hwclock and df seg fault

hwclock and df seg fault

[Martin C. Barlow]

Hey guys

I have an old SGI indy R4600 and have installed debian testing with
latest linux-mips cvs kernel. I found two problems with the programs
hwclock and df. Apart from that appears to work fine. I have included
their output. I don't know if it is a kernel or package problem. I don't
know if it as something to do with preemtible kernel which I enabled in
kernel. If anyone is interested and wants to see kernel .config, fstab
or anything else I'm happy to oblidge.

Barcelona:/var/log# hwclock
Mar 21 19:11:20 Barcelona kernel: bad: scheduling while atomic!
Mar 21 19:11:20 Barcelona kernel: Call Trace:
Mar 21 19:11:20 Barcelona kernel:  [<88030498>] schedule+0xb10/0xb18
Mar 21 19:11:20 Barcelona kernel:  [<88030490>] schedule+0xb08/0xb18
Mar 21 19:11:20 Barcelona kernel:  [<88031ef0>]
sys_sched_yield+0x154/0x1b4
Mar 21 19:11:20 Barcelona kernel:  [<88031fac>] yield+0x24/0x38
Mar 21 19:11:20 Barcelona kernel:  [<8808a2fc>] coredump_wait+0x3c/0xa4
Mar 21 19:11:20 Barcelona kernel:  [<8805d8e0>] cache_grow+0x274/0x4c0
Mar 21 19:11:20 Barcelona kernel:  [<8808a4b8>] do_coredump+0x154/0x26c
Mar 21 19:11:20 Barcelona kernel:  [<8805dd34>]
cache_alloc_refill+0x208/0x2b8
Mar 21 19:11:20 Barcelona kernel:  [<88007e14>]
indy_r4k_timer_interrupt+0x90/0x98
Mar 21 19:11:20 Barcelona kernel:  [<8801ec14>]
r4k_flush_icache_page+0xf4/0x124
Mar 21 19:11:20 Barcelona kernel:  [<882412f0>] change_floppy+0x94/0x180
Mar 21 19:11:20 Barcelona kernel:  [<88041380>]
__dequeue_signal+0x16c/0x2a0
Mar 21 19:11:20 Barcelona kernel:  [<880413b4>]
__dequeue_signal+0x1a0/0x2a0
Mar 21 19:11:20 Barcelona kernel:  [<8805e0ec>]
kmem_cache_alloc+0x94/0x9c
Mar 21 19:11:20 Barcelona kernel:  [<880414e0>] dequeue_signal+0x2c/0xa8
Mar 21 19:11:20 Barcelona kernel:  [<8802e9b8>] wake_up_state+0x18/0x2c
Mar 21 19:11:20 Barcelona kernel:  [<8804468c>]
get_signal_to_deliver+0x394/0x524
Mar 21 19:11:20 Barcelona kernel:  [<8800cf08>] do_signal+0x44/0xd74
Mar 21 19:11:20 Barcelona kernel:  [<8801dd20>]
do_page_fault+0x280/0x380
Mar 21 19:11:20 Barcelona kernel:  [<88086938>] chrdev_open+0x150/0x2cc
Mar 21 19:11:20 Barcelona kernel:  [<88079108>] dentry_open+0x1a4/0x2d0
Mar 21 19:11:20 Barcelona kernel:  [<88078f90>] dentry_open+0x2c/0x2d0
Mar 21 19:11:20 Barcelona kernel:  [<88079108>] dentry_open+0x1a4/0x2d0
Mar 21 19:11:20 Barcelona kernel:  [<88078f90>] dentry_open+0x2c/0x2d0
Mar 21 19:11:20 Barcelona kernel:  [<8807b554>] __fput+0x118/0x1c0
Mar 21 19:11:20 Barcelona kernel:  [<8807b434>] fput+0x40/0x48
Mar 21 19:11:20 Barcelona kernel:  [<8807972c>] filp_close+0x6c/0xb4
Mar 21 19:11:20 Barcelona kernel:  [<88079718>] filp_close+0x58/0xb4
Mar 21 19:11:20 Barcelona kernel:  [<8800dc70>]
do_notify_resume+0x38/0x54
Mar 21 19:11:20 Barcelona kernel:  [<88008be0>] work_notifysig+0xc/0x14
Mar 21 19:11:20 Barcelona kernel:  [<88011828>] stack_done+0x20/0x3c
Mar 21 19:11:20 Barcelona kernel:  [<88079774>] sys_close+0x0/0x12c
Mar 21 19:11:20 Barcelona kernel:
Mar 21 19:11:20 Barcelona kernel: note: hwclock[369] exited with
preempt_count 2

