similar to: [PATCH] xenbus_dev: add missing error check to watch handling

Dear xen-devel, There''s a race condition in xenfs (xenstore driver) that causes userspace utility xenstore-watch to crash. Normally, the userspace process gets an "OK" from xenfs and then the watch fires immediately after. Occasionally, this happens the other way around: the watch fires before the driver sends "OK", which confuses the xenstore-watch client. It seems

So far only the watch path was checked to be zero terminated, while the watch token was merely assumed to be. Additionally, none of the three associated memory allocations got checked for being successful. Signed-off-by: Jan Beulich <jbeulich@suse.com> --- a/drivers/xen/xenbus/xenbus_dev.c +++ b/drivers/xen/xenbus/xenbus_dev.c @@ -269,18 +269,24 @@ static ssize_t xenbus_dev_write(struct f

Hello, This patch allows userspace tools on a domU to setup a watch on the xenstore. It does this by intercepting XS_WATCH requests written to /proc/xen/xenbus and then re-submitting the request to the in-kernel xenstore interface, in linux-2.6-xen-sparse/drivers/xen/xenbus/xenbus_xs.c. When a callback occurs, an in-kernel function is invoked, which then reconstructs a response in the format

Hi guys, I updated the latest xen hg repository, and found it failed to compile, because of an missing head file xenbus_dev.h. _______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel

# HG changeset patch # User Rusty Russell <rusty@rustcorp.com.au> # Node ID b0de1894df67ac7c7d905bf61cdf0210b42752cc # Parent ba5d5bd28edf8bce89bdf9fc64047ee4f1dceded Xenstore watch rework Change watches to all fire simultaneously, removing priority argument. Watches no longer fired back to connection/domain which caused event. Fix up testsuite to match Use state enum, rather than return

While on entry to xs_unwatch, there may be an event on its way from xenstored (eg in the ring or in the local kernel), all such events will definitely come before the reply to the unwatch command. So at the point where the unwatch reply has been processed (after xs_talkv), any such now-deleted watch events will definitely have made it to libxenstore''s queue where we can remove them. As

XenStore puts in queued watch events via a thread and notifies the user. Sometimes xs_unwatch is called before all related message is read. The use case is non-threaded libevent, we have two event A and B: - Event A will destroy something and call xs_unwatch; - Event B is used to notify that a node has changed in XenStore. As the event is called one by one, event A can be handled before

On 12/18/2011 11:55 PM, Stephen Rothwell wrote: > Hi all, > > Changes since 20111216: drivers/xen/xenbus/xenbus_dev_frontend.c: In function 'xenbus_init': drivers/xen/xenbus/xenbus_dev_frontend.c:609:2: error: implicit declaration of function 'xen_domain' Full randconfig file is attached. -- ~Randy *** Remember to use Documentation/SubmitChecklist when testing your

On 12/18/2011 11:55 PM, Stephen Rothwell wrote: > Hi all, > > Changes since 20111216: drivers/xen/xenbus/xenbus_dev_frontend.c: In function 'xenbus_init': drivers/xen/xenbus/xenbus_dev_frontend.c:609:2: error: implicit declaration of function 'xen_domain' Full randconfig file is attached. -- ~Randy *** Remember to use Documentation/SubmitChecklist when testing your

It seems that xenstore watches do not work from domU userspace (they obviously do for drivers). In some preliminary debugging of this, there are two aspects of the failure: not sending the watch to xenstore and not handling watch events for userspace watches. The former can be easily fixed by adding the missing ''case XS_WATCH:'' to xenbus_dev_write() in xenbus_dev.c. The

XenStore puts in queued watch events via a thread and notifies the user. Sometimes xs_unwatch is called before all related message is read. The use case is non-threaded libevent, we have two event A and B: - Event A will destroy something and call xs_unwatch; - Event B is used to notify that a node has changed in XenStore. As the event is called one by one, event A can be handled before

XenStore puts in queued watch events via a thread and notifies the user. Sometimes xs_unwatch is called before all related message is read. The use case is non-threaded libevent, we have two event A and B: - Event A will destroy something and call xs_unwatch; - Event B is used to notify that a node has changed in XenStore. As the event is called one by one, event A can be handled before

add xenbus token support and separate watch event queues Signed-off-by: Samuel Thibault <samuel.thibault@eu.citrix.com> # HG changeset patch # User Samuel Thibault <samuel.thibault@eu.citrix.com> # Date 1200582503 0 # Node ID 112c1180720c87a0bfa99142399e3ef3f79f1711 # Parent 33d1c7f9217a1c915afe01fe822811ce69b3a90f add xenbus token support and separate watch event queues diff -r

Hi, I was reading some code and had a question in xenbus_file_write() drivers/xen/xenbus/xenbus_dev_frontend.c 461 if ((len + u->len) > sizeof(u->u.buffer)) { ^^^^^^^^^^^^ Can this addition overflow? Should the test be something like: if (len > sizeof(u->u.buffer) || len + u->len > sizeof(u->u.buffer)) { 462 /* On

Hi, I was reading some code and had a question in xenbus_file_write() drivers/xen/xenbus/xenbus_dev_frontend.c 461 if ((len + u->len) > sizeof(u->u.buffer)) { ^^^^^^^^^^^^ Can this addition overflow? Should the test be something like: if (len > sizeof(u->u.buffer) || len + u->len > sizeof(u->u.buffer)) { 462 /* On

Hi, I was reading some code and had a question in xenbus_file_write() drivers/xen/xenbus/xenbus_dev_frontend.c 461 if ((len + u->len) > sizeof(u->u.buffer)) { ^^^^^^^^^^^^ Can this addition overflow? Should the test be something like: if (len > sizeof(u->u.buffer) || len + u->len > sizeof(u->u.buffer)) { 462 /* On

A pair of patches to the xenstore library that: 1) blocks signal delivery before creating the watch wakeup thread 2) reduces the stack size of the watch wakeup thread.

Hilko, forwarding this to the mailing list. Please post patches over there. Rich. ----- Forwarded message ----- Date: Wed, 21 Jul 2010 17:09:53 +0200 From: Hilko Bengen Subject: patch: non-ASCII characters in node names Hi Richard, I was a little bit surprised when a colleague claimed that key and value names in the registry could contain non-ASCII characters. I created keys and values

libxml2-utils.c local_string_to_utf8() function could easily be reused in other places. This commit extracts it with a new parameter to allow giving the encoding of the input string and publishes it in guestfs-utils.h as guestfs_int_string_to_utf8() --- common/utils/guestfs-utils.h | 11 +++++++ common/utils/libxml2-utils.c | 69 +------------------------------------------- common/utils/utils.c

https://bugzilla.samba.org/show_bug.cgi?id=9798 Summary: rsync crash with SIGSEGV when read time out happens Product: rsync Version: 3.0.2 Platform: All OS/Version: All Status: NEW Severity: critical Priority: P5 Component: core AssignedTo: wayned at samba.org ReportedBy: vijunag at