2.4.7pre8aa1

2.4.7pre8aa1

[Andrea Arcangeli]

Diff between 2.4.7pre6aa1 and 2.4.7pre8aa1 (besides moving on top
of 2.4.7pre8).

Only in 2.4.7pre8aa1/: 00_do_swap_page-fix-1

	Account major faults for swapins. (from -ac)

Only in 2.4.7pre6aa1: 00_drop_async-io-get_bh-1
Only in 2.4.7pre8aa1/: 00_drop_async-io-get_bh-2

	Rediffed for trivial reiserfs reject (reiserfs updates are included in
	pre8).

Only in 2.4.7pre8aa1/: 00_ircomm-t39m-1

	Fix for irda so that it works with my GPRS phone.

Only in 2.4.7pre6aa1: 00_lvm-0.9.1_beta7-4.bz2
Only in 2.4.7pre8aa1/: 00_lvm-0.9.1_beta7-5.bz2

	Rediffed for trivial rejects about updates that
	were just included into this patch.

Only in 2.4.7pre6aa1: 00_lvm-0.9.1_beta7-4_rwsem-fast-path-1
Only in 2.4.7pre8aa1/: 00_lvm-0.9.1_beta7-5_rwsem-fast-path-2

	Rediffed so we solve snapshot hash lookups with the
	read lock as well (low prio optimization). Also
	#if 0 around the _pe_lock instead of any other fix,
	since the pv_move with live writes going on the lv
	is racy anyways in lvm beta7.

Only in 2.4.7pre6aa1: 00_rwsem-15
Only in 2.4.7pre8aa1/: 00_rwsem-16

	Found a race in asm xchgadd algorithm, possibly mainline
	could be affected too, will fix it and inspect mainline
	ASAP.

Only in 2.4.7pre6aa1: 00_sched-yield-1

	Merged in mainline.

Only in 2.4.7pre8aa1/: 00_via-quirks-1

	VIA fixups from -ac.

Only in 2.4.7pre6aa1: 10_prefetch-1
Only in 2.4.7pre8aa1/: 10_prefetch-2

	Rediffed for trivial rejects.

Only in 2.4.7pre6aa1/30_tux: 30_atomic-alloc-2
Only in 2.4.7pre8aa1/30_tux: 30_atomic-alloc-3

	Rediffed for trivial rejects.

Only in 2.4.7pre8aa1/30_tux: 32_tux-uml-1

	Moved the uml part of the -aa updates in the tux directory
	since they are not needed when the tux patches are not applied.

Only in 2.4.7pre6aa1: 40_blkdev-pagecache-5
Only in 2.4.7pre8aa1/: 40_blkdev-pagecache-6

	Default opens to O_LARGEFILE in blkdev_open so that mkreiserfs
	doesn't need to be recompiled and we don't break
	backwards compatibility (can be dropped in 2.5 if we want to).

Only in 2.4.7pre8aa1/: 41_blkdev-pagecache-5_drop_get_bh_async-1

	When blkdev patch is applied in combination with the
	00_drop_async-io-get_bh patch we must remeber to drop
	the get_bh from the blkdev async_io too.

Only in 2.4.7pre6aa1: 51_uml-ac-to-aa-2.bz2
Only in 2.4.7pre8aa1/: 51_uml-ac-to-aa-3.bz2

	Moved part of it in the tux directory so it can compile
	without tux (in reality I got errno compilation error
	but it's low prio and I'll sort it out later, Jeff Dike any
	hint is welcome ;).

URL:

	ftp://ftp.us.kernel.org/pub/linux/kernel/people/andrea/kernels/v2.4/2.4.7pre8aa1/
	ftp://ftp.us.kernel.org/pub/linux/kernel/people/andrea/kernels/v2.4/2.4.7pre8aa1.bz2

Andrea

Re: 2.4.7pre8aa1

[Jeff Dike]

> Only in 2.4.7pre6aa1: 51_uml-ac-to-aa-2.bz2
> Only in 2.4.7pre8aa1/: 51_uml-ac-to-aa-3.bz2
>         Moved part of it in the tux directory so it can compile
>         without tux (in reality I got errno compilation error
>         but it's low prio and I'll sort it out later, Jeff Dike any
>         hint is welcome ;).

