim.
It requires slightly more Python, but is arguably easier to read.
Backports commit 4040d995e49c5b818be79e50a18c1bf8d2354d12 from qemu
We are getting closer to the point where we could use one union
as the base or variant type within another union type (as long
as there are no collisions between any possible combination of
member names allowed across all discriminator choices). But
until we get to that point, it is worth asserting that variants
are not present in places where we are not prepared to handle
them: when exploding a type into a parameter list, we do not
expect variants. The qapi.py code is already checking this,
via the older check_type() method; but someday we hope to get
rid of that and move checking into QAPISchema*.check(). The
two asserts added here make sure any refactoring still catches
problems, and makes it locally obvious why we can iterate over
only type.members without worrying about type.variants.
Backports commit 29f6bd15eb8a55ed37b2a443f7275b3d134eb2b2 from qemu
Just specifying a custom string is simpler in basically all places that
used it, and in addition, specifying the BB or node name is something we
generally do not do in other error messages when opening a BDS, so we
should not do it here.
This changes the output for iotest 036 (to the better, in my opinion),
so the reference output needs to be changed accordingly.
Backports commit a55448b3681a880b77eaefe8b2c42912000cb481 from qemu
Qemu reports translation fault on 1st level instead of 0th level in case of
AArch64 address translation if the translation table walk is disabled or
the address is in the gap between the two regions.
Backports commit 1b4093ea6678ff79d3006db3d3abbf6990b4a59b from qemu
Starting with the ARMv7 Virtualization Extensions, the A32 and T32
instruction sets provide instructions "MSR (banked)" and "MRS
(banked)" which can be used to access registers for a mode other
than the current one:
* R<m>_<mode>
* ELR_hyp
* SPSR_<mode>
Implement the missing instructions.
Backports commit 8bfd0550be821cf27d71444e2af350de3c3d2ee3 from qemu
When &error_abort is passed in, the error reporting code
will print the current error message and then abort() the
process. Unfortunately at the time it aborts, we've not
yet appended the errno detail. This makes debugging certain
problems significantly harder as the log is incomplete.
Backports commit 20e2dec14954568848ad74e73aee9b3aeedd6584 from qemu
After reporting an error, ram_block_add was going on with the registration
of the RAMBlock. The visible effect is that it unlocked the ramlist
mutex twice.
Backports commit 39c350ee12e733070e63d64a21bd42607366ea99 from qemu
We discriminate here between opcodes that are illegal in the current
cpu mode or with illegal arguments (such as modrm.mod == 3) and
encodings that are unknown (such as an unimplemented isa extension).
Backports commit b9f9c5b41aab06479cb1695990b7cca98ef84fc7 from qemu
The patch in 7f0b714 was too simplistic, in that we wound up setting
the flag and then resetting it immediately in gen_eob.
Fixes the reported boot problem with Windows XP.
Backports commit f083d92c03e7a0741d2a9eba774a60d5a3ca772f from qemu
While ADDSEG will only be false in 16-bit mode for LEA, it can be
false even in other cases when 16-bit addresses are obtained via
the 67h prefix in 32-bit mode. In this case, gen_lea_v_seg forgets
to add a nonzero FS or GS base if CS/DS/ES/SS are all zero. This
case is pretty rare but happens when booting Windows 95/98, and
this patch fixes it.
The bug is visible since commit d6a291498, but it was introduced
together with gen_lea_v_seg and it probably could be reproduced
with a "addr16 gs movsb" instruction as early as in commit
ca2f29f555805d07fb0b9ebfbbfc4e3656530977.
Backports commit e2e02a820741ec4d96b8f313b06a2a7ed5e94fbd from qemu
In non-64-bit modes, the instruction always stores 16 bits.
But in 64-bit mode, when the destination is a register, the
instruction can write 32 or 64 bits.
Backports commit a657f79e32422634415c09f3f15c73d610297af5 from qemu
SMSW and LMSW accept register operands, but commit 1906b2a ("target-i386:
Rearrange processing of 0F 01", 2016-02-13) did not account for that.
Backports commit 880f8486503b32a29b653a3c0b3cfc5432012f38 from qemu
Under heavy workloads the lookup will likely end up with the same
MemoryRegionSection from last time. Using a pointer to cache the result,
like ram_list.mru_block, significantly reduces cost of
address_space_translate.
During address space topology update, as->dispatch will be reallocated
so the pointer is invalidated automatically.
Perf reports a visible drop on the cpu usage, because phys_page_find is
not called. Before:
2.35% qemu-system-x86_64 [.] phys_page_find
0.97% qemu-system-x86_64 [.] address_space_translate_internal
0.95% qemu-system-x86_64 [.] address_space_translate
0.55% qemu-system-x86_64 [.] address_space_lookup_region
After:
0.97% qemu-system-x86_64 [.] address_space_translate_internal
0.97% qemu-system-x86_64 [.] address_space_lookup_region
0.84% qemu-system-x86_64 [.] address_space_translate
Backports commit 729633c2bc30496073431584eb6e304776b4ebd4 from qemu
This will be shared by the next patch.
Also add a comment explaining the unobvious condition on "size.hi".
Backports commit 29cb533d8cbff1330717619780c2f1dfe764e003 from qemu
Currently the ObjectProperty iterator API works as follows:
ObjectPropertyIterator *iter;
iter = object_property_iter_init(obj);
while ((prop = object_property_iter_next(iter))) {
...
}
object_property_iter_free(iter);
This has the benefit that the ObjectPropertyIterator struct
can be opaque, but has the downside that callers need to
explicitly call a free function. It is also not in keeping
with iterator style used elsewhere in QEMU/GLib2.
This patch changes the API to use stack allocation instead:
ObjectPropertyIterator iter;
object_property_iter_init(&iter, obj);
while ((prop = object_property_iter_next(&iter))) {
...
}
Backports commit 7746abd8e9ee9db20c0b0fdb19504f163ba3cbea from qemu
When there are many instances of a given class, registering
properties against the instance is wasteful of resources. The
majority of objects have a statically defined list of possible
properties, so most of the properties are easily registerable
against the class. Only those properties which are conditionally
registered at runtime need be recorded against the klass.
Registering properties against classes also makes it possible
to provide static introspection of QOM - currently introspection
is only possible after creating an instance of a class, which
severely limits its usefulness.
This impl only supports simple scalar properties. It does not
attempt to allow child object / link object properties against
the class. There are ways to support those too, but it would
make this patch more complicated, so it is left as an exercise
for the future.
There is no equivalent to object_property_del() provided, since
classes must be immutable once they are defined.
Backports commit 16bf7f522a2ff68993f80631ed86254c71eaf5d4 from qemu
ARM GICv3 systems with large number of CPUs create lots of IRQ pins. Since
every pin is represented as a property, number of these properties becomes
very large. Every property add first makes sure there's no duplicates.
Traversing the list becomes very slow, therefore QEMU initialization takes
significant time (several seconds for e. g. 16 CPUs).
This patch replaces list with GHashTable, making lookup very fast. The only
drawback is that object_child_foreach() and object_child_foreach_recursive()
cannot add or remove properties during traversal, since GHashTableIter does
not have modify-safe version. However, the code seems not to modify objects
via these functions.
Backports commit b604a854e843505007c59d68112c654556102a20 from qemu
Commit ef701d7 screwed up handling of out-of-memory conditions.
Before the commit, we report the error and exit(1), in one place. The
commit lifts the error handling up the call chain some, to three
places. Fine. Except it uses &error_abort in these places, changing
the behavior from exit(1) to abort(), and thus undoing the work of
commit 3922825 "exec: Don't abort when we can't allocate guest
memory".
The previous two commits fixed one of the three places, another one
was fixed in commit 33e0eb5. This commit fixes the third one.
Backports commit 0bdaa3a429c6d07cd437b442a1f15f70be1addaa from qemu
Just specifying ops = NULL in some cases can be more convenient than having
two functions.
Backports commit 6d6d2abf2c2e52c0f404d0a31a963e945b0cc7ad from qemu
memory_region_present() leaks a reference to a MemoryRegion in the
case "mr == container". While fixing it, avoid reference counting
altogether for memory_region_present(), by using RCU only.
The return value could in principle be already invalid immediately
after memory_region_present returns, but presumably the caller knows
that and it's using memory_region_present to probe for devices that
are unpluggable, or something like that. The RCU critical section
is needed anyway, because it protects as->current_map.
Backports commit c6742b14fe7352059cd4954a356a8105757af31b from qemu
Very often the owner of the aliased region is the same as the owner of the alias
region itself. When this happens, the reference count can never go back to 0 and
the owner is leaked. This is for example breaking hot-unplug of virtio-pci
devices (the device cannot be plugged back again with the same id).
Another common use for alias is to transform the system I/O address space
into an MMIO regions; in this case the aliased region never dies, so there
is no problem. Otherwise the owner is always the same for aliasing
and aliased region.
I checked all calls to memory_region_init_alias introduced after commit
dfde4e6 (memory: add ref/unref calls, 2013-05-06) and they do not need the
reference in order to keep the owner of the aliased region alive.
Backports commit 52c91dac6bd891656f297dab76da51fc8bc61309 from qemu
All references to mr->ram_addr are replaced by
memory_region_get_ram_addr(mr) (except for a few assertions that are
replaced with mr->ram_block).
Backports commit 8e41fb63c5bf29ecabe0cee1239bf6230f19978a from qemu
We don't force "const" qualifiers with pointers in QEMU, but it's still
good to keep a clean function interface. Assigning to mr->ram_block is
in this sense ugly - one initializer mutating its owning object's state.
Move it to memory_region_init_*, where mr->ram_addr is assigned.
Backports commit 0a75601853c00f3729fa62c49ec0d4bb1e3d9bc1 from qemu
Previously we return RAMBlock.offset; now return the pointer to the
whole structure.
ram_block_add returns void now, error is completely passed with errp.
Backports commit 528f46af6ecd1e300db18684969104d4067b867b from qemu
Commit cbc0326b6fb9 caused SRS instructions executed from Secure
EL1 to trap to EL3 even if the specified mode was not monitor mode.
According to the ARMv8 Architecture reference manual [F6.1.203], ALL
of the following conditions need to be met for SRS to trap to EL3:
* It is executed at Secure PL1.
* The specified mode is monitor mode.
* EL3 is using AArch64.
Correct the condition governing the trap to EL3 to check the
specified mode.
Backports commit ba63cf47a93041137a94e86b7d0cd87fc896949b from qemu
System emulation only has a little-endian target; BE32 mode
is implemented by adjusting the low bits of the address
for every byte and halfword load and store. 64-bit accesses
flip the low and high words.
