... get an error without this >> patch. Do you have a testcase? ... (or within a new ipc namespace), Thanks, Manfred. Basically what I ... manage to get an oops with the mmotm I fetched this ... sem_op=1; > sop[0].sem_flg=SEM_UNDO; > > res = semop(id,sop,1 ...

linux-kernel - Serge E. Hallyn - Apr 14 2008 - 17:10

... wrote: > > >> > > >>> Hi, > > >>> > > >>> the attached patch should fix the combination of CLONE_NEWIPC ... object and then call unshare with CLONE_NEWIPC, > > >> then it's your ... semaphore id is globally unique. With > > > namespaces, this is not true ...

linux-kernel - Andrew Morton - Apr 3 2008 - 15:04

... allow sys_unshare() for multithreaded processes with CLONE_NEWIPC > >> and without CLONE_SYSVSEM, then ... threaded process unsharing the ipc namespace. > > Eric Eric, does the following ... 23 -0700 Subject: [PATCH 1/1] ipc namespaces: fix svsem unsharing ...

linux-kernel - Serge E. Hallyn - Apr 3 2008 - 15:44

... If we allow sys_unshare() for multithreaded processes with CLONE_NEWIPC > > >> and without CLONE_SYSVSEM, then we must ... 12:43:23 -0700 > Subject: [PATCH 1/1] ipc namespaces: fix svsem unsharing issue > ... , or it doesn't and we're fine. So I suspect I can ...

linux-kernel - Serge E. Hallyn - Apr 4 2008 - 00:39

... Spraul wrote: >> >>> Hi, >>> >>> the attached patch should fix the combination of CLONE_NEWIPC ... on both semid and the >>> namespace pointer. >>> >> If you start using ... semaphore id is globally unique. With > namespaces, this is not true ...

linux-kernel - Pavel Emelyanov - Apr 1 2008 - 05:44

... wrote: > >> > >>> Hi, > >>> > >>> the attached patch should fix the combination of CLONE_NEWIPC ... de-op semaphores for former task namespace. I think > that destroying this ... a semaphore id is globally unique. With > > namespaces, this is not true ...

linux-kernel - Serge E. Hallyn - Apr 1 2008 - 10:15

... can't get an error without this > patch. Do you have a testcase? The ... lookup_undo: lookup_undo doesn't check the namespace pointer, thus a simple single threaded ... .e. semaphore 100 will be accessed]. With kernel debugging (full slub debugging) ...

linux-kernel - Manfred Spraul - Apr 6 2008 - 11:11

... > the attached patch should fix the combination of CLONE_NEWIPC with > shared sysv ... based on both semid and the > namespace pointer. If you start using any ... into atomic parts, rather than mixing fixes, reworks and cleanups together? > The patch ...

linux-kernel - Pavel Emelyanov - Mar 31 2008 - 03:12

... Spraul wrote: > >> Hi, >> >> the attached patch should fix the combination of CLONE_NEWIPC with ... on both semid and the >> namespace pointer. >> > > If you start using ... a semaphore id is globally unique. With namespaces, this is not true ...

linux-kernel - Manfred Spraul - Mar 31 2008 - 12:14

... we allow sys_unshare() for multithreaded processes with CLONE_NEWIPC >> and without CLONE_SYSVSEM, then we ... for threads. > I think other namespaces developers agree with me. Serge, Suka, ... looks like about a 5 line patch or two. The effect is because ...

linux-kernel - Eric W. Biederman - Apr 1 2008 - 11:25

... we allow sys_unshare() for multithreaded processes with CLONE_NEWIPC >>> and without CLONE_SYSVSEM, then we ... threads. >> I think other namespaces developers agree with me. Serge, Suka, ... like about a 5 line patch or two. > > Probably something like ...

linux-kernel - Manfred Spraul - Apr 3 2008 - 15:40

... to-unprivileged-apps kernel > memory corrupter? If so, we should fix it asap for 2.6.25. And for > 2. ... [snip] > if (!capable(CAP_SYS_ADMIN)) > return -EPERM; > Manfred's patch doesn't come close to applying against the 2.6.26 ...

linux-kernel - Manfred Spraul - Apr 3 2008 - 15:31