All of lore.kernel.org
 help / color / mirror / Atom feed
From: "Marek Marczykowski-Górecki" <marmarek@invisiblethingslab.com>
To: xen-devel <xen-devel@lists.xen.org>
Cc: Boris Ostrovsky <boris.ostrovsky@oracle.com>,
	David Vrabel <david.vrabel@citrix.com>
Subject: Deadlock in /proc/xen/xenbus watch+read on 3.17+ (maybe earlier)
Date: Thu, 19 Mar 2015 02:19:11 +0100	[thread overview]
Message-ID: <20150319011911.GA29029@mail-itl> (raw)


[-- Attachment #1.1: Type: text/plain, Size: 2358 bytes --]

Hi,

I've hit some deadlock in kernel xenstore client exposed via
/proc/xen/xenbus. Steps to reproduce are simple:
int main() {
	struct xs_handle *xs;
	xs = xs_open(0);
	xs_watch(xs, "domid", "token");
	xs_read(xs, 0, "name", NULL);
	return 0;
}

xs_watch internally creates new thread, which uses read to wait for the
watch. And in the same time, the program tries to read some value,
but actually it hangs at sending the command (before even sending a path to be
read). Strace gives this (simplified for readability):
[pid  2494] write(3, "\4\0\0\0\0\0\0\0\0\0\0\0\f\0\0\0", 160 = 16
[pid  2494] write(3, "domid\0", 6)      = 6
[pid  2494] write(3, "token\0", 6)      = 6
[pid  2495] read(3,  <unfinished ...>
[pid  2494] futex(0x71c0d4, FUTEX_WAIT_PRIVATE, 1, NULL <unfinished ...>
[pid  2495] <... read resumed>
"\17\0\0\0\377\377\377\377\220~\255\27\f\0\0\0", 16) = 16
[pid  2495] read(3, "domid\0token\0", 12) = 12
[pid  2495] read(3, "\4\0\0\0\0\0\0\0\0\0\0\0\3\0\0\0", 16) = 16
[pid  2495] read(3, "OK\0", 3)          = 3
[pid  2495] futex(0x71c0d4, FUTEX_WAKE_OP_PRIVATE, 1, 1, 0x71c0d0,
{FUTEX_OP_SET, 0, FUTEX_OP_CMP_GT, 1} <unfinished ...>
[pid  2494] <... futex resumed> )       = 0
[pid  2495] <... futex resumed> )       = 1
[pid  2494] futex(0x71c0a8, FUTEX_WAIT_PRIVATE, 2, NULL <unfinished ...>
[pid  2495] futex(0x71c0a8, FUTEX_WAKE_PRIVATE, 1 <unfinished ...>
[pid  2494] <... futex resumed> )       = -1 EAGAIN (Resource
temporarily unavailable)
[pid  2495] <... futex resumed> )       = 0
[pid  2494] futex(0x71c0a8, FUTEX_WAKE_PRIVATE, 1 <unfinished ...>
[pid  2495] read(3,  <unfinished ...>
[pid  2494] <... futex resumed> )       = 0
[pid  2494] rt_sigaction(SIGPIPE, {SIG_DFL, [], SA_RESTORER,
0x7fc78c1488f0}, NULL, 8) = 0
[pid  2494] rt_sigaction(SIGPIPE, {SIG_IGN, [], SA_RESTORER,
0x7fc78c1488f0}, {SIG_DFL, [], SA_RESTORER, 0x7fc78c1488f0}, 8) = 0
[pid  2494] write(3, "\2\0\0\0\0\0\0\0\0\0\0\0\5\0\0\0", 16

And thats all - 2494 is waiting on write, 2495 is waiting on read.

On 3.12.x it is working. On 3.17.0 and 3.18.7 it is broken. I haven't
checked versions in the middle.

Any ideas?

-- 
Best Regards,
Marek Marczykowski-Górecki
Invisible Things Lab
A: Because it messes up the order in which people normally read text.
Q: Why is top-posting such a bad thing?

[-- Attachment #1.2: Type: application/pgp-signature, Size: 473 bytes --]

[-- Attachment #2: Type: text/plain, Size: 126 bytes --]

_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xen.org
http://lists.xen.org/xen-devel

             reply	other threads:[~2015-03-19  1:19 UTC|newest]

Thread overview: 14+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2015-03-19  1:19 Marek Marczykowski-Górecki [this message]
2015-03-19 10:38 ` Deadlock in /proc/xen/xenbus watch+read on 3.17+ (maybe earlier) Ian Campbell
2015-03-19 12:10   ` Iurii Konovalenko
2015-03-19 12:32     ` [Embedded-pv-devel] " Vitaly Chernooky
2015-03-19 13:00     ` David Vrabel
2015-03-19 13:10       ` Vitaly Chernooky
2015-03-20  4:04         ` Marek Marczykowski-Górecki
2015-03-20  9:58           ` Vitaly Chernooky
2015-03-20 10:08           ` Vitaly Chernooky
2015-03-20 10:38             ` Vitaly Chernooky
2015-03-20 10:46               ` David Vrabel
2015-03-20 11:39                 ` Vitaly Chernooky
2015-03-20 10:41             ` Marek Marczykowski-Górecki
2015-03-19 11:10 ` David Vrabel

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=20150319011911.GA29029@mail-itl \
    --to=marmarek@invisiblethingslab.com \
    --cc=boris.ostrovsky@oracle.com \
    --cc=david.vrabel@citrix.com \
    --cc=xen-devel@lists.xen.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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.