On Thu, 7 Dec 2000, Alan Cox wrote:
> Andreas is looking at a slightly older kernel, and was right for that. Every
> caller to daemonize either then did the file stuff or needed to and forgot
> so I fixed daemonize
I think, there ist still a small bug.
(This time I even checked 2.4.0-test12-pre8
> In sched.c, function daemonize, line 1216 you call exit_mm.
Yep
> time, it has to segvault. If I am not wrong at this point CLONE_VM simply
> has to be removed from kernel_thread. The kernel-thread will free his mem
> in daemonize (calling exit_mm) and the user-space-application will free
> th
On Thu, 7 Dec 2000, Tigran Aivazian wrote:
> On Thu, 7 Dec 2000, Andreas Klein wrote:
>
> > hello,
> >
> > I have found a problem in scsi.c which in present in the 2.2 and 2.4
> > series. the scsi error handler thread is created with:
> >
> > kernel_thread((int (*)(void *)) scsi_error_handler,
On Thu, 7 Dec 2000, Alan Cox wrote:
> > > > A proper way to release the references to resources is to call daemonize()
> > > > function from within the kernel thread function, which calls
> > > > exit_fs()/exit_files() internally.
> > >
> > > Nearly correct, the daemonize function does NOT call e
On Thu, 7 Dec 2000, Tigran Aivazian wrote:
> PS, Here it is, to save you time opening kernel/sched.c. The kernel is, of
> course, test12-pre7.
~~~
Before you tell me "it was not so in the earlier versions!" I am tempted
to quote a famous russian proverb
"whosoever remembereth
> > > A proper way to release the references to resources is to call daemonize()
> > > function from within the kernel thread function, which calls
> > > exit_fs()/exit_files() internally.
> >
> > Nearly correct, the daemonize function does NOT call exit_files.
>
> I do not post messages to linu
On Thu, 7 Dec 2000, Andreas Klein wrote:
> > A proper way to release the references to resources is to call daemonize()
> > function from within the kernel thread function, which calls
> > exit_fs()/exit_files() internally.
>
> Nearly correct, the daemonize function does NOT call exit_files.
I d
On Thu, 7 Dec 2000, Andreas Klein wrote:
> hello,
>
> I have found a problem in scsi.c which in present in the 2.2 and 2.4
> series. the scsi error handler thread is created with:
>
> kernel_thread((int (*)(void *)) scsi_error_handler,
> (void *) shpnt, 0);
>
>
hello,
I have found a problem in scsi.c which in present in the 2.2 and 2.4
series. the scsi error handler thread is created with:
kernel_thread((int (*)(void *)) scsi_error_handler,
(void *) shpnt, 0);
This will lead to problems, when you have to umount the file
9 matches
Mail list logo