Barcelona:/var/log# df
Filesystem           1K-blocks      Used Available Use% Mounted on
df: `/': Invalid argument
df: `/proc': Invalid argument
df: `/dev/pts': Invalid argument
df: `/dev/shm': Invalid argument
df: `/sys': Invalid argument
Barcelona:/var/log#

P.S. feel free to ignore me. It does not really worry me that they don't
work, but I thought maybe someone would want to know.

hwclock and df seg fault

[Martin C. Barlow]

Hey guys

I have an old SGI indy R4600 and have installed debian testing with
latest linux-mips cvs kernel. I found two problems with the programs
hwclock and df. Apart from that appears to work fine. I have included
their output. I don't know if it is a kernel or package problem. I don't
know if it as something to do with preemtible kernel which I enabled in
kernel. If anyone is interested and wants to see kernel .config, fstab
or anything else I'm happy to oblidge.

Barcelona:/var/log# hwclock
Mar 21 19:11:20 Barcelona kernel: bad: scheduling while atomic!
Mar 21 19:11:20 Barcelona kernel: Call Trace:
Mar 21 19:11:20 Barcelona kernel:  [<88030498>] schedule+0xb10/0xb18
Mar 21 19:11:20 Barcelona kernel:  [<88030490>] schedule+0xb08/0xb18
Mar 21 19:11:20 Barcelona kernel:  [<88031ef0>]
sys_sched_yield+0x154/0x1b4
Mar 21 19:11:20 Barcelona kernel:  [<88031fac>] yield+0x24/0x38
Mar 21 19:11:20 Barcelona kernel:  [<8808a2fc>] coredump_wait+0x3c/0xa4
Mar 21 19:11:20 Barcelona kernel:  [<8805d8e0>] cache_grow+0x274/0x4c0
Mar 21 19:11:20 Barcelona kernel:  [<8808a4b8>] do_coredump+0x154/0x26c
Mar 21 19:11:20 Barcelona kernel:  [<8805dd34>]
cache_alloc_refill+0x208/0x2b8
Mar 21 19:11:20 Barcelona kernel:  [<88007e14>]
indy_r4k_timer_interrupt+0x90/0x98
Mar 21 19:11:20 Barcelona kernel:  [<8801ec14>]
r4k_flush_icache_page+0xf4/0x124
Mar 21 19:11:20 Barcelona kernel:  [<882412f0>] change_floppy+0x94/0x180
Mar 21 19:11:20 Barcelona kernel:  [<88041380>]
__dequeue_signal+0x16c/0x2a0
Mar 21 19:11:20 Barcelona kernel:  [<880413b4>]
__dequeue_signal+0x1a0/0x2a0
Mar 21 19:11:20 Barcelona kernel:  [<8805e0ec>]
kmem_cache_alloc+0x94/0x9c
Mar 21 19:11:20 Barcelona kernel:  [<880414e0>] dequeue_signal+0x2c/0xa8
Mar 21 19:11:20 Barcelona kernel:  [<8802e9b8>] wake_up_state+0x18/0x2c
Mar 21 19:11:20 Barcelona kernel:  [<8804468c>]
get_signal_to_deliver+0x394/0x524
Mar 21 19:11:20 Barcelona kernel:  [<8800cf08>] do_signal+0x44/0xd74
Mar 21 19:11:20 Barcelona kernel:  [<8801dd20>]
do_page_fault+0x280/0x380
Mar 21 19:11:20 Barcelona kernel:  [<88086938>] chrdev_open+0x150/0x2cc
Mar 21 19:11:20 Barcelona kernel:  [<88079108>] dentry_open+0x1a4/0x2d0
Mar 21 19:11:20 Barcelona kernel:  [<88078f90>] dentry_open+0x2c/0x2d0
Mar 21 19:11:20 Barcelona kernel:  [<88079108>] dentry_open+0x1a4/0x2d0
Mar 21 19:11:20 Barcelona kernel:  [<88078f90>] dentry_open+0x2c/0x2d0
Mar 21 19:11:20 Barcelona kernel:  [<8807b554>] __fput+0x118/0x1c0
Mar 21 19:11:20 Barcelona kernel:  [<8807b434>] fput+0x40/0x48
Mar 21 19:11:20 Barcelona kernel:  [<8807972c>] filp_close+0x6c/0xb4
Mar 21 19:11:20 Barcelona kernel:  [<88079718>] filp_close+0x58/0xb4
Mar 21 19:11:20 Barcelona kernel:  [<8800dc70>]
do_notify_resume+0x38/0x54
Mar 21 19:11:20 Barcelona kernel:  [<88008be0>] work_notifysig+0xc/0x14
Mar 21 19:11:20 Barcelona kernel:  [<88011828>] stack_done+0x20/0x3c
Mar 21 19:11:20 Barcelona kernel:  [<88079774>] sys_close+0x0/0x12c
Mar 21 19:11:20 Barcelona kernel:
Mar 21 19:11:20 Barcelona kernel: note: hwclock[369] exited with
preempt_count 2

