[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Google
  Web www.spinics.net

Re: TUX development? Anyone? Bad EIP



On Fri, Feb 15, 2002 at 03:56:34PM +0200, Alex Antropoff wrote:
> On Fri, Feb 15, 2002 at 02:21:58PM +0100, Roy Sigurd Karlsbakk wrote:
> > hi all
> > 
> > After tux2-full-2.4.17-final-A1 was released, I've submitted a few bug
> > reports. However; noone have picked them up.
> Possible tt was 'case nobody can reproduce your bugs and never see
> such bugs before. 

Hi all.

Try to reproduce these 8)



I mounted a dir on tmpfs partition with "mount --bind" under TUX ftproot,
"mount --bind /mnt/tmpfs/foo ~www/mp3/foo" ...
When I tried to fetch a file from dir ~www/mp3/foo thru TUX ftp-server,
I got these:

Unable to handle kernel NULL pointer dereference at virtual address 00000000
 printing eip:
00000000
*pde = 00000000
Oops: 0000
CPU:    0
EIP:    0010:[<00000000>]    Tainted: P
EFLAGS: 00010206
eax: 00000000   ebx: c1242c00   ecx: c1129f80   edx: c7caf06c
esi: c1450ed8   edi: c207b4d0   ebp: cd3d3f44   esp: cd3d3f14
ds: 0018   es: 0018   ss: 0018
Process async IO 0/3 (pid: 21010, stackpage=cd3d3000)
Stack: c01248c3 c7caf06c c1242c00 c7caf000 c7caf000 00000001 00001000 00000001
       00000000 00000000 00000000 c207b420 cd3d3f78 d0b3feb9 c7caf06c c7caf08c
       c7caf0d8 d0b3fab8 00000000 c7caf000 c7caf000 00000001 00000000 c5364980
Call Trace: [<c01248c3>] [<d0b3feb9>] [<d0b3fab8>] [<d0b46e95>] [<d0b3e4b4>]
   [<d0b3f658>] [<c01054bc>]

[ksymoops]
Code:  Bad EIP value.
Trace; c01248c2 <do_generic_file_read+292/414>
Trace; d0b3feb8 <[tux]generic_send_file+ec/1d4>
Trace; d0b3fab8 <[tux]sock_send_actor+0/278>
Trace; d0b46e94 <[tux]ftp_send_file+24/178>
Trace; d0b3e4b4 <[tux]tux_schedule_atom+58/60>
Trace; d0b3f658 <[tux]cachemiss_thread+ec/1d8>
Trace; c01054bc <kernel_thread+28/38>

...

now when I try to read the file which caused TUX to barf, it fails, with
backtrace of...:

Trace; c01241d4 <__lock_page+6c/94>
Trace; c0124212 <lock_page+16/1c>
Trace; c01242fa <__find_lock_page_helper+56/7c>
Trace; c0124332 <__find_lock_page+12/18>
Trace; c012d1f8 <shmem_getpage_locked+4c/2c0>
Trace; c012d4bc <shmem_getpage+50/a0>
Trace; c012da3e <do_shmem_file_read+42/e8>
Trace; c012db3c <shmem_file_read+58/70>
Trace; c0130238 <sys_read+98/d4>
Trace; c0106baa <system_call+32/38>

I can read other files on /mnt/tmpfs/foo normally.

tuxstat shows "5086K /      0K (0.0%) ..." for the barfed file.

...

I don't know exact version of TUX, but net/tux/*.c files are modified
20011217 ;-) and kernel is 2.4.16 with low-latency patch.

9:14pm  up 25 days, 54 min,  1 user,  load average: 2.22, 2.11, 2.05

...

then I tried to restart tux, but as I expected, it did not succeed:

kernel: kernel BUG at main.c:277!
kernel: invalid operand: 0000
kernel: CPU:    0
kernel: EIP:    0010:[<d0b49cc5>]    Tainted: P
kernel: EFLAGS: 00010202
kernel: eax: 0000001a   ebx: 00000001   ecx: c0335120   edx: 000b954c
kernel: esi: 00000001   edi: 0804c560   ebp: ccdf9f50   esp: ccdf9d3c
kernel: ds: 0018   es: 0018   ss: 0018
kernel: Process tux (pid: 3053, stackpage=ccdf9000)
kernel: Stack: d0b55c48 00000115 00000001 00000001 0804c560 00000000 d0b58428 d0b58428 
kernel:        00000006 c8782f00 00000026 cf246500 00000003 00000000 00000000 ccdf9d90 
kernel:        c01a9755 ccdf9e28 cff3b800 00000037 00000000 00000301 00001000 00006414 
kernel: Call Trace: [<d0b55c48>] [<d0b58428>] [<d0b58428>] [<c01a9755>] [<c0131ea0>] 
kernel:    [<c013209f>] [<c01a21ff>] [<c01a2a09>] [<c01a9d53>] [<c012a9a0>] [<c012a729>] 
kernel:    [<c0121aae>] [<c0121fb6>] [<c0111a2a>] [<c01118a4>] [<c0122391>] [<d0b4b437>] 
kernel:    [<c012fe5b>] [<c025449b>] [<c0106bab>] 
kernel: Code: 0f 0b 83 c4 08 8d b6 00 00 00 00 8d bd 0c fe ff ff 89 bd 00 

