mirror of https://lore.kernel.org/lkml/
 help / color / mirror / Atom feed
* 2.5.74 kernel oops in OSS emu10k1
@ 2003-07-07  8:45 Laurent Canet
  0 siblings, 0 replies; only message in thread
From: Laurent Canet @ 2003-07-07  8:45 UTC (permalink / raw)
  To: linux-kernel, rui.p.m.sousa

[-- Attachment #1: Type: text/plain, Size: 2806 bytes --]

Hi,

While loading emu10k1 module from 2.5.74 I got an oops:

Unable to handle kernel paging request at virtual address 000040a8
 printing eip:
e085d0e5
*pde = 00000000
Oops: 0002 [#1]
CPU:    0
EIP:    0060:[<e085d0e5>]    Not tainted
EFLAGS: 00010246
EIP is at emu10k1_midi_init+0x15/0x170 [emu10k1]
eax: 00000002   ebx: 0000000c   ecx: c03d9669   edx: c03d966a
esi: 00000000   edi: dd280000   ebp: dd2cdeb4   esp: dd2cdea0
ds: 007b   es: 007b   ss: 0068
Process insmod (pid: 122, threadinfo=dd2cc000 task=dd5b06e0)
Stack: e0865f00 ffffffff dff66454 00000000 dd280000 dd2cde00 e085ed90 0000000c
       ffffffed dff66400 e0865e40 dd280000 80261102 dd2cdef0 c026fd1d dff66400
       e0865de4 dff66400 e0865e40 dff66454 dd2cdf08 c026fe60 e0865e40 dff66400
Call Trace:
 [<e085ed90>] emu10k1_probe+0x2c0/0x3a0 [emu10k1]
 [<c026fd1d>] pci_device_probe_static+0x2d/0x50
 [<c026fe60>] __pci_device_probe+0x20/0x40
 [<c026fea1>] pci_device_probe+0x21/0x40
 [<c02c28b8>] bus_match+0x38/0x70
 [<c02c29a3>] driver_attach+0x43/0x60
 [<c02c2c1e>] bus_add_driver+0x6e/0x90
 [<c02c2ff6>] driver_register+0x36/0x40
 [<c027012a>] pci_register_driver+0x6a/0xa0
 [<e083a01b>] emu10k1_init_module+0x1b/0x45 [emu10k1]
 [<c012d7c8>] sys_init_module+0xf8/0x200
 [<c0108ea7>] syscall_call+0x7/0xb
                                                                                                                              
Code: 89 83 9c 40 00 00 83 c4 08 85 c0 7d 14 68 20 26 86 e0 e8 d4
---------

After some debugging, I found that the error comes from sound/emu10k1/main.c
in emu10k1_mixer_init()

	char s[32];
....
	sprintf(s, "driver/emu10k1/%s/ac97", card->pci_dev->slot_name);

here I have "driver/emu10k1/0000:00:11.0/ac97" which is too long for the 32 
bytes buffer. 
To fix this bug, we just need to add some bytes to the buffer. I think that
48 would be enough. patch follows

diff -Naur linux-2.5.74-orig/sound/oss/emu10k1/main.c linux-2.5.74-modified/sound/oss/emu10k1/main.c
--- linux-2.5.74-orig/sound/oss/emu10k1/main.c  2003-07-07 10:35:17.000000000 +0200
+++ linux-2.5.74-modified/sound/oss/emu10k1/main.c      2003-07-07 10:31:25.000000000 +0200
@@ -221,7 +221,7 @@
  
 static int __devinit emu10k1_mixer_init(struct emu10k1_card *card)
 {
-       char s[32];
+       char s[48];
  
        struct ac97_codec *codec = &card->ac97;
        card->ac97.dev_mixer = register_sound_mixer(&emu10k1_mixer_fops, -1);
@@ -288,7 +288,7 @@
  
 static void __devinit emu10k1_mixer_cleanup(struct emu10k1_card *card)
 {
-       char s[32];
+       char s[48];
  
        if (!card->is_aps) {
                sprintf(s, "driver/emu10k1/%s/ac97", card->pci_dev->slot_name);



(Maybe we could avoid using the stack here, I already seen emu10k1 in the top
stack users list) 

-- 
Laurent Canet
----------------------------


[-- Attachment #2: patch-emu10k1-oops-2.5.74 --]
[-- Type: application/octet-stream, Size: 791 bytes --]

diff -Naur linux-2.5.74-orig/sound/oss/emu10k1/main.c linux-2.5.74-modified/sound/oss/emu10k1/main.c
--- linux-2.5.74-orig/sound/oss/emu10k1/main.c  2003-07-07 10:35:17.000000000 +0200
+++ linux-2.5.74-modified/sound/oss/emu10k1/main.c      2003-07-07 10:31:25.000000000 +0200
@@ -221,7 +221,7 @@
  
 static int __devinit emu10k1_mixer_init(struct emu10k1_card *card)
 {
-       char s[32];
+       char s[48];
  
        struct ac97_codec *codec = &card->ac97;
        card->ac97.dev_mixer = register_sound_mixer(&emu10k1_mixer_fops, -1);
@@ -288,7 +288,7 @@
  
 static void __devinit emu10k1_mixer_cleanup(struct emu10k1_card *card)
 {
-       char s[32];
+       char s[48];
  
        if (!card->is_aps) {
                sprintf(s, "driver/emu10k1/%s/ac97", card->pci_dev->slot_name);


^ permalink raw reply	[flat|nested] only message in thread

only message in thread, other threads:[~2003-07-07  8:30 UTC | newest]

Thread overview: (only message) (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2003-07-07  8:45 2.5.74 kernel oops in OSS emu10k1 Laurent Canet

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox

all inboxes | Powered by JetHome®