15.11.2021 18:48, Dan Carpenter пишет:
> On Mon, Nov 15, 2021 at 05:34:47PM +0300, Dmitry Osipenko wrote:
>> 15.11.2021 15:44, Dan Carpenter пишет:
>>> On Mon, Nov 15, 2021 at 01:34:18AM +0300, Dmitry Osipenko wrote:
>>>> 15.11.2021 01:23, Dmitry Osipenko пишет:
>>>>> + vde->secure_bo = tegra_vde_alloc_bo(vde, DMA_FROM_DEVICE, 4096);
>>>>> + if (!vde->secure_bo) {
>>>>> + dev_err(dev, "Failed to allocate secure BO\n");
>>>>> + goto err_pm_runtime;
>>>>> + }
>>>>
>>>> My eye just caught that by accident err variable isn't assigned to
>>>> -ENOMEM here. I'll make v2 shortly.
>>>
>>> Smatch has a check for this so we would have caught it. :)
>>
>> Whish smatch was built-in into kernel and I could simply run "make
>> smatch". On the other hand, I know that you're periodically checking
>> upstream with smatch and patching the found bugs, so maybe it's fine
>> as-is.
>
> I run it every day, and it's integrated into the zero day bot and Mauro
> runs it. So someone would have caught it.
Very nice, I haven't noticed that 0-day runs smatch. I assume the smatch
reports from the bots are private to you and Mauro since I never saw
them in my inbox and on LKML, correct?
