Gentoo Forums
Gentoo Forums
Gentoo Forums
Quick Search: in
Kernel paging request failure w/gentoo-alpha-1.4rc1-test4
View unanswered posts
View posts from last 24 hours

 
Reply to topic    Gentoo Forums Forum Index Gentoo on Alternative Architectures
View previous topic :: View next topic  
Author Message
mmcgreal
n00b
n00b


Joined: 04 Feb 2003
Posts: 7
Location: St Louis, MO

PostPosted: Tue Feb 04, 2003 11:15 am    Post subject: Kernel paging request failure w/gentoo-alpha-1.4rc1-test4 Reply with quote

Hi All,

I've just installed gentoo-alpha-1.4rc1-test4 on my Alphaserver 1200, SMP. The kernel that comes w/this distro is called 2.4.19-gentoo-r2. Every time I execute the following command:

Code:

    echo | grep blah | sed 's/blah//'


I get a segmentation fault on the command line, and a "Unable to handle kernel paging request at virtual address 000006aaaaaaa000" on my console. Has anyone else run into this? It only happens when using grep (or egrep or fgrep) and piping the results to sed. I wonder if it has something to do w/the Superpage patch that is applied to this kernel source. Here's the oops message that gets logged (it's the same every time except for the pid):

Code:

Feb  4 05:07:39 wolfman kernel: CPU 1 sed(3842): Oops 0
Feb  4 05:07:39 wolfman kernel: pc = [<fffffc00008687e8>]  ra = [<fffffc0000868a70>]  ps = 0000    Not tainted
Feb  4 05:07:39 wolfman kernel: v0 = 0000000000000007  t0 = 0000000000000001  t1 = 0000000000000000
Feb  4 05:07:39 wolfman kernel: t2 = 0000000000000000  t3 = fffffc0000acb4a8  t4 = 000006aaaaaaa000
Feb  4 05:07:39 wolfman kernel: t5 = 0000000000000000  t6 = 0000000000000000  t7 = fffffc0029788000
Feb  4 05:07:39 wolfman kernel: s0 = 0000020000000000  s1 = fffffc0000acb4a8  s2 = 0000000000002000
Feb  4 05:07:39 wolfman kernel: s3 = fffffc0000acb4b4  s4 = 0000020000002000  s5 = 0000000000000001
Feb  4 05:07:39 wolfman kernel: s6 = fffffc002ae544c0
Feb  4 05:07:39 wolfman kernel: a0 = fffffc002ae544c0  a1 = 0000000000000001  a2 = 000006aaaaaaa000
Feb  4 05:07:39 wolfman kernel: a3 = 0000000000000000  a4 = 0000000000000001  a5 = fffffc0000a9c900
Feb  4 05:07:39 wolfman kernel: t8 = 0000000000000000  t9 = 0000004c1c7fb45a  t10= 9800000000000000
Feb  4 05:07:39 wolfman kernel: t11= fffffc0000acb5d8  pv = fffffc0000868910  at = fffffc0000af15c0
Feb  4 05:07:39 wolfman kernel: gp = fffffc0000af15c0  sp = fffffc002978bb88
Feb  4 05:07:39 wolfman kernel: Trace:fffffc0000849504 fffffc000084f780 fffffc0000832b68 fffffc000083a4e0 fffffc0000842a94 fffffc000081befc fffffc0000813968
Feb  4 05:07:39 wolfman kernel: Code: 48330721  47ff0407  403f0001  ec200012  40f20405  a49da7a8 <a4250000> 442c1001


If anyone could provide me w/some clues on how to decipher this oops message, I would be very grateful. Anyway I just thought I'd check around to see if anyone else has had this problem. I'm going to try to figure out how to glean information from this oops message. Also worth a try will be recompiling grep and sed w/out the optimization flags, and probobly compiling a kernel w/out the superpage patch as well.

Thanks for any help,
Martin
_________________
"Come, my songs, let us speak of perfection--
We shall get ourselves rather disliked."
:Ezra Pound, Salvationists
Back to top
View user's profile Send private message
mmcgreal
n00b
n00b


Joined: 04 Feb 2003
Posts: 7
Location: St Louis, MO

PostPosted: Wed Feb 05, 2003 10:14 am    Post subject: Reply with quote

FYI, since I've replaced sed w/ssed (supersed) the paging errors have apparently been eradicated. But the question still remains: what was the actual cause of the problem...

Here's the facts as I know them:
The original /bin/sed was a statically linked binary (one of the only ones on the whole system).
The s0 register in the oops message is the address of /lib/ld-linux.so.2, the run-time linker:

Code:

wolfman:~- find /proc -name maps | xargs grep 0000020000000000 |head -1
/proc/1/maps:0000020000000000-0000020000018000 r-xp 0000000000000000 08:22 1451       /lib/ld-2.3.1.so


The s1 and t3 registers contain the address of super_page_order():
Code:

wolfman:~- grep fffffc0000acb4a8 /proc/ksyms
fffffc0000acb4a8 super_page_order_R__ver_super_page_order


Ok.... this is obviously significant information, but I don't know what to do with it. I'll have a look at super_page_order() to see if there's any reason why it would mess up a pipe between a dynamically linked binary and a static one, but unfortunately my knowledge of Linux kernel internals and ELF is still a bit weak at this point. If anyone wants to guide me in the right direction, I would appreciate it!