Barcelona:/var/log# df
Filesystem           1K-blocks      Used Available Use% Mounted on
df: `/': Invalid argument
df: `/proc': Invalid argument
df: `/dev/pts': Invalid argument
df: `/dev/shm': Invalid argument
df: `/sys': Invalid argument
Barcelona:/var/log#

P.S. feel free to ignore me. It does not really worry me that they don't
work, but I thought maybe someone would want to know.

Re: hwclock and df seg fault

[Ralf Baechle]

On Sat, Mar 20, 2004 at 09:05:40PM +1100, Martin C. Barlow wrote:

> I have an old SGI indy R4600 and have installed debian testing with
> latest linux-mips cvs kernel. I found two problems with the programs
> hwclock and df. Apart from that appears to work fine. I have included
> their output. I don't know if it is a kernel or package problem. I don't
> know if it as something to do with preemtible kernel which I enabled in
> kernel. If anyone is interested and wants to see kernel .config, fstab
> or anything else I'm happy to oblidge.

Standard flame - what kernel version?

I checked in the last fixes for the preemptible kernel less than two days
ago so if your kernel is older than that it's time to update :-)

  Ralf

Re: hwclock and df seg fault

[Thiemo Seufer]

Martin C. Barlow wrote:
[snip]
> Barcelona:/var/log# hwclock
> Mar 21 19:11:20 Barcelona kernel: bad: scheduling while atomic!
[snip]
> Mar 21 19:11:20 Barcelona kernel: note: hwclock[369] exited with
> preempt_count 2

So this was with CONFIG_PREEMPT, I guess. Does it happen also without
that?


Thiemo

RE: hwclock and df seg fault

[Martin C. Barlow]

I downloaded the kernel from cvs today (less than 6 hours ago). 
I will recompile without preemptible and let you know the result. 
The version is 2.6.4

marty

-----Original Message-----
From: Ralf Baechle [mailto:ralf@linux-mips.org] 
Sent: Saturday, 20 March 2004 11:22 PM
To: Martin C. Barlow
Cc: linux-mips@linux-mips.org
Subject: Re: hwclock and df seg fault


On Sat, Mar 20, 2004 at 09:05:40PM +1100, Martin C. Barlow wrote:

> I have an old SGI indy R4600 and have installed debian testing with 
> latest linux-mips cvs kernel. I found two problems with the programs 
> hwclock and df. Apart from that appears to work fine. I have included 
> their output. I don't know if it is a kernel or package problem. I 
> don't know if it as something to do with preemtible kernel which I 
> enabled in kernel. If anyone is interested and wants to see kernel 
> .config, fstab or anything else I'm happy to oblidge.

Standard flame - what kernel version?

