MSVC is confused by initializer list and default parameter, and reports
an ambiguous call.
test/cctest/test-assembler-arm64.cc(12208): error C2668: 'v8::internal::Clobber': ambiguous call to overloaded function
test-utils-arm64.h(251): note: could be 'void v8::internal::Clobber(v8::internal::MacroAssembler *,v8::internal::CPURegList)'
test-utils-arm64.h(241): note: or 'void v8::internal::Clobber(v8::internal::MacroAssembler *,v8::internal::RegList,const uint64_t)'
Solution is to construct with explicit type.
Bug: v8:13312
Change-Id: I66f5ba48bcdf6eb30035beaf7214a3d26fc9f18b
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3913034
Reviewed-by: Igor Sheludko <ishell@chromium.org>
Commit-Queue: Jakob Kummerow <jkummerow@chromium.org>
Reviewed-by: Jakob Kummerow <jkummerow@chromium.org>
Cr-Commit-Position: refs/heads/main@{#83406}
clang/clang-cl compiled happily (probably included transitively this
header), but not MSVC.
Bug: v8:13312
Change-Id: I69b6c15f76d8ef13e4fac33f733717429ba96f71
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3913033
Reviewed-by: Jakob Kummerow <jkummerow@chromium.org>
Commit-Queue: Jakob Kummerow <jkummerow@chromium.org>
Cr-Commit-Position: refs/heads/main@{#83402}
Now that we require C++17 support, we can just use the standard
static_assert without message, instead of our STATIC_ASSERT macro.
R=leszeks@chromium.org
Bug: v8:12425
Change-Id: I1d4e39c310b533bcd3a4af33d027827e6c083afe
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3647353
Reviewed-by: Leszek Swirski <leszeks@chromium.org>
Reviewed-by: Hannes Payer <hpayer@chromium.org>
Commit-Queue: Clemens Backes <clemensb@chromium.org>
Cr-Commit-Position: refs/heads/main@{#80524}
This prevents ambiguity errors in C++20 due to ADL when casting types in
std::, which gains std::bit_cast<>().
Bug: chromium:1284275
Change-Id: I25046d1952a9304852e481ad8b84049c6769c289
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3625838
Auto-Submit: Peter Kasting <pkasting@chromium.org>
Reviewed-by: Adam Klein <adamk@chromium.org>
Reviewed-by: Michael Lippautz <mlippautz@chromium.org>
Commit-Queue: Adam Klein <adamk@chromium.org>
Cr-Commit-Position: refs/heads/main@{#80378}
This is a reland of commit 9d31f8663a
There were issues with --future flag implications on M1.
Original change's description:
> [rwx][mac] Support fast W^X permission switching on Apple Silicon (M1)
>
> ... for V8 code space. The feature is currently disabled.
>
> In order to use fast W^X permission switching we must allocate
> executable pages with readable writable executable permissions (RWX).
> However, MacOS on ARM64 ("Apple M1"/Apple Silicon) prohibits further
> permission changing of RWX memory pages. This means that the code page
> headers must be allocated with RWX permissions too because otherwise
> it wouldn't be possible to allocate a large code page over the freed
> regular code page and vice versa.
>
> When enabled, the new machinery works as follows:
>
> 1) when memory region is reserved for allocating executable pages, the
> whole region is committed with RWX permissions and then decommitted,
> 2) since reconfiguration of RWX page permissions is not allowed on
> MacOS on ARM64 ("Apple M1"/Apple Silicon), there must be no attempts
> to change them,
> 3) the request to set RWX permissions in the executable page region
> just recommits the pages without changing permissions (see (1), they
> were already allocated as RWX and then discarded),
> 4) in order to make executable pages inaccessible one must use
> OS::DiscardSystemPages() instead of OS::DecommitPages() or
> setting permissions to kNoAccess because the latter two are not
> allowed by the MacOS (see (2)).
> 5) since code space page headers are allocated as RWX pages it's also
> necessary to switch between W^X modes when updating the data in the
> page headers (i.e. when marking, updating stats, wiring pages in
> lists, etc.). The new CodePageHeaderModificationScope class is used
> in the respective places. On unrelated configurations it's a no-op.
>
> The fast permission switching can't be used for V8 configuration with
> enabled pointer compression and disabled external code space because
> a) the pointer compression cage has to be reserved with MAP_JIT flag
> which is too expensive,
> b) in case of shared pointer compression cage if the code range will
> be deleted while the cage is still alive then attempt to configure
> permissions of pages that were previously set to RWX will fail.
>
> This also CL extends the unmapper unit tests with permissions tracking
> for discarded pages.
>
> Bug: v8:12797
> Change-Id: Idb28cbc481306477589eee9962d2e75167d87c61
> Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3579303
> Reviewed-by: Nico Hartmann <nicohartmann@chromium.org>
> Reviewed-by: Clemens Backes <clemensb@chromium.org>
> Reviewed-by: Michael Lippautz <mlippautz@chromium.org>
> Commit-Queue: Igor Sheludko <ishell@chromium.org>
> Cr-Commit-Position: refs/heads/main@{#80238}
Bug: v8:12797
Change-Id: I0fe86666f31bad37d7074e217555c95900d2afba
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3610433
Reviewed-by: Nico Hartmann <nicohartmann@chromium.org>
Reviewed-by: Michael Lippautz <mlippautz@chromium.org>
Commit-Queue: Igor Sheludko <ishell@chromium.org>
Reviewed-by: Clemens Backes <clemensb@chromium.org>
Cr-Commit-Position: refs/heads/main@{#80259}
This reverts commit 9d31f8663a.
Reason for revert: crashes on Mac/arm64 bots:
https://ci.chromium.org/ui/p/v8/builders/ci/V8%20Mac%20-%20arm64%20-%20debug/5923/overview
Original change's description:
> [rwx][mac] Support fast W^X permission switching on Apple Silicon (M1)
>
> ... for V8 code space. The feature is currently disabled.
>
> In order to use fast W^X permission switching we must allocate
> executable pages with readable writable executable permissions (RWX).
> However, MacOS on ARM64 ("Apple M1"/Apple Silicon) prohibits further
> permission changing of RWX memory pages. This means that the code page
> headers must be allocated with RWX permissions too because otherwise
> it wouldn't be possible to allocate a large code page over the freed
> regular code page and vice versa.
>
> When enabled, the new machinery works as follows:
>
> 1) when memory region is reserved for allocating executable pages, the
> whole region is committed with RWX permissions and then decommitted,
> 2) since reconfiguration of RWX page permissions is not allowed on
> MacOS on ARM64 ("Apple M1"/Apple Silicon), there must be no attempts
> to change them,
> 3) the request to set RWX permissions in the executable page region
> just recommits the pages without changing permissions (see (1), they
> were already allocated as RWX and then discarded),
> 4) in order to make executable pages inaccessible one must use
> OS::DiscardSystemPages() instead of OS::DecommitPages() or
> setting permissions to kNoAccess because the latter two are not
> allowed by the MacOS (see (2)).
