Thread (19 messages) 19 messages, 5 authors, 2012-06-28

Re: net: nfc: BUG and panic in accept() on 3.5-rc2

From: Sasha Levin <hidden>
Date: 2012-06-25 15:04:58
Also in: linux-wireless, lkml

On Mon, Jun 11, 2012 at 7:25 PM, Dave Jones [off-list ref] wrote:
On Mon, Jun 11, 2012 at 06:56:50PM +0200, Sasha Levin wrote:

 > > > > > > What's trinity ?
 > > > > > > Also, if this one is reproducible, would you mind sharing some details about
 > > > > > > how we could reproduce it ?
 > > > > >
 > > > > > Well, bugfix should be trivial enough ;)
 > > > > Yep, I looked at the code only after looking at Sasha's report.
 > > > >
 > > > > Thanks for the patch, do you mind if I add your SOB to it ?
 > > >
 > > > I  would prefer making sure it fixes the bug first ;)
 > > Sure, although your patch makes sense regardless of that :)
 > > I'll still wait for Sasha to confirm that it fixes his crash.
 >
 > I don't have a direct way of reproducing it, but I've put it in the test
 > vm and the fuzzer is running, I'll let you know tomorrow if it happened
 > again.

You might be able to trigger it faster by using -P PF_NFC, which will
force trinity to only use NFC sockets.

sidenote: most protocols trigger the module to be auto-loaded when a socket
is created. This doesn't seem to happen with nfc, making me need to manually
modprobe it first. Intentional ?
It fixes the bug, wasn't been able to reproduce it since then.
Keyboard shortcuts
hback out one level
jnext message in thread
kprevious message in thread
ldrill in
Escclose help / fold thread tree
?toggle this help