I checked in the last fixes for the preemptible kernel less than two
days ago so if your kernel is older than that it's time to update :-)

  Ralf

RE: hwclock and df seg fault

[Martin C. Barlow]

I downloaded the kernel from cvs today (less than 6 hours ago). 
I will recompile without preemptible and let you know the result. 
The version is 2.6.4

marty

-----Original Message-----
From: Ralf Baechle [mailto:ralf@linux-mips.org] 
Sent: Saturday, 20 March 2004 11:22 PM
To: Martin C. Barlow
Cc: linux-mips@linux-mips.org
Subject: Re: hwclock and df seg fault


On Sat, Mar 20, 2004 at 09:05:40PM +1100, Martin C. Barlow wrote:

> I have an old SGI indy R4600 and have installed debian testing with 
> latest linux-mips cvs kernel. I found two problems with the programs 
> hwclock and df. Apart from that appears to work fine. I have included 
> their output. I don't know if it is a kernel or package problem. I 
> don't know if it as something to do with preemtible kernel which I 
> enabled in kernel. If anyone is interested and wants to see kernel 
> .config, fstab or anything else I'm happy to oblidge.

Standard flame - what kernel version?

I checked in the last fixes for the preemptible kernel less than two
days ago so if your kernel is older than that it's time to update :-)

  Ralf

RE: hwclock and df seg fault

[Martin C. Barlow]

Thiemo

With PREEMP turned off, the hwclock command works again. Looks like the
new scheduler may have a little problem.
The df problem is still there. This may be just a user space
mis-configuration error. I'll see if I can find the problem.
Let me know if you need any dumps.

marty

-----Original Message-----
From: Thiemo Seufer [mailto:ica2_ts@csv.ica.uni-stuttgart.de] 
Sent: Saturday, 20 March 2004 11:22 PM
To: Martin C. Barlow
Cc: linux-mips@linux-mips.org
Subject: Re: hwclock and df seg fault


Martin C. Barlow wrote:
[snip]
> Barcelona:/var/log# hwclock
> Mar 21 19:11:20 Barcelona kernel: bad: scheduling while atomic!
[snip]
> Mar 21 19:11:20 Barcelona kernel: note: hwclock[369] exited with 
> preempt_count 2

So this was with CONFIG_PREEMPT, I guess. Does it happen also without
that?


Thiemo

RE: hwclock and df seg fault

[Martin C. Barlow]

Thiemo

With PREEMP turned off, the hwclock command works again. Looks like the
new scheduler may have a little problem.
The df problem is still there. This may be just a user space
mis-configuration error. I'll see if I can find the problem.
Let me know if you need any dumps.

marty

-----Original Message-----
From: Thiemo Seufer [mailto:ica2_ts@csv.ica.uni-stuttgart.de] 
Sent: Saturday, 20 March 2004 11:22 PM
To: Martin C. Barlow
Cc: linux-mips@linux-mips.org
Subject: Re: hwclock and df seg fault


Martin C. Barlow wrote:
[snip]
> Barcelona:/var/log# hwclock
> Mar 21 19:11:20 Barcelona kernel: bad: scheduling while atomic!
[snip]
> Mar 21 19:11:20 Barcelona kernel: note: hwclock[369] exited with 
> preempt_count 2

So this was with CONFIG_PREEMPT, I guess. Does it happen also without
that?


Thiemo

RE: hwclock and df seg fault

[Martin C. Barlow]

I did a search for this problem and found that the sparc64 guys had a
similar problem that they solved. I don't know if the problem is
similar.
http://lists.debian.org/debian-sparc/2003/debian-sparc-200311/msg00068.h
tml

They solved it by analysing an strace.  I did the same. Looking at it,
it looks like some linking is failing and then later on SYS_4255(), the
critical function is failing.

