2.5.46-mm1

2.5.46-mm1

[Andrew Morton]

url: http://www.zip.com.au/~akpm/linux/patches/2.5/2.5.46/2.5.46-mm1/

It wasn't clear whether it was useful or desirable to keep these patchsets
turning over.  But it will be helpful to keep them as a marshalling point
for people to see what is queued up, to get some additional testing and
stabilisation and for people to sync up against.  And also to keep things
like shared pagetables and dcache-rcu under test.

2.5.46-mm1 includes various fixes to things, Bill's hugetlb rework,
dcache-rcu and shared pagetables.

Also the patches which make the address_space's private_lock and page_lock
irq-safe.  So Badari can run set_page_dirty() from interrupts...


linus.patch
  cset-1.895.1.7-to-1.925.txt.gz

kgdb.patch

net-timer-init.patch

genksyms-hurts.patch
  fix exporting of per-cpu symbols for modversions

misc.patch
  misc fixes

writev-bad-seg-fix.patch
  Fix readv/writev return value

wli-01-iowait.patch
  SMP iowait stats

wli-02-zap_hugetlb_resources.patch
  hugetlb: fix zap_hugetlb_resources()

wli-03-remove-unlink_vma.patch
  hugetlb: remove unlink_vma()

wli-04-internalize-hugetlb-init.patch
  hugetlb: internalize hugetlb init

wli-05-sysctl-cleanup.patch
  hugetlb: remove sysctl.c intrusion

wli-06-cleanup-proc.patch
  hugetlb: remove /proc/ intrusion

wli-07-hugetlb-static.patch
  hugetlb: make private functions static

msec-fix.patch
  Fix math underflow in disk accounting

touch_buffer-fix.patch
  buffer_head refcounting fixes and cleanup

mbcache-atomicity-fix.patch

pgalloc-accounting-fix.patch
  fix page alloc/free accounting

htree-fix.patch
  fix ext3-htree buffer_head leak

irq-save-vm-locks.patch
  make mapping->page_lock irq-safe

irq-safe-private-lock.patch
  make mapping->private_lock irq-safe

akpm-deadline.patch
  deadline scheduler tweaks

dcache_rcu.patch
  Use RCU for dcache

page-reservation.patch
  Page reservation API

resurrect-incremental-min.patch
  strengthen the `incremental min' logic in the page allocator

wli-show_free_areas.patch
  show_free_areas extensions

shpte-ng.patch
  pagetable sharing for ia32

Re: 2.5.46-mm1 3 uninitialized timers during boot, ipv6 related?

[Helge Hafting]

I see these aren't dangerous, but I guess they're there for
someone to see.  So here they are.
Seems they have something to do with ipv6.

VFS: Mounted root (ext2 filesystem) readonly.
Mounted devfs on /dev
Freeing unused kernel memory: 308k freed
Adding 502700k swap on /dev/ide/host0/bus0/target0/lun0/part5. 
Priority:-1 extents:1
NTFS volume version 1.2.
Uninitialised timer!
This is just a warning.  Your computer is OK
function=0xc02b53dc, data=0xc1293aec
Call Trace:
 [<c011ebc0>] check_timer_failed+0x40/0x4c
 [<c02b53dc>] igmp6_timer_handler+0x0/0x50
 [<c011eff1>] del_timer+0x15/0x74
 [<c02b529e>] igmp6_join_group+0x8a/0x104
 [<c02b455c>] igmp6_group_added+0xb0/0xbc
 [<c02bcb99>] fl_create+0x1b9/0x1f8
 [<c02b48a9>] ipv6_dev_mc_inc+0x281/0x294
 [<c02a5a1e>] addrconf_join_solict+0x3a/0x44
 [<c02a6e73>] addrconf_dad_start+0x13/0x12c
 [<c02a6844>] addrconf_add_linklocal+0x28/0x40
 [<c02a68f5>] addrconf_dev_config+0x99/0xa4
 [<c02a69de>] addrconf_notify+0x52/0xc0
 [<c0121d5a>] notifier_call_chain+0x1e/0x38
 [<c024f976>] dev_open+0xa2/0xac
 [<c0250a55>] dev_change_flags+0x51/0x104
 [<c0281420>] devinet_ioctl+0x2bc/0x598
 [<c0283827>] inet_ioctl+0x7b/0xb8
 [<c02c3a9b>] packet_ioctl+0x173/0x190
 [<c0249943>] sock_ioctl+0x1ef/0x218
 [<c014bbf9>] sys_ioctl+0x21d/0x274
 [<c0108943>] syscall_call+0x7/0xb

