Bug 183508 – Use alignas instead of compiler-specific attributes

• Status: RESOLVED
• Resolution: FIXED
• Component: New Bugs

[Yusuke Suzuki]

Use alignas instead of compiler-specific attributes

[Yusuke Suzuki]

Created attachment 335408 [details]
Patch

[Yusuke Suzuki]

Created attachment 335411 [details]
Patch

[Yusuke Suzuki]

Created attachment 335413 [details]
Patch

[Yusuke Suzuki]

Created attachment 335419 [details]
Patch

[Mark Lam]

Comment on attachment 335419 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=335419&action=review

r=me

> Source/JavaScriptCore/jit/JIT.h:172
> +    // performance fluctuations. Try forcing alignment in an attempt to stabalize this.

nit: /stabalize/stabilize/.  I know this came from an old comment, but might as well fix it.

> Source/WTF/wtf/Gigacage.cpp:35
> +alignas(void*) char g_gigacageBasePtrs[GIGACAGE_BASE_PTRS_SIZE];

This disagrees with the bmalloc definition which has it as alignas(GIGACAGE_BASE_PTRS_SIZE).  Unless you have a reason to not do so, let's make it consistent with the bmalloc one.

> Source/WTF/wtf/Gigacage.h:37
> +alignas(void*) extern WTF_EXPORTDATA char g_gigacageBasePtrs[GIGACAGE_BASE_PTRS_SIZE];

Ditto.

[Yusuke Suzuki]

Comment on attachment 335419 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=335419&action=review

>> Source/JavaScriptCore/jit/JIT.h:172
>> +    // performance fluctuations. Try forcing alignment in an attempt to stabalize this.
> 
> nit: /stabalize/stabilize/.  I know this came from an old comment, but might as well fix it.

Nice, fixed.

>> Source/WTF/wtf/Gigacage.cpp:35
>> +alignas(void*) char g_gigacageBasePtrs[GIGACAGE_BASE_PTRS_SIZE];
> 
> This disagrees with the bmalloc definition which has it as alignas(GIGACAGE_BASE_PTRS_SIZE).  Unless you have a reason to not do so, let's make it consistent with the bmalloc one.

OK, changed.

>> Source/WTF/wtf/Gigacage.h:37
>> +alignas(void*) extern WTF_EXPORTDATA char g_gigacageBasePtrs[GIGACAGE_BASE_PTRS_SIZE];
> 
> Ditto.

Fixed.

[Yusuke Suzuki]

Committed r230380: <https://trac.webkit.org/changeset/230380>

[Radar WebKit Bug Importer]

<rdar://problem/39267802>

[Yusuke Suzuki]

Comment on attachment 335419 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=335419&action=review

>>> Source/WTF/wtf/Gigacage.cpp:35
>>> +alignas(void*) char g_gigacageBasePtrs[GIGACAGE_BASE_PTRS_SIZE];
>> 
>> This disagrees with the bmalloc definition which has it as alignas(GIGACAGE_BASE_PTRS_SIZE).  Unless you have a reason to not do so, let's make it consistent with the bmalloc one.
> 
> OK, changed.

I've reverted this part to void* since some of the compilers do not support such a large alignment.
Here, what we want is pointer size alignment.

>>> Source/WTF/wtf/Gigacage.h:37
>>> +alignas(void*) extern WTF_EXPORTDATA char g_gigacageBasePtrs[GIGACAGE_BASE_PTRS_SIZE];
>> 
>> Ditto.
> 
> Fixed.

Ditto.

[Yusuke Suzuki]

Committed r230381: <https://trac.webkit.org/changeset/230381>

[Yusuke Suzuki]

Committed r230388: <https://trac.webkit.org/changeset/230388>

[Yusuke Suzuki]

Hmmm, we enabled warnings on MSVC, which easily leads to build failures.
I'll just disable JIT alignas for MSVC.

[Yusuke Suzuki]

Reopening to attach new patch.

[Yusuke Suzuki]

Created attachment 337481 [details]
Patch

Attempt to fix

[Yusuke Suzuki]

Committed r230403: <https://trac.webkit.org/changeset/230403>

[youenn fablet]

Reopening to attach new patch.

[youenn fablet]

Created attachment 337517 [details]
Patch

[youenn fablet]

Wrong bug id

[Mark Lam]

*** Bug 183617 has been marked as a duplicate of this bug. ***