On Wed, Sep 15, 2004 at 06:11:04 -0000, matrix reloaded wrote: > I have also observed some conditions like this, but this generally happens when > kernel data structures are inconsistent and things like this. Example, if you build a > filesystem and try to insert the module and then mount a device having that fs, then > if doing some read you get a NULL pointer access, because of which the module is > in inconsistent state, then after that running "df" command hangs and you get the > same situation. You can produce conditions like this if you don't return anything in > the read operations for a filesystem's read_super function and there are lot many > ways to get this... Noticed the sunrpc and nfs kernel threads? If a process tries to access a NFS-mounted file and the server is not responding, it get's blocked in D state. If the share was mounted with soft option, the operation will time out after some time. Unfortunately, it's not possible to fix this without rather heavy changes in kernel. > On Tue, 14 Sep 2004 Learner wrote : > >Hi , > > > > How to kill processes hanging in a blocked state as > >shown below :- > > > >100 D root 3810 1 0 78 0 - 355 end > > 06:51 ? 00:00:00 df > > > >The stack trace is as below :- > > > >df D C5AE9DA0 160 3810 1 > >4241 31874 (NOTLB) > >Call Trace: [<f89db2a4>] __rpc_execute [sunrpc] 0x204 > >(0xc5ae9d48) > >[<f89d7596>] rpc_call_setup_Rsmp_f8ecb729 [sunrpc] > >0x46 (0xc5ae9d64) > >[<f89d7479>] rpc_call_sync_Rsmp_f3c0f1eb [sunrpc] 0x69 > >(0xc5ae9d70) > >[<f89d748a>] rpc_call_sync_Rsmp_f3c0f1eb [sunrpc] 0x7a > >(0xc5ae9d90) > >[<f89d7680>] call_reserveresult [sunrpc] 0x0 > >(0xc5ae9de4) > >[<f89da410>] rpc_run_timer [sunrpc] 0x0 (0xc5ae9e04) > >[<f89fa019>] nfs3_proc_statfs [nfs] 0x59 (0xc5ae9e40) > >[<f89eda97>] nfs_statfs [nfs] 0x37 (0xc5ae9e90) > >[<c0144759>] vfs_statfs [kernel] 0x59 (0xc5ae9f2c) > >[<c01447cb>] sys_statfs [kernel] 0x3b (0xc5ae9f44) > >[<c01073e3>] system_call [kernel] 0x33 (0xc5ae9fc0) > > > > Even a kill -9 as root does not eliminate these > >processes . > > > > Also the stack trace shows that it is not > >in "sleep_uninterruptible" . > > > > Any pointers would help . > > > >Regards > > > > > > > >_______________________________ > >Do you Yahoo!? > >Declare Yourself - Register online to vote today! > >http://vote.yahoo.com > > > >-- > >Kernelnewbies: Help each other learn about the Linux kernel. > >Archive: http://mail.nl.linux.org/kernelnewbies/ > >FAQ: http://kernelnewbies.org/faq/ > > ------------------------------------------------------------------------------- Jan 'Bulb' Hudec <bulb@xxxxxx>
Attachment:
signature.asc
Description: Digital signature