Uninitialised timer!
This is just a warning.  Your computer is OK
function=0xc02b53dc, data=0xc1293a1c
Call Trace:
 [<c011ebc0>] check_timer_failed+0x40/0x4c
 [<c02b53dc>] igmp6_timer_handler+0x0/0x50
 [<c011eff1>] del_timer+0x15/0x74
 [<c02b529e>] igmp6_join_group+0x8a/0x104
 [<c02b455c>] igmp6_group_added+0xb0/0xbc
 [<c02bcb99>] fl_create+0x1b9/0x1f8
 [<c02b48a9>] ipv6_dev_mc_inc+0x281/0x294
 [<c02a5a1e>] addrconf_join_solict+0x3a/0x44
 [<c02a6e73>] addrconf_dad_start+0x13/0x12c
 [<c02a6844>] addrconf_add_linklocal+0x28/0x40
 [<c02a68f5>] addrconf_dev_config+0x99/0xa4
 [<c02a69de>] addrconf_notify+0x52/0xc0
 [<c0121d5a>] notifier_call_chain+0x1e/0x38
 [<c024f976>] dev_open+0xa2/0xac
 [<c0250a55>] dev_change_flags+0x51/0x104
 [<c0281420>] devinet_ioctl+0x2bc/0x598
 [<c0283827>] inet_ioctl+0x7b/0xb8
 [<c0249943>] sock_ioctl+0x1ef/0x218
 [<c014bbf9>] sys_ioctl+0x21d/0x274
 [<c0108943>] syscall_call+0x7/0xb

Uninitialised timer!
This is just a warning.  Your computer is OK
function=0xc02b53dc, data=0xc1293b54
Call Trace:
 [<c011ebc0>] check_timer_failed+0x40/0x4c
 [<c02b53dc>] igmp6_timer_handler+0x0/0x50
 [<c011eff1>] del_timer+0x15/0x74
 [<c02b529e>] igmp6_join_group+0x8a/0x104
 [<c02b455c>] igmp6_group_added+0xb0/0xbc
 [<c02bcb99>] fl_create+0x1b9/0x1f8
 [<c02b48a9>] ipv6_dev_mc_inc+0x281/0x294
 [<c02a5a1e>] addrconf_join_solict+0x3a/0x44
 [<c02a6e73>] addrconf_dad_start+0x13/0x12c
 [<c02a6844>] addrconf_add_linklocal+0x28/0x40
 [<c02a68f5>] addrconf_dev_config+0x99/0xa4
 [<c02a69de>] addrconf_notify+0x52/0xc0
 [<c0121d5a>] notifier_call_chain+0x1e/0x38
 [<c024f976>] dev_open+0xa2/0xac
 [<c0250a55>] dev_change_flags+0x51/0x104
 [<c0281420>] devinet_ioctl+0x2bc/0x598
 [<c0283827>] inet_ioctl+0x7b/0xb8
 [<c0249943>] sock_ioctl+0x1ef/0x218
 [<c014bbf9>] sys_ioctl+0x21d/0x274
 [<c0108943>] syscall_call+0x7/0xb

eth0: no IPv6 routers present

Re: 2.5.46-mm1 3 uninitialized timers during boot, ipv6 related?

[Dipankar Sarma]

On Wed, Nov 06, 2002 at 02:40:05PM +0100, Helge Hafting wrote:
> Uninitialised timer!
> This is just a warning.  Your computer is OK
> function=0xc02b53dc, data=0xc1293aec
> Call Trace:
>  [<c011ebc0>] check_timer_failed+0x40/0x4c
>  [<c02b53dc>] igmp6_timer_handler+0x0/0x50
>  [<c011eff1>] del_timer+0x15/0x74
>  [<c02b529e>] igmp6_join_group+0x8a/0x104
>  [<c02b455c>] igmp6_group_added+0xb0/0xbc
>  [<c02bcb99>] fl_create+0x1b9/0x1f8
>  [<c02b48a9>] ipv6_dev_mc_inc+0x281/0x294
>  [<c02a5a1e>] addrconf_join_solict+0x3a/0x44
>  [<c02a6e73>] addrconf_dad_start+0x13/0x12c
>  [<c02a6844>] addrconf_add_linklocal+0x28/0x40
>  [<c02a68f5>] addrconf_dev_config+0x99/0xa4
>  [<c02a69de>] addrconf_notify+0x52/0xc0
>  [<c0121d5a>] notifier_call_chain+0x1e/0x38
>  [<c024f976>] dev_open+0xa2/0xac
>  [<c0250a55>] dev_change_flags+0x51/0x104
>  [<c0281420>] devinet_ioctl+0x2bc/0x598
>  [<c0283827>] inet_ioctl+0x7b/0xb8
>  [<c02c3a9b>] packet_ioctl+0x173/0x190
>  [<c0249943>] sock_ioctl+0x1ef/0x218
>  [<c014bbf9>] sys_ioctl+0x21d/0x274
>  [<c0108943>] syscall_call+0x7/0xb

