699101 - NO_MAC_JEMALLOC can't work on 10.5, so display a warning when it's specified

Status: RESOLVED FIXED

(Core :: Memory Allocator, defect)

Description

[Justin Lebar (not reading bugmail)]

Patch in a moment.

Comment 1

[Justin Lebar (not reading bugmail)]

 * memalign is unavailable on Leopard, so we can't dynamically do this there.
 * However, we don't use jemalloc on Leopard, so we can ignore this.
 */
#if defined(MOZ_MEMORY_DARWIN) && !defined(__i386__)

So now we *are* using jemalloc on Leopard.

When does MEMALIGN get called?  If MEMALIGN is called and we're not using jemalloc, what the heck are we supposed to do?  Crash?

Comment 2

[Justin Lebar (not reading bugmail)]

I should have given more context in comment 1.  It's

#if defined(MOZ_MEMORY_DARWIN) && !defined(__i386__)
#define DARWIN_ONLY(A) if (!osx_use_jemalloc) { A; }
#else
#define DARWIN_ONLY(A)
#endif

Comment 3

[Justin Lebar (not reading bugmail)]

Attachment: Patch? v1

Comment 4

[Justin Lebar (not reading bugmail)]

So, memalign is apparently called on 10.5.  I guess this means that we can't dynamically turn off jemalloc there.  Not a huge deal; this is only for testing, anyway.

I'll post a patch fixing the comment.

Comment 5

[Justin Lebar (not reading bugmail)]

Attachment: Patch v2

Comment 6

[Justin Lebar (not reading bugmail)]

https://hg.mozilla.org/integration/mozilla-inbound/rev/8a385ef6f650

Comment 7

[Matt Brubeck (:mbrubeck)]

https://hg.mozilla.org/mozilla-central/rev/8a385ef6f650

Comment 8

[u279076]

Is there a testcase QA can use to verify this fix?

Comment 9

[Justin Lebar (not reading bugmail)]

(In reply to Anthony Hughes, Mozilla QA (irc: ashughes) from comment #8)
> Is there a testcase QA can use to verify this fix?

No.