Re: [PATCH] reset: tegra-bpmp: Revert Handle errors in BPMP response

[Jon Hunter <jonathanh@nvidia.com>]

On 15/11/2021 17:21, Mikko Perttunen wrote:
> On 11/12/21 1:27 PM, Jon Hunter wrote:
>> Commit c045ceb5a145 ("reset: tegra-bpmp: Handle errors in BPMP
>> response") fixed an issue in the Tegra BPMP error handling but has
>> exposed an issue in the Tegra194 HDA driver and now resetting the
>> Tegra194 HDA controller is failing. For now revert the commit
>> c045ceb5a145 ("reset: tegra-bpmp: Handle errors in BPMP response")
>> while a fix for the Tegra HDA driver is created.
>>
>> Fixes: c045ceb5a145 ("reset: tegra-bpmp: Handle errors in BPMP response")
>> Signed-off-by: Jon Hunter <jonathanh@nvidia.com>
>> ---
>>   drivers/reset/tegra/reset-bpmp.c | 9 +--------
>>   1 file changed, 1 insertion(+), 8 deletions(-)
>>
>> diff --git a/drivers/reset/tegra/reset-bpmp.c 
>> b/drivers/reset/tegra/reset-bpmp.c
>> index 4c5bba52b105..24d3395964cc 100644
>> --- a/drivers/reset/tegra/reset-bpmp.c
>> +++ b/drivers/reset/tegra/reset-bpmp.c
>> @@ -20,7 +20,6 @@ static int tegra_bpmp_reset_common(struct 
>> reset_controller_dev *rstc,
>>       struct tegra_bpmp *bpmp = to_tegra_bpmp(rstc);
>>       struct mrq_reset_request request;
>>       struct tegra_bpmp_message msg;
>> -    int err;
>>       memset(&request, 0, sizeof(request));
>>       request.cmd = command;
>> @@ -31,13 +30,7 @@ static int tegra_bpmp_reset_common(struct 
>> reset_controller_dev *rstc,
>>       msg.tx.data = &request;
>>       msg.tx.size = sizeof(request);
>> -    err = tegra_bpmp_transfer(bpmp, &msg);
>> -    if (err)
>> -        return err;
>> -    if (msg.rx.ret)
>> -        return -EINVAL;
>> -
>> -    return 0;
>> +    return tegra_bpmp_transfer(bpmp, &msg);
>>   }
>>   static int tegra_bpmp_reset_module(struct reset_controller_dev *rstc,
>>
> 
> Is there some reason the subject and commit message does not follow 
> normal revert convention (as done by 'git revert')?


Some maintainers prefer that the patch is not titled "Revert ..." and 
want the patch subject to follow the normal sub-system prefix. If you 
search through the patches for 'revert' in the subject you will see a 
mixture. I prefer this approach.

Jon

-- 
nvpublic