Hope that helps someone to understand.

marty

Barcelona:~# strace df
execve("/bin/df", ["df"], [/* 16 vars */]) = 0
uname({sys="Linux", node="Barcelona", ...}) = 0
brk(0)                                  = 0x10001000
access("/etc/ld.so.nohwcap", F_OK)      = -1 ENOENT (No such file or
directory)
old_mmap(NULL, 4096, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS,
-1, 0) = 0x2aac2000
open("/etc/ld.so.preload", O_RDONLY)    = -1 ENOENT (No such file or
directory)
open("/etc/ld.so.cache", O_RDONLY)      = 3
fstat64(3, {st_mode=S_IFREG|0644, st_size=9507, ...}) = 0
old_mmap(NULL, 9507, PROT_READ, MAP_PRIVATE, 3, 0) = 0x2aac4000
close(3)                                = 0
access("/etc/ld.so.nohwcap", F_OK)      = -1 ENOENT (No such file or
directory)
open("/lib/libc.so.6", O_RDONLY)        = 3
read(3, "\177ELF\1\2\1\0\0\0\0\0\0\0\0\0\0\3\0\10\0\0\0\1\0\1PD"...,
512) = 512
lseek(3, 588, SEEK_SET)                 = 588
read(3, "\0\0\0\4\0\0\0\20\0\0\0\1GNU\0\0\0\0\0\0\0\0\2\0\0\0\4"..., 32)
= 32
fstat64(3, {st_mode=S_IFREG|0644, st_size=1692012, ...}) = 0
old_mmap(NULL, 1872896, PROT_READ|PROT_EXEC, MAP_PRIVATE, 3, 0) =
0x2ab02000
mprotect(0x2ac7f000, 312320, PROT_NONE) = 0
old_mmap(0x2acbe000, 45056, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_FIXED,
3, 0x17c000) = 0x2acbe000
old_mmap(0x2acc9000, 9216, PROT_READ|PROT_WRITE,
MAP_PRIVATE|MAP_FIXED|MAP_ANONYMOUS, -1, 0) = 0x2acc9000
close(3)                                = 0
munmap(0x2aac4000, 9507)                = 0
brk(0)                                  = 0x10001000
brk(0x10022000)                         = 0x10022000
brk(0)                                  = 0x10022000
open("/etc/mtab", O_RDONLY)             = 3
fstat64(3, {st_mode=S_IFREG|0644, st_size=162, ...}) = 0
old_mmap(NULL, 4096, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS,
-1, 0) = 0x2aac3000
read(3, "/dev/sda1 / ext3 rw,errors=remou"..., 4096) = 162
read(3, "", 4096)                       = 0
close(3)                                = 0
munmap(0x2aac3000, 4096)                = 0
fstat64(1, {st_mode=S_IFCHR|0600, st_rdev=makedev(136, 0), ...}) = 0
old_mmap(NULL, 4096, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS,
-1, 0) = 0x2aac3000
write(1, "Filesystem           1K-blocks  "..., 67Filesystem
1K-blocks      Used Available Use% Mounted on
) = 67
SYS_4255()                              = -1 EINVAL (Invalid argument)
write(2, "df: ", 4df: )                     = 4
write(2, "`/\'", 3`/')                     = 3
write(2, ": Invalid argument", 18: Invalid argument)      = 18
write(2, "\n", 1
)                       = 1
SYS_4255()                              = -1 EINVAL (Invalid argument)
write(2, "df: ", 4df: )                     = 4
write(2, "`/proc\'", 7`/proc')                 = 7
write(2, ": Invalid argument", 18: Invalid argument)      = 18
write(2, "\n", 1
)                       = 1
SYS_4255()                              = -1 EINVAL (Invalid argument)
write(2, "df: ", 4df: )                     = 4
write(2, "`/sys\'", 6`/sys')                  = 6
write(2, ": Invalid argument", 18: Invalid argument)      = 18
write(2, "\n", 1
)                       = 1


Barcelona:~# df
Filesystem           1K-blocks      Used Available Use% Mounted on
df: `/': Invalid argument
df: `/proc': Invalid argument
df: `/sys': Invalid argument
df: `/dev/pts': Invalid argument
df: `/dev/shm': Invalid argument

Barcelona:~# uname -a
Linux Barcelona 2.6.4 #3 Mon Mar 22 00:14:42 EST 2004 mips GNU/Linux

Barcelona:~# cat /proc/version
Linux version 2.6.4 (root@Barcelona) (gcc version 3.3.3 (Debian)) #3 Mon
Mar 22 00:14:42 EST 2004

