Re: [BUG] skge: a possible sleep-in-atomic bug in skge_remove

On Wed, 13 Dec 2017 15:42:56 +0800
Jia-Ju Bai [off-list ref] wrote:

On 2017/12/13 13:18, Stephen Hemminger wrote:

quoted

On Tue, 12 Dec 2017 20:57:01 -0500 (EST)
David Miller [off-list ref] wrote:
 

quoted

From: Stephen Hemminger <stephen@networkplumber.org>
Date: Tue, 12 Dec 2017 10:22:40 -0800
 

quoted

On Tue, 12 Dec 2017 08:34:45 -0500 (EST)
David Miller [off-list ref] wrote:
     

quoted

From: Jia-Ju Bai <redacted>
Date: Tue, 12 Dec 2017 16:38:12 +0800
     

quoted

According to drivers/net/ethernet/marvell/skge.c, the driver may sleep
under a spinlock.
The function call path is:
skge_remove (acquire the spinlock)
   free_irq --> may sleep

I do not find a good way to fix it, so I only report.
This possible bug is found by my static analysis tool (DSAC) and
checked by my code review.  

This was added by:

commit a9e9fd7182332d0cf5f3e601df3e71dd431b70d7
Author: Stephen Hemminger [off-list ref]
Date:   Tue Sep 27 13:41:37 2011 -0400

     skge: handle irq better on single port card

I think the free_irq() can be moved below the unlock.

Stephen, please take a look.  

The IRQ was being free twice.
How did you see it, I really doubt any multi-port SKGE cards
still exist.  

He sees it by reading the code, please take a look at this
and move the free_irq() out of the spin locked section since
it can sleep.  

Thanks, I was hoping for some automated static analysis tool.  

This bug was found by an automated static analysis tool named DSAC, 
which is written by myself.
Then I manually checked driver source code, and finally sent the bug report.

Thanks.
Would it be possible to put tool in tools directory and then have
it automated by kbuild robot?