Bug 4171

Summary:	bttv seems broken here
Product:	Alternate Trees	Reporter:	JJ Luza (jjluza)
Component:	mm	Assignee:	Andrew Morton (akpm)
Status:	RESOLVED CODE_FIX
Severity:	normal
Priority:	P2
Hardware:	i386
OS:	Linux
Kernel Version:	2.6.11-rc3-mm1	Subsystem:
Regression:	---	Bisected commit-id:
Attachments:	Fix the i2c bug

Description JJ Luza 2005-02-05 13:36:53 UTC

Distribution: Debian Sid    
Hardware Environment: Nforce2 + gforce4 + leadtek winfast TV 2000 (Bt878)  
Software Environment: kernel module is snd_bt87x (and tuner and bttv), and tv 
app is xdtv. 
Problem Description: I can't watch tv with this kernel (it worked with 
2.6.10). There is no image or sound with xdtv, and I get these errors sent by 
kernel when I try to access the bttv device : 
 
Feb  5 17:43:57 client kernel: tuner: i2c i/o error: rc == -121 (should be 4) 
Feb  5 17:44:05 client last message repeated 4 times 
Feb  5 17:44:48 client kernel: bttv0: timeout: drop=328 irq=3696/3696,  
risc=1654f044, bits: HSYNC OFLOW FDSR 
Feb  5 17:44:48 client kernel: bttv0: reset, reinitialize 
Feb  5 17:44:48 client kernel: bttv0: PLL: 28636363 => 35468950 . ok 
Feb  5 17:44:48 client kernel: tuner: i2c i/o error: rc == -121 (should be 4) 
Feb  5 17:45:26 client kernel: bttv0: timeout: drop=608 irq=4243/4243,  
risc=1653406c, bits: HSYNC OFLOW FDSR 
Feb  5 17:45:26 client kernel: bttv0: reset, reinitialize 
Feb  5 17:45:26 client kernel: bttv0: PLL: 28636363 => 35468950 . ok 
Feb  5 17:45:26 client kernel: tuner: i2c i/o error: rc == -121 (should be 4) 
Feb  5 17:46:06 client kernel: bttv0: timeout: drop=889 irq=4815/4815,  
risc=16534034, bits: HSYNC OFLOW FDSR 
Feb  5 17:46:06 client kernel: bttv0: reset, reinitialize 
Feb  5 17:46:06 client kernel: bttv0: PLL: 28636363 => 35468950 . ok 
Feb  5 17:46:06 client kernel: tuner: i2c i/o error: rc == -121 (should be 4) 
Feb  5 17:47:10 client kernel: tuner: i2c i/o error: rc == -121 (should be 4) 
Feb  5 17:47:17 client last message repeated 3 times 
Feb  5 17:47:26 client kernel: bttv0: timeout: drop=1497 irq=5994/5994,  
risc=1bfec03c, bits: VSYNC HSYNC OFLOW FBUS FDSR 
Feb  5 17:47:26 client kernel: bttv0: reset, reinitialize 
Feb  5 17:47:26 client kernel: bttv0: PLL: 28636363 => 35468950 . ok 
Feb  5 17:47:26 client kernel: tuner: i2c i/o error: rc == -121 (should be 4) 
Feb  5 17:48:41 client kernel: bttv0: timeout: drop=2035 irq=7073/7073,  
risc=140a6044, bits: HSYNC OFLOW FDSR 
Feb  5 17:48:41 client kernel: bttv0: reset, reinitialize 
Feb  5 17:48:41 client kernel: bttv0: PLL: 28636363 => 35468950 . ok 
Feb  5 17:48:41 client kernel: tuner: i2c i/o error: rc == -121 (should be 4) 
Feb  5 17:48:52 client kernel: bttv0: timeout: drop=2118 irq=7234/7234,  
risc=138f303c, bits: HSYNC OFLOW FDSR 
Feb  5 17:48:52 client kernel: bttv0: reset, reinitialize 
Feb  5 17:48:52 client kernel: bttv0: PLL: 28636363 => 35468950 . ok 
Feb  5 17:48:52 client kernel: tuner: i2c i/o error: rc == -121 (should be 4) 
Feb  5 17:48:59 client kernel: bttv0: timeout: drop=2170 irq=7332/7332,  
risc=1bfec03c, bits: VSYNC HSYNC OFLOW FBUS FDSR 
Feb  5 17:48:59 client kernel: bttv0: reset, reinitialize 
Feb  5 17:48:59 client kernel: bttv0: PLL: 28636363 => 35468950 . ok 
Feb  5 17:48:59 client kernel: tuner: i2c i/o error: rc == -121 (should be 4) 
 
 
Steps to reproduce:    
Runs xdtv (that's it) ... and look at kernel log.

Comment 1 JJ Luza 2005-02-06 04:14:55 UTC

for people who have this problem too, just reverse these patches :   
http://www.kernel.org/pub/linux/kernel/people/akpm/patches/2.6/2.6.11-rc3/2.6.11-rc3-mm1/broken-out/changes-to-the-i2c-driver-to-support-a-non-blocking-interface.patch   
http://www.kernel.org/pub/linux/kernel/people/akpm/patches/2.6/2.6.11-rc3/2.6.11-rc3-mm1/broken-out/modify-the-i801-i2c-driver-to-use-the-non-blocking-interface.patch   
   
then it works  
  
  
I leave it open for developpers who would like to add some information.  
and thanks to them

Comment 2 JJ Luza 2005-02-10 02:12:44 UTC

Created attachment 4533 [details]
Fix the i2c bug

Ok, there is finally a fix for this bug which can be applied on the top of mm 
tree.
Thanks to Corey Minyard for this patch and his great work on the kernel.