Backports commit e334bd3190f6c4ca12f1d40d316dc471c70009ab from qemu
Since this is not a high-performance path, just use a helper to
flip the E bit and force a lookup in the hash table since the
flags have changed.
Backports commit 9886ecdf31165de2d4b8bccc1a220bd6ac8bc192 from qemu
Introduce a tbflags for endianness, set based upon the CPUs current
endianness. This in turn propagates through to the disas endianness
flag.
Backports commit 91cca2cda9823b1e7a049cb308a05104b5076cba from qemu
Introduce a disas flag for setting the CPU data endianness. This allows
control of the endianness from the CPU state rather than hard-coding it
to TARGET_WORDS_BIGENDIAN.
Backports commit dacf0a2ff7d39ab12bd90f2f5496a3889facd54a from qemu
Implement SCTLR.EE bit which controls data endianess for exceptions
and page table translations. SCTLR.EE is mirrored to the CPSR.E bit
on exception entry.
Backports commit 73462dddf670c32c45c8ea359658092b0365b2d4 from qemu
endian with address manipulations on subword accesses (to give the
illusion of BE). But user-mode cannot tell the difference and is
already implemented as straight BE. So handle the difference in the
endianess query, where USER mode is BE and system is not.
Backports commit b2e62d9a7b9a2eb10e451a57813bad168376e122 from qemu
There is a CPU data endianness test that is used to drive the
virtio_big_endian test.
Move this up to the header so it can be more generally used for endian
tests. The KVM specific cpu_syncronize_state call is left behind in the
virtio specific function.
Rename it arm_cpu-data_is_big_endian() to more accurately capture that
this is for data accesses only.
Backports commit ed50ff7875d61a75517c92deb0444d73fbbca878 from qemu
bswap_code is a CPU property of sorts ("is the iside endianness the
opposite way round to TARGET_WORDS_BIGENDIAN?") but it is not the
actual CPU state involved here which is SCTLR.B (set for BE32
binaries, clear for BE8).
Replace bswap_code with SCTLR.B, and pass that to arm_ld*_code.
The next patches will make data fetches honor both SCTLR.B and
CPSR.E appropriately.
Backports commit f9fd40ebe4f55e0048e002925b8d65e66d56e7a7 from qemu
In helper.c the expression
(env->uncached_cpsr & CPSR_M) != CPSR_USER
is always true; the right hand side was supposed to be ARM_CPU_MODE_USR
(an error in commit cb01d391).
Since the incorrect expression was always true, this just meant that
commit cb01d391 had no effect.
However simply changing the RHS here would reveal a logic error: if
the mode is USR we wish to completely ignore the attempt to set the
mode bits, which means that we must clear the CPSR_M bits from mask
to avoid the uncached_cpsr bits being updated at the end of the
function.
Move the condition into the correct place in the code, fix its RHS
constant, and add a comment about the fact that we must be doing a
gdbstub write if we're in user mode.
Backports commit 8c4f0eb94cc65ee32a12feba88d0b32e3665d5ea from qemu
The target-dependant type TCGv must be defined in "tcg/tcg.h" before
including the tracing helper wrappers in "tcg/tcg-op.h".
It also makes more sense to define it here, where other TCG types are
defined too.
Backports commit 5d4e1a1081d3f1ec2908ff0eaebe312389971ab4 from qemu
Adds the 'TCGv_env' type for pointers to 'CPUArchState' objects. The
tracing infrastructure later needs to differentiate between regular
pointers and pointers to vCPUs.
Also changes all targets to use the new 'TCGv_env' type instead of the
generic 'TCGv_ptr'. As of now, the change is merely cosmetic ('TCGv_env'
translates into 'TCGv_ptr'), but that could change in the future to
enforce the difference.
Note that a 'TCGv_env' type (for 'CPUState') is not added, since all
helpers currently receive the architecture-specific
pointer ('CPUArchState').
Backports commit 1bcea73e13b2b059d0cb3301aeaca43e5656ef57 from qemu
The v8 ARM ARM defines that unused spaces in the ID_AA64* system
register ranges are Reserved and must RAZ, rather than being UNDEF.
Implement this.
In particular, ARM v8.2 adds a new feature register ID_AA64MMFR2,
and newer versions of the Linux kernel will attempt to read this,
which causes them not to boot up on versions of QEMU missing this fix.
Since the encoding .opc0 = 3, .opc1 = 0, .crn = 0, .crm = 2, .opc2 = 6
is actually defined in ARMv8 (as ID_MMFR4), we give it an entry in
the ARMCPU struct so CPUs can override it, though since none do
this too will just RAZ.
Backports commit e20d84c1407d43d5a2e2ac95dbb46db3b0af8f9f from qemu
Mark CNTHP_TVAL_EL2 as ARM_CP_NO_RAW due to the register not
having any underlying state. This fixes an issue with booting
KVM enabled kernels when EL2 is on.
Backports commit d44ec156300a149b386a14d3ab349d3b83b66b8c from qemu
Implement the performance monitor register traps controlled
by MDCR_EL3.TPM and MDCR_EL2.TPM. Most of the performance
registers already have an access function to deal with the
user-enable bit, and the TPM checks can be added there. We
also need a new access function which only implements the
TPM checks for use by the few not-EL0-accessible registers
and by PMUSERENR_EL0 (which is always EL0-readable).
Backports commit 1fce1ba985d9c5c96e5b9709e1356d1814b8fa9e from qemu
Fix two issues with our implementation of the SDCR:
* it is only present from ARMv8 onwards
* it does not contain several of the trap bits present in its 64-bit
counterpart the MDCR_EL3
Put the register description in the right place so that it does not
get enabled for ARMv7 and earlier, and give it a write function so that
we can mask out the bits which should not be allowed to have an effect
if EL3 is 32-bit.
Backports commit a8d64e735182cbbb5dcc98f41656b118c45e57cc from qemu
If HCR.TGE is 1 then mode changes via CPS and MSR from Monitor to
NonSecure PL1 modes are illegal mode changes. Implement this check
in bad_mode_switch().
(We don't currently implement HCR.TGE, but this is the only missing
check from the v8 ARM ARM G1.9.3 and so it's worth adding now; the
rest of the HCR.TGE checks can be added later as necessary.)
Backports commit 10eacda787ac9990dc22d4437b289200c819712c from qemu
Mode switches from Hyp to any other mode via the CPS and MRS
instructions are illegal mode switches (though obviously switching
via exception return is valid). Add this check to bad_mode_switch().
Backports commit af393ffc6da116b9dd4c70901bad1f4cafb1773d from qemu
In v8, the illegal mode changes which are UNPREDICTABLE in v7 are
given architected behaviour:
* the mode field is unchanged
* PSTATE.IL is set (so any subsequent instructions will UNDEF)
* any other CPSR fields are written to as normal
This is pretty much the same behaviour we picked for our
UNPREDICTABLE handling, with the exception that for v8 we
need to set the IL bit.
Backports commit 81907a582901671c15be36a63b5063f88f3487e2 from qemu
In v8 trying to switch mode to Mon from Secure EL1 is an
illegal mode switch. (In v7 this is impossible as all secure
modes except User are at EL3.) We can handle this case by
making a switch to Mon valid only if the current EL is 3,
which then gives the correct answer whether EL3 is AArch32
or AArch64.
Backports commit 58ae2d1f037fae1d90eed4522053a85d79edfbec from qemu
We don't actually support Hyp mode yet, but add the correct
checks for it to the bad_mode_switch() function for completeness.
Backports commit e6c8fc07b4fce0729bb747770756835f4b0ca7f4 from qemu
QEMU doesn't implement the NSACR.RFR bit, which is a permitted
IMPDEF in choice in ARMv7 and the only permitted choice in ARMv8.
Add a comment to bad_mode_switch() to note that this is why
FIQ is always a valid mode regardless of the CPU's Secure state.
Backports commit 52ff951b4f63a29593650a15efdf82f63d6d962d from qemu
The only case where we can attempt a cpsr_write() mode switch from
User is from the gdbstub; all other cases are handled in the
calling code (notably translate.c). Architecturally attempts to
alter the mode bits from user mode are simply ignored (and not
treated as a bad mode switch, which in v8 sets CPSR.IL). Make
mode switches from User ignored in cpsr_write() as well, for
consistency.
Backports commit cb01d3912c8b000ed26d5fe95f6c194b3e3ba7a6 from qemu
Raw CPSR writes should skip the architectural checks for whether
we're allowed to set the A or F bits and should also not do
the switching of register banks if the mode changes. Handle
this inside cpsr_write(), which allows us to drop the "manually
set the mode bits to avoid the bank switch" code from all the
callsites which are using CPSRWriteRaw.
This fixes a bug in 32-bit KVM handling where we had forgotten
the "manually set the mode bits" part and could thus potentially
trash the register state if the mode from the last exit to userspace
differed from the mode on this exit.
Backports commit f8c88bbcda76d5674e4bb125471371b41d330df8 from qemu
Add an argument to cpsr_write() to indicate what kind of CPSR
write is being requested, since the exact behaviour should
differ for the different cases.
Backports commit 50866ba5a2cfe922aaf3edb79f6eac5b0653477a from qemu
The rules for setting the CPSR on a 32-bit exception return are
subtly different from those for setting the CPSR via an instruction
like MSR or CPS. (In particular, in Hyp mode changing the mode bits
is not valid via MSR or CPS.) Split the exception-return case into
its own helper for setting CPSR, so we can eventually handle them
differently in the helper function.
Backports commit 235ea1f5c89abf30e452539b973b0dbe43d3fe2b from qemu
MIPS Release 6 provides multi-threading features which replace
pre-R6 MT Module. CP0.Config3.MT is always 0 in R6, instead there is new
CP0.Config5.VP (Virtual Processor) bit which indicates presence of
multi-threading support which includes CP0.GlobalNumber register and
DVP/EVP instructions.
Backports commit 01bc435b44b8802cc4697faa07d908684afbce4e from qemu
The xsave and xrstor helpers are accessing the x86_ext_save_areas array
using a bit mask instead of a bit position. Provide two sets of XSTATE_*
definitions and use XSTATE_*_BIT when a bit position is requested.
Backports commit cfc3b074de4b4ccee2540edbf8cfdb026dc19943 from qemu
these two functions consume too much cpu overhead to
find the RAMBlock by ram address.
After this patch, we can pass the RAMBlock pointer
to them so that they don't need to find the RAMBlock
anymore most of the time. We can get better performance
in address translation processing.
Backports commit 3655cb9c7375a595a8051ec677c515b24d5c1fe6 from qemu
Each RAM memory region has a unique corresponding RAMBlock.
In the current realization, the memory region only stored
the ram_addr which means the offset of RAM address space,
We need to qurey the global ram.list to find the ram block
by ram_addr if we want to get the ram block, which is very
expensive.