> 5) since code space page headers are allocated as RWX pages it's also
> necessary to switch between W^X modes when updating the data in the
> page headers (i.e. when marking, updating stats, wiring pages in
> lists, etc.). The new CodePageHeaderModificationScope class is used
> in the respective places. On unrelated configurations it's a no-op.
>
> The fast permission switching can't be used for V8 configuration with
> enabled pointer compression and disabled external code space because
> a) the pointer compression cage has to be reserved with MAP_JIT flag
> which is too expensive,
> b) in case of shared pointer compression cage if the code range will
> be deleted while the cage is still alive then attempt to configure
> permissions of pages that were previously set to RWX will fail.
>
> This also CL extends the unmapper unit tests with permissions tracking
> for discarded pages.
>
> Bug: v8:12797
> Change-Id: Idb28cbc481306477589eee9962d2e75167d87c61
> Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3579303
> Reviewed-by: Nico Hartmann <nicohartmann@chromium.org>
> Reviewed-by: Clemens Backes <clemensb@chromium.org>
> Reviewed-by: Michael Lippautz <mlippautz@chromium.org>
> Commit-Queue: Igor Sheludko <ishell@chromium.org>
> Cr-Commit-Position: refs/heads/main@{#80238}
Bug: v8:12797
Change-Id: Ic07948e036db36326d464a2a901d052aa060a406
No-Presubmit: true
No-Tree-Checks: true
No-Try: true
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3611665
Commit-Queue: Rubber Stamper <rubber-stamper@appspot.gserviceaccount.com>
Bot-Commit: Rubber Stamper <rubber-stamper@appspot.gserviceaccount.com>
Auto-Submit: Adam Klein <adamk@chromium.org>
Cr-Commit-Position: refs/heads/main@{#80239}
... for V8 code space. The feature is currently disabled.
In order to use fast W^X permission switching we must allocate
executable pages with readable writable executable permissions (RWX).
However, MacOS on ARM64 ("Apple M1"/Apple Silicon) prohibits further
permission changing of RWX memory pages. This means that the code page
headers must be allocated with RWX permissions too because otherwise
it wouldn't be possible to allocate a large code page over the freed
regular code page and vice versa.
When enabled, the new machinery works as follows:
1) when memory region is reserved for allocating executable pages, the
whole region is committed with RWX permissions and then decommitted,
2) since reconfiguration of RWX page permissions is not allowed on
MacOS on ARM64 ("Apple M1"/Apple Silicon), there must be no attempts
to change them,
3) the request to set RWX permissions in the executable page region
just recommits the pages without changing permissions (see (1), they
were already allocated as RWX and then discarded),
4) in order to make executable pages inaccessible one must use
OS::DiscardSystemPages() instead of OS::DecommitPages() or
setting permissions to kNoAccess because the latter two are not
allowed by the MacOS (see (2)).
5) since code space page headers are allocated as RWX pages it's also
necessary to switch between W^X modes when updating the data in the
page headers (i.e. when marking, updating stats, wiring pages in
lists, etc.). The new CodePageHeaderModificationScope class is used
in the respective places. On unrelated configurations it's a no-op.
The fast permission switching can't be used for V8 configuration with
enabled pointer compression and disabled external code space because
a) the pointer compression cage has to be reserved with MAP_JIT flag
which is too expensive,
b) in case of shared pointer compression cage if the code range will
be deleted while the cage is still alive then attempt to configure
permissions of pages that were previously set to RWX will fail.
This also CL extends the unmapper unit tests with permissions tracking
for discarded pages.
Bug: v8:12797
Change-Id: Idb28cbc481306477589eee9962d2e75167d87c61
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3579303
Reviewed-by: Nico Hartmann <nicohartmann@chromium.org>
Reviewed-by: Clemens Backes <clemensb@chromium.org>
Reviewed-by: Michael Lippautz <mlippautz@chromium.org>
Commit-Queue: Igor Sheludko <ishell@chromium.org>
Cr-Commit-Position: refs/heads/main@{#80238}
This is a reland of commit 4d8e1846a7
One of the Mac arm64 bots failed to link an exported thread_local
static variable (crbug/1316800).
Original change's description:
> [rwx][mac] Introduce RwxMemoryWriteScope
>
> ... as a single bottleneck that encapsulates the semantics and
> implementation of fast per-thread W^X permission switching supported
> by Apple Silicon (arm64 M1).
> On other architectures this class is a no-op.
>
> Bug: v8:12797
> Change-Id: Ica842ff9f843e20b7f61fd7e80591e7a1fd29771
> Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3586986
> Reviewed-by: Jakob Kummerow <jkummerow@chromium.org>
> Commit-Queue: Igor Sheludko <ishell@chromium.org>
> Cr-Commit-Position: refs/heads/main@{#79994}
Bug: v8:12797
Change-Id: Ifbd15c233bb343f11daa89b1328b5bf65c4806f4
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3591332
Reviewed-by: Leszek Swirski <leszeks@chromium.org>
Reviewed-by: Clemens Backes <clemensb@chromium.org>
Commit-Queue: Igor Sheludko <ishell@chromium.org>
Cr-Commit-Position: refs/heads/main@{#80011}
This reverts commit 4d8e1846a7.
Reason for revert: Blocks V8 roll, crbug.com/1316800
Original change's description:
> [rwx][mac] Introduce RwxMemoryWriteScope
>
> ... as a single bottleneck that encapsulates the semantics and
> implementation of fast per-thread W^X permission switching supported
> by Apple Silicon (arm64 M1).
> On other architectures this class is a no-op.
>
> Bug: v8:12797
> Change-Id: Ica842ff9f843e20b7f61fd7e80591e7a1fd29771
> Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3586986
> Reviewed-by: Jakob Kummerow <jkummerow@chromium.org>
> Commit-Queue: Igor Sheludko <ishell@chromium.org>
> Cr-Commit-Position: refs/heads/main@{#79994}
Bug: v8:12797
Change-Id: I81792567839e72b4147d009c0845b0c0de003eb0
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3590752
Bot-Commit: Rubber Stamper <rubber-stamper@appspot.gserviceaccount.com>
Commit-Queue: Deepti Gandluri <gdeepti@chromium.org>
Owners-Override: Deepti Gandluri <gdeepti@chromium.org>
Cr-Commit-Position: refs/heads/main@{#80007}
... as a single bottleneck that encapsulates the semantics and
implementation of fast per-thread W^X permission switching supported
by Apple Silicon (arm64 M1).
