This enables parallel TCG code generation. However, we do not take
advantage of it yet since tb_lock is still held during tb_gen_code.
In user-mode we use a single TCG context; see the documentation
added to tcg_region_init for the rationale.
Note that targets do not need any conversion: targets initialize a
TCGContext (e.g. defining TCG globals), and after this initialization
has finished, the context is cloned by the vCPU threads, each of
them keeping a separate copy.
TCG threads claim one entry in tcg_ctxs[] by atomically increasing
n_tcg_ctxs. Do not be too annoyed by the subsequent atomic_read's
of that variable and tcg_ctxs; they are there just to play nice with
analysis tools such as thread sanitizer.
Note that we do not allocate an array of contexts (we allocate
an array of pointers instead) because when tcg_context_init
is called, we do not know yet how many contexts we'll use since
the bool behind qemu_tcg_mttcg_enabled() isn't set yet.
Previous patches folded some TCG globals into TCGContext. The non-const
globals remaining are only set at init time, i.e. before the TCG
threads are spawned. Here is a list of these set-at-init-time globals
under tcg/:
Only written by tcg_context_init:
- indirect_reg_alloc_order
- tcg_op_defs
Only written by tcg_target_init (called from tcg_context_init):
- tcg_target_available_regs
- tcg_target_call_clobber_regs
- arm: arm_arch, use_idiv_instructions
- i386: have_cmov, have_bmi1, have_bmi2, have_lzcnt,
have_movbe, have_popcnt
- mips: use_movnz_instructions, use_mips32_instructions,
use_mips32r2_instructions, got_sigill (tcg_target_detect_isa)
- ppc: have_isa_2_06, have_isa_3_00, tb_ret_addr
- s390: tb_ret_addr, s390_facilities
- sparc: qemu_ld_trampoline, qemu_st_trampoline (build_trampolines),
use_vis3_instructions
Only written by tcg_prologue_init:
- 'struct jit_code_entry one_entry'
- aarch64: tb_ret_addr
- arm: tb_ret_addr
- i386: tb_ret_addr, guest_base_flags
- ia64: tb_ret_addr
- mips: tb_ret_addr, bswap32_addr, bswap32u_addr, bswap64_addr
Backports commit 3468b59e18b179bc63c7ce934de912dfa9596122 from qemu
This is groundwork for supporting multiple TCG contexts.
The naive solution here is to split code_gen_buffer statically
among the TCG threads; this however results in poor utilization
if translation needs are different across TCG threads.
What we do here is to add an extra layer of indirection, assigning
regions that act just like pages do in virtual memory allocation.
(BTW if you are wondering about the chosen naming, I did not want
to use blocks or pages because those are already heavily used in QEMU).
We use a global lock to serialize allocations as well as statistics
reporting (we now export the size of the used code_gen_buffer with
tcg_code_size()). Note that for the allocator we could just use
a counter and atomic_inc; however, that would complicate the gathering
of tcg_code_size()-like stats. So given that the region operations are
not a fast path, a lock seems the most reasonable choice.
The effectiveness of this approach is clear after seeing some numbers.
I used the bootup+shutdown of debian-arm with '-tb-size 80' as a benchmark.
Note that I'm evaluating this after enabling per-thread TCG (which
is done by a subsequent commit).
* -smp 1, 1 region (entire buffer):
qemu: flush code_size=83885014 nb_tbs=154739 avg_tb_size=357
qemu: flush code_size=83884902 nb_tbs=153136 avg_tb_size=363
qemu: flush code_size=83885014 nb_tbs=152777 avg_tb_size=364
qemu: flush code_size=83884950 nb_tbs=150057 avg_tb_size=373
qemu: flush code_size=83884998 nb_tbs=150234 avg_tb_size=373
qemu: flush code_size=83885014 nb_tbs=154009 avg_tb_size=360
qemu: flush code_size=83885014 nb_tbs=151007 avg_tb_size=370
qemu: flush code_size=83885014 nb_tbs=151816 avg_tb_size=367
That is, 8 flushes.