Now, we store the RAMBlock pointer into memory region
structure. So, if we know the mr, we can easily get the
RAMBlock.
Backports commit 58eaa2174e99d9a05172d03fd2799ab8fd9e6f60 from qemu
Commit 757e725b58c57d added a number of #include "qemu/osdep.h"
files to the tcg-target.c files (as they were named at the time).
These are unnecessary because these files are not standalone C
files, and the tcg/tcg.c file which includes them will have
already included osdep.h on their behalf. Remove the unneeded
include directives.
Backports commit c3b7f66800fbf9f47fddbcf2e2cd30ea932e0aae from qemu
Rename the per-architecture tcg-target.c files to tcg-target.inc.c.
This makes it clearer that they are not intended to be standalone
C files, but are instead #included into another source file.
Backports commit ce151109813e2770fd3cee2f37bfa2cdd01a12b9 from qemu
Since we've not got liveness analysis for indirect bases,
placing them at the end of the call-saved registers makes
it more likely that it'll stay live.
Backports commit 91478cefaaf2fa678e56df8635b34957f4d5d565 from qemu
That is, global_mem registers whose base is another global_mem
register, rather than a fixed register.
Backports commit b3915dbbdcdb2e04753f3d34a1b0865eea005069 from qemu
A previous patch patch changed the type of REG from int
to enum TCGReg, which provokes the following bug in clang:
https://llvm.org/bugs/show_bug.cgi?id=16154
Backports commit 869938ae2a284fe730cb6f807ea0f9e324e0f87c from qemu
NEED_CPU_H is the define we use to distinguish per-target object
compilation from common object compilation. For the former, we must
also include config-target.h so that the .c files see the necessary
CONFIG_ constants.
Backports commit b1e34d1c3a9059e87719634bfc4db53174d63e14 from qemu
For C++ before C++11, <stdint.h> requires definition of the macros
__STDC_CONSTANT_MACROS, __STDC_LIMIT_MACROS and __STDC_FORMAT_MACROS
in order to enable definition of various macros by the header file.
Define these in osdep.h, so that we get the right header file
definitions whether osdep.h is being used by plain C, C++11 or
older C++.
In particular libvixl's header files depend on this and won't
compile if osdep.h is included before them otherwise.
Backports commit 79f56d82f805b170fa2be8c04b682117be56483f from qemu
We now do not use the int_fast*_t types anywhere in QEMU, so we can
remove the compatibility definitions we were providing for the
benefit of ancient Solaris versions.
Backports commit 50fe4df8ee6aba63ae51457bad40ba26e3c9746f from qemu
Use the plain 'int' type rather than 'int_fast16_t' for handling
exponents. Exponents don't need to be exactly 16 bits, so using int16_t
for them would confuse more than it clarified.
This should be a safe change because int_fast16_t semantics
permit use of 'int' (and on 32-bit glibc that is what you get).
Backports commit 0c48262d4772d40677364199372fb6ffcf487558 from qemu
Use the plain 'int' type rather than 'int_fast16_t' for shift counts
in the various shift related functions, since we don't actually care
about the size of the integer at all here, and using int16_t would
be confusing.
This should be a safe change because int_fast16_t semantics
permit use of 'int' (and on 32-bit glibc that is what you get).
Backports commit 07d792d2b08669bf6a97cbf590496078c4621068 from qemu
Make the functions which convert floating point to 16 bit integer
return int16_t rather than int_fast16_t, and correspondingly use
int_fast16_t in their internal implementations where appropriate.
(These functions are used only by the ARM target.)
Backports commit 0bb721d7217ed4a1abb44f521c5c7ec185062d58 from qemu
The r4k_tlb_t structure uses the uint_fast*_t types. Most of these
uses are in bitfields and are thus pointless, because the bitfield
itself specifies the width of the type; just use 'unsigned int'
instead. (On glibc uint_fast16_t is defined as either 32 or 64 bits,
so we know the code is not reliant on it being exactly 16 bits.)
There is also one use of uint_fast8_t, which we replace with uint8_t,
because both are exactly 8 bits on glibc and this is the only
place outside the softfloat code which uses an int_fast*_t type.
Backports commit d783f78933b212537ece77c7ec66866cc2bc0f4d from qemu
There's no reason to do two malloc's for a flat union; let's just
inline the branch struct directly into the C union branch of the
flat union.
Surprisingly, fewer clients were actually using explicit references
to the branch types in comparison to the number of flat unions
thus modified.
This lets us reduce the hack in qapi-types:gen_variants() added in
the previous patch; we no longer need to distinguish between
alternates and flat unions.
The change to unboxed structs means that u.data (added in commit
cee2dedb) is now coincident with random fields of each branch of
the flat union, whereas beforehand it was only coincident with
pointers (since all branches of a flat union have to be objects).
Note that this was already the case for simple unions - but there
we got lucky. Remember, visit_start_union() blindly returns true
for all visitors except for the dealloc visitor, where it returns
the value !!obj->u.data, and that this result then controls
whether to proceed with the visit to the variant. Pre-patch,
this meant that flat unions were testing whether the boxed pointer
was still NULL, and thereby skipping visit_end_implicit_struct()
and avoiding a NULL dereference if the pointer had not been
allocated. The same was true for simple unions where the current
branch had pointer type, except there we bypassed visit_type_FOO().
But for simple unions where the current branch had scalar type, the
contents of that scalar meant that the decision to call
visit_type_FOO() was data-dependent - the reason we got lucky there
is that visit_type_FOO() for all scalar types in the dealloc visitor
is a no-op (only the pointer variants had anything to free), so it
did not matter whether the dealloc visit was skipped. But with this
patch, we would risk leaking memory if we could skip a call to
visit_type_FOO_fields() based solely on a data-dependent decision.
But notice: in the dealloc visitor, visit_type_FOO() already handles
a NULL obj - it was only the visit_type_implicit_FOO() that was
failing to check for NULL. And now that we have refactored things to
have the branch be part of the parent struct, we no longer have a
separate pointer that can be NULL in the first place. So we can just
delete the call to visit_start_union() altogether, and blindly visit
the branch type; there is no change in behavior except to the dealloc
visitor, where we now unconditionally visit the branch, but where that
visit is now always safe (for a flat union, we can no longer
dereference NULL, and for a simple union, visit_type_FOO() was already
safely handling NULL on pointer types).
Unfortunately, simple unions are not as easy to switch to unboxed
layout; because we are special-casing the hidden implicit type with
a single 'data' member, we really DO need to keep calling another
layer of visit_start_struct(), with a second malloc; although there
are some cleanups planned for simple unions in later patches.
visit_start_union() and gen_visit_implicit_struct() are now unused.
Drop them.
Note that after this patch, the only remaining use of
visit_start_implicit_struct() is for alternate types; the next patch
will do further cleanup based on that fact.
Backports commit 544a3731591f5d53e15f22de00ce5ac758d490b3 from qemu
There's no reason to do two malloc's for an alternate type visiting
a QAPI struct; let's just inline the struct directly as the C union
branch of the struct.
Surprisingly, no clients were actually using the struct member prior
to this patch outside of the testsuite; an earlier patch in the series
added some testsuite coverage to make the effect of this patch more
obvious.
In qapi.py, c_type() gains a new is_unboxed flag to control when we
are emitting a C struct unboxed within the context of an outer
struct (different from our other two modes of usage with no flags
for normal local variable declarations, and with is_param for adding
'const' in a parameter list). I don't know if there is any more
pythonic way of collapsing the two flags into a single parameter,
as we never have a caller setting both flags at once.
Ultimately, we want to also unbox branches for QAPI unions, but as
that touches a lot more client code, it is better as separate
patches. But since unions and alternates share gen_variants(), I
had to hack in a way to test if we are visiting an alternate type
for setting the is_unboxed flag: look for a non-object branch.
This works because alternates have at least two branches, with at
most one object branch, while unions have only object branches.
The hack will go away in a later patch.
The generated code difference to qapi-types.h is relatively small:
| struct BlockdevRef {
| QType type;
| union { /* union tag is @type */
| void *data;
|- BlockdevOptions *definition;
|+ BlockdevOptions definition;
| char *reference;
| } u;
| };
The corresponding spot in qapi-visit.c calls visit_type_FOO(), which
first calls visit_start_struct() to allocate or deallocate the member
and handle a layer of {} from the JSON stream, then visits the
members. To peel off the indirection and the memory management that
comes with it, we inline this call, then suppress allocation /
deallocation by passing NULL to visit_start_struct(), and adjust the
member visit:
| switch ((*obj)->type) {
| case QTYPE_QDICT:
|- visit_type_BlockdevOptions(v, name, &(*obj)->u.definition, &err);
|+ visit_start_struct(v, name, NULL, 0, &err);
|+ if (err) {
|+ break;
|+ }
|+ visit_type_BlockdevOptions_fields(v, &(*obj)->u.definition, &err);
|+ error_propagate(errp, err);
|+ err = NULL;
|+ visit_end_struct(v, &err);
| break;
| case QTYPE_QSTRING:
| visit_type_str(v, name, &(*obj)->u.reference, &err);
The visit of non-object fields is unchanged.
Backports commit becceedc4d9bc1435099c90a0514945a89844d3a from qemu
We have several instances of methods that do an early exit if
output is not needed, then log that output is being generated,
and finally produce the output; see qapi-types.py:gen_object()
and qapi-visit.py:gen_visit_implicit_struct(). The odd man
out was gen_visit_fields_decl(); rearrange it to be more like
the others. No semantic change or difference to generated code.
Backports commit 2208d64998c5f867ccee7eeee298971685bf822d from qemu
Right now, we emit the branches of union types as a boxed pointer,
and it suffices to have a forward declaration of the type. However,
a future patch will swap things to directly use the branch type,
instead of hiding it behind a pointer. For this to work, the
compiler needs the full definition of the type, not just a forward
declaration, prior to the union that is including the branch type.
This patch just adds topological sorting to hoist all types
mentioned in a branch of a union to be fully declared before the
union itself. The sort is always possible, because we do not
allow circular union types that include themselves as a direct
branch (it is, however, still possible to include a branch type
that itself has a pointer to the union, for a type that can
indirectly recursively nest itself - that remains safe, because
that the member of the branch type will remain a pointer, and the
QMP representation of such a type adds another {} for each recurring
layer of the union type).
Backports commit 1de5d4ca0752138034305f3d4e8fe17ef6503569 from qemu
We were passing 'Foo **obj' to the internal helper function, but
all uses within the helper were via reads of '*obj'. Refactor
things to pass one less level of indirection, by having the
callers dereference before calling.
