[Webkit-unassigned] [Bug 209236] REGRESSION(r249808): [GTK] Crash in JSC Config::permanentlyFreeze() on architecture ppc64el

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=209236

--- Comment #32 from Mark Lam <mark.lam at apple.com> ---
(In reply to Michael Catanzaro from comment #30)
> Hey Mark, how was ConfigSizeToProtect chosen?
> 
> #if !OS(WINDOWS)
> constexpr size_t ConfigSizeToProtect = 16 * KB;
> #else
> constexpr size_t ConfigSizeToProtect = 4 * KB;
> #endif
> 
> Is it supposed to match page size (in which case most Linux architectures
> should be using 4 KB rather than 16 KB)? Or is it desired to be exactly 16
> KB everywhere regardless of page size unless page size is bigger than 16 KB?
> Why is Windows using 4 KB while everything else uses 16 KB?
> 
> Same question applies to the blockSize in MarkedBlock.h. In that case, the
> code is a bit more clear, and I guess the desired block size is min(16 KB,
> page size)?

Look at JSC::Config::ensureSize.  It is padded so that sizeof(JSC::Config) is a multiple of OS page size.  ConfigSizeToProtect is shone to be that rounded up multiple of OS page size.

-- 
You are receiving this mail because:
You are the assignee for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.webkit.org/pipermail/webkit-unassigned/attachments/20200319/3440d784/attachment.htm>