[PATCH 1/3] tests: add xetpriority.test

[Dmitry V. Levin]

On Wed, Mar 09, 2016 at 11:27:37AM +0800, Fei, Jie/费 杰 wrote:
> On 03/08/2016 06:36 AM, Dmitry V. Levin wrote:
> >On Mon, Mar 07, 2016 at 11:29:49AM +0800, Fei, Jie/费 杰 wrote:
> >>On 03/04/2016 10:46 PM, Dmitry V. Levin wrote:
> >>>On Fri, Mar 04, 2016 at 05:10:46PM +0800, Fei Jie wrote:
> >>>[...]
> >>>>* linux/dummy.h: (sys_getpriority, sys_setpriority): Change to
> >>>>printargs_d.
> >>>Why?
> >>>
> >>I added this to make strace print xetpriority in the following format:
> >>getpriority(0, 17207) = 20
> >>instead of
> >>getpriority(PRIO_PROCESS, 18446744069414605786) = 20.
> >Why do you think that
> >	getpriority(0, 17207) = 20
> >is better than
> >	getpriority(PRIO_PROCESS, 17207) = 20
> >?
> >
> >I'd agree if you suggested to print 2nd argument of getpriority and
> >setpriority using %d format, but what's wrong with current symbolic
> >representation of their 1st argument?
> >
> My original idea was to print the 2nd argument in %d format by adding
> printargs_d, but it changed the 1st argument's format.
> Would it be helpful by modifying SYS_FUNC(xetpriority) in resource.c as 
> follows?
> 
> SYS_FUNC(getpriority)
> {
>     printxval(priorities, tcp->u_arg[0], "PRIO_???");
>     tprintf(", %d", (int) tcp->u_arg[1]);
> 
>     return RVAL_DECODED;
> }
> 
> SYS_FUNC(setpriority)
> {
>     printxval(priorities, tcp->u_arg[0], "PRIO_???");
>     tprintf(", %d, %d", (int) tcp->u_arg[1], (int) tcp->u_arg[2]);
> 
>     return RVAL_DECODED;
> }

Yes, I think that's OK.  Please reflect this change in the commit message.


-- 
ldv
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 819 bytes
Desc: not available
URL: <http://lists.strace.io/pipermail/strace-devel/attachments/20160309/ca79d584/attachment.bin>