linux-kernel - Re: [PATCH 1/3] remoteproc: k3-r5: Fix checks in k3_r5_rproc_{mbox

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: <31b95870-2405-4f1f-bdee-3d635e169c13@ti.com>
Date: Mon, 30 Dec 2024 09:36:49 +0530
From: Beleswar Prasad Padhi <b-padhi@...com>
To: Hari Nagalla <hnagalla@...com>, <andersson@...nel.org>,
        <mathieu.poirier@...aro.org>
CC: <afd@...com>, <u-kumar1@...com>, <s-vadapalli@...com>, <srk@...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 1/3] remoteproc: k3-r5: Fix checks in
 k3_r5_rproc_{mbox_callback/kick}

Hi Hari,

On 27/12/24 20:08, Hari Nagalla wrote:
> On 12/24/24 03:14, Beleswar Padhi wrote:
>>   /**
>> @@ -194,8 +196,11 @@ static void k3_r5_rproc_mbox_callback(struct 
>> mbox_client *client, void *data)
>>       const char *name = kproc->rproc->name;
>>       u32 msg = omap_mbox_message(data);
>>   -    /* Do not forward message from a detached core */
>> -    if (kproc->rproc->state == RPROC_DETACHED)
>> +    /*
>> +     * Do not forward messages from a detached core, except when the 
>> core
>> +     * is in the process of being attached in IPC-only mode.
>> +     */
>> +    if (!kproc->core->is_attach_ongoing && kproc->rproc->state == 
>> RPROC_DETACHED)
>>           return;
> Instead of introducing a new state variable, is it possible to use 
> device virtio status? 


See below related comment.


> And i wonder what if you remove this conditional check altogether? If 
> the device is detached and with no virtio queues, does not the mailbox 
> message gets dropped?


This check is necessary for mailbox level communication between cores. 
Some Mbox messages directly use payloads like 
RP_MBOX_ECHO_REQUEST/RP_MBOX_CRASH etc. which do not rely on virtqueue 
read/writes for communication (see omap_remoteproc.h). In that case, 
mailbox message won't be dropped even if virtio queues are not 
initialized. IMO, when we say core is detached in "IPC-only" mode, this 
mbox communication should also not happen. What do you think?