linux-kernel - Re: [PATCH 5/6] soundwire: qcom: update register read/write 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]

Date:   Tue, 2 Feb 2021 10:19:46 +0000
From:   Srinivas Kandagatla <srinivas.kandagatla@...aro.org>
To:     Pierre-Louis Bossart <pierre-louis.bossart@...ux.intel.com>,
        vkoul@...nel.org
Cc:     yung-chuan.liao@...ux.intel.com, sanyog.r.kale@...el.com,
        linux-kernel@...r.kernel.org, alsa-devel@...a-project.org
Subject: Re: [PATCH 5/6] soundwire: qcom: update register read/write routine



On 01/02/2021 16:42, Pierre-Louis Bossart wrote:
> 
> 
> On 2/1/21 9:50 AM, Srinivas Kandagatla wrote:
>>
>>
>> On 29/01/2021 19:33, Pierre-Louis Bossart wrote:
>>>
>>>
>>> On 1/29/21 11:32 AM, Srinivas Kandagatla wrote:
>>>> In the existing code every soundwire register read and register write
>>>> are kinda blocked. Each of these are using a special command id that
>>>
>>> what does 'kinda blocked' mean?
>>
>> I meant read/writes are waiting for completion interrupt!
>>
>>>
>>>> generates interrupt after it successfully finishes. This is really
>>>> overhead, limiting and not really necessary unless we are doing
>>>> something special.
>>>>
>>>> We can simply read/write the fifo that should also give exactly
>>>> what we need! This will also allow to read/write registers in
>>>> interrupt context, which was not possible with the special
>>>> command approach.
>>>
>>> This is really unclear, sorry.
>>
>> If read/writes are waiting for an interrupt, it becomes difficult to 
>> read or write to any registers from same interrupt handler!
> 
> Well, yes, you need to handle the complete() at a lower level than the 
> code that initiates the transactions otherwise you self-deadlock.
> 
> IIRC in the Intel initial code, the complete was in the handler and the 
> register IOs in the thread.
> 
Yes, we did the same in previous version of the code, however with this 
patch reading/writing fifo directly without need of completion should 
remove that need of another thread!
>>
>>
>>>
>>>> +    if (id != SWR_BROADCAST_CMD_ID) {
>>>> +        if (id < 14)
>>>> +            id += 1;
>>>> +        else
>>>> +            id = 0;
>>>
>>> that is really odd. if id=13 (group2) then id becomes 14 (master 
>>> address). A comment is really needed here.
>>
>> This is magic value for each fifo read or write, so that we can verify 
>> that them by comparing with this magic value!
>>
>> This has nothing to do with device number!
> 
> You should probably add a comment here then, or use a #define instead of 
> the 14 which threw me off.

I agree!
> 
>>
>>>
>>>> +    if (cmd_id == SWR_BROADCAST_CMD_ID) {
>>>> +        /*
>>>> +         * sleep for 10ms for MSM soundwire variant to allow broadcast
>>>> +         * command to complete.
>>>
>>> that's also super-odd. There is nothing in SoundWire that makes any 
>>> difference between a regular and a broadcast command. they all 
>>> complete in the same time (a frame).
>>>> +         */
>>>> +        ret = wait_for_completion_timeout(&swrm->broadcast, (2 * 
>>>> HZ/10));
>>>
>>> is this 10ms really or dependent on CONFIG_HZ?
> 
> comment missed?
Not intentionally :-)


I should probably to use msecs_to_jiffies here to keep it inline with 
the comment!

--srini
> 
>>>
>>>> +        if (!ret)
>>>> +            ret = SDW_CMD_IGNORED;
>>>> +        else
>>>> +            ret = SDW_CMD_OK;
>>>
>>> no CMD_FAILED support?
>>
>> Qcom controllers does not provide that information if the command is 
>> ignored or failed by any means!
>>
>> That was the behavior from the starting of this driver.
> 
> ah yes, now I remember this.