mirror of https://lore.kernel.org/lkml/
 help / color / mirror / Atom feed
From: Sergey Vlasov <vsu@altlinux.ru>
To: Linus Torvalds <torvalds@osdl.org>
Cc: Harald Welte <laforge@gnumonks.org>,
	linux-usb-devel@lists.sourceforge.net, vendor-sec@lst.de,
	linux-kernel@vger.kernel.org, greg@kroah.com,
	security@linux.kernel.org
Subject: Re: [linux-usb-devel] Re: [Security] [vendor-sec] [BUG/PATCH/RFC] Oops while completing async USB via usbdevio
Date: Tue, 27 Sep 2005 20:00:30 +0400	[thread overview]
Message-ID: <20050927160029.GA20466@master.mivlgu.local> (raw)
In-Reply-To: <Pine.LNX.4.58.0509270746200.3308@g5.osdl.org>

[-- Attachment #1: Type: text/plain, Size: 2290 bytes --]

On Tue, Sep 27, 2005 at 07:53:30AM -0700, Linus Torvalds wrote:
> On Sun, 25 Sep 2005, Harald Welte wrote:
> > 
> > async_completed() calls send_sig_info(), which in turn does a
> > spin_lock(&tasklist_lock) to protect itself from task_struct->sighand
> > from going away.  However, the call to
> > "spin_lock_irqsave(task_struct->sighand->siglock)" causes an oops,
> > because "sighand" has disappeared.
> 
> And the real bug is that you're buggering up the system in the first 
> place.
> 
> You don't save "current". You save "pid", and then you send a signal using 
> that and kill_proc_info(). End of story, bug gone. And it works with 
> threaded programs too, which the old thing didn't work at all with.

And then a process calls USBDEVFS_SUBMITURB and immediately exits; its
pid gets reused by a completely different process (maybe even
root-owned), then the urb completes, and kill_proc_info() sends the
signal to the unsuspecting process.

> I refuse to apply this patch - Greg, don't even _try_ to sneak this in 
> through a git merge. What a horribly broken thing to do: why would USB 
> _ever_ need to know about things like tasklist_lock, and internal signal 
> handling functions and rules like "p->sighand"?

Hmm, then probably send_sig_info() should check for non-NULL
p->sighand after taking tasklist_lock?  Otherwise all uses of
send_sig_info() for non-current tasks are unsafe.

"git grep -w send_sig_info" shows that most callers call
send_sig_info() for the current task, except these:

arch/sparc/kernel/traps.c:      send_sig_info(SIGFPE, &info, fpt);
arch/sparc64/kernel/sys_sunos32.c:      send_sig_info(SIGSYS, &info, current);
drivers/char/drm/drm_irq.c:                     send_sig_info( vbl_sig->info.si_signo, &vbl_sig->info, vbl_sig->task );
drivers/usb/core/devio.c:               send_sig_info(as->signr, &sinfo, as->task);
drivers/usb/core/inode.c:                       send_sig_info(ds->discsignr, &sinfo, ds->disctask);
drivers/usb/gadget/file_storage.c:                      send_sig_info(SIGUSR1, SEND_SIG_FORCED, thread_task);
kernel/signal.c:        return send_sig_info(sig, (void*)(long)(priv != 0), p);

BTW, all other callers of send_sig_info() are under
arch/{alpha,arm,sparc,sparc64}/ - 23 total.

[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]

  reply	other threads:[~2005-09-27 16:00 UTC|newest]

Thread overview: 48+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2005-09-25 15:13 Harald Welte
2005-09-27  8:04 ` [vendor-sec] " Greg KH
2005-09-27  9:13   ` Greg KH
     [not found]     ` <20050927110319.GD1980@piware.de>
2005-09-27 12:22       ` [vendor-sec] " Greg KH
2005-09-27 12:48   ` [vendor-sec] " Christoph Hellwig
2005-09-27 12:57     ` Greg KH
2005-09-27 12:59       ` Christoph Hellwig
2005-09-27 13:09         ` Greg KH
2005-09-27 15:27           ` David Brownell
2005-09-27 14:53 ` [Security] " Linus Torvalds
2005-09-27 16:00   ` Sergey Vlasov [this message]
2005-09-27 16:09     ` [linux-usb-devel] " Linus Torvalds
2005-09-27 16:52       ` Sergey Vlasov
2005-09-27 17:02         ` Linus Torvalds
2005-09-30 10:47           ` Harald Welte
2005-09-30 14:56             ` Linus Torvalds
2005-09-30 18:44               ` Chris Wright
2005-09-30 19:27                 ` Linus Torvalds
2005-09-30 20:38                   ` Chris Wright
2005-09-30 22:08                   ` Harald Welte
2005-09-30 22:16                     ` Linus Torvalds
2005-10-10 17:44                       ` Harald Welte
2005-10-10 18:07                         ` Chris Wright
2005-10-11  9:45                           ` Harald Welte
2005-10-11 23:10                             ` [vendor-sec] " Greg KH
2005-10-11 23:44                               ` Linus Torvalds
2005-10-12  7:24                                 ` Harald Welte
2005-10-13  5:51                             ` Horms
2005-10-11 13:57                           ` Bernd Petrovitsch
2005-10-10 18:19                         ` Linus Torvalds
2005-10-10 22:47                           ` Chris Wright
2005-10-10 20:03                         ` [linux-usb-devel] " Alan Stern
2005-10-11  8:28                           ` Harald Welte
2005-10-11 17:37                           ` Paul Jackson
2005-10-11 17:58                             ` linux-os (Dick Johnson)
2005-10-11 19:13                               ` Alan Stern
2005-10-11 20:02                                 ` [Security] " Alan Cox
2005-09-27 17:20         ` PID reuse safety for userspace apps (Re: [linux-usb-devel] Re: [Security] [vendor-sec] [BUG/PATCH/RFC] Oops while completing async USB via usbdevio) Solar Designer
2005-09-27 20:34           ` Alan Cox
2005-09-27 20:42             ` Linus Torvalds
2005-09-27 21:16               ` Solar Designer
2005-09-27 21:03             ` Solar Designer
2005-09-27 16:58       ` [linux-usb-devel] Re: [Security] [vendor-sec] [BUG/PATCH/RFC] Oops while completing async USB via usbdevio Alan Cox
2005-09-27 16:59         ` Linus Torvalds
2005-09-27 20:35           ` Alan Cox
2005-10-13 23:00         ` Pete Zaitcev
2005-10-13 23:16           ` Linus Torvalds
2005-10-13 23:56             ` Pete Zaitcev

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=20050927160029.GA20466@master.mivlgu.local \
    --to=vsu@altlinux.ru \
    --cc=greg@kroah.com \
    --cc=laforge@gnumonks.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-usb-devel@lists.sourceforge.net \
    --cc=security@linux.kernel.org \
    --cc=torvalds@osdl.org \
    --cc=vendor-sec@lst.de \
    /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®