linux-kernel - Re: [Question] Calling request_firmware under the spinlocks in file advansys.c

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite for Android: free password hash cracker in your pocket

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Message-ID: <8ec2185d-203e-71fa-c274-65faf338820f@suse.de>
Date:   Tue, 27 Sep 2016 17:16:11 +0200
From:   Hannes Reinecke <hare@...e.de>
To:     Vaishali Thakkar <vaishali.thakkar@...cle.com>,
        Matthew Wilcox <matthew@....cx>,
        Hannes Reinecke <hare@...e.com>,
        "James E.J. Bottomley" <jejb@...ux.vnet.ibm.com>,
        "Martin K. Petersen" <martin.petersen@...cle.com>,
        linux-scsi@...r.kernel.org, linux-kernel@...r.kernel.org
Subject: Re: [Question] Calling request_firmware under the spinlocks in file
 advansys.c

On 09/27/2016 01:26 PM, Vaishali Thakkar wrote:
> 
> 
> On Tuesday 13 September 2016 02:48 PM, Vaishali Thakkar wrote:
>> Hi,
>>
>> In the file drivers/scsi/advansys.c we are calling function AdvISR at 2 instances
>> [in the function advansys_reset and advansys_interrupt] while holding spinlock.
>> Function AdvISR eventually calls request_firmware following this sequence of
>> routines: 
>>
>> AdvISR -> adv_async_callback -> AdvResetChipAndSB -> AdvInitAsc3550Driver ->
>> request_firmware
>>
>> According to the definition of request_firmware it should be called from user
>> context where sleeping is allowed. And usually sleeping under the spin lock is
>> not allowed. Is it really necessary to call AdvISR under spinlocks here? Are
>> we taking care of sleeping related concern of request_firmware or am I
>> overlooking something here?
> 
> Hi,
> 
> Any comments on this?
> 
AdvISR is the main interrupt handling routine, for which we definitely
will want to have interupts disabled. Plus the SCSI parallel drivers
(ab-) use the interrupt routine to do all sorts of things, and are
thereby well versed to keep interrupts disabled for an extented amount
of time.
However, we don't really have a good way of handling a request firmware
here; we probably would need to implement a workqueue to handle this
properly ...

Cheers,

Hannes
-- 
Dr. Hannes Reinecke		   Teamlead Storage & Networking
hare@...e.de			               +49 911 74053 688
SUSE LINUX GmbH, Maxfeldstr. 5, 90409 Nürnberg
GF: F. Imendörffer, J. Smithard, J. Guild, D. Upmanyu, G. Norton
HRB 21284 (AG Nürnberg)