Re: [Xen-devel] [PATCH 0/4] enhance lock debugging

[Jan Beulich <JBeulich@xxxxxxxx>]

On 08.08.2019 11:45, Andrew Cooper wrote:
> On 08/08/2019 10:08, Jan Beulich wrote:
>> On 08.08.2019 10:33, Andrew Cooper wrote:
>>> On 08/08/2019 05:50, Juergen Gross wrote:
>>>> On 07.08.19 20:11, Andrew Cooper wrote:
>>>>> <snip>
>>>>> Its not exactly the easiest to dump to follow.
>>>>>
>>>>> First of all - I don't see why the hold/block time are printed like
>>>>> that.  It
>>>>> might be a holdover from the 32bit build, pre PRId64 support.  They
>>>>> should
>>>>> probably use PRI_stime anyway.
>>>> Fine with me.
>>>>
>>>>> The domid rendering is unfortunate.  Ideally we'd use %pd but that would
>>>>> involve rearranging the logic to get a struct domain* in hand.
>>>>> Seeing as
>>>>> you're the last person to modify this code, how hard would that be to
>>>>> do?
>>>> It would completely break the struct type agnostic design.
>>> Ok.  As an alternatively, how about %pdr which takes a raw domid?  It
>>> would be a trivial adjustment in the vsnprintf code, and could plausibly
>>> be useful elsewhere where we have a domid and not a domain pointer.
>> At the risk of upsetting / annoying you:
> 
> Yes you really have
> 
>> A domid_t would again
>> better be formatted via %od (and without the need to take the
>> address of a respective variable). In the case here it would
>> have the minor additional benefit of conserving on format string
>> length.
> 
> There are concrete reasons why it is a terrible idea, because none of
> this has anything to do with octal, and that %od has a well defined
> meaning which is not related to domid's.

I'm curious to learn what well defined meaning %od has.

>  There is also a very good
> reason why all the magic is hidden behind one single formatter.
> 
> You seem hell bent on making it actively confusing and complicated to
> write and read printk()'s, and I am not willing to lumber anyone
> developing on Xen with this cognitive complexity.
> 
> I am stick to death of having the same over and over, so let me stop any
> further wasting of time and be absolutely crystal clear.
> 
> NACK to any form of overloading %o

In which case, if I took a similar position for the PCI SBDF
formatting using %pp, we'd be in a dead end. Waste of time or not
- while you have made crystal clear why you personally dislike
overloading %o, afaic you've not provided any non-subjective
(i.e. truly technical) reasons, which would be what might help
convince me of sticking to just %p overloading.

Jan
_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxxxxxxxxx
https://lists.xenproject.org/mailman/listinfo/xen-devel