On other architectures this class is a no-op.
Bug: v8:12797
Change-Id: Ica842ff9f843e20b7f61fd7e80591e7a1fd29771
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3586986
Reviewed-by: Jakob Kummerow <jkummerow@chromium.org>
Commit-Queue: Igor Sheludko <ishell@chromium.org>
Cr-Commit-Position: refs/heads/main@{#79994}
Callee saved registers do not include the LR anymore, so we can
now remove the last place where we pass a non-default template
argument to PushCPURegList/PopCPURegList (in the code generator).
This makes the template argument redundant, so we can remove the
template altogether.
Change-Id: I07f0c0a10840817df8a5afc1dc74330e290ce5bf
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3571816
Reviewed-by: Jakob Linke <jgruber@chromium.org>
Commit-Queue: Georgia Kouveli <georgia.kouveli@arm.com>
Cr-Commit-Position: refs/heads/main@{#79842}
Modernise the RegList interface to be a proper class, rather than a
typedef to an integer, and add proper methods onto it rather than ad-hoc
bit manipulation.
In particular, this makes RegList typesafe, adding a DoubleRegList for
DoubleRegisters.
The Arm64 CPURegList isn't updated to use (or extend) the new RegList
interface, because of its weird type-erasing semantics (it can store
Registers and VRegisters). Maybe in the future we'll want to get rid of
CPURegList entirely and use RegList/DoubleRegList directly.
Change-Id: I3cb2a4d386cb92a4dcd2edbdd3fba9ef71f354d6
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3516747
Commit-Queue: Leszek Swirski <leszeks@chromium.org>
Reviewed-by: Jakob Gruber <jgruber@chromium.org>
Reviewed-by: Camillo Bruni <cbruni@chromium.org>
Reviewed-by: Clemens Backes <clemensb@chromium.org>
Cr-Commit-Position: refs/heads/main@{#79460}
NONE clashes with the PropertyAttributes::NONE, which is defined in
v8::internal namespace. PropertyAttributes have too many call sites
and depend on using the enums as masks, making it hard to convert
to an enum class. So we are changing the name instead.
Bug: v8:12244
Change-Id: Iec0be12c626549cca137aceeaee0e30fafab8b05
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3284003
Commit-Queue: Zhi An Ng <zhin@chromium.org>
Reviewed-by: Andreas Haas <ahaas@chromium.org>
Reviewed-by: Maya Lekova <mslekova@chromium.org>
Cr-Commit-Position: refs/heads/main@{#77956}
Removing prefetch operations as per the vote in the github issue:
https://github.com/WebAssembly/simd/pull/352
Bug:v8:11168
Change-Id: Ia72684e68ce886f8f26a7d3b5bea601be416dfab
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2771758
Reviewed-by: Jakob Kummerow <jkummerow@chromium.org>
Reviewed-by: Maya Lekova <mslekova@chromium.org>
Reviewed-by: Zhi An Ng <zhin@chromium.org>
Commit-Queue: Deepti Gandluri <gdeepti@chromium.org>
Cr-Commit-Position: refs/heads/master@{#73578}
This requires a small fix in {Push,Pop}CalleeSavedRegisters, where
the return address was signed/authenticated at the wrong point,
which meant the stack pointer used as modifier was different from
the one the StackFrameIterator expected.
Bug: v8:10026
Change-Id: Idebd2ee8f07312b5e99dd2ea5181fc7a7e4a87bc
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2667861
Commit-Queue: Georgia Kouveli <georgia.kouveli@arm.com>
Reviewed-by: Jakob Gruber <jgruber@chromium.org>
Cr-Commit-Position: refs/heads/master@{#72542}
Prototype these 6 instructions on arm64:
- f64x2.convert_low_i32x4_s
- f64x2.convert_low_i32x4_u
- i32x4.trunc_sat_f64x2_s_zero
- i32x4.trunc_sat_f64x2_u_zero
- f32x4.demote_f64x2_zero
- f64x2.promote_low_f32x4
Drive-by fix:
- f64x2.promote_low_f32x4 accesses out of bounds for the global, the
result only has 2 doubles
- fcvtn in simulator needs to clear top bits of the Q reg
Bug: v8:11265
Change-Id: Icfb3338942f0d0374448fdcfef3847a6e3ce8ff6
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2644066
Reviewed-by: Deepti Gandluri <gdeepti@chromium.org>
Reviewed-by: Jakob Kummerow <jkummerow@chromium.org>
Commit-Queue: Zhi An Ng <zhin@chromium.org>
Cr-Commit-Position: refs/heads/master@{#72373}
Prototype 2 prefetch instructions (temporal and non-temporal) on arm64
and interpreter.
Add prfm to assembler, and use MiscField to encode the two versions.
Small tweak to simulator to handle these new instructions (no-op).
The implementation in the interpreter just pops the memory index and
does nothing.
Simple test cases added for these 2 new instructions, as well as a
prefetch with OOB index, which should not trap.
Bug: v8:11168
Change-Id: Ieced8081615d07f950d6d4c1128d1bc6a75839fd
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2543167
Reviewed-by: Bill Budge <bbudge@chromium.org>
Reviewed-by: Jakob Kummerow <jkummerow@chromium.org>
Reviewed-by: Tobias Tebbi <tebbi@chromium.org>
Commit-Queue: Zhi An Ng <zhin@chromium.org>
Cr-Commit-Position: refs/heads/master@{#71353}
This is a reland of fbfa9bf4ec
The arm64 was missing proper codegen for CFI, thus sizes were off.
Original change's description:
> Reland "[deoptimizer] Change deopt entries into builtins"
>
> This is a reland of 7f58ced72e
>
> It fixes the different exit size emitted on x64/Atom CPUs due to
> performance tuning in TurboAssembler::Call. Additionally, add
> cctests to verify the fixed size exits.
>
> Original change's description:
> > [deoptimizer] Change deopt entries into builtins
> >
> > While the overall goal of this commit is to change deoptimization
> > entries into builtins, there are multiple related things happening:
> >
> > - Deoptimization entries, formerly stubs (i.e. Code objects generated
> > at runtime, guaranteed to be immovable), have been converted into
> > builtins. The major restriction is that we now need to preserve the
> > kRootRegister, which was formerly used on most architectures to pass
> > the deoptimization id. The solution differs based on platform.
> > - Renamed DEOPT_ENTRIES_OR_FOR_TESTING code kind to FOR_TESTING.
> > - Removed heap/ support for immovable Code generation.
> > - Removed the DeserializerData class (no longer needed).
> > - arm64: to preserve 4-byte deopt exits, introduced a new optimization
> > in which the final jump to the deoptimization entry is generated
> > once per Code object, and deopt exits can continue to emit a
> > near-call.
