mirror of https://lore.kernel.org/lkml/
 help / color / mirror / Atom feed
From: Dawson Engler <engler@csl.stanford.edu>
To: linux-kernel@vger.kernel.org
Cc: engler@csl.stanford.edu (Dawson Engler)
Subject: [CHECKER] deadlock in 2.5.62/fs/lockd/svc*.c?
Date: Sat, 22 Mar 2003 02:10:28 -0800 (PST)	[thread overview]
Message-ID: <200303221010.h2MAAS506370@csl.stanford.edu> (raw)

Hi All,

here's a more interesting potential deadlock in  the 
	2.5.62/fs/lockd/svc*.c
code that requires the miscordination of three threads.  It might be
a false because of 
        1. infeasible callchain path or 

        2. the two struct nlm_file.f_sema pointers cannot actually point
	to the same object.

It requires three threads: 
        thread 1: acquires nlm_file_sema then tries to get file->f_sema
        thread 2: acquires file->f_sema and tries to get nlm_host_sema
        thread 3: acquires nlm_host_sema and tries to get nlm_file_sema

Any feedback on this one would be great.

Dawson


BUG: ERROR: 3 thread global-global deadlock.
   <&nlm_file_sema>-><&nlm_host_sema> occurred 4 times
   <&nlm_host_sema>-><&nlm_file_sema> occurred 5 times


Plausible chain
   thread 1: <&nlm_file_sema,struct nlm_file.f_sema>
    depth = 3:
        2.5.62/fs/lockd/svcsubs.c:nlm_traverse_files:218

        down(&nlm_file_sema);

           ->2.5.62/fs/lockd/svcsubs.c:nlm_traverse_files:218
           	->nlm_traverse_files:224
           	->nlm_inspect_file:202
           	->end=fs/lockd/svclock.c:nlmsvc_traverse_blocks:274:down

        	down(&file->f_sema);

Seems like a plausible chain.
   thread 2: <struct nlm_file.f_sema,&nlm_host_sema>
    depth = 4:
        2.5.62/fs/lockd/svclock.c:nlmsvc_lock:309

        /* Lock file against concurrent access */
        down(&file->f_sema);


           ->2.5.62/fs/lockd/svclock.c:nlmsvc_lock:309
           	->nlmsvc_lock:351
           	->nlmsvc_create_block:179
           		->2.5.62/fs/lockd/host.c:nlmclnt_lookup_host:44
           		    ->end=nlm_lookup_host:74:down

        			/* Lock hash table */
        			down(&nlm_host_sema);

Seems reasonable.
  <&nlm_host_sema>-><&nlm_file_sema> =
    depth = 4:
        2.5.62/fs/lockd/host.c:nlm_lookup_host:74
           /* Lock hash table */
           down(&nlm_host_sema);

           ->2.5.62/fs/lockd/host.c:nlm_lookup_host:74
           ->nlm_lookup_host:77
           ->nlm_gc_hosts:319
           ->/u2/engler/mc/oses/linux/linux-2.5.62/fs/lockd/svcsubs.c:nlmsvc_mark_resources:279
           ->end=nlm_traverse_files:218:down
           ->2.5.62/fs/lockd/svcsubs.c:nlm_traverse_files:218

        	down(&nlm_file_sema);



                 reply	other threads:[~2003-03-22  9:59 UTC|newest]

Thread overview: [no followups] expand[flat|nested]  mbox.gz  Atom feed

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=200303221010.h2MAAS506370@csl.stanford.edu \
    --to=engler@csl.stanford.edu \
    --cc=linux-kernel@vger.kernel.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox

all inboxes | Powered by JetHome®