This is the mail archive of the
libc-alpha@sourceware.org
mailing list for the glibc project.
Re: [patch v1] malloc: fix set_max_fast "impossibly small" value
- From: Carlos O'Donell <carlos at redhat dot com>
- To: Joseph Myers <joseph at codesourcery dot com>
- Cc: DJ Delorie <dj at redhat dot com>, libc-alpha at sourceware dot org
- Date: Mon, 4 Nov 2019 10:07:03 -0500
- Subject: Re: [patch v1] malloc: fix set_max_fast "impossibly small" value
- References: <xnr22tevec.fsf@greed.delorie.com> <c80aa002-7d86-0f72-db64-8e27441b4314@redhat.com> <alpine.DEB.2.21.1910311654030.18779@digraph.polyomino.org.uk>
On 10/31/19 12:56 PM, Joseph Myers wrote:
> On Wed, 30 Oct 2019, Carlos O'Donell wrote:
>
>> Which on i386 is overriden here:
>> sysdeps/i386/malloc-alignment.h:
>> 22 #define MALLOC_ALIGNMENT 16
>>
>> This is because we have to account for 16-byte aligned vector types.
>
> Rather, to account for 16-byte-aligned _Float128 and _Decimal128. Vector
> types are expected to require aligned_alloc, but any type that's a basic
> type as defined in ISO C has a fundamental alignment and so malloc must
> allocate memory suitably aligned for it.
Thanks for the clarification.
--
Cheers,
Carlos.