> > - arm,ia32,x64: change to fixed-size deopt exits. This reduces exit
> > sizes by 4/8, 5, and 5 bytes, respectively.
> >
> > On arm the deopt exit size is reduced from 12 (or 16) bytes to 8 bytes
> > by using the same strategy as on arm64 (recalc deopt id from return
> > address). Before:
> >
> > e300a002 movw r10, <id>
> > e59fc024 ldr ip, [pc, <entry offset>]
> > e12fff3c blx ip
> >
> > After:
> >
> > e59acb35 ldr ip, [r10, <entry offset>]
> > e12fff3c blx ip
> >
> > On arm64 the deopt exit size remains 4 bytes (or 8 bytes in same cases
> > with CFI). Additionally, up to 4 builtin jumps are emitted per Code
> > object (max 32 bytes added overhead per Code object). Before:
> >
> > 9401cdae bl <entry offset>
> >
> > After:
> >
> > # eager deoptimization entry jump.
> > f95b1f50 ldr x16, [x26, <eager entry offset>]
> > d61f0200 br x16
> > # lazy deoptimization entry jump.
> > f95b2b50 ldr x16, [x26, <lazy entry offset>]
> > d61f0200 br x16
> > # the deopt exit.
> > 97fffffc bl <eager deoptimization entry jump offset>
> >
> > On ia32 the deopt exit size is reduced from 10 to 5 bytes. Before:
> >
> > bb00000000 mov ebx,<id>
> > e825f5372b call <entry>
> >
> > After:
> >
> > e8ea2256ba call <entry>
> >
> > On x64 the deopt exit size is reduced from 12 to 7 bytes. Before:
> >
> > 49c7c511000000 REX.W movq r13,<id>
> > e8ea2f0700 call <entry>
> >
> > After:
> >
> > 41ff9560360000 call [r13+<entry offset>]
> >
> > Bug: v8:8661,v8:8768
> > Change-Id: I13e30aedc360474dc818fecc528ce87c3bfeed42
> > Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2465834
> > Commit-Queue: Jakob Gruber <jgruber@chromium.org>
> > Reviewed-by: Ross McIlroy <rmcilroy@chromium.org>
> > Reviewed-by: Tobias Tebbi <tebbi@chromium.org>
> > Reviewed-by: Ulan Degenbaev <ulan@chromium.org>
> > Cr-Commit-Position: refs/heads/master@{#70597}
>
> Tbr: ulan@chromium.org, tebbi@chromium.org, rmcilroy@chromium.org
> Bug: v8:8661,v8:8768,chromium:1140165
> Change-Id: Ibcd5c39c58a70bf2b2ac221aa375fc68d495e144
> Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2485506
> Reviewed-by: Jakob Gruber <jgruber@chromium.org>
> Reviewed-by: Tobias Tebbi <tebbi@chromium.org>
> Commit-Queue: Jakob Gruber <jgruber@chromium.org>
> Cr-Commit-Position: refs/heads/master@{#70655}
Tbr: ulan@chromium.org, tebbi@chromium.org, rmcilroy@chromium.org
Bug: v8:8661
Bug: v8:8768
Bug: chromium:1140165
Change-Id: I471cc94fc085e527dc9bfb5a84b96bd907c2333f
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2488682
Reviewed-by: Jakob Gruber <jgruber@chromium.org>
Commit-Queue: Jakob Gruber <jgruber@chromium.org>
Cr-Commit-Position: refs/heads/master@{#70672}
This is a reland of 7f58ced72e
It fixes the different exit size emitted on x64/Atom CPUs due to
performance tuning in TurboAssembler::Call. Additionally, add
cctests to verify the fixed size exits.
Original change's description:
> [deoptimizer] Change deopt entries into builtins
>
> While the overall goal of this commit is to change deoptimization
> entries into builtins, there are multiple related things happening:
>
> - Deoptimization entries, formerly stubs (i.e. Code objects generated
> at runtime, guaranteed to be immovable), have been converted into
> builtins. The major restriction is that we now need to preserve the
> kRootRegister, which was formerly used on most architectures to pass
> the deoptimization id. The solution differs based on platform.
> - Renamed DEOPT_ENTRIES_OR_FOR_TESTING code kind to FOR_TESTING.
> - Removed heap/ support for immovable Code generation.
> - Removed the DeserializerData class (no longer needed).
> - arm64: to preserve 4-byte deopt exits, introduced a new optimization
> in which the final jump to the deoptimization entry is generated
> once per Code object, and deopt exits can continue to emit a
> near-call.
> - arm,ia32,x64: change to fixed-size deopt exits. This reduces exit
> sizes by 4/8, 5, and 5 bytes, respectively.
>
> On arm the deopt exit size is reduced from 12 (or 16) bytes to 8 bytes
> by using the same strategy as on arm64 (recalc deopt id from return
> address). Before:
>
> e300a002 movw r10, <id>
> e59fc024 ldr ip, [pc, <entry offset>]
> e12fff3c blx ip
>
> After:
>
> e59acb35 ldr ip, [r10, <entry offset>]
> e12fff3c blx ip
>
> On arm64 the deopt exit size remains 4 bytes (or 8 bytes in same cases
> with CFI). Additionally, up to 4 builtin jumps are emitted per Code
> object (max 32 bytes added overhead per Code object). Before:
>
> 9401cdae bl <entry offset>
>
> After:
>
> # eager deoptimization entry jump.
> f95b1f50 ldr x16, [x26, <eager entry offset>]
> d61f0200 br x16
> # lazy deoptimization entry jump.
> f95b2b50 ldr x16, [x26, <lazy entry offset>]
> d61f0200 br x16
> # the deopt exit.
> 97fffffc bl <eager deoptimization entry jump offset>
>
> On ia32 the deopt exit size is reduced from 10 to 5 bytes. Before:
>
> bb00000000 mov ebx,<id>
> e825f5372b call <entry>
>
> After:
>
> e8ea2256ba call <entry>
>
> On x64 the deopt exit size is reduced from 12 to 7 bytes. Before:
>
> 49c7c511000000 REX.W movq r13,<id>
> e8ea2f0700 call <entry>
>
> After:
>
> 41ff9560360000 call [r13+<entry offset>]
>
> Bug: v8:8661,v8:8768
> Change-Id: I13e30aedc360474dc818fecc528ce87c3bfeed42
> Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2465834
> Commit-Queue: Jakob Gruber <jgruber@chromium.org>
> Reviewed-by: Ross McIlroy <rmcilroy@chromium.org>
> Reviewed-by: Tobias Tebbi <tebbi@chromium.org>
> Reviewed-by: Ulan Degenbaev <ulan@chromium.org>
> Cr-Commit-Position: refs/heads/master@{#70597}
Tbr: ulan@chromium.org, tebbi@chromium.org, rmcilroy@chromium.org
Bug: v8:8661,v8:8768,chromium:1140165
Change-Id: Ibcd5c39c58a70bf2b2ac221aa375fc68d495e144
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2485506
Reviewed-by: Jakob Gruber <jgruber@chromium.org>
Reviewed-by: Tobias Tebbi <tebbi@chromium.org>
Commit-Queue: Jakob Gruber <jgruber@chromium.org>
Cr-Commit-Position: refs/heads/master@{#70655}
This reverts commit 7f58ced72e.
