linux-kernel - Re: [PATCH v12 04/36] remoteproc: k3-m4: Don't assert reset in detach routine

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: Windows password security audit tool. GUI, reports in PDF.

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

Message-ID: <c746e97a-7792-4dfb-902b-7f1e705cb413@collabora.com>
Date: Tue, 20 May 2025 10:07:48 +0100
From: Martyn Welch <martyn.welch@...labora.com>
To: Beleswar Prasad Padhi <b-padhi@...com>,
 Mathieu Poirier <mathieu.poirier@...aro.org>
Cc: andersson@...nel.org, afd@...com, hnagalla@...com, u-kumar1@...com,
 jm@...com, jan.kiszka@...mens.com, christophe.jaillet@...adoo.fr,
 jkangas@...hat.com, eballetbo@...hat.com, linux-remoteproc@...r.kernel.org,
 linux-kernel@...r.kernel.org
Subject: Re: [PATCH v12 04/36] remoteproc: k3-m4: Don't assert reset in detach
 routine



On 17/05/2025 14:23, Beleswar Prasad Padhi wrote:
> 
> On 5/16/2025 9:15 PM, Mathieu Poirier wrote:
>> On Tue, May 13, 2025 at 11:14:38AM +0530, Beleswar Padhi wrote:
>>> The rproc_detach() function invokes __rproc_detach() before
>>> rproc_unprepare_device(). The __rproc_detach() function sets the
>>> rproc->state to "RPROC_DETACHED".
>>>
>>> However, the TI K3 M4 driver erroneously looks for "RPROC_ATTACHED"
>>> state in its .unprepare ops to identify IPC-only mode; which leads to
>>> resetting the rproc in detach routine.
>>>
>>> Therefore, correct the IPC-only mode detection logic to look for
>>> "RPROC_DETACHED" in k3_m4_rproc_unprepare() function.
>>>
>> This driver has been upstream for 9 whole months, it is hard for me to 
>> believe
>> this but was just noticed.  Martyn from Collabora should be CC'ed on 
>> this, and I
>> will also need the required R-b/T-b tags.
> 
> 
> Cc: Martyn Welch martyn.welch@...labora.com
> 
> Requesting Andrew/Judith for review and test too.
> 

It's been a while since I used this, IIRC the project I thought we 
needed this for went in a different direction almost as soon as I'd 
managed to get the driver upstream...  ...But I've spent some time to 
review as best as I can:

Reviewed-by: Martyn Welch <martyn.welch@...labora.com>

>>
>> Typically bug fixes are not part of refactoring exercises.
> 
> 
> Typically, yes. But the refactor depends on this fix. This 
> k3_m4_rproc_unprepare() function is entirely refactored to common driver 
> in [PATCH v12 26/36].
> 
> So, If the refactor is picked without this patch fix, the mainline M4 
> driver would be fixed, but the older stable kernels would always have 
> this bug. Let me know what you think.
> 
> Thanks,
> Beleswar
> 
>>   I suggest to apply
>> this set without this patch - you can then work on fixing this bug.
>>
>> Thanks,
>> Mathieu
>>
>>> Fixes: ebcf9008a895 ("remoteproc: k3-m4: Add a remoteproc driver for 
>>> M4F subsystem")
>>> Signed-off-by: Beleswar Padhi <b-padhi@...com>
>>> ---
>>> v12: Changelog:
>>> 1. New patch. Fixup a state detection logic.
>>>
>>>   drivers/remoteproc/ti_k3_m4_remoteproc.c | 2 +-
>>>   1 file changed, 1 insertion(+), 1 deletion(-)
>>>
>>> diff --git a/drivers/remoteproc/ti_k3_m4_remoteproc.c b/drivers/ 
>>> remoteproc/ti_k3_m4_remoteproc.c
>>> index a16fb165fcedd..6cd50b16a8e82 100644
>>> --- a/drivers/remoteproc/ti_k3_m4_remoteproc.c
>>> +++ b/drivers/remoteproc/ti_k3_m4_remoteproc.c
>>> @@ -228,7 +228,7 @@ static int k3_m4_rproc_unprepare(struct rproc 
>>> *rproc)
>>>       int ret;
>>>       /* If the core is going to be detached do not assert the module 
>>> reset */
>>> -    if (rproc->state == RPROC_ATTACHED)
>>> +    if (rproc->state == RPROC_DETACHED)
>>>           return 0;
>>>       ret = kproc->ti_sci->ops.dev_ops.put_device(kproc->ti_sci,
>>> -- 
>>> 2.34.1
>>>