linux-kernel - Re: tasks getting stuck on mmap

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]

Message-ID: <AANLkTik9qUf_F0NDoW2SBjWGHk2=ZdvkLrCORu3gGa2o@mail.gmail.com>
Date:	Tue, 17 Aug 2010 14:08:33 -0700
From:	Linus Torvalds <torvalds@...ux-foundation.org>
To:	Tony Luck <tony.luck@...el.com>
Cc:	David Miller <davem@...emloft.net>,
	Michel Lespinasse <walken@...gle.com>,
	David Howells <dhowells@...hat.com>,
	Andrew Morton <akpm@...ux-foundation.org>,
	linux-kernel@...r.kernel.org
Subject: Re: tasks getting stuck on mmap_sem?

On Tue, Aug 17, 2010 at 1:46 PM, Tony Luck <tony.luck@...el.com> wrote:
>
> Setting up for "git pull" would have been messy - but "git grep" is
> multi-threaded in newer versions of git. So I've set a loop to keep
> running "git grep" over and over.  Load average is above 7, so
> there must be some threading going on, even though I only see
> one "git" process in "top(1)".  Seems to be running fine so far.

I'm not sure how page-fault intensive that 'git grep' thing is - I
think it does mainly open/read/close on the files and probably doesn't
take that many faults once malloc() has created the main heap. In
contrast, the delta compute code (that david hit) is probably going to
do a lot more page faults due to being way more data-intensive (many
more mmap/munmap calls - it's how git tends to like to map git
objects).

So a more fault-intensive - and still threaded and all local - git
load is likely to do something like

   git repack -adf

and do it on something smaller than the whole kernel (git itself is
probably a good test-case, and probably isn't that far off from a
medium-sized kernel "pull" request).

                 Linus
--
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/