Re: [PATCH] blktrace: disallow -o when using multiple devices

Re: [PATCH] blktrace: disallow -o when using multiple devices

[Brett Russ]

On 09/16/2010 08:33 AM, Alan D. Brunelle wrote:
> Document that "-o" does not work when specyfing multiple devices to
> blktrace, also: enforce this by stopping blktrace when one tries do
> do this.
>
> The technical reason why "-o" doesn't work with multiple devices is
> because we use multiple threads of execution - one per device/CPU pair -
> and each of them opens a file named "<prefix>.blktrace.<cpu>". With the
> "-o" all of the"<prefix>" values are the same - so multiple threads
> open the same file and try to do output. Not good. Without the "-o"
> we get unique files named: "<device>.blktrace.<cpu>"  - as the tuple
> (<device>,<cpu>) is unique.

Alan,

-Wouldn't it be a good idea to just add <prefix> before the filename(s) 
that would have been created in either the single or multiple device 
case?  I.e. in the multiple device case, how about 
"<prefix>.<device>.blktrace.<cpu>"?  That way the program always does 
what the user wants, no special case documentation required.

-I just tested that the blktrace "-D" option (output dir, create if 
doesn't exist) works for the multiple device case as long as the -o 
option isn't also used.  So, this is another alternative.

-I found the following inconsistencies between the doc and the command 
line help, these would probably be a separate patch:
--blktrace
---doc is missing -[Dhlps]
---doc still contains deprecated option -k
--blkparse
---doc is missing -[aAD]
---doc still contains deprecated option -m


Thanks,
Brett

Re: [PATCH] blktrace: disallow -o when using multiple devices

[Brett Russ]

On 09/16/2010 10:20 AM, Brett Russ wrote:
> -Wouldn't it be a good idea to just add<prefix>  before the filename(s)
> that would have been created in either the single or multiple device
> case?  I.e. in the multiple device case, how about
> "<prefix>.<device>.blktrace.<cpu>"?  That way the program always does
> what the user wants, no special case documentation required.

I just now read in full your second email on the original thread where 
you mentioned doing this solution, sorry for the redundancy.  You 
expressed reservations about it, wondering what else it might break. 
I'd suggest we make the change and test it, rather than having a "-o" 
option with limited utility?  I imagine the biggest question mark is how 
blkparse handles it.

-Brett

Re: [PATCH] blktrace: disallow -o when using multiple devices

[Alan D. Brunelle]

On 09/16/2010 05:17 PM, Brett Russ wrote:
> On 09/16/2010 10:20 AM, Brett Russ wrote:
>> -Wouldn't it be a good idea to just add<prefix> before the filename(s)
>> that would have been created in either the single or multiple device
>> case? I.e. in the multiple device case, how about
>> "<prefix>.<device>.blktrace.<cpu>"? That way the program always does
>> what the user wants, no special case documentation required.
>
> I just now read in full your second email on the original thread where
> you mentioned doing this solution, sorry for the redundancy. You
> expressed reservations about it, wondering what else it might break. I'd
> suggest we make the change and test it, rather than having a "-o" option
> with limited utility? I imagine the biggest question mark is how
> blkparse handles it.
>
> -Brett

I'm not so sure about this Brett - people may have scripts out there 
that depend on the current semantics. I'd be OK with doing it if Jens 
thought this was a minor concern...

Regards,
Alan

Re: [PATCH] blktrace: disallow -o when using multiple devices

[Brett Russ]

On 09/20/2010 10:35 AM, Alan D. Brunelle wrote:
> On 09/16/2010 05:17 PM, Brett Russ wrote:
>> On 09/16/2010 10:20 AM, Brett Russ wrote:
>>> -Wouldn't it be a good idea to just add<prefix>  before the filename(s)
>>> that would have been created in either the single or multiple device
>>> case? I.e. in the multiple device case, how about
>>> "<prefix>.<device>.blktrace.<cpu>"? That way the program always does
>>> what the user wants, no special case documentation required.
>>
>> I just now read in full your second email on the original thread where
>> you mentioned doing this solution, sorry for the redundancy. You
>> expressed reservations about it, wondering what else it might break. I'd
>> suggest we make the change and test it, rather than having a "-o" option
>> with limited utility? I imagine the biggest question mark is how
>> blkparse handles it.
>>
>> -Brett
>
> I'm not so sure about this Brett - people may have scripts out there
> that depend on the current semantics. I'd be OK with doing it if Jens
> thought this was a minor concern...

As long as the change only affected output of multiple devices with 
"-o", which doesn't work today, then I'd think no scripts could be affected.

Thanks,
Brett