* -smp 8, 32 regions (80/32 MB per region) [i.e. this patch]:
qemu: flush code_size=76328008 nb_tbs=141040 avg_tb_size=356
qemu: flush code_size=75366534 nb_tbs=138000 avg_tb_size=361
qemu: flush code_size=76864546 nb_tbs=140653 avg_tb_size=361
qemu: flush code_size=76309084 nb_tbs=135945 avg_tb_size=375
qemu: flush code_size=74581856 nb_tbs=132909 avg_tb_size=375
qemu: flush code_size=73927256 nb_tbs=135616 avg_tb_size=360
qemu: flush code_size=78629426 nb_tbs=142896 avg_tb_size=365
qemu: flush code_size=76667052 nb_tbs=138508 avg_tb_size=368
Again, 8 flushes. Note how buffer utilization is not 100%, but it
is close. Smaller region sizes would yield higher utilization,
but we want region allocation to be rare (it acquires a lock), so
we do not want to go too small.
* -smp 8, static partitioning of 8 regions (10 MB per region):
qemu: flush code_size=21936504 nb_tbs=40570 avg_tb_size=354
qemu: flush code_size=11472174 nb_tbs=20633 avg_tb_size=370
qemu: flush code_size=11603976 nb_tbs=21059 avg_tb_size=365
qemu: flush code_size=23254872 nb_tbs=41243 avg_tb_size=377
qemu: flush code_size=28289496 nb_tbs=52057 avg_tb_size=358
qemu: flush code_size=43605160 nb_tbs=78896 avg_tb_size=367
qemu: flush code_size=45166552 nb_tbs=82158 avg_tb_size=364
qemu: flush code_size=63289640 nb_tbs=116494 avg_tb_size=358
qemu: flush code_size=51389960 nb_tbs=93937 avg_tb_size=362
qemu: flush code_size=59665928 nb_tbs=107063 avg_tb_size=372
qemu: flush code_size=38380824 nb_tbs=68597 avg_tb_size=374
qemu: flush code_size=44884568 nb_tbs=79901 avg_tb_size=376
qemu: flush code_size=50782632 nb_tbs=90681 avg_tb_size=374
qemu: flush code_size=39848888 nb_tbs=71433 avg_tb_size=372
qemu: flush code_size=64708840 nb_tbs=119052 avg_tb_size=359
qemu: flush code_size=49830008 nb_tbs=90992 avg_tb_size=362
qemu: flush code_size=68372408 nb_tbs=123442 avg_tb_size=368
qemu: flush code_size=33555560 nb_tbs=59514 avg_tb_size=378
qemu: flush code_size=44748344 nb_tbs=80974 avg_tb_size=367
qemu: flush code_size=37104248 nb_tbs=67609 avg_tb_size=364
That is, 20 flushes. Note how a static partitioning approach uses
the code buffer poorly, leading to many unnecessary flushes.
Backports commit e8feb96fcc6c16eab8923332e86ff4ef0e2ac276 from qemu
This cleanup makes the number of objects depending on qapi/error.h
drop from 1910 (out of 4743) to 1612 in my "build everything" tree.
While there, separate #include from file comment with a blank line,
and drop a useless comment on why qemu/osdep.h is included first.
Backports commit e688df6bc4549f28534cdb001f168b8caae55b0c from qemu
Normally we create an address space for that CPU and pass that address
space into the function. Let's just do it inside to unify address space
creations. It'll simplify my next patch to rename those address spaces.
Backports commit 80ceb07a83375e3a0091591f96bd47bce2f640ce from qemu
The patch enables handling atomic code in the guest. This should be
preferably done in cpu_handle_exception(), but the current assumptions
regarding when we can execute atomic sections cause a deadlock.
The current mechanism discards the flags which were set in atomic
execution. We ensure they are properly saved by calling the
cc->cpu_exec_enter/leave() functions around the loop.
As we are running cpu_exec_step_atomic() from the outermost loop we
need to avoid an abort() when single stepping over atomic code since
debug exception longjmp will point to the the setlongjmp in
cpu_exec(). We do this by setting a new jmp_env so that it jumps back
here on an exception.
Backports relevant parts of commit 08e73c48b053566bfe0c994f154f73991cd0ff0e from qemu
There are now only two uses of the global exit_request left.
The first ensures we exit the run_loop when we first start to process
pending work and in the kick handler. This is just as easily done by
setting the first_cpu->exit_request flag.
