Re: Defect in XSH pthread_attr_getschedpolicy()

[Dave Butenhof <yyyyyyyyyyyyyy@xxxxxx>]

yyyyyyy@xxxxxxxxxx wrote:

> 	Defect report from : Ulrich Drepper , Red Hat
>
> (Please direct followup comments direct to yyyyyyyyyyyyyy@xxxxxxxxxxxxx)
>
> @ page 989 line 31443 section pthread_attr_getschedpolicy() objection {ptattr-2}
>
> Problem:
>
> Defect code :  2. Omission
>
> The standard does not mention anywhere the default value of the value retrieved via pthread_attr_getschedpolicy() after a call to
> pthread_attr_init().
>
> The pthread_attr_getguardsize() man page does contain the appropriate information.
>
> I would guess the default value is SCHED_OTHER.
>
> Action:
>
> Add new paragraph after line 31442:
>
>  The default value of the /schedpolicy/ attribute is <t>SCHED_OTHER<t>
>  
Again, this is a scheduling attribute, where a single default would not 
be appropriate both for an embedded realtime system and for a multiuser 
timeshare system. The default values were intended to be implementation 
defined.

Like many parts of POSIX, this is a compromise position. Like all good 
compromises, it has the essential attribute that each side knows the 
other sides are just as uncomfortable with the result as they are. ;-)

But then, another way to look at it is that most applications don't need 
to worry about this at all. The implementation should choose defaults 
that support its normal policies. On a multiuser timeshare system, the 
default should clearly be a timeshare policy (unspecified by POSIX!) 
with a priority that contends fairly with other processes and threads on 
the system. Most applications will never need anything else. And if they 
do, they'd almost certainly need to set it manually anyway.

--
/--------------------[ yyyyyyyyyyyyyy@xxxxxx ]--------------------\
| Hewlett-Packard Company       Tru64 UNIX & VMS Thread Architect |
|     My book: http://www.awl.com/cseng/titles/0-201-63392-2/     |
\----[ http://homepage.mac.com/dbutenhof/Threads/Threads.html ]---/