Reason for revert: Segfaults on Atom_x64 https://ci.chromium.org/p/v8-internal/builders/ci/v8_linux64_atom_perf/5686?
Original change's description:
> [deoptimizer] Change deopt entries into builtins
>
> While the overall goal of this commit is to change deoptimization
> entries into builtins, there are multiple related things happening:
>
> - Deoptimization entries, formerly stubs (i.e. Code objects generated
> at runtime, guaranteed to be immovable), have been converted into
> builtins. The major restriction is that we now need to preserve the
> kRootRegister, which was formerly used on most architectures to pass
> the deoptimization id. The solution differs based on platform.
> - Renamed DEOPT_ENTRIES_OR_FOR_TESTING code kind to FOR_TESTING.
> - Removed heap/ support for immovable Code generation.
> - Removed the DeserializerData class (no longer needed).
> - arm64: to preserve 4-byte deopt exits, introduced a new optimization
> in which the final jump to the deoptimization entry is generated
> once per Code object, and deopt exits can continue to emit a
> near-call.
> - arm,ia32,x64: change to fixed-size deopt exits. This reduces exit
> sizes by 4/8, 5, and 5 bytes, respectively.
>
> On arm the deopt exit size is reduced from 12 (or 16) bytes to 8 bytes
> by using the same strategy as on arm64 (recalc deopt id from return
> address). Before:
>
> e300a002 movw r10, <id>
> e59fc024 ldr ip, [pc, <entry offset>]
> e12fff3c blx ip
>
> After:
>
> e59acb35 ldr ip, [r10, <entry offset>]
> e12fff3c blx ip
>
> On arm64 the deopt exit size remains 4 bytes (or 8 bytes in same cases
> with CFI). Additionally, up to 4 builtin jumps are emitted per Code
> object (max 32 bytes added overhead per Code object). Before:
>
> 9401cdae bl <entry offset>
>
> After:
>
> # eager deoptimization entry jump.
> f95b1f50 ldr x16, [x26, <eager entry offset>]
> d61f0200 br x16
> # lazy deoptimization entry jump.
> f95b2b50 ldr x16, [x26, <lazy entry offset>]
> d61f0200 br x16
> # the deopt exit.
> 97fffffc bl <eager deoptimization entry jump offset>
>
> On ia32 the deopt exit size is reduced from 10 to 5 bytes. Before:
>
> bb00000000 mov ebx,<id>
> e825f5372b call <entry>
>
> After:
>
> e8ea2256ba call <entry>
>
> On x64 the deopt exit size is reduced from 12 to 7 bytes. Before:
>
> 49c7c511000000 REX.W movq r13,<id>
> e8ea2f0700 call <entry>
>
> After:
>
> 41ff9560360000 call [r13+<entry offset>]
>
> Bug: v8:8661,v8:8768
> Change-Id: I13e30aedc360474dc818fecc528ce87c3bfeed42
> Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2465834
> Commit-Queue: Jakob Gruber <jgruber@chromium.org>
> Reviewed-by: Ross McIlroy <rmcilroy@chromium.org>
> Reviewed-by: Tobias Tebbi <tebbi@chromium.org>
> Reviewed-by: Ulan Degenbaev <ulan@chromium.org>
> Cr-Commit-Position: refs/heads/master@{#70597}
TBR=ulan@chromium.org,rmcilroy@chromium.org,jgruber@chromium.org,tebbi@chromium.org
# Not skipping CQ checks because original CL landed > 1 day ago.
Bug: v8:8661,v8:8768,chromium:1140165
Change-Id: I3df02ab42f6e02233d9f6fb80e8bb18f76870d91
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2485504
Reviewed-by: Jakob Gruber <jgruber@chromium.org>
Commit-Queue: Jakob Gruber <jgruber@chromium.org>
Cr-Commit-Position: refs/heads/master@{#70649}
This fixes a bug that made a test fail in mjsunit/wasm/return-call.js
(the CFI bot does not run the tests with --variants=extra, hence why
it didn't catch it).
It also introduces --sim-abort-on-bad-auth, a debug flag for the arm64
simulator that stops a program as soon as an authentication error
appears, to make debugging easier.
Change-Id: Ibee731ab788aff45301d268ef05256b82f5e4613
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2473833
Commit-Queue: Jakob Gruber <jgruber@chromium.org>
Reviewed-by: Thibaud Michaud <thibaudm@chromium.org>
Reviewed-by: Jakob Gruber <jgruber@chromium.org>
Cr-Commit-Position: refs/heads/master@{#70601}
While the overall goal of this commit is to change deoptimization
entries into builtins, there are multiple related things happening:
- Deoptimization entries, formerly stubs (i.e. Code objects generated
at runtime, guaranteed to be immovable), have been converted into
builtins. The major restriction is that we now need to preserve the
kRootRegister, which was formerly used on most architectures to pass
the deoptimization id. The solution differs based on platform.
- Renamed DEOPT_ENTRIES_OR_FOR_TESTING code kind to FOR_TESTING.
- Removed heap/ support for immovable Code generation.
- Removed the DeserializerData class (no longer needed).
- arm64: to preserve 4-byte deopt exits, introduced a new optimization
in which the final jump to the deoptimization entry is generated
once per Code object, and deopt exits can continue to emit a
near-call.
- arm,ia32,x64: change to fixed-size deopt exits. This reduces exit
sizes by 4/8, 5, and 5 bytes, respectively.