Barcelona:~# dpkg -l coreutils libc6
ii  coreutils                 5.0.91-2                  The GNU core
utilities
ii  libc6                     2.3.2.ds1-11              GNU C Library:
Shared libraries and Timezone data

-----Original Message-----
From: Ralf Baechle [mailto:ralf@linux-mips.org] 
Sent: Saturday, 20 March 2004 11:22 PM
To: Martin C. Barlow
Cc: linux-mips@linux-mips.org
Subject: Re: hwclock and df seg fault

I checked in the last fixes for the preemptible kernel less than two
days ago so if your kernel is older than that it's time to update :-)

  Ralf

RE: hwclock and df seg fault

[Martin C. Barlow]

I did a search for this problem and found that the sparc64 guys had a
similar problem that they solved. I don't know if the problem is
similar.
http://lists.debian.org/debian-sparc/2003/debian-sparc-200311/msg00068.h
tml

They solved it by analysing an strace.  I did the same. Looking at it,
it looks like some linking is failing and then later on SYS_4255(), the
critical function is failing.

Hope that helps someone to understand.

marty

Barcelona:~# strace df
execve("/bin/df", ["df"], [/* 16 vars */]) = 0
uname({sys="Linux", node="Barcelona", ...}) = 0
brk(0)                                  = 0x10001000
access("/etc/ld.so.nohwcap", F_OK)      = -1 ENOENT (No such file or
directory)
old_mmap(NULL, 4096, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS,
-1, 0) = 0x2aac2000
open("/etc/ld.so.preload", O_RDONLY)    = -1 ENOENT (No such file or
directory)
open("/etc/ld.so.cache", O_RDONLY)      = 3
fstat64(3, {st_mode=S_IFREG|0644, st_size=9507, ...}) = 0
old_mmap(NULL, 9507, PROT_READ, MAP_PRIVATE, 3, 0) = 0x2aac4000
close(3)                                = 0
access("/etc/ld.so.nohwcap", F_OK)      = -1 ENOENT (No such file or
directory)
open("/lib/libc.so.6", O_RDONLY)        = 3
read(3, "\177ELF\1\2\1\0\0\0\0\0\0\0\0\0\0\3\0\10\0\0\0\1\0\1PD"...,
512) = 512
lseek(3, 588, SEEK_SET)                 = 588
read(3, "\0\0\0\4\0\0\0\20\0\0\0\1GNU\0\0\0\0\0\0\0\0\2\0\0\0\4"..., 32)
= 32
fstat64(3, {st_mode=S_IFREG|0644, st_size=1692012, ...}) = 0
old_mmap(NULL, 1872896, PROT_READ|PROT_EXEC, MAP_PRIVATE, 3, 0) =
0x2ab02000
mprotect(0x2ac7f000, 312320, PROT_NONE) = 0
old_mmap(0x2acbe000, 45056, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_FIXED,
3, 0x17c000) = 0x2acbe000
old_mmap(0x2acc9000, 9216, PROT_READ|PROT_WRITE,
MAP_PRIVATE|MAP_FIXED|MAP_ANONYMOUS, -1, 0) = 0x2acc9000
close(3)                                = 0
munmap(0x2aac4000, 9507)                = 0
brk(0)                                  = 0x10001000
brk(0x10022000)                         = 0x10022000
brk(0)                                  = 0x10022000
open("/etc/mtab", O_RDONLY)             = 3
fstat64(3, {st_mode=S_IFREG|0644, st_size=162, ...}) = 0
old_mmap(NULL, 4096, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS,
-1, 0) = 0x2aac3000
read(3, "/dev/sda1 / ext3 rw,errors=remou"..., 4096) = 162
read(3, "", 4096)                       = 0
close(3)                                = 0
munmap(0x2aac3000, 4096)                = 0
fstat64(1, {st_mode=S_IFCHR|0600, st_rdev=makedev(136, 0), ...}) = 0
old_mmap(NULL, 4096, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS,
-1, 0) = 0x2aac3000
write(1, "Filesystem           1K-blocks  "..., 67Filesystem
1K-blocks      Used Available Use% Mounted on
) = 67
SYS_4255()                              = -1 EINVAL (Invalid argument)
write(2, "df: ", 4df: )                     = 4
write(2, "`/\'", 3`/')                     = 3
write(2, ": Invalid argument", 18: Invalid argument)      = 18
write(2, "\n", 1
)                       = 1
SYS_4255()                              = -1 EINVAL (Invalid argument)
write(2, "df: ", 4df: )                     = 4
write(2, "`/proc\'", 7`/proc')                 = 7
write(2, ": Invalid argument", 18: Invalid argument)      = 18
write(2, "\n", 1
)                       = 1
SYS_4255()                              = -1 EINVAL (Invalid argument)
write(2, "df: ", 4df: )                     = 4
write(2, "`/sys\'", 6`/sys')                  = 6
write(2, ": Invalid argument", 18: Invalid argument)      = 18
write(2, "\n", 1
)                       = 1


