Re: 2.5.67-mm2

Re: 2.5.67-mm2

[Shane Shrybman]

Hi,

2.5.67-mm2 hangs just after "ok, booting the kernel", 2.5.67-mm1 worked
fine with the same config. This on a uniproc, k7, ide machine.

shane

Re: 2.5.67-mm2

[Joshua Kwan]

[-- Attachment #1: Type: text/plain, Size: 300 bytes --]

On Sat, Apr 12, 2003 at 10:17:36PM -0400, Shane Shrybman wrote:
> 2.5.67-mm2 hangs just after "ok, booting the kernel", 2.5.67-mm1 worked
> fine with the same config. This on a uniproc, k7, ide machine.

I'm seeing this too. pentium4+ide+radeonfb

-Josh

-- 
New PGP public key: 0x27AFC3EE

[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]

Re: 2.5.67-mm2

[Felipe Alfaro Solana]

On Sun, 2003-04-13 at 04:17, Shane Shrybman wrote:
> Hi,
> 
> 2.5.67-mm2 hangs just after "ok, booting the kernel", 2.5.67-mm1 worked
> fine with the same config. This on a uniproc, k7, ide machine.

I'm experiencing exactly the same as you... It worked fine with
2.5.67-mm1 and 2.5.67-bk3.
-- 
Please AVOID sending me WORD, EXCEL or POWERPOINT attachments.
See http://www.fsf.org/philosophy/no-word-attachments.html
Linux Registered User #287198

Re: 2.5.67-mm2

[Alistair Strachan]

Joshua Kwan wrote:

[snip]
> 
> Unable to handle kernel NULL pointer dereference at virtual address
> 00000040
>  printing eip:
> c0223cf4
> *pde = 00000000
> Oops: 0002 [#1]
> CPU:    0
> EIP:    0060:[<c0223cf4>]    Not tainted VLI
> EFLAGS: 00010282
> EIP is at devclass_add_driver+0x34/0x8a
> eax: 00000040   ebx: fffffff8   ecx: fffffff0   edx: ffff0001
> esi: 00000040   edi: 00000000   ebp: c03af298   esp: c1297f50
> ds: 0007b   es: 007b   ss: 0068
> Process swapper: (pid: 1, threadinfo=c1296000 task=c1294080)
> Stack: c03af500 c03af2b4 c03ae7c4 00000000 c03ae780 c0223620 c03af298
> c03af280
>        00000001 00000000 00000000 c0223a72 c03af298 c0223a72 c03af048
>        fffffffc 00000034 c026d15d c03af298 c01afc02 c03af048 c0402ebe
>        c03af020 0000003c
> Call Trace:
>  [<c0223620>] bus_add_driver+0x82/0xd2
>  [<c0223a72>] driver_register+0x38/0x3c
>  [<c0223a72>] driver_register+0x38/0x3c
>  [<c026d15d>] usb_register+0x73/0xae
>  [<c01afc02>] pci_register_driver+0x47/0x57
>  [<c0402ebe>] uhci_hcd_init+0xd9/0x137
>  [<c0402f2b>] hid_init+0xf/0x21
>  [<c03ec6e0>] do_initcalls+0x28/0x94
>  [<c0129ce8>] init_workqueues+0xf/0x27
>  [<c0105099>] init+0x36/0x195
>  [<c0105063>] init+0x0/0x195
>  [<c0106ffd>] kernel_thread_helper+0x5/0xb
> 
> Code: 24 0c 89 6c 24 10 89 74 24 08 8b 6c 24 18 31 ff 8b 45 08 89 04 
24 e8
> f7 02 00 00 85 c0 89 c3 74 4a 8d 70 48 ba 01 00 ff ff 98 f0 <0f> c1 10 
85
> d2 0f 85 f 03 00 00 89 2c 24 e8 c0 fe ff ff 85 c0
>  <0>Kernel panic: Attempted to kill init!
> 
> This doesn't happen with 2.5.67{-bk4}. I looked at the USB changes for
> 2.5.67-mm2 and they are just about the same as in -bk4. I didn't look
> at every line, though.

I get the same thing on an mm2 boot. Are you certain it isn't a -bk4 
bug? kobject, bus_add_driver and friends have all been touched by greg 
in bk, and I can't see anything immediately obvious in the new -mm 
patches (-mm1 works fine).

I'll try with just the linus drop now.

Cheers,
Alistair.