On arm the deopt exit size is reduced from 12 (or 16) bytes to 8 bytes
by using the same strategy as on arm64 (recalc deopt id from return
address). Before:
e300a002 movw r10, <id>
e59fc024 ldr ip, [pc, <entry offset>]
e12fff3c blx ip
After:
e59acb35 ldr ip, [r10, <entry offset>]
e12fff3c blx ip
On arm64 the deopt exit size remains 4 bytes (or 8 bytes in same cases
with CFI). Additionally, up to 4 builtin jumps are emitted per Code
object (max 32 bytes added overhead per Code object). Before:
9401cdae bl <entry offset>
After:
# eager deoptimization entry jump.
f95b1f50 ldr x16, [x26, <eager entry offset>]
d61f0200 br x16
# lazy deoptimization entry jump.
f95b2b50 ldr x16, [x26, <lazy entry offset>]
d61f0200 br x16
# the deopt exit.
97fffffc bl <eager deoptimization entry jump offset>
On ia32 the deopt exit size is reduced from 10 to 5 bytes. Before:
bb00000000 mov ebx,<id>
e825f5372b call <entry>
After:
e8ea2256ba call <entry>
On x64 the deopt exit size is reduced from 12 to 7 bytes. Before:
49c7c511000000 REX.W movq r13,<id>
e8ea2f0700 call <entry>
After:
41ff9560360000 call [r13+<entry offset>]
Bug: v8:8661,v8:8768
Change-Id: I13e30aedc360474dc818fecc528ce87c3bfeed42
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2465834
Commit-Queue: Jakob Gruber <jgruber@chromium.org>
Reviewed-by: Ross McIlroy <rmcilroy@chromium.org>
Reviewed-by: Tobias Tebbi <tebbi@chromium.org>
Reviewed-by: Ulan Degenbaev <ulan@chromium.org>
Cr-Commit-Position: refs/heads/master@{#70597}
CodeKind::OPTIMIZED_CODE -> TURBOFAN
Kinds are now more fine-grained and distinguish between TF, TP, NCI.
CodeKind::STUB -> DEOPT_ENTRIES_OR_FOR_TESTING
Code stubs (like builtins, but generated at runtime) were removed from
the codebase years ago, this is the last remnant. This kind is used
only for deopt entries (which should be converted into builtins) and
for tests.
Change-Id: I67beb15377cb60f395e9b051b25f3e5764982e93
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2440335
Auto-Submit: Jakob Gruber <jgruber@chromium.org>
Commit-Queue: Mythri Alle <mythria@chromium.org>
Reviewed-by: Mythri Alle <mythria@chromium.org>
Cr-Commit-Position: refs/heads/master@{#70234}
Apple's upcoming arm64 devices will prevent rwx access to memory,
but in turn provide a new per-thread way to switch between write
and execute permissions. This patch puts that system to use for
the WebAssembly subsystem.
The approach relies on CodeSpaceWriteScope objects for now. That
isn't optimal for background threads (which could stay in "write"
mode permanently instead of toggling), but its simplicity makes
it a good first step.
Background:
https://developer.apple.com/documentation/apple_silicon/porting_just-in-time_compilers_to_apple_silicon
Bug: chromium:1117591
Change-Id: I3b60f0efd34c0fed924dfc71ee2c7805801c5d42
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2378307
Commit-Queue: Jakob Kummerow <jkummerow@chromium.org>
Reviewed-by: Michael Lippautz <mlippautz@chromium.org>
Reviewed-by: Thibaud Michaud <thibaudm@chromium.org>
Cr-Commit-Position: refs/heads/master@{#69791}
With the new Turbofan variants (NCI and Turboprop), we need a way to
distinguish between them both during and after compilation. We
initially introduced CompilationTarget to track the variant during
compilation, but decided to reuse the code kind as the canonical spot to
store this information instead.
Why? Because it is an established mechanism, already available in most
of the necessary spots (inside the pipeline, on Code objects, in
profiling traces).
This CL removes CompilationTarget and adds a new
NATIVE_CONTEXT_INDEPENDENT kind, plus helper functions to determine
various things about a given code kind (e.g.: does this code kind
deopt?).
As a (very large) drive-by, refactor both Code::Kind and
AbstractCode::Kind into a new CodeKind enum class.
Bug: v8:8888
Change-Id: Ie858b9a53311b0731630be35cf5cd108dee95b39
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2336793
Commit-Queue: Jakob Gruber <jgruber@chromium.org>
Reviewed-by: Clemens Backes <clemensb@chromium.org>
Reviewed-by: Ross McIlroy <rmcilroy@chromium.org>
Reviewed-by: Dominik Inführ <dinfuehr@chromium.org>
Reviewed-by: Georg Neis <neis@chromium.org>
Cr-Commit-Position: refs/heads/master@{#69244}
The C++ code uses the A instruction key for return address signing,
which is the default for Clang and GCC when the -mbranch-protection
option is used (although this can be configured to use the B key).
Using the B key for JS means that it's not possible to use an A key
signing gadget to replace a return address signed with the B key and
vice-versa. This should offer a degree of separation from the C++ side.
Bug: v8:10026
Change-Id: Ia9dcc7ae7096c96b4a271efbe25fc02940f6fc8e
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2242953
Reviewed-by: Ross McIlroy <rmcilroy@chromium.org>
Reviewed-by: Georg Neis <neis@chromium.org>
Commit-Queue: Georgia Kouveli <georgia.kouveli@arm.com>
Cr-Commit-Position: refs/heads/master@{#68360}
This includes the instruction opcode, its use in TF, its support in the
simulator and the detection of the associated CPU feature.
The instruction can be tested in the simulator with the new
--sim-arm64-optional-features flag.
Change-Id: I6047fa16696394fe0ced4535f7788d2c8716a18c
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2222348
Reviewed-by: Ross McIlroy <rmcilroy@chromium.org>
Reviewed-by: Georg Neis <neis@chromium.org>
Commit-Queue: Georg Neis <neis@chromium.org>
Cr-Commit-Position: refs/heads/master@{#68261}
Generate a BTI instruction at each target of an indirect branch
(BR/BLR). An indirect branch that doesn't jump to a BTI instruction
will generate an exception on a BTI-enabled core. On cores that do
not support the BTI extension, the BTI instruction is a NOP.
Targets of indirect branch instructions include, among other things,
function entrypoints, exception handlers and jump tables. Lazy deopt
exits can potentially be reached through an indirect branch when an
exception is thrown, so they also get an additional BTI instruction.
Bug: v8:10026
Change-Id: I0ebf51071f1b604f60f524096e013dfd64fcd7ff
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1967315
Commit-Queue: Georgia Kouveli <georgia.kouveli@arm.com>
Reviewed-by: Ross McIlroy <rmcilroy@chromium.org>
Reviewed-by: Georg Neis <neis@chromium.org>
Reviewed-by: Clemens Backes <clemensb@chromium.org>
Cr-Commit-Position: refs/heads/master@{#66751}
Changes the lambda capture for this test so that referenced variables
are captured automatically. This means MSVC and Clang can compile the
same test without compiler-specific #ifdef's.
Bug: v8:10012
Change-Id: Ife4ee050e227478842c0f8def92adb0857082ccf
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2066984
Reviewed-by: Jakob Kummerow <jkummerow@chromium.org>
Commit-Queue: Richard Townsend <richard.townsend@arm.com>
Cr-Commit-Position: refs/heads/master@{#66414}
There are cases where we push FP and LR to the stack even though we have
already created a frame and already stored them.
