Re: spinlock under fork()?
Re: spinlock under fork()?
- Subject: Re: spinlock under fork()?
- From: Terry Lambert <email@hidden>
- Date: Wed, 17 Sep 2008 19:32:00 -0700
On Sep 17, 2008, at 6:52 PM, Nathan Herring wrote:
Fair enough. What’s the official behavior of malloc_zone_register
supposed to be? If it’s supposed to not handle multiple-
registrations, then I can post a documentation bug to better
document that that’s the behavior and that furthermore it needn’t be
called after malloc_create_zone. If it’s supposed to handle multiple-
registrations (i.e., politely does nothing), then there’s clearly a
bug either in it or somewhere in the Safari-used code that is
revealed by it, and that needs to be radared.
On 9/17/08 6:20 PM, "Brian Mastenbrook" <email@hidden> wrote:
malloc_create_zone calls malloc_zone_register itself. That comment
is a bit confusing, and I'm not sure under what circumstances anyone
would want to call malloc_zone_register themselves. I'm also not
sure why this leads to a spinlock, but clearly something doesn't
like having the same zone in the zone list twice.
Please file a different bug: "malloc_zone_register(0 of an already
registered zone should result in an error".
-- Terry _______________________________________________
Do not post admin requests to the list. They will be ignored.
Darwin-dev mailing list (email@hidden)
Help/Unsubscribe/Update your Subscription:
This email sent to email@hidden