Hi John,
Thanks for having a look.
John Kacur <jkacur@xxxxxxxxxx> writes:
> On Tue, 28 Sep 2021, Punit Agrawal wrote:
>
>> From: Punit Agrawal <punit1.agrawal@xxxxxxxxxxxxx>
>>
>> On systems where cpu idle is disabled to reduce latencies,
>> "/dev/cpu_dma_latency" does not exist and leads to the following
>> exception report from python when using hwlatdetect.py -
>>
>> FileNotFoundError: [Errno 2] No such file or directory: '/dev/cpu_dma_latency
>>
>> Update hwlatdetect to check whether the file exists before attemping
>> to write values to it.
>>
>> Reported-by: Suresh Hegde <suresh.c11@xxxxxxxxxxxxxxxx>
>> Signed-off-by: Punit Agrawal <punit1.agrawal@xxxxxxxxxxxxx>
>> ---
>> v1 -> v2:
>> * Correct Suresh's email
>>
>> src/hwlatdetect/hwlatdetect.py | 7 +++++--
>> 1 file changed, 5 insertions(+), 2 deletions(-)
>>
>> diff --git a/src/hwlatdetect/hwlatdetect.py b/src/hwlatdetect/hwlatdetect.py
>> index 12228f45f852..89537a6e371f 100755
>> --- a/src/hwlatdetect/hwlatdetect.py
>> +++ b/src/hwlatdetect/hwlatdetect.py
>> @@ -216,8 +216,11 @@ class Detector(object):
>> # use c_states_on() to close the file descriptor and re-enable c-states
>> #
>> def c_states_off(self):
>> - self.dma_latency_handle = os.open("/dev/cpu_dma_latency", os.O_WRONLY)
>> - os.write(self.dma_latency_handle, b'\x00\x00\x00\x00')
>> + if os.path.exists("/dev/cpu_dma_latency"):
>> + self.dma_latency_handle = os.open("/dev/cpu_dma_latency", os.O_WRONLY)
>> + os.write(self.dma_latency_handle, b'\x00\x00\x00\x00')
>> + else:
>> + self.dma_latency_handle = None
>> debug("c-states disabled")
>>
>> def c_states_on(self):
>> --
>> 2.32.0
>>
>>
>
> Thanks! I have a couple of nit-picky suggestions.
>
> If you initialize self.dma_latency_handle = None in the init method of
> class Detector, then you can drop the else part of your check to see if
> the file exists.
I can do that - but then I wonder if there is any harm in initialising
the file handle in the init method itself too. That way
c_states_[off|on]() can focus on doing their thing after checking the
handle is not null.
I'll send a patch in sometime addressing this and below comments.
>
> Also, in the situation you describe, the debug messages should probably be
> hoisted up under the "if" part, and that goes for method c_states_on()
> too, as they are not relevant.
>
> You might as well fix the comments above the code too,
> openinging -> opening
> writeing -> writing
> And edit that one line not to over flow 80 chars.
>
> Thanks!
>
> John