For an example of the generated code change:
|-static void visit_type_BalloonInfo_fields(Visitor *v, BalloonInfo **obj, Error **errp)
|+static void visit_type_BalloonInfo_fields(Visitor *v, BalloonInfo *obj, Error **errp)
| {
| Error *err = NULL;
|
|- visit_type_int(v, "actual", &(*obj)->actual, &err);
|+ visit_type_int(v, "actual", &obj->actual, &err);
| error_propagate(errp, err);
| }
|
|@@ -261,7 +261,7 @@ void visit_type_BalloonInfo(Visitor *v,
| if (!*obj) {
| goto out_obj;
| }
|- visit_type_BalloonInfo_fields(v, obj, &err);
|+ visit_type_BalloonInfo_fields(v, *obj, &err);
| out_obj:
The refactoring will also make it easier to reuse the helpers in
a future patch when implicit structs are stored directly in the
parent struct rather than boxed through a pointer.
Backports commit 655519030b5d20967ae3afa1fe91ef5ad4406065 from qemu
gen_visit_union() is now just like gen_visit_struct(). Rename
it to gen_visit_object(), use it for structs, and drop
gen_visit_struct(). Output is unchanged.
Backports commit 59d9e84cc94b8268f13ec184acfb997e8f352593 from qemu
We initially created the static visit_type_FOO_fields() helper
function for reuse of code - we have cases where the initial
setup for a visit has different allocation (depending on whether
the fields represent a stand-alone type or are embedded as part
of a larger type), but where the actual field visits are
identical once a pointer is available.
Up until the previous patch, visit_type_FOO_fields() was only
used for structs (no variants), so it was covering every field
for each type where it was emitted.
Meanwhile, the code for visiting unions looks like:
static visit_type_U_fields() {
visit base;
visit local_members;
}
visit_type_U() {
visit_start_struct();
visit_type_U_fields();
visit variants;
visit_end_struct();
}
which splits the fields of the union visit across two functions.
Move the code to visit variants to live inside visit_type_U_fields(),
while making it conditional on having variants so that all other
instances of the helper function remain unchanged. This is also
a step closer towards unifying struct and union visits, and towards
allowing one union type to be the branch of another flat union.
The resulting diff to the generated code is a bit hard to read,
but it can be verified that it touches only union types, and that
the end result is the following general structure:
static visit_type_U_fields() {
visit base;
visit local_members;
visit variants;
}
visit_type_U() {
visit_start_struct();
visit_type_U_fields();
visit_end_struct();
}
Backports commit 9a5cd424d5f06fb5293eb264456d89343c557558 from qemu
For a simple union SU, gen_visit_union() generates a visit of its
single tag member, like this:
visit_type_SUKind(v, "type", &(*obj)->type, &err);
For a flat union FU with base B, it generates a visit of its base
fields:
visit_type_B_fields(v, (B **)obj, &err);
Instead, we can simply visit the common members using the same fields
visit function we use for structs, generated with
gen_visit_struct_fields(). This function visits the base if any, then
the local members.
For a simple union SU, visit_type_SU_fields() contains exactly the old
tag member visit, because there is no base, and the tag member is the
only member. For instance, the code generated for qapi-schema.json's
KeyValue changes like this:
+static void visit_type_KeyValue_fields(Visitor *v, KeyValue **obj, Error **errp)
+{
+ Error *err = NULL;
+
+ visit_type_KeyValueKind(v, "type", &(*obj)->type, &err);
+ if (err) {
+ goto out;
+ }
+
+out:
+ error_propagate(errp, err);
+}
+
void visit_type_KeyValue(Visitor *v, const char *name, KeyValue **obj, Error **errp)
{
Error *err = NULL;
@@ -4863,7 +4911,7 @@ void visit_type_KeyValue(Visitor *v, con
if (!*obj) {
goto out_obj;
}
- visit_type_KeyValueKind(v, "type", &(*obj)->type, &err);
+ visit_type_KeyValue_fields(v, obj, &err);
if (err) {
goto out_obj;
}
For a flat union FU, visit_type_FU_fields() contains exactly the old
base fields visit, because there is a base, but no members. For
instance, the code generated for qapi-schema.json's CpuInfo changes
like this:
static void visit_type_CpuInfoBase_fields(Visitor *v, CpuInfoBase **obj, Error **errp);
+static void visit_type_CpuInfo_fields(Visitor *v, CpuInfo **obj, Error **errp)
+{
+ Error *err = NULL;
+
+ visit_type_CpuInfoBase_fields(v, (CpuInfoBase **)obj, &err);
+ if (err) {
+ goto out;
+ }
+
+out:
+ error_propagate(errp, err);
+}
+
static void visit_type_CpuInfoX86_fields(Visitor *v, CpuInfoX86 **obj, Error **errp)
...
@@ -3485,7 +3509,7 @@ void visit_type_CpuInfo(Visitor *v, cons
if (!*obj) {
goto out_obj;
}
- visit_type_CpuInfoBase_fields(v, (CpuInfoBase **)obj, &err);
+ visit_type_CpuInfo_fields(v, obj, &err);
if (err) {
goto out_obj;
}
As you see, the generated code grows a bit, but in practice, it's lost
in the noise: qapi-schema.json's qapi-visit.c gains roughly 1%.
This simplification became possible with commit 441cbac "qapi-visit:
Convert to QAPISchemaVisitor, fixing bugs". It's a step towards
unifying gen_struct() and gen_union().
Backports commit d7445b57f4342d50c25b5bc1746048fa3720511b from qemu
The whole point of an alternate is to allow some type-safety while
still accepting more than one JSON type. Meanwhile, the 'any'
type exists to bypass type-safety altogether. The two are
incompatible: you can't accept every type, and still tell which
branch of the alternate to use for the parse; fix this to give a
sane error instead of a Python stack trace.
Note that other types that can't be alternate members are caught
earlier, by check_type().
Backports commit 46534309e667fd860720f983c2c9aefe0354340d from qemu
Empty unions serve no purpose, and while we compile with gcc
which permits them, strict C99 forbids them. We happen to inject
a dummy 'void *data' member into the C unions that represent QAPI
unions and alternates, but we want to get rid of that member (it
pollutes the namespace for no good reason), which would leave us
with an empty union if the user didn't provide any branches. While
empty structs make sense in QAPI, empty unions don't add any
expressiveness to the QMP language. So prohibit them at parse
time. Update the documentation and testsuite to match.
Note that the documentation already mentioned that alternates
should have "two or more JSON data types"; so this also fixes
the code to enforce that. However, we have existing uses of a
union type with only one branch, so the 2-or-more strictness
is intentionally limited to alternates.
Backports commit 02a57ae32b08e8981b59979b80e682c9a153e94d from qemu
When we added support for a user-specified prefix for an enum
type (commit 351d36e), we forgot to teach the qapi-visit code
to honor that prefix in the case of using a prefixed enum as
the discriminator for a flat union. While there is still some
on-list debate on whether we want to keep prefixes, we should
at least make it work as long as it is still part of the code
base.
Backports commit 9d3524b39e1fe5f3bb7a990ad7841e469e954a3b from qemu
Make get_r13_banked() raise an exception at runtime for the
corner case of SRS from System mode, so that we can UNDEF it;
this brings us in to line with the ARM ARM's set of permitted
CONSTRAINED UNPREDICTABLE choices.
Backports commit f01377f591fe15c652f947646c4a69a7d4a71ad9 from qemu
The user-mode versions of get/set_r13_banked() exist just to assert
if they're ever called -- the translate time code should never
emit calls to them because SRS from user mode always UNDEF.
There's no code in the softmmu versions that can't compile in
CONFIG_USER_ONLY, and the assertion is not particularly useful,
so combine the two functions rather than having completely split
versions under ifdefs.
Backports commit d86d57d4fe683c99823f625f941eff26c07c72c3 from qemu
Move get/set_r13_banked() from helper.c to op_helper.c. This will
let us add exception-raising code to them, and also puts them
in the same file as get/set_user_reg(), which makes some conceptual
sense.
(The original reason for the helper.c/op_helper.c split was that
only op_helper.c had access to the CPU env pointer; this distinction
has not been true for a long time, though, and so the split is
now rather arbitrary.)
Backports commit 72309cee482868d6c4711931c3f7e02ab9dec229 from qemu
target-arm: Move bank_number() into internals.h
Move bank_number()'s implementation into internals.h, so
it's available in the user-mode-only compile as well.
Backports commit c766568d3604082c6fd45cbabe42c48e4861a13f from qemu
The SRS instruction is:
* UNDEFINED in Hyp mode
* UNPREDICTABLE in User or System mode
* UNPREDICTABLE if the specified mode isn't accessible
* trapped to EL3 if EL3 is AArch64 and we are at Secure EL1
Clean up the code to handle all these cases cleanly, including
picking UNDEF as our choice of UNPREDICTABLE behaviour rather
blindly trusting the mode field passed in the instruction.
As part of this, move the check for IS_USER into gen_srs()
itself rather than having it done by the caller.
The exception is that we don't UNDEF for calls from System
mode, which need a runtime check. This will be dealt with in
the following commits.
Backports commit cbc0326b6fb905f80b7cef85b24571f7ebb62077 from qemu
If access to FPEXC32_EL2 is trapped by CPTR_EL2.TFP or CPTR_EL3.TFP,
this should be reported with a syndrome register indicating an
FP access trap, not one indicating a system register access trap.
Backports commit f2cae6092767aaf418778eada15be444c23883be from qemu
Implement trapping of the "debug ROM" registers, which are controlled
by MDCR_EL2.TDRA for EL2 but by the more general MDCR_EL3.TDA for EL3.
Backports commit 91b0a23865558e2ce9c2e7042d404e8bf2e4b817 from qemu
Implement the traps to EL2 and EL3 controlled by the bits
MDCR_EL2.TDOSA MDCR_EL3.TDOSA. These can configurably trap
accesses to the "powerdown debug" registers.
Backports commit 187f678d5c28251dba2b44127e59966b14518ef7 from qemu
We weren't quite implementing the handling of SCR.SMD correctly.
The condition governing whether the SMD bit should apply only
for NS state is "is EL3 is AArch32", not "is the current EL AArch32".
Fix the condition, and clarify the comment both to reflect this and
to expand slightly on what's going on for the v7-no-Virtualization case.
Backports commit f096e92b6385fd87e8ea948ad3af70faf752c13a from qemu
Correct some corner cases we were getting wrong for
CNTFRQ access rights:
* should UNDEF from 32-bit Secure EL1
* only writable from the highest implemented exception level,
which might not be EL1 now
To clarify the code, provide a new utility function
arm_highest_el() which returns the highest implemented
exception level.
Backports commit 755026728abb19fba70e6b4396a27fa2e7550d74 from qemu
This is immediately usable by lea and multi-byte nop,
and will be required to implement parts of the mpx spec.
