This is the mail archive of the libc-help@sourceware.org mailing list for the glibc project.

Index Nav:	[Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav:	[Date Prev] [Date Next]	[Thread Prev] [Thread Next]
Other format:	[Raw text]

Re: pthread_exit behavior from main

From: "Carlos O'Donell" <carlos at systemhalted dot org>
To: Siddhesh Poyarekar <siddhesh dot poyarekar at gmail dot com>
Cc: Godmar Back <godmar at gmail dot com>, Bharath Ramesh <bramesh at vt dot edu>, libc-help at sourceware dot org
Date: Thu, 4 Oct 2012 22:30:10 -0400
Subject: Re: pthread_exit behavior from main
References: <20121004220540.GI3182@vt.edu><CAE2sS1gGmQc=oGe29TQ2VDyoihM90s7Z_K6Zg3LGdTrn0T_75g@mail.gmail.com><CAB4+JY+tUXDQX8+vhJ_S-_Jq_GynNMoKOfMZ3bTgP+4juXh1Fw@mail.gmail.com><CAAHN_R2kQMPY-AZfv+k=Y0QeC=jAdL3j5UjDFbX+=u5qnhkf1w@mail.gmail.com>

On Thu, Oct 4, 2012 at 10:15 PM, Siddhesh Poyarekar
<siddhesh.poyarekar@gmail.com> wrote:
> (Sorry, accidentally did a Reply instead of Reply All)
>
> On 5 October 2012 07:11, Godmar Back <godmar@gmail.com> wrote:
>> Just out of curiosity, why do you madvise it with MADV_DONTNEED rather
>> than munmapping it outright?
>
> That's a performance tweak.  The pthread library caches a specific
> number of stacks so that they can be reused by new threads, thus
> avoiding the need for repeated syscalls.

Right, I'd forgotten all about that.

When we hit the 40MB static limit in allocatestate we call __free_stacks
which *does* call munmap for the stack.

Cheers,
Carlos.

References:
- pthread_exit behavior from main
  - From: Bharath Ramesh
- Re: pthread_exit behavior from main
  - From: Carlos O'Donell
- Re: pthread_exit behavior from main
  - From: Godmar Back
- Re: pthread_exit behavior from main
  - From: Siddhesh Poyarekar

Index Nav:	[Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav:	[Date Prev] [Date Next]	[Thread Prev] [Thread Next]