Avoid doing this, by cleaning up the caller- and callee-saved
register lists, and passing a LinkRegisterStatus argument to
TruncateDoubleToI.
Change-Id: I652e7ec346a99e94617f1ec6b13935f5d2f73e7f
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2060505
Commit-Queue: Georgia Kouveli <georgia.kouveli@arm.com>
Reviewed-by: Ross McIlroy <rmcilroy@chromium.org>
Cr-Commit-Position: refs/heads/master@{#66330}
This is a reland of 137bfe47c9
Original change's description:
> [arm64] Protect return addresses stored on stack
>
> This change uses the Arm v8.3 pointer authentication instructions in
> order to protect return addresses stored on the stack. The generated
> code signs the return address before storing on the stack and
> authenticates it after loading it. This also changes the stack frame
> iterator in order to authenticate stored return addresses and re-sign
> them when needed, as well as the deoptimizer in order to sign saved
> return addresses when creating new frames. This offers a level of
> protection against ROP attacks.
>
> This functionality is enabled with the v8_control_flow_integrity flag
> that this CL introduces.
>
> The code size effect of this change is small for Octane (up to 2% in
> some cases but mostly much lower) and negligible for larger benchmarks,
> however code size measurements are rather noisy. The performance impact
> on current cores (where the instructions are NOPs) is single digit,
> around 1-2% for ARES-6 and Octane, and tends to be smaller for big
> cores than for little cores.
>
> Bug: v8:10026
> Change-Id: I0081f3938c56e2f24d8227e4640032749f4f8368
> Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1373782
> Commit-Queue: Georgia Kouveli <georgia.kouveli@arm.com>
> Reviewed-by: Ross McIlroy <rmcilroy@chromium.org>
> Reviewed-by: Georg Neis <neis@chromium.org>
> Cr-Commit-Position: refs/heads/master@{#66239}
Bug: v8:10026
Change-Id: Id1adfa2e6c713f6977d69aa467986e48fe67b3c2
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2051958
Reviewed-by: Georg Neis <neis@chromium.org>
Reviewed-by: Ross McIlroy <rmcilroy@chromium.org>
Commit-Queue: Georgia Kouveli <georgia.kouveli@arm.com>
Cr-Commit-Position: refs/heads/master@{#66254}
This reverts commit 137bfe47c9.
Reason for revert: https://ci.chromium.org/p/v8/builders/ci/V8%20Arm%20-%20debug/13072
Original change's description:
> [arm64] Protect return addresses stored on stack
>
> This change uses the Arm v8.3 pointer authentication instructions in
> order to protect return addresses stored on the stack. The generated
> code signs the return address before storing on the stack and
> authenticates it after loading it. This also changes the stack frame
> iterator in order to authenticate stored return addresses and re-sign
> them when needed, as well as the deoptimizer in order to sign saved
> return addresses when creating new frames. This offers a level of
> protection against ROP attacks.
>
> This functionality is enabled with the v8_control_flow_integrity flag
> that this CL introduces.
>
> The code size effect of this change is small for Octane (up to 2% in
> some cases but mostly much lower) and negligible for larger benchmarks,
> however code size measurements are rather noisy. The performance impact
> on current cores (where the instructions are NOPs) is single digit,
> around 1-2% for ARES-6 and Octane, and tends to be smaller for big
> cores than for little cores.
>
> Bug: v8:10026
> Change-Id: I0081f3938c56e2f24d8227e4640032749f4f8368
> Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1373782
> Commit-Queue: Georgia Kouveli <georgia.kouveli@arm.com>
> Reviewed-by: Ross McIlroy <rmcilroy@chromium.org>
> Reviewed-by: Georg Neis <neis@chromium.org>
> Cr-Commit-Position: refs/heads/master@{#66239}
TBR=rmcilroy@chromium.org,mstarzinger@chromium.org,neis@chromium.org,georgia.kouveli@arm.com
Change-Id: I57d5928949b0d403774550b9bf7dc0b08ce4e703
No-Presubmit: true
No-Tree-Checks: true
No-Try: true
Bug: v8:10026
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2051952
Reviewed-by: Nico Hartmann <nicohartmann@chromium.org>
Commit-Queue: Nico Hartmann <nicohartmann@chromium.org>
Cr-Commit-Position: refs/heads/master@{#66242}
This change uses the Arm v8.3 pointer authentication instructions in
order to protect return addresses stored on the stack. The generated
code signs the return address before storing on the stack and
authenticates it after loading it. This also changes the stack frame
iterator in order to authenticate stored return addresses and re-sign
them when needed, as well as the deoptimizer in order to sign saved
return addresses when creating new frames. This offers a level of
protection against ROP attacks.
This functionality is enabled with the v8_control_flow_integrity flag
that this CL introduces.
The code size effect of this change is small for Octane (up to 2% in
some cases but mostly much lower) and negligible for larger benchmarks,
however code size measurements are rather noisy. The performance impact
on current cores (where the instructions are NOPs) is single digit,
around 1-2% for ARES-6 and Octane, and tends to be smaller for big
cores than for little cores.
Bug: v8:10026
Change-Id: I0081f3938c56e2f24d8227e4640032749f4f8368
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1373782
Commit-Queue: Georgia Kouveli <georgia.kouveli@arm.com>
Reviewed-by: Ross McIlroy <rmcilroy@chromium.org>
Reviewed-by: Georg Neis <neis@chromium.org>
Cr-Commit-Position: refs/heads/master@{#66239}
This removes {CPURegister::Is} and {CPURegister::is}, and just uses
{CPURegister::operator==} instead.
Drive-by: Use DCHECK_EQ and DCHECK_NE where possible.
R=mstarzinger@chromium.org
Bug: v8:9810
Change-Id: I03aad8b4223bd4ae37d468326a734f7a5c3c8061
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1916202
Reviewed-by: Jakob Gruber <jgruber@chromium.org>
Reviewed-by: Michael Starzinger <mstarzinger@chromium.org>
Commit-Queue: Clemens Backes <clemensb@chromium.org>
Cr-Commit-Position: refs/heads/master@{#64956}
Just use the default {is_valid} method provided by RegisterBase.
Drive-by: Also rename {CPURegList::IsValid} to {CPURegList::is_valid}.
R=mstarzinger@chromium.org
Bug: v8:9810
Change-Id: Ia3bc3c477e6329d63ffd00bca59762d9a6cf2fcc
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1916201
Commit-Queue: Clemens Backes <clemensb@chromium.org>
Reviewed-by: Michael Starzinger <mstarzinger@chromium.org>
Cr-Commit-Position: refs/heads/master@{#64952}
Windows SDK defines `near` and `far` as macro in minwindef.h, so they cannot be
used as variable name if Windows SDK header file is included for Windows build.
