Re: 2.5 Problem Report Status

[Robert Love]

On Sat, 2002-10-05 at 12:57, Thomas Molina wrote:

>    open                   04 Oct 2002 scheduling while atomic oops
>    6. http://marc.theaimsgroup.com/?l=linux-kernel&m=103270005902896&w=2
> 
> This appears to be a long-running problem.  Is it related to the group of 
> problems below involving "function might sleep while holding a lock" or is 
> it a scheduling system problem?

This is the same thing as all those "sleeping while atomic"
(might_sleep) bugs below.  It is just a debugging check.  It does the
same check as might_sleep but during schedule().

If you had specific culprits (i.e. foo() calls bar() which schedules
while foo() holds the baz lock) would be very useful.  Otherwise listing
this as a problem is not useful.

>    open                   29 Sep 2002 Oracle 9.2 goes OOM on startup
>   14. http://marc.theaimsgroup.com/?l=linux-kernel&m=103333545310595&w=2
> 
> This problem was reported for 2.5.39.  I have seen neither a followup, nor 
> a reference to a fix.  Does this problem still exist in 2.5.40?

Should be fixed in bk.

>    open                   2.5.40      init_irq() function doing unsafe 
>                                       things inside ide_lock
>   24. http://marc.theaimsgroup.com/?l=linux-kernel&m=103316967724891&w=2
> 
> Might sleep while holding a lock.

Is this still not fixed?  Ugh.

BTW, I like the fact you are listing specific atomicity issues.  Thank
you.  It is a lot more useful than just saying there are "sleeping while
atomic" bugs.

	Robert Love