Barcelona:~# df
Filesystem           1K-blocks      Used Available Use% Mounted on
df: `/': Invalid argument
df: `/proc': Invalid argument
df: `/sys': Invalid argument
df: `/dev/pts': Invalid argument
df: `/dev/shm': Invalid argument

Barcelona:~# uname -a
Linux Barcelona 2.6.4 #3 Mon Mar 22 00:14:42 EST 2004 mips GNU/Linux

Barcelona:~# cat /proc/version
Linux version 2.6.4 (root@Barcelona) (gcc version 3.3.3 (Debian)) #3 Mon
Mar 22 00:14:42 EST 2004

Barcelona:~# dpkg -l coreutils libc6
ii  coreutils                 5.0.91-2                  The GNU core
utilities
ii  libc6                     2.3.2.ds1-11              GNU C Library:
Shared libraries and Timezone data

-----Original Message-----
From: Ralf Baechle [mailto:ralf@linux-mips.org] 
Sent: Saturday, 20 March 2004 11:22 PM
To: Martin C. Barlow
Cc: linux-mips@linux-mips.org
Subject: Re: hwclock and df seg fault

I checked in the last fixes for the preemptible kernel less than two
days ago so if your kernel is older than that it's time to update :-)

  Ralf

Re: hwclock and df seg fault

[Ralf Baechle]

The RTC bug was trivial, patch below and in CVS,

  Ralf

Index: drivers/char/ds1286.c
===================================================================
RCS file: /home/cvs/linux/drivers/char/ds1286.c,v
retrieving revision 1.9
diff -u -r1.9 ds1286.c
--- drivers/char/ds1286.c	22 Jan 2004 02:15:40 -0000	1.9
+++ drivers/char/ds1286.c	20 Mar 2004 17:09:39 -0000
@@ -254,7 +254,7 @@
 
 	ds1286_status |= RTC_IS_OPEN;
 
-	spin_lock_irq(&ds1286_lock);
+	spin_unlock_irq(&ds1286_lock);
 	return 0;
 
 out_busy:

Re: hwclock and df seg fault

[Ralf Baechle]

On Sun, Mar 21, 2004 at 03:16:14AM +1100, Martin C. Barlow wrote:

> With PREEMP turned off, the hwclock command works again. Looks like the
> new scheduler may have a little problem.
> The df problem is still there. This may be just a user space
> mis-configuration error. I'll see if I can find the problem.
> Let me know if you need any dumps.

No, the different locking code used for the preemptible kernel contains
assertions which simply detected the bug in the RTC driver.

  Ralf

Re: hwclock and df seg fault

[Jun Sun]

On Sun, Mar 21, 2004 at 03:16:14AM +1100, Martin C. Barlow wrote:
> Thiemo
> 
> With PREEMP turned off, the hwclock command works again. Looks like the
> new scheduler may have a little problem.
> The df problem is still there. This may be just a user space
> mis-configuration error. I'll see if I can find the problem.
> Let me know if you need any dumps.
> 

df works fine here with PREEMPT, both UP and SMP.

There are still some holes in PREEMPT, but mostly are theorectical ones
and you won't hit thme easily.

Jun