This patch should fix this up.

Thanks
Dipankar


--- linux-2.5.46-base/net/ipv6/mcast.c	Mon Oct 21 02:12:47 2002
+++ linux-2.5.46-mm/net/ipv6/mcast.c	Wed Nov  6 22:01:32 2002
@@ -296,6 +296,7 @@
 	}
 
 	memset(mc, 0, sizeof(struct ifmcaddr6));
+	init_timer(&mc->mca_timer);
 	mc->mca_timer.function = igmp6_timer_handler;
 	mc->mca_timer.data = (unsigned long) mc;
 

Re: 2.5.46-mm1 3 uninitialized timers during boot, ipv6 related?

[Andrew Morton]

Helge Hafting wrote:
> 
> I see these aren't dangerous, but I guess they're there for
> someone to see.  So here they are.
> Seems they have something to do with ipv6.
> 

Dave has a bunch of additional fixes for net/*.  Expect them to emerge
from kernel.org real soon.

Re: 2.5.46-mm1

[Paul P Komkoff Jr]

Replying to Andrew Morton:
> stabilisation and for people to sync up against.  And also to keep things
> like shared pagetables and dcache-rcu under test.

Why sharepte is dependent on highmem now ?

I thought I will benefit from it on forkloads on lowmem too ...

-- 
Paul P 'Stingray' Komkoff 'Greatest' Jr /// (icq)23200764 /// (http)stingr.net
  When you're invisible, the only one really watching you is you (my keychain)

Re: 2.5.46-mm1

[Andrew Morton]

Paul P Komkoff Jr wrote:
> 
> Replying to Andrew Morton:
> > stabilisation and for people to sync up against.  And also to keep things
> > like shared pagetables and dcache-rcu under test.
> 
> Why sharepte is dependent on highmem now ?
> 

Senility?   Just delete line 725 of arch/i386/Kconfig

Re: 2.5.46-mm1

[Dave McCracken]

[-- Attachment #1: Type: text/plain, Size: 605 bytes --]


--On Wednesday, November 06, 2002 20:12:50 +0300 Paul P Komkoff Jr
<i@stingr.net> wrote:

> Why sharepte is dependent on highmem now ?

It's not supposed to be.  I'm guessing it's a conversion error in the move
to Kconfig.  A patch to fix it is attached.

> I thought I will benefit from it on forkloads on lowmem too ...

It's definitely a benefit for all sizes of memory.

Dave McCracken

======================================================================
Dave McCracken          IBM Linux Base Kernel Team      1-512-838-3059
dmccr@us.ibm.com                                        T/L   678-3059

[-- Attachment #2: shpte-2.5.46-mm1-1.diff --]
[-- Type: text/plain, Size: 425 bytes --]

--- 2.5.46-mm1/arch/i386/Kconfig	2002-11-06 13:17:20.000000000 -0600
+++ 2.5.46-mm1-shsent/arch/i386/Kconfig	2002-11-06 11:38:50.000000000 -0600
@@ -722,7 +722,6 @@
 
 config SHAREPTE
 	bool "Share 3rd-level pagetables between processes"
-	depends on HIGHMEM4G || HIGHMEM64G
 	help
 	  Normally each address space has its own complete page table for all
 	  its mappings.  This can mean many mappings of a set of shared data

Re: 2.5.46-mm1

[Bill Davidsen]

On Wed, 6 Nov 2002, Andrew Morton wrote:

> 
> url: http://www.zip.com.au/~akpm/linux/patches/2.5/2.5.46/2.5.46-mm1/
> 
> It wasn't clear whether it was useful or desirable to keep these patchsets
> turning over.  But it will be helpful to keep them as a marshalling point
> for people to see what is queued up, to get some additional testing and
> stabilisation and for people to sync up against.  And also to keep things
> like shared pagetables and dcache-rcu under test.

