Enlarging Kernel Stack (doesn't this question ever go away?)

Enlarging Kernel Stack (doesn't this question ever go away?)

[David C. Chiu]

What other thing needs to be done in addition to modifying
INIT_TASK_SIZE in include/linux/sched.h? This is for the 405GP kernel.
TIA.

** Sent via the linuxppc-embedded mail list. See http://lists.linuxppc.org/

Re: Enlarging Kernel Stack (doesn't this question ever go away?)

[Tom Rini]

On Mon, Jun 24, 2002 at 01:50:07PM -0700, David C. Chiu wrote:

> What other thing needs to be done in addition to modifying
> INIT_TASK_SIZE in include/linux/sched.h? This is for the 405GP kernel.

Why are you trying to enlarge the stack?  This question never goes away
since it's almost always wrong to enlarge the stack...

--
Tom Rini (TR1265)
http://gate.crashing.org/~trini/

** Sent via the linuxppc-embedded mail list. See http://lists.linuxppc.org/

Re: Enlarging Kernel Stack (doesn't this question ever go away?)

[Dan Malek]

Tom Rini wrote:

> Why are you trying to enlarge the stack?  This question never goes away
> since it's almost always wrong to enlarge the stack...

That's correct......you should be allocating space with vmalloc() and
just hanging on to a pointer that you vfree() later.


	-- Dan


** Sent via the linuxppc-embedded mail list. See http://lists.linuxppc.org/

RE: Enlarging Kernel Stack (doesn't this question ever go away?)

[David C. Chiu]

Thanks for the tip, however the problem isn't that we are allocating
large stack variables but we have deeply nested calls...

-----Original Message-----
From: Dan Malek [mailto:dan@embeddededge.com]
Sent: Monday, June 24, 2002 3:06 PM
To: Tom Rini
Cc: David C. Chiu; linuxppc-embedded@lists.linuxppc.org
Subject: Re: Enlarging Kernel Stack (doesn't this question ever go
away?)


Tom Rini wrote:

> Why are you trying to enlarge the stack?  This question never goes
away
> since it's almost always wrong to enlarge the stack...

That's correct......you should be allocating space with vmalloc() and
just hanging on to a pointer that you vfree() later.


	-- Dan

** Sent via the linuxppc-embedded mail list. See http://lists.linuxppc.org/

Re: Enlarging Kernel Stack (doesn't this question ever go away?)

[Dan Malek]

David C. Chiu wrote:

> Thanks for the tip, however the problem isn't that we are allocating
> large stack variables but we have deeply nested calls...


Hmmmm.....Don't do that, then :-)  The stack plus thread structures
consist of two pages.  I don't know how many places may still assume
that in the code.  Due to both physical and virtual access to the pages,
I think they may also have to be physically contiguous (they used to be),
so you will have to allocate 4 pages as the next step.

Good Luck.


	-- Dan


** Sent via the linuxppc-embedded mail list. See http://lists.linuxppc.org/