>>EIP; d0b49cc4 <[tux]user_req_startup+54/220>   <=====
Trace; d0b55c48 <[tux]not_found+1888/2b9e>
Trace; d0b58428 <[tux]tux_proto_http+28/60>
Trace; d0b58428 <[tux]tux_proto_http+28/60>
Trace; c01a9754 <journal_end+28/30>
Trace; c0131ea0 <getblk+18/44>
Trace; c013209e <bread+22/6c>
Trace; c01a21fe <is_tree_node+46/50>
Trace; c01a2a08 <search_by_key+800/c14>
Trace; c01a9d52 <check_journal_end+1fe/22c>
Trace; c012a9a0 <__alloc_pages+34/164>
Trace; c012a728 <_alloc_pages+18/20>
Trace; c0121aae <do_wp_page+12e/160>
Trace; c0121fb6 <handle_mm_fault+7a/b4>
Trace; c0111a2a <do_page_fault+186/4cc>
Trace; c01118a4 <do_page_fault+0/4cc>
Trace; c0122390 <__vma_link+60/b0>
Trace; d0b4b436 <[tux]__sys_tux+8a/a3c>
Trace; c012fe5a <filp_close+52/60>
Trace; c025449a <sys_tux+62/88>
Trace; c0106baa <system_call+32/38>
Code;  d0b49cc4 <[tux]user_req_startup+54/220>
00000000 <_EIP>:
Code;  d0b49cc4 <[tux]user_req_startup+54/220>   <=====
   0:   0f 0b                     ud2a      <=====
Code;  d0b49cc6 <[tux]user_req_startup+56/220>
   2:   83 c4 08                  add    $0x8,%esp
Code;  d0b49cc8 <[tux]user_req_startup+58/220>
   5:   8d b6 00 00 00 00         lea    0x0(%esi),%esi
Code;  d0b49cce <[tux]user_req_startup+5e/220>
   b:   8d bd 0c fe ff ff         lea    0xfffffe0c(%ebp),%edi
Code;  d0b49cd4 <[tux]user_req_startup+64/220>
  11:   89 bd 00 00 00 00         mov    %edi,0x0(%ebp)

...

then a couple other "issues"...

1) http dir indexer doesn't generate valid href's (e.g. no ' ' --> "%20")
   http://www.faqs.org/rfcs/rfc2396.html
2) ftp dirs of mode 111, owner root:root are readable (I get dir listing)
   (I'd hope to get no dir listing, but chdir should succeed)
3) tuxstat truncates the filename to first ' ' char, that is, if
   downloading file "foo bar.txt", tuxstat shows only "foo".
4) does TUX force max tcp/ip send window size to ~5840 bytes?

---

Then, hoping Ingo sees this (which I think was not case with the
private email I sent to him): is SCHED_FIFO borked with O(1) K3
scheduler or did I break it by myself? The prob I have is:
when "rtc_latencytest 1024" is running and I run command "ls"
in xterm, rtc_latencytest doesn't get run for 1200 ms:
http://iki.fi/safari/O1/O1-ls.png
I now run 2.4.18-pre9-ac2 etc...

For comparison, this with 2.4.16+low-lat:
http://iki.fi/safari/O1/stock.png

If this O1 stuff is (too) offtopic, where could I ask about this?

-- 
Safari - safari@iki.fi - PGP key 0x427E7914 - http://iki.fi/safari/
  The UNIX Guru's View of Sex: "unzip ; strip ; touch ; finger ; \
  mount ; fsck ; more ; yes ; umount ; sleep"





[Older Fedora Users Mail]     [Home]     [Fedora Legacy]     [Fedora Desktop]     [iPod Nano]     [ATA RAID]     [Fedora Bible]     [Fedora Marketing]     [Fedora Mentors]     [Fedora Packaging]     [Fedora SELinux]     [Big List of Linux Books]     [Yosemite News]     [Yosemite Photos]     [KDE Users]     [Fedora Tools]     [Fedora Docs]

Powered by Linux

Google
  Web www.spinics.net