Backports commit a074ce42a3186bd9f96ef541bb2e01419181dae3 from qemu
This includes XSAVE, XRSTOR, XGETBV, XSETBV, which are all related,
as well as the associate cpuid bits.
Backports commit 19dc85dba23c0db1ca932c62e453c37e00761628 from qemu
Rather than nesting tests of OP, MOD, and RM, decode them all at once
with a switch. Also, add some missing #UD checks for e.g. incorrect
LOCK prefix.
Backports commit 121f3157887f92268a3d6169e2d4601f9292020b from qemu
Rather than nesting tests of OP, MOD, and RM, decode them
all at once with a switch. Fixes incorrect decoding of
AMD Pacifica extensions (aka vmrun et al) via op==2 path.
Backports commit 1906b2af7c2345037d9b2fdf484b457b5acd09d1 from qemu
The description of object_property_get_int() stated that on an error
it returns NULL. This is not the case and the function will return -1
if an error occurs. Update the commented documentation accordingly.
Backports commit b29b47e9b35017428904e0e934700877dfaabe73 from qemu
ARM stops before access to a location covered by watchpoint. Also, QEMU
watchpoint fire is not necessarily an architectural watchpoint match.
Unfortunately, that is hardly possible to ignore a fired watchpoint in
debug exception handler. So move watchpoint check from debug exception
handler to the dedicated watchpoint checking callback.
Backports commit 3826121d9298cde1d29ead05910e1f40125ee9b0 from qemu
When QEMU watchpoint matches, that is not definitely an architectural
watchpoint match yet. If it is a stop-before-access watchpoint then that
is hardly possible to ignore it after throwing a TCG exception.
A special callback is introduced to check for architectural watchpoint
match before raising a TCG exception.
Backports commit 568496c0c0f1863a4bc18539962cd8d81baa4e30 from qemu
All Thumb Neon and VFP instructions are 32 bits, so the IL
bit in the syndrome register should be set. Pass false to the
syn_* function's is_16bit argument rather than s->thumb
so we report the correct IL bit.
Backports commit 7d197d2db5e99e4c8b20f6771ddc7303acaa1c89 from qemu
All Thumb coprocessor instructions are 32 bits, so the IL
bit in the syndrome register should be set. Pass false to the
syn_* function's is_16bit argument rather than s->thumb
so we report the correct IL bit.
Backports commit 4df322593037d2700f72dfdfb967300b7ad2e696 from qemu
In syndrome register values, the IL bit indicates the instruction
length, and is 1 for 4-byte instructions and 0 for 2-byte
instructions. All A64 and A32 instructions are 4-byte, but
Thumb instructions may be either 2 or 4 bytes long. Unfortunately
we named the parameter to the syn_* functions for constructing
syndromes "is_thumb", which falsely implies that it should be
set for all Thumb instructions, rather than only the 16-bit ones.
Fix the functions to name the parameter 'is_16bit' instead.
Backports commit fc05f4a62c568b607ec3fe428a419bb38205b570 from qemu
Enable EL3 support for our Cortex-A53 and Cortex-A57 CPU models.
We have enough implemented now to be able to run real world code
at least to some extent (I can boot ARM Trusted Firmware to the
point where it pulls in OP-TEE and then falls over because it
doesn't have a UEFI image it can chain to).
Backports commit 3ad901bc2b98f5539af9a7d4aef140a6d8fa6442 from qemu
Implement some corner cases of the behaviour of the NSACR
register on ARMv8:
* if EL3 is AArch64 then accessing the NSACR from Secure EL1
with AArch32 should trap to EL3
* if EL3 is not present or is AArch64 then reads from NS EL1 and
NS EL2 return constant 0xc00
It would in theory be possible to implement all these with
a single reginfo definition, but for clarity we use three
separate definitions for the three cases and install the
right one based on the CPU feature flags.
Backports commit 2f027fc52d4b444a47cb05a9c96697372a6b57d2 from qemu
System registers might have access requirements which need to
be described via a CPAccessFn and which differ for reads and
writes. For this to be possible we need to pass the access
function a parameter to tell it whether the access being checked
is a read or a write.
Backports commit 3f208fd76bcc91a8506681bb8472f2398fe6f487 from qemu
The arm_generate_debug_exceptions() function as originally implemented
assumes no EL2 or EL3. Since we now have much more of an implementation
of those now, fix this assumption.
Backports commit 533e93f1cf12c570aab45f14663dab6fb8ea3ffc from qemu
The registers MVBAR and SCR should have the behaviour of trapping to
EL3 if accessed from Secure EL1, but we were incorrectly implementing
them to UNDEF (which would trap to EL1). Fix this by using the new
access_trap_aa32s_el1() access function.
Backports commit efe4a274083f61484a8f1478d93f229d43aa8095 from qemu
Implement the MDCR_EL3 register (which is SDCR for AArch32).
For the moment we implement it as reads-as-written.
Backports commit 5513c3abed8e5fabe116830c63f0d3fe1f94bd21 from qemu
(pde & 0x1fe000) is a 32-bit integer; when shifting it
into bits 39-32 the result is zero. Fix it by making the
mask (and thus the result of the AND) a 64-bit integer.
Reported by Coverity.
Backports commit 388ee48a88e684e719660a2cae9c21897b94fa37 from qemu
All references to cpu_T are done with a constant index. It aids
readability to decompose the array into two scalar variables.
Backports commit 1d1cc4d0f481b2939c7e9f6606e571b2fc81971a from qemu
Merge gen_op_addl_A0_im and gen_op_addq_A0_im into gen_add_A0_im
and clean up the ifdef.
Replace the one remaining user of gen_op_addl_A0_im with gen_add_A0_im.
Backports commit 4e85057b92d214decf10045d3d4faa2faf33d100 from qemu
Unify the code across stack pointer widths. Fix the note about
not updating ESP before the potential exception.
Backports commit 2045f04c3ae030bda650f84035f114bbd84909a9 from qemu
Use gen_lea_v_seg for centralized segment base knowledge. Unify
code across 32- and 64-bit. Fix note about "must save state"
before using the out-of-line helpers.
Backports commit 743e398e2fbf2f7183bf7a53c9d011fabcaa1770 from qemu
More centralization of handling of segment bases.
Also fixes the note about 16-bit wrap around not fully handled.
Backports commit d37ea0c04723f3e15fde55fe97cff6278159929b from qemu
Having segs[].base as a register significantly improves code
generation for real and protected modes, particularly for TBs
that have multiple memory references where the segment base
can be held in a hard register through the TB.
Backports commit 3558f8055f37a34762b7a2a0f02687e6eeab893d from qemu
I.e. gen_push_v, gen_pop_T0, gen_stack_A0.
More centralization of handling of segment bases.
Backports commit 77ebcad04f3659fa7eb799928fdd68280fac720d from qemu
Add forgotten zero-extension in the TARGET_X86_64, !CODE64, ss32 case;
use this new function to implement gen_string_movl_A0_EDI,
gen_string_movl_A0_ESI, gen_add_A0_ds_seg.
Backports commit ca2f29f555805d07fb0b9ebfbbfc4e3656530977 from qemu
This condition is true in the common case, so we can cut out the body of
the function. In addition, this makes it easier for the compiler to do
at least partial inlining, even if it decides that fully inlining the
function is unreasonable.
Backports commit 8bafcb21643a39a5b29109f8bd5ee5a6f0f6850b from qemu
Similar to error_abort, but doesn't report where the error was
created, and terminates the process with exit(1) rather than abort().
Backports commit a29a37b994ca3c5a1d39fa0e8934f7e0f2cf57ef from qemu
Don't claim error_report_err() always reports to stderr. It actually
reports to the current monitor when we have one.
Clarify intended use of error_abort and error_fatal.
Backports commit 10303f04b98efa76e638b9ae4632688f56f088fc from qemu
Commit 86f4b687 broke compilation on MIPS and SPARC, which have a
preprocessor pollution of '#define mips 1' and '#define sparc 1',
respectively. Treat it the same way as we do for the pollution with
'unix', so that QMP remains backwards compatible and only the C code
needs to use the alternative 'q_mips', 'q_sparc' spelling.
Backports commit 86ae191163d48ff4ff9d9996868e6cfe92a82449 from qemu
The previous commit documented an inconsistency in how we are
using the stack of qmp-output-visitor. Normally, pushing a
single top-level object puts the object on the stack twice:
once as the root, and once as the current container being
appended to; but popping that struct only pops once. However,
qmp_ouput_add() was trying to either set up the added object
as the new root (works if you parse two top-level scalars in a
row: the second replaces the first as the root) or as a member
of the current container (works as long as you have an open
container on the stack; but if you have popped the first
top-level container, it then resolves to the root and still
tries to add into that existing container).
Fix the stupidity by not tracking two separate things in the
stack. Drop the now-useless qmp_output_first() and
qmp_output_last() while at it.
Saved for a later patch: we still are rather sloppy in that
qmp_output_get_object() can be called in the middle of a parse,
rather than requiring that a visit is complete.
Backports commit 455ba08afde784466420449d01c6458f88349d55 from qemu
Commit 6c2f9a15 ensured that we would not return NULL when the
caller used an output visitor but had nothing to visit. But
in doing so, it added a FIXME about a reference count leak
that could abort qemu in the (unlikely) case of SIZE_MAX such
visits (more plausible on 32-bit). (Although that commit
suggested we might fix it in time for 2.5, we ran out of time;
fortunately, it is unlikely enough to bite that it was not
worth worrying about during the 2.5 release.)
This fixes things by documenting the internal contracts, and
explaining why the internal function can return NULL and only
the public facing interface needs to worry about qnull(),
thus avoiding over-referencing the qnull_ global object.
It does not, however, fix the stupidity of the stack mixing
up two separate pieces of information; add a FIXME to explain
that issue, which will be fixed shortly in a future patch.
Backports commit a86156401559cb4401cf9ecc704faeab6fc8bb19 from qemu
The only way that qmp_input_pop() will set errp is if a dictionary
was the most recent thing pushed. Since we don't have any
push(struct)/pop(list) or push(list)/pop(struct) mismatches (such
a mismatch is a programming bug), we therefore cannot set errp
inside qmp_input_end_list(). Make this obvious by
using &error_abort. A later patch will then remove the errp
parameter of qmp_input_pop(), but that will first require the
larger task of splitting visit_end_struct().
Backports commit bdd8e6b5d8a9def83d491a3f41c10424fc366258 from qemu
visit_start_struct() and visit_type_enum() had a 'kind' argument
that was usually set to either the stringized version of the
corresponding qapi type name, or to NULL (although some clients
didn't even get that right). But nothing ever used the argument.
It's even hard to argue that it would be useful in a debugger,
as a stack backtrace also tells which type is being visited.