Re: 2.5.67-mm2

[Andrew Morton]

Alistair Strachan <alistair@devzero.co.uk> wrote:
>
> > EIP is at devclass_add_driver+0x34/0x8a
> ...
> 
> I get the same thing on an mm2 boot. Are you certain it isn't a -bk4 
> bug? kobject, bus_add_driver and friends have all been touched by greg 
> in bk, and I can't see anything immediately obvious in the new -mm 
> patches (-mm1 works fine).
> 
> I'll try with just the linus drop now.

It's a bk bug.  This might make it boot:

 drivers/base/class.c |    2 +-
 1 files changed, 1 insertion(+), 1 deletion(-)

diff -puN drivers/base/class.c~a drivers/base/class.c
--- 25/drivers/base/class.c~a	2003-04-13 13:04:47.000000000 -0700
+++ 25-akpm/drivers/base/class.c	2003-04-13 13:04:52.000000000 -0700
@@ -105,7 +105,7 @@ int devclass_add_driver(struct device_dr
 	struct device_class * cls = get_devclass(drv->devclass);
 	int error = 0;
 
-	if (cls) {
+	if (cls && cls->subsys) {
 		down_write(&cls->subsys.rwsem);
 		pr_debug("device class %s: adding driver %s:%s\n",
 			 cls->name,drv->bus->name,drv->name);

_

Re: 2.5.67-mm2

[Robert Love]

On Sun, 2003-04-13 at 16:05, Andrew Morton wrote:

> It's a bk bug.  This might make it boot:

Yah, I needed a similar patch to make 2.5.67-mm2 boot.  Not sure if its
hiding the real problem or not, but it works.

Note the difference between mine and your's, though.  I think you need
it.

	Robert Love


 drivers/base/class.c |    4 ++--
 1 files changed, 2 insertions(+), 2 deletions(-)


diff -urN linux-2.5.67/drivers/base/class.c linux/drivers/base/class.c
--- linux-2.5.67/drivers/base/class.c	2003-04-07 13:33:00.000000000 -0400
+++ linux/drivers/base/class.c	2003-04-13 16:32:52.000000000 -0400
@@ -103,12 +105,12 @@
 	struct device_class * cls = get_devclass(drv->devclass);
 	int error = 0;
 
-	if (cls) {
+	if (cls && &cls->subsys) {
 		down_write(&cls->subsys.rwsem);
 		pr_debug("device class %s: adding driver %s:%s\n",
 			 cls->name,drv->bus->name,drv->name);
 		error = devclass_drv_link(drv);
-		
+
 		if (!error)
 			list_add_tail(&drv->class_list,&cls->drivers.list);
 		up_write(&cls->subsys.rwsem);

Re: 2.5.67-mm2

[Greg KH]

On Sun, Apr 13, 2003 at 04:45:24PM -0400, Robert Love wrote:
> On Sun, 2003-04-13 at 16:05, Andrew Morton wrote:
> 
> > It's a bk bug.  This might make it boot:
> 
> Yah, I needed a similar patch to make 2.5.67-mm2 boot.  Not sure if its
> hiding the real problem or not, but it works.

This patch is just hiding the real problem.  The real problem is in the
input class patch that I sent to Linus.  Either the input core needs to
be marked as a subsys_init(), or that patch needs to be reverted.

I'll work on fixing this up on Monday...

greg k-h

Re: 2.5.67-mm2

[Greg KH]

On Sun, Apr 13, 2003 at 05:49:26PM -0700, Greg KH wrote:
> 
> I'll work on fixing this up on Monday...

David Miller pointed out the proper fix, and here it is.  Let me know if
this works for everyone or not.

thanks,

greg k-h


# Input: change input_init() to be a subsys initcall
#
# This fixes oopses when it and the hid core are compiled into the kernel.

diff -Nru a/drivers/input/input.c b/drivers/input/input.c
--- a/drivers/input/input.c	Mon Apr 14 11:16:02 2003
+++ b/drivers/input/input.c	Mon Apr 14 11:16:02 2003
@@ -717,5 +717,5 @@
 	devclass_unregister(&input_devclass);
 }
 
-module_init(input_init);
+subsys_initcall(input_init);
 module_exit(input_exit);

Re: 2.5.67-mm2

[Alistair Strachan]

On Monday 14 April 2003 19:21, Greg KH wrote:
> On Sun, Apr 13, 2003 at 05:49:26PM -0700, Greg KH wrote:
> > I'll work on fixing this up on Monday...
>
> David Miller pointed out the proper fix, and here it is.  Let me know
> if this works for everyone or not.