This is the patch I sent to Alan a while back which works around the problem.

rmk suggested a better way which I'll add at some point.

				Jeff


diff -Naur -X exclude-files ac_cur/arch/um/Makefile ac/arch/um/Makefile
--- ac_cur/arch/um/Makefile	Mon Jul  9 13:05:03 2001
+++ ac/arch/um/Makefile	Mon Jul  9 13:26:21 2001
@@ -20,6 +20,8 @@
 LINK_PROFILE = $(PROFILE) -Wl,--wrap,__monstartup
 endif
 
+CFLAGS := $(subst -fno-common,,$(CFLAGS))
+
 SUBDIRS += $(ARCH_DIR)/fs $(ARCH_DIR)/drivers $(ARCH_DIR)/kernel \
 	$(ARCH_DIR)/sys-$(SUBARCH)

Re: 2.4.7pre8aa1

[Jeff Dike]

andrea@suse.de said:
> works fine thanks! Of course I agree with rmk it would be better not
> to disable -fno-common but this is ok for now ;) 

Yeah, it's temporary.  rmk's idea was to use the link script to toss errno.o
out of the final binary.

> (after all we would
> catch any potential important name collision during the compiles of
> the other targets)

Agreed.  -fno-common is definitely good.  The only conflict is errno, but 
when Arjan first started playing with -fno-common, he found a couple of UML
bugs.  

				Jeff

Re: 2.4.7pre8aa1

[Andrea Arcangeli]

On Thu, Jul 19, 2001 at 06:45:38PM -0400, Jeff Dike wrote:
> > Only in 2.4.7pre6aa1: 51_uml-ac-to-aa-2.bz2
> > Only in 2.4.7pre8aa1/: 51_uml-ac-to-aa-3.bz2
> >         Moved part of it in the tux directory so it can compile
> >         without tux (in reality I got errno compilation error
> >         but it's low prio and I'll sort it out later, Jeff Dike any
> >         hint is welcome ;).
> 
> This is the patch I sent to Alan a while back which works around the problem.
> 
> rmk suggested a better way which I'll add at some point.
> 
> 				Jeff
> 
> 
> diff -Naur -X exclude-files ac_cur/arch/um/Makefile ac/arch/um/Makefile
> --- ac_cur/arch/um/Makefile	Mon Jul  9 13:05:03 2001
> +++ ac/arch/um/Makefile	Mon Jul  9 13:26:21 2001
> @@ -20,6 +20,8 @@
>  LINK_PROFILE = $(PROFILE) -Wl,--wrap,__monstartup
>  endif
>  
> +CFLAGS := $(subst -fno-common,,$(CFLAGS))
> +
>  SUBDIRS += $(ARCH_DIR)/fs $(ARCH_DIR)/drivers $(ARCH_DIR)/kernel \
>  	$(ARCH_DIR)/sys-$(SUBARCH)

works fine thanks! Of course I agree with rmk it would be better not to
disable -fno-common but this is ok for now ;) (after all we would catch
any potential important name collision during the compiles of the other
targets)

Andrea

Re: 2.4.7pre8aa1

[Andrea Arcangeli]

On Fri, Jul 20, 2001 at 03:37:49AM +0200, Andrea Arcangeli wrote:
> On Thu, Jul 19, 2001 at 06:45:38PM -0400, Jeff Dike wrote:
> > > Only in 2.4.7pre6aa1: 51_uml-ac-to-aa-2.bz2
> > > Only in 2.4.7pre8aa1/: 51_uml-ac-to-aa-3.bz2
> > >         Moved part of it in the tux directory so it can compile
> > >         without tux (in reality I got errno compilation error
> > >         but it's low prio and I'll sort it out later, Jeff Dike any
> > >         hint is welcome ;).
> > 
> > This is the patch I sent to Alan a while back which works around the problem.
> > 
> > rmk suggested a better way which I'll add at some point.
> > 
> > 				Jeff
> > 
> > 
> > diff -Naur -X exclude-files ac_cur/arch/um/Makefile ac/arch/um/Makefile
> > --- ac_cur/arch/um/Makefile	Mon Jul  9 13:05:03 2001
> > +++ ac/arch/um/Makefile	Mon Jul  9 13:26:21 2001
> > @@ -20,6 +20,8 @@
> >  LINK_PROFILE = $(PROFILE) -Wl,--wrap,__monstartup
> >  endif
> >  
> > +CFLAGS := $(subst -fno-common,,$(CFLAGS))
> > +
> >  SUBDIRS += $(ARCH_DIR)/fs $(ARCH_DIR)/drivers $(ARCH_DIR)/kernel \
> >  	$(ARCH_DIR)/sys-$(SUBARCH)
> 
> works fine thanks! Of course I agree with rmk it would be better not to