Therefore, drop the 'kind' argument as dead.
Backports commit 337283dffbb5ad5860ed00408a5fd0665c21be07 from qemu
As explained in the previous patches, matching argument order of
'name, &value' to JSON's "name":value makes sense. However,
while the last two patches were easy with Coccinelle, I ended up
doing this one all by hand. Now all the visitor callbacks match
the main interface.
The compiler is able to enforce that all clients match the changed
interface in visitor-impl.h, even where two pointers are being
swapped, because only one of the two pointers is const (if that
were not the case, then C's looseness on treating 'char *' like
'void *' would have made review a bit harder).
Backports commit 0b2a0d6bb2446060944061e53e87d0c7addede79 from qemu
Commit 4e27e819 introduced optional visitor callbacks for all
sorts of int types, but no visitor has supplied any of the
callbacks for sizes less than 64 bits. In other words, the
generic implementation based on using type_[u]int64() followed
by bounds-checking works just fine. In the interest of
simplicity, it's easier to make the visitor callback interface
not have to worry about the other sizes.
Adding some helper functions minimizes the boilerplate required
to correct FIXMEs added earlier with regards to questionable
reuse of errp, particularly now that we can guarantee from a
single file audit that value is unchanged if an error is set.
Backports commit 04e070d217b4414f1f91aa8ad25fc0ae7ca0be93 from qemu
Similar to the previous patch, it's nice to have all functions
in the tree that involve a visitor and a name for conversion to
or from QAPI to consistently stick the 'name' parameter next
to the Visitor parameter.
Done by manually changing include/qom/object.h and qom/object.c,
then running this Coccinelle script and touching up the fallout
(Coccinelle insisted on adding some trailing whitespace).
@ rule1 @
identifier fn;
typedef Object, Visitor, Error;
identifier obj, v, opaque, name, errp;
@@
void fn
- (Object *obj, Visitor *v, void *opaque, const char *name,
+ (Object *obj, Visitor *v, const char *name, void *opaque,
Error **errp) { ... }
@@
identifier rule1.fn;
expression obj, v, opaque, name, errp;
@@
fn(obj, v,
- opaque, name,
+ name, opaque,
errp)
Backports commit d7bce9999df85c56c8cb1fcffd944d51bff8ff48 from qemu
JSON uses "name":value, but many of our visitor interfaces were
called with visit_type_FOO(v, &value, name, errp). This can be
a bit confusing to have to mentally swap the parameter order to
match JSON order. It's particularly bad for visit_start_struct(),
where the 'name' parameter is smack in the middle of the
otherwise-related group of 'obj, kind, size' parameters! It's
time to do a global swap of the parameter ordering, so that the
'name' parameter is always immediately after the Visitor argument.
Additional reason in favor of the swap: the existing include/qjson.h
prefers listing 'name' first in json_prop_*(), and I have plans to
unify that file with the qapi visitors; listing 'name' first in
qapi will minimize churn to the (admittedly few) qjson.h clients.
Later patches will then fix docs, object.h, visitor-impl.h, and
those clients to match.
Done by first patching scripts/qapi*.py by hand to make generated
files do what I want, then by running the following Coccinelle
script to affect the rest of the code base:
$ spatch --sp-file script `git grep -l '\bvisit_' -- '**/*.[ch]'`
I then had to apply some touchups (Coccinelle insisted on TAB
indentation in visitor.h, and botched the signature of
visit_type_enum() by rewriting 'const char *const strings[]' to
the syntactically invalid 'const char*const[] strings'). The
movement of parameters is sufficient to provoke compiler errors
if any callers were missed.
// Part 1: Swap declaration order
@@
type TV, TErr, TObj, T1, T2;
identifier OBJ, ARG1, ARG2;
@@
void visit_start_struct
-(TV v, TObj OBJ, T1 ARG1, const char *name, T2 ARG2, TErr errp)
+(TV v, const char *name, TObj OBJ, T1 ARG1, T2 ARG2, TErr errp)
{ ... }
@@
type bool, TV, T1;
identifier ARG1;
@@
bool visit_optional
-(TV v, T1 ARG1, const char *name)
+(TV v, const char *name, T1 ARG1)
{ ... }
@@
type TV, TErr, TObj, T1;
identifier OBJ, ARG1;
@@
void visit_get_next_type
-(TV v, TObj OBJ, T1 ARG1, const char *name, TErr errp)
+(TV v, const char *name, TObj OBJ, T1 ARG1, TErr errp)
{ ... }
@@
type TV, TErr, TObj, T1, T2;
identifier OBJ, ARG1, ARG2;
@@
void visit_type_enum
-(TV v, TObj OBJ, T1 ARG1, T2 ARG2, const char *name, TErr errp)
+(TV v, const char *name, TObj OBJ, T1 ARG1, T2 ARG2, TErr errp)
{ ... }
@@
type TV, TErr, TObj;
identifier OBJ;
identifier VISIT_TYPE =~ "^visit_type_";
@@
void VISIT_TYPE
-(TV v, TObj OBJ, const char *name, TErr errp)
+(TV v, const char *name, TObj OBJ, TErr errp)
{ ... }
// Part 2: swap caller order
@@
expression V, NAME, OBJ, ARG1, ARG2, ERR;
identifier VISIT_TYPE =~ "^visit_type_";
@@
(
-visit_start_struct(V, OBJ, ARG1, NAME, ARG2, ERR)
+visit_start_struct(V, NAME, OBJ, ARG1, ARG2, ERR)
|
-visit_optional(V, ARG1, NAME)
+visit_optional(V, NAME, ARG1)
|
-visit_get_next_type(V, OBJ, ARG1, NAME, ERR)
+visit_get_next_type(V, NAME, OBJ, ARG1, ERR)
|
-visit_type_enum(V, OBJ, ARG1, ARG2, NAME, ERR)
+visit_type_enum(V, NAME, OBJ, ARG1, ARG2, ERR)
|
-VISIT_TYPE(V, OBJ, NAME, ERR)
+VISIT_TYPE(V, NAME, OBJ, ERR)
)
Backports commit 51e72bc1dd6ace6e91d675f41a1f09bd00ab8043 from qemu
No need to repeat 'struct Visitor' when we already have it in
typedefs.h. Omitting the redundant 'struct' also makes a later
patch easier to search for all object property callbacks that
are associated with a Visitor.
Backports commit 4fa45492c3387c0fa51e8e81160ac9a7814f44a2 from qemu
C compilers are allowed to represent enums as a smaller type
than int, if all enum values fit in the smaller type. There
are even compiler flags that force the use of this smaller
representation, although using them changes the ABI of a
binary. Therefore, our generated code for visit_type_ENUM()
(for all qapi enums) was wrong for casting Enum* to int* when
calling visit_type_enum().
It appears that no one has been using compiler ABI switches
for qemu, because if they had, we are potentially dereferencing
beyond bounds or even risking a SIGBUS on platforms where
unaligned pointer dereferencing is fatal. But it is still
better to avoid the practice entirely, and just use the correct
types.
This matches the fix for alternate qapi types, done earlier in
commit 0426d53 "qapi: Simplify visiting of alternate types",
with generated code changing as:
| void visit_type_QType(Visitor *v, QType *obj, const char *name, Error **errp)
| {
|- visit_type_enum(v, (int *)obj, QType_lookup, "QType", name, errp);
|+ int value = *obj;
|+ visit_type_enum(v, &value, QType_lookup, "QType", name, errp);
|+ *obj = value;
| }
Backports commit 395a233f7c089f23e3c0d43ce34c709dc5acd7de from qemu
The generated code can call visit_end_union() without having called
visit_start_union(). Example:
if (!*obj) {
goto out_obj;
}
visit_type_CpuInfoBase_fields(v, (CpuInfoBase **)obj, &err);
if (err) {
goto out_obj; // if we go from here...
}
if (!visit_start_union(v, !!(*obj)->u.data, &err) || err) {
goto out_obj;
}
switch ((*obj)->arch) {
[...]
}
out_obj:
// ... then *obj is true, and ...
error_propagate(errp, err);
err = NULL;
if (*obj) {
// we end up here
visit_end_union(v, !!(*obj)->u.data, &err);
}
error_propagate(errp, err);
Harmless only because no visitor implements end_union(). Clean it up
anyway, by deleting the function as useless.
Messed up since we have visit_end_union (commit cee2ded).
Backports commit 7c91aabd8964cfdf637f302c579c95401f21ce92 from qemu
Inside the generated code between visit_start_struct() and
visit_end_struct(), we were blindly setting the error into
the caller's errp parameter. But a future patch to split
visit_end_struct() will require that we take action based
on whether an error has occurred, which requires us to track
all actions through a local err. Rewrite the visits to be
more in line with the other generated calls.
Generated code changes look like:
| visit_start_struct(v, (void **)obj, "Abort", name, sizeof(Abort), &err);
|- if (!err) {
|- if (*obj) {
|- visit_type_Abort_fields(v, obj, errp);
|- }
|- visit_end_struct(v, &err);
|+ if (err) {
|+ goto out;
| }
|+ if (!*obj) {
|+ goto out_obj;
|+ }
|+ visit_type_Abort_fields(v, obj, &err);
|+ error_propagate(errp, err);
|+ err = NULL;
|+out_obj:
|+ visit_end_struct(v, &err);
|+out:
| error_propagate(errp, err);
| }
Backports commit 92b09babc11b60458c28cfe37eaa314de50e6241 from qemu
All other successful clients of visit_start_struct() were paired
with an unconditional visit_end_struct(); but the generated
code for events was relying on qmp_output_visitor_cleanup() to
work on an incomplete visit. Alter the code to guarantee that
the struct is completed, which will make a future patch to
split visit_end_struct() easier to reason about. While at it,
drop some assertions and comments that are not present in other
uses of the qmp output visitor, and pass NULL rather than "" as
the 'kind' parameter (matching most other uses where obj is NULL).
The changes to the generated code look like:
| qmp = qmp_event_build_dict("DEVICE_TRAY_MOVED");
|
| qov = qmp_output_visitor_new();
|- g_assert(qov);
|-
| v = qmp_output_get_visitor(qov);
|- g_assert(v);
|
|- /* Fake visit, as if all members are under a structure */
|- visit_start_struct(v, NULL, "", "DEVICE_TRAY_MOVED", 0, &err);
|+ visit_start_struct(v, NULL, NULL, "DEVICE_TRAY_MOVED", 0, &err);
| if (err) {
| goto out;
| }
| visit_type_str(v, (char **)&device, "device", &err);
| if (err) {
|- goto out;
|+ goto out_obj;
| }
| visit_type_bool(v, &tray_open, "tray-open", &err);
| if (err) {
|- goto out;
|+ goto out_obj;
| }
|- visit_end_struct(v, &err);
|+out_obj:
|+ visit_end_struct(v, err ? NULL : &err);
| if (err) {
| goto out;
| }
|
| obj = qmp_output_get_qobject(qov);
|- g_assert(obj != NULL);
|+ g_assert(obj);
|
| qdict_put_obj(qmp, "data", obj);
| emit(QAPI_EVENT_DEVICE_TRAY_MOVED, qmp, &err);
Note that the 'goto out_obj' with no intervening code before the
label, as well as the construct of 'err ? NULL : &err', are both
a bit unusual but also temporary; they get fixed in a later patch
that splits visit_end_struct() to drop its errp parameter by moving
some checking before the label. But until that time, this was the
simplest way to avoid the appearance of passing a possibly-set
error to visit_end_struct(), even though actual code inspection
shows that visit_end_struct() for a QMP output visitor will never
set an error.