The second use is in the round robin kick routine. The global
exit_request ensured every vCPU would set its local exit_request and
cause a full exit of the loop. Now the iothread isn't being held while
running we can just rely on the kick handler to push us out as intended.
We lightly re-factor the main vCPU thread to ensure cpu->exit_requests
cause us to exit the main loop and process any IO requests that might
come along. As an cpu->exit_request may legitimately get squashed
while processing the EXCP_INTERRUPT exception we also check
cpu->queued_work_first to ensure queued work is expedited as soon as
possible.
Backports commit e5143e30fb87fbf179029387f83f98a5a9b27f19 from qemu
We know there will be cases where MTTCG won't work until additional work
is done in the front/back ends to support. It will however be useful to
be able to turn it on.
As a result MTTCG will default to off unless the combination is
supported. However the user can turn it on for the sake of testing.
Backports commit 8d4e9146b3568022ea5730d92841345d41275d66 from qemu
When we cannot emulate an atomic operation within a parallel
context, this exception allows us to stop the world and try
again in a serial context.
Backports commit fdbc2b5722f6092e47181a947c90fd4bdcc1c121 from qemu
Also backports parts of commit 02d57ea115b7669f588371c86484a2e8ebc369be
CPUState is a fairly common pointer to pass to these helpers. This means
if you need other arguments for the async_run_on_cpu case you end up
having to do a g_malloc to stuff additional data into the routine. For
the current users this isn't a massive deal but for MTTCG this gets
cumbersome when the only other parameter is often an address.
This adds the typedef run_on_cpu_func for helper functions which has an
explicit CPUState * passed as the first parameter. All the users of
run_on_cpu and async_run_on_cpu have had their helpers updated to use
CPUState where available.
Backports commit e0eeb4a21a3ca4b296220ce4449d8acef9de9049 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
Clean up includes so that osdep.h is included first and headers
which it implies are not included manually.
This commit was created with scripts/clean-includes.
Backports commit 7b31bbc2e68605ab2f10dc609dd54cf4c7b5f49a from qemu
Add a MemoryRegion property, which if set is used to construct
the CPU's initial (default) AddressSpace.
Backports commit 6731d864f80938e404dc3e5eb7f6b76b891e3e43 from qemu
Allow multiple calls to cpu_address_space_init(); each
call adds an entry to the cpu->ases array at the specified
index. It is up to the target-specific CPU code to actually use
these extra address spaces.
Since this multiple AddressSpace support won't work with
KVM, add an assertion to avoid confusing failures.
Backports commit 12ebc9a76dd7702aef0a3618717a826c19c34ef4 from qemu
Rather than setting cpu->as unconditionally in cpu_exec_init
(and then having target-i386 override this later), don't set
it until the first call to cpu_address_space_init.
This requires us to initialise the address space for
both TCG and KVM (KVM doesn't need the AS listener but
it does require cpu->as to be set).
For target CPUs which don't set up any address spaces (currently
everything except i386), add the default address_space_memory
in qemu_init_vcpu().
Backports commit 56943e8cc14b7eeeab67d1942fa5d8bcafe3e53f from qemu
Remove un-needed usages of ENV_GET_CPU() by converting the APIs to use
CPUState pointers and retrieving the env_ptr as minimally needed.
Scripted conversion for target-* change:
for I in target-*/cpu.h; do
sed -i \
's/\(^int cpu_[^_]*_exec(\)[^ ][^ ]* \*s);$/\1CPUState *cpu);/' \
$I;
done
Backports commit ea3e9847408131abc840240bd61e892d28459452 from qemu
The sole caller of this function navigates the cpu->env_ptr only for
this function to take it back the cpu pointer straight away. Pass in
cpu pointer instead and grab the env pointer locally in the function.
Removes a core code usage of ENV_GET_CPU().
Backports commit 3d57f7893c90d911d786cb2c622b0926fc808b57 from qemu
ARM and probably the rest of the arches have significant memory leaks as
they have no release interface.
Additionally, DrMemory does not have 64-bit support and thus I can't
test the 64-bit version under Windows. Under Linux valgrind supports
both 32-bit and 64-bit but there are different macros and code for Linux
and Windows.