Thanks
Martin
_________________
"Come, my songs, let us speak of perfection--
We shall get ourselves rather disliked."
:Ezra Pound, Salvationists
Back to top
View user's profile Send private message
mmcgreal
n00b
n00b


Joined: 04 Feb 2003
Posts: 7
Location: St Louis, MO

PostPosted: Wed Feb 05, 2003 10:43 am    Post subject: Reply with quote

Ok, I've finally run ksymoops. Here's the output:

Code:

Unable to handle kernel paging request at virtual address 000006aaaaaaa000
CPU 1 sed(3780): Oops 0
pc = [<fffffc000085c908>]  ra = [<fffffc000085cb90>]  ps = 0000    Not tainted
Using defaults from ksymoops -t elf64-alpha -a alpha
v0 = 0000000000000007  t0 = 0000000000000001  t1 = 0000000000000000
t2 = 0000000000000000  t3 = fffffc0000a7e100  t4 = 000006aaaaaaa000
t5 = 0000000000000000  t6 = 0000000000000000  t7 = fffffc002b0d0000
s0 = 0000020000000000  s1 = fffffc0000a7e100  s2 = 0000000000002000
s3 = fffffc0000a7e10c  s4 = 0000020000002000  s5 = 0000000000000001
s6 = fffffc002f13e500
a0 = fffffc002f13e500  a1 = 0000000000000001  a2 = 000006aaaaaaa000
a3 = 0000000000000000  a4 = 0000000000000001  a5 = fffffc0000a55480
t8 = 0000000000000000  t9 = 000000534c37396c  t10= a600000000000000
t11= fffffc0000a7e230  pv = fffffc000085ca30  at = fffffc0000aa0a20
gp = fffffc0000aa0a20  sp = fffffc002b0d3b88
Trace:fffffc000083d1d4 fffffc00008433b0 fffffc0000826c98 fffffc000082e3a0 fffffc0000836864 fffffc000081bc1c fffffc0000813818
Code: 48330721  47ff0407  403f0001  ec200012  40f20405  a49da798 <a4250000> 442c1001


>>RA;  fffffc000085cb90 <adj_sp_range+160/1e0>

>>PC;  fffffc000085c908 <adj_sp_pte+a8/1d0>   <=====

Trace; fffffc000083d1d4 <zap_page_range+74/370>
Trace; fffffc00008433b0 <exit_mmap+110/220>
Trace; fffffc0000826c98 <mmput+88/190>
Trace; fffffc000082e3a0 <do_exit+190/5c0>
Trace; fffffc0000836864 <sig_exit+114/120>
Trace; fffffc000081bc1c <do_signal+2fc/490>
Trace; fffffc0000813818 <signal_return+18/20>

Code;  fffffc000085c8f0 <adj_sp_pte+90/1d0>
0000000000000000 <_PC>:
Code;  fffffc000085c8f0 <adj_sp_pte+90/1d0>
   0:   21 07 33 48       sll  t0,a3,t0
Code;  fffffc000085c8f4 <adj_sp_pte+94/1d0>
   4:   07 04 ff 47       clr  t6
Code;  fffffc000085c8f8 <adj_sp_pte+98/1d0>
   8:   01 00 3f 40       addl t0,zero,t0
Code;  fffffc000085c8fc <adj_sp_pte+9c/1d0>
   c:   12 00 20 ec       ble  t0,58 <_PC+0x58>
Code;  fffffc000085c900 <adj_sp_pte+a0/1d0>
  10:   05 04 f2 40       addq t6,a2,t4
Code;  fffffc000085c904 <adj_sp_pte+a4/1d0>
  14:   98 a7 9d a4       ldq  t3,-22632(gp)
Code;  fffffc000085c908 <adj_sp_pte+a8/1d0>   <=====
  18:   00 00 25 a4       ldq  t0,0(t4)   <=====
Code;  fffffc000085c90c <adj_sp_pte+ac/1d0>
  1c:   01 10 2c 44       and  t0,0x60,t0


This definately has to do w/the super page patch...
_________________
"Come, my songs, let us speak of perfection--
We shall get ourselves rather disliked."
:Ezra Pound, Salvationists
Back to top
View user's profile Send private message
Display posts from previous:   
Reply to topic    Gentoo Forums Forum Index Gentoo on Alternative Architectures All times are GMT
Page 1 of 1

 
Jump to:  
You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot vote in polls in this forum