For what it's worth, the last mm kernel which booted on my old P-II IDE
test machine was 44-mm2. With 44-mm6 and this one I get an oops on boot.
Unfortunately it isn't written to disk, scrolls off the console, and
leaves the machine totally dead to anything less than a reset. I will try
2.5.46 base after I apply the few patches needed to make it compile, and
send you my config if you think it will help.

2.5.44 and all 4[45]-ac? builds booted and ran.

This is just a "problem present" notice, I have no decent info for
debugging.

-- 
bill davidsen <davidsen@tmr.com>
  CTO, TMR Associates, Inc
Doing interesting things with little computers since 1979.

Re: 2.5.46-mm1

[Martin J. Bligh]

> For what it's worth, the last mm kernel which booted on my old P-II IDE
> test machine was 44-mm2. With 44-mm6 and this one I get an oops on boot.
> Unfortunately it isn't written to disk, scrolls off the console, and
> leaves the machine totally dead to anything less than a reset. I will try

Any chance of setting up a serial console? They're very handy for 
things like this ...

M.

Re: 2.5.46-mm1

[Andrew Morton]

"Martin J. Bligh" wrote:
> 
> > For what it's worth, the last mm kernel which booted on my old P-II IDE
> > test machine was 44-mm2. With 44-mm6 and this one I get an oops on boot.
> > Unfortunately it isn't written to disk, scrolls off the console, and
> > leaves the machine totally dead to anything less than a reset. I will try
> 
> Any chance of setting up a serial console? They're very handy for
> things like this ...
> 

"vga=extended" gets you 50 rows, which is usually enough.

Re: 2.5.46-mm1

[Martin J. Bligh]

>> > For what it's worth, the last mm kernel which booted on my old P-II IDE
>> > test machine was 44-mm2. With 44-mm6 and this one I get an oops on boot.
>> > Unfortunately it isn't written to disk, scrolls off the console, and
>> > leaves the machine totally dead to anything less than a reset. I will try
>> 
>> Any chance of setting up a serial console? They're very handy for
>> things like this ...
> 
> "vga=extended" gets you 50 rows, which is usually enough.

Depends if it keeps booting afterwards, or your pen skills are
just bad (like mine ;-))

M.

Re: 2.5.46-mm1

[Bill Davidsen]

On Thu, 7 Nov 2002, Martin J. Bligh wrote:

> > For what it's worth, the last mm kernel which booted on my old P-II IDE
> > test machine was 44-mm2. With 44-mm6 and this one I get an oops on boot.
> > Unfortunately it isn't written to disk, scrolls off the console, and
> > leaves the machine totally dead to anything less than a reset. I will try
> 
> Any chance of setting up a serial console? They're very handy for 
> things like this ...

Certainly not with a real serial terminal ;-) As far as going through a
crossover cable to another system, it's doable, but finding the time, the
cable, etc... maybe next week. I just wanted to get the base report out in
case someone had more info, or if it was a known problem. 

I will try to capture it, but I'll try any new versions which come out
before then as well.

-- 
bill davidsen <davidsen@tmr.com>
  CTO, TMR Associates, Inc
Doing interesting things with little computers since 1979.

Re: 2.5.46-mm1

[Bill Davidsen]

On Thu, 7 Nov 2002, Andrew Morton wrote:

> "Martin J. Bligh" wrote:
> > 
> > > For what it's worth, the last mm kernel which booted on my old P-II IDE
> > > test machine was 44-mm2. With 44-mm6 and this one I get an oops on boot.
> > > Unfortunately it isn't written to disk, scrolls off the console, and
> > > leaves the machine totally dead to anything less than a reset. I will try
> > 
> > Any chance of setting up a serial console? They're very handy for
> > things like this ...
> > 
> 
> "vga=extended" gets you 50 rows, which is usually enough.

Doesn't seem to do anything, nor does vga=773 which I remember from the
days of using VESA modes. Serial is definitely the one of choice, this
video card is not cooperating. I'll bring a cable from the office Monday.

Stock 2.5.46 does the same thing, but so did 44 and 44-ac? is running
happily. I'll look. 

-- 
bill davidsen <davidsen@tmr.com>
  CTO, TMR Associates, Inc
Doing interesting things with little computers since 1979.