I should have said "compiles" fine (not "works" fine :).

__initdata is broken in uml and the kernel deadlocks because the wait
list is empty in complete() despite wait_for_completion actually
registered correctly into it. This because wait_for_completion runs in
a different address space than complete() and the virtual memory is not
shared across the two address spaces (it is not rempped in a MAP_SHARED
so it generates a cow). The registration is basically only private to
the entity that is registrating and it will never get visible to the
waker task that will do nothing. This is a severe bug not just for the
completion code in 2.4.7. here the fix (plus the ptrace bit):

--- ./arch/um/kernel/ptrace.c.~1~	Fri Jul 20 17:23:16 2001
+++ ./arch/um/kernel/ptrace.c	Fri Jul 20 17:43:56 2001
@@ -43,7 +43,7 @@
 	if (request == PTRACE_ATTACH) {
 		if (child == current)
 			goto out_tsk;
-		if ((!child->dumpable ||
+		if ((!ptrace_dumpable(child) ||
 		     (current->uid != child->euid) ||
 		     (current->uid != child->suid) ||
 		     (current->uid != child->uid) ||
--- ./arch/um/link.ld.in.~1~	Fri Jul 20 17:23:16 2001
+++ ./arch/um/link.ld.in	Mon Jul 23 03:06:40 2001
@@ -45,7 +45,6 @@
   . = ALIGN(4096);		/* Init code and data */
   __init_begin = .;
   .text.init : { *(.text.init) }
-  .data.init : { *(.data.init) }
   . = ALIGN(16);
   __setup_start = .;
   .setup.init : { *(.setup.init) }
@@ -92,6 +91,7 @@
     . = ALIGN(16384);		/* init_task */
     *(.data.init_task)
     *(.data)
+    *(.data.init)
     *(.gnu.linkonce.d*)
     CONSTRUCTORS
   }


(I didn't update the init_begin/end logic since there's no logic at all
in uml because we don't free that memory anyways so there's no side
effect because of that change)

> disable -fno-common but this is ok for now ;) (after all we would catch
> any potential important name collision during the compiles of the other
> targets)
> 
> Andrea


Andrea

Re: 2.4.7pre8aa1

[Jeff Dike]

andrea@suse.de said:
> I should have said "compiles" fine (not "works" fine :). 

2.4.6 "works" fine :-)

I just finished fixing 2.4.7.  My fixes are similar to yours (I moved 
stext rather than moving .data.init).

> __initdata is broken in uml and the kernel deadlocks because the wait
> list is empty in complete() despite wait_for_completion actually
> registered correctly into it. This because wait_for_completion runs in
> a different address space than complete() and the virtual memory is
> not shared across the two address spaces (it is not rempped in a
> MAP_SHARED so it generates a cow). The registration is basically only
> private to the entity that is registrating and it will never get
> visible to the waker task that will do nothing. This is a severe bug
> not just for the completion code in 2.4.7.

Nice debugging, btw.  I spent an hour this morning chasing that problem.

My 2.4.7 patch is available from the usual places:

       http://ftp.nl.linux.org/pub/uml/uml-patch-2.4.7-1.bz2
       http://uml-pub.ists.dartmouth.edu/uml/uml-patch-2.4.7-1.bz2
       http://prdownloads.sourceforge.net/user-mode-linux/uml-patch-2.4.7-1.bz2

				Jeff