Re: Large USB storage devices cause df to hang and cpu load reaches 100%

Matthew Dharm wrote:
> It would be interesting to turn on CONFIG_USB_STORAGE_DEBUG and capture the
> debugging output.
>
> Matt
>
> On Tue, Jul 31, 2007 at 08:37:31PM +0100, Kostas Peletidis wrote:
>   
>> Matthew Dharm wrote:
>>     
>>> Did you change any USB-related compile options between the two builds?
>>>
>>> Do you have usb device suspend enabled?
>>>
>>> Can you tell (via top) where the 100% CPU usage is coming from (i.e. df,
>>> usb-storage thread, syswait, etc)?
>>>
>>> Matt
>>>
>>>  
>>>       
>> Hi Matt,
>>
>> The most relevant USB configuration option I changed seems to be 
>> "CONFIG_USB_DEVICE_CLASS=y". I have also added module support for 
>> CONFIG_USB_USBNET but I don't think this is related. Please find 
>> attached the kernel configuration.
>>
>> USB suspend is disabled (# CONFIG_USB_SUSPEND is not set) and when I run 
>> top and press Shift+P I see that df is at the top of the list, followed 
>> by Xorg and  usb-storage. The load average reaches 1.17. Interestingly 
>> enough, when df is at the top of the list I see roughly "90%wa" which I 
>> presume means the CPU spends most of the time waiting. Here are the 
>> times derived from "time df":
>>
>> real    1m4.612s
>> user    0m0.006s
>> sys     0m2.867s
>>
>> It is worth mentioning that the problem happens only the first time I 
>> run df. Subsequent runs of df work normally. If I unmount the disk and 
>> remount it I see the same problem.
>>
>> Regards,
>> Kostas
>>     
>
>
>   
>> -------------------------------------------------------------------------
>> This SF.net email is sponsored by: Splunk Inc.
>> Still grepping through log files to find problems?  Stop.
>> Now Search log events and configuration files using AJAX and a browser.
>> Download your FREE copy of Splunk now >>  http://get.splunk.com/
>> _______________________________________________
>> [email protected]
>> To unsubscribe, use the last form field at:
>> https://lists.sourceforge.net/lists/listinfo/linux-usb-devel
>>     
>
>
>   
I enabled usb storage debugging and now I get a lot of messages. It 
seems to me that part of the traced information is lost e.g. there is a 
message "kernel: <>[  125.105000] usb-stora transfer result 0x0" which 
seems to have some characters missing.

I noticed that now top shows roughly 33%us and 67%sy usage(no waiting 
percentage this time, however the problem is still there) probably 
because the debugging code uses a lot of cpu time.Apart from that I see 
quite a lot of "usb-storage: Attempting to get CSW..." messages.

I don't know what to look for so perhaps I missed some more important 
messages. Can you please have a look at the attached file? Thank you.

Regards,
Kostas

-------------------------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc.
Still grepping through log files to find problems?  Stop.
Now Search log events and configuration files using AJAX and a browser.
Download your FREE copy of Splunk now >>  http://get.splunk.com/

_______________________________________________
[email protected]
To unsubscribe, use the last form field at:
https://lists.sourceforge.net/lists/listinfo/linux-usb-devel