linux-kernel - Re: [PATCH 2.6.27-rc5 resubmit] Fix itimer/many thread hang.

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date:	Thu, 18 Sep 2008 12:23:53 +0200
From:	Ingo Molnar <mingo@...e.hu>
To:	Frank Mayhar <fmayhar@...gle.com>
Cc:	linux-kernel <linux-kernel@...r.kernel.org>,
	Roland McGrath <roland@...hat.com>,
	Thomas Gleixner <tglx@...utronix.de>,
	Alexey Dobriyan <adobriyan@...il.com>,
	Andrew Morton <akpm@...ux-foundation.org>,
	Oleg Nesterov <oleg@...sign.ru>
Subject: Re: [PATCH 2.6.27-rc5 resubmit] Fix itimer/many thread hang.


* Frank Mayhar <fmayhar@...gle.com> wrote:

> I finally finished this.  Yeah, there's a visible size difference, 
> although it's small:
> 
> bobble ~/build/linux-2.6>size vmlinux
>    text    data     bss     dec     hex filename
> 6417840  824160  578416 7820416  775480 vmlinux
> 
> bobble ~/build/up-simplify-tree>size vmlinux
>    text    data     bss     dec     hex filename
> 6417999  824160  578416 7820575  77551f vmlinux
> 
> For a difference of 159 bytes.  Not big, but nonzero.  The delta patch 
> (relative to a tree containing all suggested changes _except_ this 
> simplification) follows:

i'd not worry about those 160 bytes - getting this stuff to work fine is 
far more important. Details around threading seem to be one of the 
slowest converging technological details of Linux.

a silly question:

> +#else /* CONFIG_SMP */
> +
> +/*
> + * Allocate the thread_group_cputime structure appropriately for UP kernels
> + * and fill in the current values of the fields.  Called from copy_signal()
> + * via thread_group_cputime_clone_thread() when adding a second or subsequent
> + * thread to a thread group.  Assumes interrupts are enabled when called.
> + */
> +int thread_group_cputime_alloc(struct task_struct *tsk)
> +{

why is there _any_ assymetry needed between UP and SMP? These days we 
just write straight code for SMP, and UP is just a single-CPU 
special-case of it. _Sometimes_ if it's really worth it we do some UP 
special cases but it's the exception, not the rule.

	Ingo
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/