Backports commit a16e3e5c5825c90887a863513916f93eeec16c55 from qemu
Commit 5cdc8831 reworked gen_params() to be simpler, but forgot
to clean up a now-unused errp named argument.
No change to generated code.
Backports commit e408311546b633b232312450d9017f4db21df0dc from qemu
PEP 8 calls for it, because it's forward compatible with Python 3.
Supported since Python 2.6, which we require (commit fec2103).
Backports commit 291928a80f39670929fcce222acd5e21d1434692 from qemu
It should be fairly obvious that qapi base classes need to
form an acyclic graph, since QMP cannot specify the same
key more than once, while base classes are included as flat
members alongside other members added by the child. But the
old check_member_clash() parser function was not prepared to
check for this, and entered an infinite recursion (at least
until Python gives up, complaining about nesting too deep).
Now that check_member_clash() has been recently removed,
attempts at self-inheritance trigger an assertion failure
introduced by commit ac88219a. The obvious fix is to turn
the assertion into a conditional.
This patch includes both the tests (base-cycle-direct and
base-cycle-indirect) and the fix, since the .err file output
for the unfixed case is not useful (particularly when it was
warning about unbounded recursion, as that limit may be
platform-specific).
We don't need to worry about cycles in flat unions (neither
the base type nor the type of a variant can be a union) nor
in alternates (alternate branches cannot themselves be an
alternate). But if we later allow a union type as a variant,
we will still be okay, as QAPISchemaObjectTypeVariants.check()
triggers the same QAPISchemaObjectType.check() that will
detect any loops.
Likewise, we need not worry about the case of diamond
inheritance where the same class is used for a flat union base
class and one of its variants; either both uses will introduce
a collision in trying to insert the same member name twice, or
the shared type is empty and changes nothing.
Backports commit bac5429ccb4f41d421ec641b11f1852c8420fdb7 from qemu
With the recent commit 'qapi: Detect collisions in C member
names', we have two different locations for detecting clashes -
one at parse time, and another at QAPISchema*.check() time.
Remove all of the ad hoc parser checks, and delete associated
code (for example, the global check_member_clash() method is
no longer needed).
Testing this showed that the test union-bad-branch wasn't adding
much: union-clash-branches also exposes the error message when
branches collide, and we've recently fixed things to avoid an
implicit collision with max. Likewise, the error for
enum-clash-member changes to report our new detection of
upper case in a value name, unless we modify the test to use
all lower case.
The wording of several error messages has changed, but the
change is generally an improvement rather than a regression.
No change to generated code.
Backports commit 01cfbaa4c36ecd9f1c7bcbad50c92758e1d147c4 from qemu
We document that members of enums and objects should be
'lower-case', although we were not enforcing it. We have to
whitelist a few pre-existing entities that violate the norms.
Add three new tests to expose the new error message, each of
which first uses the whitelisted name 'UuidInfo' to prove the
whitelist works, then triggers the failure (this is the same
pattern used in the existing returns-whitelist.json test).
Note that by adding this check, we have effectively forbidden
an entity with a case-insensitive clash of member names, for
any entity that is not on the whitelist (although there is
still the possibility to clash via '-' vs. '_').
Not done here: a future patch should also add naming convention
support and whitelist exceptions for command, event, and type
names.
The additions to QAPISchemaMember.check_clash() check whether
info['name'] is in the whitelist (the top-most entity name at
the point 'info' tracks), rather than self.owner (the type,
possibly implicit, that directly owns the member), because it
is easier to maintain the whitelist by the names actually in
the user's .json file, rather than worrying about the names
of implicit types.
Backports commit 893e1f2c5170d54316f1dcf3fefae679175622fc from qemu
Rather than using just an array of strings, make enum.values be
an array of the new QAPISchemaMember type, and add a helper
member_names() method to get back at the original list of names.
Likewise, creating an enum requires wrapping strings, via a new
QAPISchema._make_enum_members() method. The benefit of wrapping
enum members in a QAPISchemaMember Python object is that we now
share the existing code for C name clash detection (although the
code is not yet active until a later commit removes the earlier
ad hoc parser checks).
In a related change, the QAPISchemaMember._pretty_owner() method
needs to learn about one more implicit type name: the generated
enum associated with a simple union.
In the interest of keeping the changes of this patch local to one
file, the visitor interface still passes just a list of names
rather than the full list of QAPISchemaMember instances. We may
want to revisit this in the future, if the consistency with
visit_object_type() is worth it.
Backports commit 93bda4dd461358b4fc05dfd8e2d6419cdd574789 from qemu
We want to share some clash detection code between enum values
and object type members. To assist with that, split off part
of QAPISchemaObjectTypeMember into a new base class
QAPISchemaMember that tracks name, owner, and common clash
detection code; while the former keeps the additional fields
for type and optional flag.
Backports commit d44f9ac80c43e34b1522cde8829f0ab371f086ca from qemu
For less code, reflect the determined boolean value of an optional
visit back to the caller instead of making the caller read the
boolean after the fact.
The resulting generated code has the following diff:
|- visit_optional(v, &has_fdset_id, "fdset-id");
|- if (has_fdset_id) {
|+ if (visit_optional(v, &has_fdset_id, "fdset-id")) {
| visit_type_int(v, &fdset_id, "fdset-id", &err);
| if (err) {
| goto out;
| }
| }
Backports commit 29637a6ee913df8fcdf371426ee48956b945b618 from qemu
None of the visitor callbacks would set an error when testing
if an optional field was present; make this part of the interface
contract by eliminating the errp argument.
The resulting generated code has a nice diff:
|- visit_optional(v, &has_fdset_id, "fdset-id", &err);
|- if (err) {
|- goto out;
|- }
|+ visit_optional(v, &has_fdset_id, "fdset-id");
| if (has_fdset_id) {
| visit_type_int(v, &fdset_id, "fdset-id", &err);
| if (err) {
| goto out;
| }
| }
Backports commit 5cdc8831a795fb8452d7e34f644202fd724e122a from qemu
The QMP input visitor allows integral values to be assigned by
promotion to a QTYPE_QFLOAT. However, when parsing an alternate,
we did not take this into account, such that an alternate that
accepts 'number' and some other type, but not 'int', would reject
integral values.
With this patch, we now have the following desirable table:
alternate has case selected for
'int' 'number' QTYPE_QINT QTYPE_QFLOAT
no no error error
no yes 'number' 'number'
yes no 'int' error
yes yes 'int' 'number'
While it is unlikely that we will ever use 'number' in an
alternate other than in the testsuite, it never hurts to be
more precise in what we allow.
Backports commit d00341af384665d259af475b14c96bb8414df415 from qemu
Now that alternates no longer use an implicit tag, we can
inline _make_implicit_tag() into its one caller,
_def_union_type().
No change to generated code.
Backports commit 9d3f3494c5b941774e2c3e7639332d53bbe6f8be from qemu
Previously, the generated code in qapi-types.c initialized all
enum lookup tables first, prior to any other definitions. But
there are no topological sorting requirements that mandate this
layout, so we can drop the QAPISchemaGenTypeVisitor._fwdefn
field and just generate all definitions in visitation order.
The generated code shows some churn due to reordering, but it
is still fairly straightforward to follow (all the deletions
occur in one hunk, and all the deleted lines are re-inserted
in the same order later in the same files, just spread across
multiple insertion points).
Backports commit 0b2e84ba774651656771ed697dee8825759dffa9 from qemu
Previously, working with alternates required two lookup arrays
and some indirection: for type Foo, we created Foo_qtypes[]
which maps each qtype to a value of the generated FooKind enum,
then look up that value in FooKind_lookup[] like we do for other
union types.
This has a couple of subtle bugs. First, the generator was
creating a call with a parameter '(int *) &(*obj)->type' where
type is an enum type; this is unsafe if the compiler chooses
to store the enum type in a different size than int, where
assigning through the wrong size pointer can corrupt data or
cause a SIGBUS.
Related bug, not not fixed in this patch: qapi-visit.py's
gen_visit_enum() generates a cast of its enum * argument to
int *. Marked FIXME.
Second, since the values of the FooKind enum start at zero, all
entries of the Foo_qtypes[] array that were not explicitly
initialized will map to the same branch of the union as the
first member of the alternate, rather than triggering a desired
failure in visit_get_next_type(). Fortunately, the bug seldom
bites; the very next thing the input visitor does is try to
parse the incoming JSON with the wrong parser, which normally
fails; the output visitor is not used with a C struct in that
state, and the dealloc visitor has nothing to clean up (so
there is no leak).
However, the second bug IS observable in one case: parsing an
integer causes unusual behavior in an alternate that contains
at least a 'number' member but no 'int' member, because the
'number' parser accepts QTYPE_QINT in addition to the expected
QTYPE_QFLOAT (that is, since 'int' is not a member, the type
QTYPE_QINT accidentally maps to FooKind 0; if this enum value
is the 'number' branch the integer parses successfully, but if
the 'number' branch is not first, some other branch tries to
parse the integer and rejects it). A later patch will worry
about fixing alternates to always parse all inputs that a
non-alternate 'number' would accept, for now this is still
marked FIXME in the updated test-qmp-input-visitor.c, to
merely point out that new undesired behavior of 'ans' matches
the existing undesired behavior of 'asn'.
This patch fixes the default-initialization bug by deleting the
indirection, and modifying get_next_type() to directly assign a
QTypeCode parameter. This in turn fixes the type-casting bug,
as we are no longer casting a pointer to enum to a questionable
size. There is no longer a need to generate an implicit FooKind
enum associated with the alternate type (since the QMP wire
format never uses the stringized counterparts of the C union
member names). Since the updated visit_get_next_type() does not
know which qtypes are expected, the generated visitor is
modified to generate an error statement if an unexpected type is
encountered.