Bug: chromium:893460
Change-Id: I5ed1076b965979b8e4e09958c1b6f0a698ec8d4f
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1783839
Reviewed-by: Sigurd Schneider <sigurds@chromium.org>
Commit-Queue: Tom Tan <Tom.Tan@microsoft.com>
Cr-Commit-Position: refs/heads/master@{#63559}
The `Instruction::IsValidImmPCOffset()` method was taking an `offset` argument
in numbers of *instructions* while we were passing it numbers of *bytes*. See
`Instruction::IsTargetInImmPCOffsetRange()` and
`MacroAssembler::NeedExtraInstructionsOrRegisterBranch()`.
As a result, we were 4 times too conservative when computing branch ranges going
backwards, forcing us to generate the following sequence for TBZ more often than
needed:
```
TBNZ <skip>
B <target>
skip:
```
This happened rarely for loops, but a lot when doing an early return from
out-of-line calls to write barriers. Since out-of-line code is easily out of
range of 8K, although the real range of TBZ is 32K.
This fixes it by changing this method to take a byte offset instead of
instructions, as this is more intuitive and in line with similar methods. For
instance, `Instruction::ImmPcOffset()` returns an offset in bytes.
The tests are adapted so that they would have caught such a bug:
* TEST(far_branch_backward):
This test used to only check the code worked if the branch was very far away,
but it didn't test the range was correct. So this test was changed to check
each branch type separately, and test in-range and out-of-range cases
separately too.
* TEST(far_branch_veneer_broken_link_chain):
Because of the backwards range bug, this test wasn't actually testing what it
should. The idea of the test is to make sure the MacroAssembler can still cope
when the chain of links is broken after a veneer was emitted. But no veneers
were ever emitted.
Change-Id: Iddb5c683a71147455175f38fa7ae57da0a3e7337
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1781058
Reviewed-by: Ulan Degenbaev <ulan@chromium.org>
Reviewed-by: Sigurd Schneider <sigurds@chromium.org>
Commit-Queue: Sigurd Schneider <sigurds@chromium.org>
Cr-Commit-Position: refs/heads/master@{#63518}
This is a reland of ac79b539ec
This CL adds a missing BlockPoolsScope to guard a RequestHeapObject
call. This fixes a latend bug that the original land flushed out.
Original change's description:
> [arm64] Refactor constant pool implementation
>
> This refactors the constant pool handling for arm64. The immediate goal
> is to allow 32bit compressed pointers in the pool. The mediate goal is
> to unify the implementation with the arm constant pool, which will be
> done in a follow-up CL.
>
> Bug: v8:8054
> Change-Id: I74db4245e5e1025f2e4de4144090fa4ce25883ab
> Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1645316
> Reviewed-by: Michael Starzinger <mstarzinger@chromium.org>
> Reviewed-by: Jakob Gruber <jgruber@chromium.org>
> Commit-Queue: Sigurd Schneider <sigurds@chromium.org>
> Cr-Commit-Position: refs/heads/master@{#62209}
TBR=mstarzinger@chromium.org,jgruber@chromium.org,georgia.kouveli@arm.com
Bug: v8:8054
Change-Id: I1e3ab13619a48caad33d77ed8bed86782f9d9674
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1664054
Reviewed-by: Sigurd Schneider <sigurds@chromium.org>
Commit-Queue: Sigurd Schneider <sigurds@chromium.org>
Cr-Commit-Position: refs/heads/master@{#62237}
This reverts commit ac79b539ec.
Reason for revert: https://ci.chromium.org/p/v8/builders/ci/V8%20Linux%20-%20arm64%20-%20sim/18611
Original change's description:
> [arm64] Refactor constant pool implementation
>
> This refactors the constant pool handling for arm64. The immediate goal
> is to allow 32bit compressed pointers in the pool. The mediate goal is
> to unify the implementation with the arm constant pool, which will be
> done in a follow-up CL.
>
> Bug: v8:8054
> Change-Id: I74db4245e5e1025f2e4de4144090fa4ce25883ab
> Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1645316
> Reviewed-by: Michael Starzinger <mstarzinger@chromium.org>
> Reviewed-by: Jakob Gruber <jgruber@chromium.org>
> Commit-Queue: Sigurd Schneider <sigurds@chromium.org>
> Cr-Commit-Position: refs/heads/master@{#62209}
TBR=mstarzinger@chromium.org,sigurds@chromium.org,jgruber@chromium.org,georgia.kouveli@arm.com
Change-Id: Iff03e81a2e70d125ef2c06b6ff3aff8d0e3688ef
No-Presubmit: true
No-Tree-Checks: true
No-Try: true
Bug: v8:8054
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1662293
Reviewed-by: Sigurd Schneider <sigurds@chromium.org>
Commit-Queue: Sigurd Schneider <sigurds@chromium.org>
Cr-Commit-Position: refs/heads/master@{#62211}
This refactors the constant pool handling for arm64. The immediate goal
is to allow 32bit compressed pointers in the pool. The mediate goal is
to unify the implementation with the arm constant pool, which will be
done in a follow-up CL.
Bug: v8:8054
Change-Id: I74db4245e5e1025f2e4de4144090fa4ce25883ab
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1645316
Reviewed-by: Michael Starzinger <mstarzinger@chromium.org>
Reviewed-by: Jakob Gruber <jgruber@chromium.org>
Commit-Queue: Sigurd Schneider <sigurds@chromium.org>
Cr-Commit-Position: refs/heads/master@{#62209}
This is pre-work for a refactoring that changes how heap objects
are handled in the assembler: Currently, we put the handle location in
the constant pool, and replace these with the actual heap object when
we copy the code from the assembler's buffer to the heap.
In the future, we will put a small index in the constant pool, which
will ultimately enable 32bit constant pool slots for compressed heap
objects. This small index will be fixed up when we copy the code to
the heap.
This CL makes the assembler tests copy the code to the heap, which
ensures that the fix-up phase is actually run.
Change-Id: I80cd69dc57414a3bd0a27f8d558616aadcae05a2
Bug: v8:7703
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1647166
Commit-Queue: Sigurd Schneider <sigurds@chromium.org>
Reviewed-by: Ulan Degenbaev <ulan@chromium.org>
Cr-Commit-Position: refs/heads/master@{#62112}
This fixes the assembler tests on Windows arm64, that were previously crashing.
Bug: v8:9235
Change-Id: I616e6eb00bb71d70195787f80f8b54a9ae33abe6
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1634919
Commit-Queue: Martyn Capewell <martyn.capewell@arm.com>
Reviewed-by: Jakob Gruber <jgruber@chromium.org>
Cr-Commit-Position: refs/heads/master@{#61981}
