Re: rc9 + orinoco WPA patchset: BUG: scheduling while atomic loading firmware with PCMCIA adapter

Previous message: [thread] [date] [author]
Next message: [thread] [date] [author]

Andrey Borzenkov wrote:

Thanks for identifying the problem. The Agere case looks good - a few suggestions for the Symbol case though:


Please initialise pda to NULL here...


And move the allocation to


 ... here.


Then kfree(pda) here. We're done with it now.


So this isn't needed.


nor this.


or these.

 
But you absolutely have to kfree(pda) here. I would prefer the label be something like 'abort' (what we are doing), rather than 'free' - but there's merit in keeping with what you have called in in orinoco_dl_firmware which already has an 'abort' label.

 
The net effect of the above suggestion is that we won't allocate memory when programming the primary firmware (which is just before we do the secondary).

Dan Williams wrote:

Using fw->pda_size should be fine. The value comes from a const static, set to 0x100 for Symbol.


Regards,

Dave.
--
Previous message: [thread] [date] [author]
Next message: [thread] [date] [author]

Messages in current thread:
Re: rc9 + orinoco WPA patchset: BUG: scheduling while atom ..., Dave, (Thu Oct 9, 3:47 pm)
Re: rc9 + orinoco WPA patchset: BUG: scheduling while atom ..., Andrey Borzenkov, (Fri Oct 10, 12:41 am)
Re: rc9 + orinoco WPA patchset: BUG: scheduling while atom ..., Andrey Borzenkov, (Fri Oct 10, 10:14 am)
Re: rc9 + orinoco WPA patchset: BUG: scheduling while atom ..., Andrey Borzenkov, (Fri Oct 10, 10:26 am)
Re: rc9 + orinoco WPA patchset: BUG: scheduling while atom ..., Dominik Brodowski, (Fri Oct 17, 1:44 pm)