Callers now have to know the QTYPE_* mapping when looking at the
discriminator; but so far, only the testsuite was even using the
C struct of an alternate types. I considered the possibility of
keeping the internal enum FooKind, but initialized differently
than most generated arrays, as in:
typedef enum FooKind {
FOO_KIND_A = QTYPE_QDICT,
FOO_KIND_B = QTYPE_QINT,
} FooKind;
to create nicer aliases for knowing when to use foo->a or foo->b
when inspecting foo->type; but it turned out to add too much
complexity, especially without a client.
There is a user-visible side effect to this change, but I
consider it to be an improvement. Previously,
the invalid QMP command:
{"execute":"blockdev-add", "arguments":{"options":
{"driver":"raw", "id":"a", "file":true}}}
failed with:
{"error": {"class": "GenericError",
"desc": "Invalid parameter type for 'file', expected: QDict"}}
(visit_get_next_type() succeeded, and the error comes from the
visit_type_BlockdevOptions() expecting {}; there is no mention of
the fact that a string would also work). Now it fails with:
{"error": {"class": "GenericError",
"desc": "Invalid parameter type for 'file', expected: BlockdevRef"}}
(the error when the next type doesn't match any expected types for
the overall alternate).
Backports commit 0426d53c6530606bf7641b83f2b755fe61c280ee from qemu
What's more meta than using qapi to define qapi? :)
Convert QType into a full-fledged[*] builtin qapi enum type, so
that a subsequent patch can then use it as the discriminator
type of qapi alternate types. Fortunately, the judicious use of
'prefix' in the qapi definition avoids churn to the spelling of
the enum constants.
To avoid circular definitions, we have to flip the order of
inclusion between "qobject.h" vs. "qapi-types.h". Back in commit
28770e0, we had the latter include the former, so that we could
use 'QObject *' for our implementation of 'any'. But that usage
also works with only a forward declaration, whereas the
definition of QObject requires QType to be a complete type.
[*] The type has to be builtin, rather than declared in
qapi/common.json, because we want to use it for alternates even
when common.json is not included. But since it is the first
builtin enum type, we have to add special cases to qapi-types
and qapi-visit to only emit definitions once, even when two
qapi files are being compiled into the same binary (the way we
already handled builtin list types like 'intList'). We may
need to revisit how multiple qapi files share common types,
but that's a project for another day.
Backports commit 7264f5c50cc1be0f1406e3ebb45aedcca02f603a from qemu
The name QType matches our CODING_STYLE conventions for type names
in CamelCase. It also matches the fact that we are already naming
all the enum members with a prefix of QTYPE, not QTYPE_CODE. And
doing the rename will also make it easier for the next patch to use
QAPI for providing the enum, which also wants CamelCase type names.
Backports commit 1310a3d3bd9301ff5a825287638cfab24c2c6689 from qemu
The QObject hierarchy is small enough, and unlikely to grow further
(since we only use it to map to JSON and already cover all JSON
types), that we can simplify things by not tracking a separate
vtable, but just inline the code element of the vtable QType
directly into QObject (renamed to type), and track a separate array
of destroy functions. We can drop qnull_destroy_obj() in the
process.
The remaining QObject subclasses must export their destructor.
This also has the nice benefit of moving the typename 'QType'
out of the way, so that the next patch can repurpose it for a
nicer name for 'qtype_code'.
The various objects are still the same size (so no change in cache
line pressure), but now have less indirection (although I didn't
bother benchmarking to see if there is a noticeable speedup, as
we don't have hard evidence that this was in a performance hotspot
in the first place).
A future patch could drop the refcnt size to 32 bits for a smaller
struct on 64-bit architectures, if desired (we have limits on the
largest JSON that we are willing to parse, and will probably never
need to take full advantage of a 64-bit refcnt).
Backports commit 55e1819c509b3d9c10a54678b9c585bbda13889e from qemu
I'm going to fix the JSON parser to recognize null. The obvious
representation of JSON null as (QObject *)NULL doesn't work, because
the parser already uses it as an error value. Perhaps we should
change it to free NULL for null, but that's more than I can do right
now. Create a special null QObject instead.
The existing QDict, QList, and QString all represent something that
is a pointer in C and could therefore be associated with NULL. But
right now, all three of these sub-types are always non-null once
created, so the new null sentinel object is intentionally unrelated
to them.
Backports commit 481b002cc81ed7fc7b06e32e9d4d495d81739d14 from qemu
When munging enum values, the fact that we were passing the entire
prefix + value through camel_to_upper() meant that enum values
spelled with CamelCase could be turned into CAMEL_CASE. However,
this provides a potential collision (both OneTwo and One-Two would
munge into ONE_TWO) for enum types, when the same two names are
valid side-by-side as QAPI member names. By changing the generation
of enum constants to always be prefix + '_' + c_name(value,
False).upper(), and ensuring that there are no case collisions (in
the next patches), we no longer have to worry about names that
would be distinct as QAPI members but collide as variant tag names,
without having to think about what munging the heuristics in
camel_to_upper() will actually perform on an enum value.
Making the change will affect enums that did not follow coding
conventions, using 'CamelCase' rather than desired 'lower-case'.
Thankfully, there are only two culprits: InputButton and ErrorClass.
We already tweaked ErrorClass to make it an alias of QapiErrorClass,
where only the alias needs changing rather than the whole tree. So
the bulk of this change is modifying INPUT_BUTTON_WHEEL_UP to the
new INPUT_BUTTON_WHEELUP (and likewise for WHEELDOWN). That part
of this commit may later need reverting if we rename the enum
constants from 'WheelUp' to 'wheel-up' as part of moving
x-input-send-event to a stable interface; but at least we have
documentation bread crumbs in place to remind us (commit 513e7cd),
and it matches the fact that SDL constants are also spelled
SDL_BUTTON_WHEELUP.
Backports commit d20a580bc0eac9d489884f6d2ed28105880532b6 from qemu
The qapi enum ErrorClass is unusual that it uses 'CamelCase' names,
contrary to our documented convention of preferring 'lower-case'.
However, this enum is entrenched in the API; we cannot change
what strings QMP outputs. Meanwhile, we want to simplify how
c_enum_const() is used to generate enum constants, by moving away
from the heuristics of camel_to_upper() to a more straightforward
c_name(N).upper() - but doing so will rename all of the ErrorClass
constants and cause churn to all client files, where the new names
are aesthetically less pleasing (ERROR_CLASS_DEVICENOTFOUND looks
like we can't make up our minds on whether to break between words).
So as always in computer science, solve the problem by some more
indirection: rename the qapi type to QapiErrorClass, and add a
new enum ErrorClass in error.h whose members are aliases of the
qapi type, but with the spelling expected elsewhere in the tree.
Then, when c_enum_const() changes the munging, we only have to
adjust the one alias spot.
Backports commit f22a28b898322c01b0463a8b7ec551d72bc61a5b from qemu
Now that we no longer collide with an implicit _MAX enum member,
we no longer need to reject it in the ad hoc parser, and can
remove several tests that are no longer needed.
Backports commit 04e0639d4e77b6d55491d396c8aa13929ee8ed9a from qemu
Now that we guarantee the user doesn't have any enum values
beginning with a single underscore, we can use that for our
own purposes. Renaming ENUM_MAX to ENUM__MAX makes it obvious
that the sentinel is generated.
This patch was mostly generated by applying a temporary patch:
|diff --git a/scripts/qapi.py b/scripts/qapi.py
|index e6d014b..b862ec9 100644
|--- a/scripts/qapi.py
|+++ b/scripts/qapi.py
|@@ -1570,6 +1570,7 @@ const char *const %(c_name)s_lookup[] = {
| max_index = c_enum_const(name, 'MAX', prefix)
| ret += mcgen('''
| [%(max_index)s] = NULL,
|+// %(max_index)s
| };
| ''',
| max_index=max_index)
then running:
$ cat qapi-{types,event}.c tests/test-qapi-types.c |
sed -n 's,^// \(.*\)MAX,s|\1MAX|\1_MAX|g,p' > list
$ git grep -l _MAX | xargs sed -i -f list
The only things not generated are the changes in scripts/qapi.py.
Rejecting enum members named 'MAX' is now useless, and will be dropped
in the next patch.
Backports commit 7fb1cf1606c78c9d5b538f29176fd5a101726a9d from qemu
We already documented that qapi names should match specific
patterns (such as starting with a letter unless it was an enum
value or a downstream extension). Tighten that from a suggestion
into a hard requirement, which frees up names beginning with a
single underscore for qapi internal usage.
The tighter regex doesn't forbid everything insane that a user
could provide (for example, a user could name a type 'Foo-lookup'
to collide with the generated 'Foo_lookup[]' for an enum 'Foo'),
but does a good job at protecting the most obvious uses, and
also happens to reserve single leading underscore for later use.
The handling of enum values starting with a digit is tricky:
commit 9fb081e introduced a subtle bug by using c_name() on
a munged value, which would allow an enum to include the
member 'q-int' in spite of our reservation. Furthermore,
munging with a leading '_' would fail our tighter regex. So
fix it by only munging for leading digits (which are never
ticklish in c_name()) and by using a different prefix (I
picked 'D', although any letter should do).
Add new tests, reserved-member-underscore and reserved-enum-q,
to demonstrate the tighter checking.
Backports commit 59a92feedc6927e0e1ff87fdaccfb4dd42ad4c84 from qemu
Commit cbc95538 removed unused start_handle() and end_handle(),
but forgot to remove their declarations.
Backports commit 7549457200ec3871ee827765f4d3bbc8d903b2ec from qemu
The method c_name() is supposed to do two different actions: munge
'-' into '_', and add a 'q_' prefix to ticklish names. But it did
these steps out of order, making it possible to submit input that
is not ticklish until after munging, where the output then lacked
the desired prefix.
The failure is exposed easily if you have a compiler that recognizes
C11 keywords, and try to name a member '_Thread-local', as it would
result in trying to compile the declaration 'uint64_t _Thread_local;'
which is not valid. However, this name violates our conventions
(ultimately, want to enforce that no qapi names start with single
underscore), so the test is slightly weaker by instead testing
'wchar-t'; the declaration 'uint64_t wchar_t;' is valid in C (where
wchar_t is only a typedef) but would fail with a C++ compiler (where
it is a keyword).
Fix things by reversing the order of actions within c_name().
Backports commit c43567c12042cf401b039bfc94a5f85e1cc1e796 from qemu
Detect attempts to declare two object members that would result
in the same C member name, by keying the 'seen' dictionary off
of the C name rather than the qapi name. It also requires passing
info through the check_clash() methods.
This addresses a TODO and fixes the previously-broken
args-name-clash test. The resulting error message demonstrates
the utility of the .describe() method added previously. No change
to generated code.
Backports commit 27b60ab93bd1d5d8c85f009aac7a97ffd2c53c86 from qemu
