Messages in this thread | | | Date | Tue, 07 May 2013 17:41:06 +0200 | From | Jiri Slaby <> | Subject | Re: [PATCH 01/15] Char: lp, protect LPGETSTATUS with port_mutex |
| |
On 05/07/2013 04:48 PM, Arnd Bergmann wrote: > On Tuesday 07 May 2013, Jiri Slaby wrote: >> From: "salina@us.ibm.com" <salina@us.ibm.com> >> >> The patch fixes a problem in the lp driver that can cause oopses as >> follows: >> process A: calls lp_write, which in turn calls >> parport_ieee1284_write_compat, and that invokes >> parport_wait_peripheral >> process B: meanwhile does an ioctl(LPGETSTATUS), which call >> lp_release_parport when done. This function will set >> physport->cad = NULL. >> process A: parport_wait_peripheral tries to dereference >> physport->cad and dies >> >> So, protect that code with the port_mutex in order to protect against >> simultaneous calls to lp_read/lp_write. >> >> Similar protection is probably required for ioctl(LPRESET)... >> >> This patch was done by IBM a while back and we (at suse) have that >> since at least 2004 in our repos. Let's make it upstream. > > Hmm, it seems the driver has changed a bit since 2004, at least when > I added the lp_mutex to lp_open()/lp_ioctl(). It's probably worth > taking a look at the bigger picture now, to combine lp_mutex with > lp_table[minor].port_mutex. I don't see any reason why we can't always > use the per-device mutex.
Yeah, it looks sensible to me too to get rid of the lp_mutex, another BKL left-over. However I don't have the hardware, the patch I attached was taken from our tree and tested, at least some time ago. Patches to clean that mess up welcome.
thanks, -- js suse labs
| |