2.4.19pre7aa3

2.4.19pre7aa3

[Andrea Arcangeli]

URL:

	http://www.us.kernel.org/pub/linux/kernel/people/andrea/kernels/v2.4/2.4.19pre7aa3.gz
	http://www.us.kernel.org/pub/linux/kernel/people/andrea/kernels/v2.4/2.4.19pre7aa3/

Diff between 2.4.19pre7aa2 and 2.4.19pre7aa3:

Only in 2.4.19pre7aa3: 00_alpha-numa-VALID_PAGE-1

	Fix VALID_PAGE for virt_to_page outside the ram space.
	Bug found by Russell King. Alternate cleaner (but
	more invasive) fix for 2.5 is availabe from Roman Zippel.
	Doesn't worth to change the VALID_PAGE logic in 2.4 IMHO.

Only in 2.4.19pre7aa2: 00_get_pid-no-deadlock-and-boosted-2
Only in 2.4.19pre7aa3: 00_get_pid-no-deadlock-and-boosted-3

	s/set_pid/__set_pid/ noticed by Russell King.

Only in 2.4.19pre7aa3: 00_initrd-free-1

	/dev/root.old is been moved under /old after / is been
	moved to /old. Found and fixed by Michael Schroeder.

Only in 2.4.19pre7aa2: 00_max_bytes-3
Only in 2.4.19pre7aa3: 00_max_bytes-4

	Export symbol so intermezzo can compile.

Only in 2.4.19pre7aa3: 00_pre7ac3-p4-xeon-1

	Recognize p4-xeon. Extracted from the diff between
	2.4.19pre7ac2 and 2.4.19pre7ac3.

Only in 2.4.19pre7aa2: 00_rcu-poll-4
Only in 2.4.19pre7aa3: 00_rcu-poll-5

	Drop a buggy optimization that could lead not to reach the quiescent
	state in the cpu runing rcu_prepare_polling(). From Paul McKenney.
	Harmless bug inside -aa because it's used only in the rmmod path.

Only in 2.4.19pre7aa3: 00_timer_bh-deadlock-1

	Forward port to 2.4 this old longstanding anti deadlock timer fix
	that is been merged in 2.2 ages ago.

	ftp://ftp.us.kernel.org/pub/linux/kernel/people/andrea/patches/v2.2/2.2.14/timer_bh-deadlock-3.gz

Only in 2.4.19pre7aa2: 00_vm86-2
Only in 2.4.19pre7aa3: 00_vm86-3

	vm86 updates from 2.4.19pre7ac3.

Only in 2.4.19pre7aa3: 00_wake_up_page-1

	Reintroduced wake_up_page (not deadlock prone anymore), for modules
	that were waking up pages.

Only in 2.4.19pre7aa2: 50_uml-patch-2.4.18-18.gz
Only in 2.4.19pre7aa3: 50_uml-patch-2.4.18-21.gz
Only in 2.4.19pre7aa3: 53_uml-page-struct-updates-1
Only in 2.4.19pre7aa2: 55_uml-page_address-2
Only in 2.4.19pre7aa3: 55_uml-page_address-3

	New version from Jeff at user-mode-linux.sourceforge.net.

Only in 2.4.19pre7aa2: 81_x86_64-arch-4.gz
Only in 2.4.19pre7aa3: 81_x86_64-arch-5.gz
Only in 2.4.19pre7aa2: 85_x86_64-mmx-xmm-init-2
Only in 2.4.19pre7aa3: 85_x86_64-mmx-xmm-init-3
Only in 2.4.19pre7aa2: 87_x86_64-dec_and_lock-2

	Merged latest cvs.x86-64.org CVS updates.

Only in 2.4.19pre7aa2: 91_zone_start_pfn-1
Only in 2.4.19pre7aa3: 91_zone_start_pfn-2

	Avoid some gcc warning while compiling with alpha target.

Andrea

Re: 2.4.19pre7aa3

[Christoph Hellwig]

On Tue, Apr 30, 2002 at 08:31:54PM +0200, Andrea Arcangeli wrote:
> Only in 2.4.19pre7aa3: 00_wake_up_page-1
> 
> 	Reintroduced wake_up_page (not deadlock prone anymore), for modules
> 	that were waking up pages.

For what module?  (Don't say a agp/drm upgrade!)  As the person who invented
wake_up_page I can't really see a good reason for it anymore.  Every single
caller should have used unlock_page() instead.

Re: 2.4.19pre7aa3

[Russell King]

On Tue, Apr 30, 2002 at 08:31:54PM +0200, Andrea Arcangeli wrote:
> Only in 2.4.19pre7aa2: 00_get_pid-no-deadlock-and-boosted-2
> Only in 2.4.19pre7aa3: 00_get_pid-no-deadlock-and-boosted-3
> 
> 	s/set_pid/__set_pid/ noticed by Russell King.

s/set_pid/set_bit/g

-- 
Russell King (rmk@arm.linux.org.uk)                The developer of ARM Linux
             http://www.arm.linux.org.uk/personal/aboutme.html

Re: 2.4.19pre7aa3

[rwhron]

Extended changelog for 2.4.19pre7aa3 at
http://home.earthlink.net/~rwhron/kernel/andrea/2.4.19pre7aa3.html

Another new page is benchmarks on the latest releases from 
various kernel trees:
http://home.earthlink.net/~rwhron/kernel/latest.html

The "latest" page will get updated very regularly.
The purpose of the "latest" page isn't to say kernel
X is better than Y.  It may be useful for people who 
like to experiment with patches, hunting for what gives 
a kernel it's edge.  Each kernel excels in different 
areas.  

If there are any hunters out there, you may want to
look at the pages by tree and the big page too.  Some
of the tests vary between runs, and others are pretty
consistent.  

Have fun.
-- 
Randy Hron

block a thread in kernel, thanks alot

[Wanghong Yuan]

Hi,

How can I expicitly block a thread in kernel?  For example, I can use
wake_up_process(task) to wakeup (and put to runqueue). How to block task, if
I do not want it to run? Thanks a lot

Re: 2.4.19pre7aa3

[Andrea Arcangeli]

On Tue, Apr 30, 2002 at 08:20:10PM +0100, Christoph Hellwig wrote:
> On Tue, Apr 30, 2002 at 08:31:54PM +0200, Andrea Arcangeli wrote:
> > Only in 2.4.19pre7aa3: 00_wake_up_page-1
> > 
> > 	Reintroduced wake_up_page (not deadlock prone anymore), for modules
> > 	that were waking up pages.
> 
> For what module?  (Don't say a agp/drm upgrade!)  As the person who invented

that's a drm upgrade indeed.

> wake_up_page I can't really see a good reason for it anymore.  Every single
> caller should have used unlock_page() instead.

Agreed, thanks for noticing.

Andrea