Patching 2.5.67 to -mm3, then reverting Andrew's workaround and applying 
your fix still allows me to boot. Therefore it is fixed, here.

Cheers,
Alistair.

Re: 2.5.67-mm2

[Robert Love]

On Mon, 2003-04-14 at 14:21, Greg KH wrote:

> David Miller pointed out the proper fix, and here it is.  Let me know if
> this works for everyone or not.

Ja, bestatigt.

Danke,

	Robert Love

Re: 2.5.67-mm2

[Alistair Strachan]

On Sunday 13 April 2003 21:05, Andrew Morton wrote:
>
> It's a bk bug.  This might make it boot:
>
>  drivers/base/class.c |    2 +-
>  1 files changed, 1 insertion(+), 1 deletion(-)
>
> diff -puN drivers/base/class.c~a drivers/base/class.c
> --- 25/drivers/base/class.c~a	2003-04-13 13:04:47.000000000 -0700
> +++ 25-akpm/drivers/base/class.c	2003-04-13 13:04:52.000000000 -0700
> @@ -105,7 +105,7 @@ int devclass_add_driver(struct device_dr
>  	struct device_class * cls = get_devclass(drv->devclass);
>  	int error = 0;
>
> -	if (cls) {
> +	if (cls && cls->subsys) {
>  		down_write(&cls->subsys.rwsem);
>  		pr_debug("device class %s: adding driver %s:%s\n",
>  			 cls->name,drv->bus->name,drv->name);
>

Ah yes. Works fine, thanks.

Cheers,
Alistair.

Re: 2.5.67-mm2

[Shane Shrybman]

Hi,

The boot hang just after the  'ok, booting the kernel message' with
2.5.67-mm2 is reproducible and the problem is not seen with just 2.5.67
and the linus.patch. I also took a wild guess and tried backing out
bootmem-speedup.patch and mem_map-init-arch-hooks.patch but that didn't
help either.

shane

2.5.67-mm2

[Andrew Morton]

ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/patches/2.5/2.5.67/2.5.67-mm2/

. Lots of misc saved-up things.

. I've changed the 32-bit dev_t patch to provide a 12:20 split rather than
  16:16.  This patch is starting to drag a bit and unless someone stops me I
  might just go submit the thing.




Changes since 2.5.67-mm1:


 linus.patch

 Latest -bk

-wait_on_buffer-debug-fix.patch
-gcc-295-required.patch
-remove-nr_reverse_maps.patch
-rmap-search-speedup.patch
-rmap-tweaks.patch
-page-lock-is-spin_lock.patch
-paride-remove-blk_queue_empty.patch
-jbd-warning-fix.patch
-earlier-keyboard-init.patch
-panic-on-oops.patch
-epoll-cross-thread-deletion-fix.patch
-mbcache-missing-brelse.patch
-nfs-read-corruption-fix.patch
-MS_ASYNC-more-async.patch
-tasklist_lock-docco-fix.patch
-posix-timer-hang-fix.patch
-compound-page-fix.patch
-fadvise-file-leak.patch
-conntrack-use-after-free-fix.patch

 Merged

+p4-oprofile-fix.patch

 Try to fix oprofile on P4-HT.  Doesn't work yet.

+flush_workqueue-hang-fix.patch

 Fix a hang with delayed work and workqueue flushing

+ppc64-update-fixes.patch

 Build fixes

+kobject-leak-fix.patch

 Plug an error-path memleak

+radix_tree_delete-api-cleanup.patch

 Return a more useful value from radix_tree_delete()

+gen_rtc-compile-fix.patch

 Fix up gen_rtc.c

-ptrace-flush.patch

 Dropped - it was never right and it conflicted with the flush_page_to_ram()
 removal patch.

+sched_idle-typo-fix.patch

 Use the right priority array

+ext3-quota-deadlock-fix.patch

 Fix a lock ranking problem with ext3 and quotas

+dont-clear-PG_uptodate-on-ENOSPC.patch

 Don't mark a page non-uptodate if writeout ran out of disk space.

+stack-protection-fix.patch

 Use the right permissions on the stack segment

+sparc-PTE_FILE_MAX_BITS-fix.patch

 Teach remap_file_pages() about sparc32.

+bootmem-speedup.patch

 Solve some boot-time search complexity problems

+mem_map-init-arch-hooks.patch

 Let ia64 get at mem_map[] initialisation.  For virtually-addressed
 mem_map[].

+posix-timer-hang-fix-2.patch

 Stuff from George.

+tty-modem-control-api.patch

 Internal API for diddling RTS/CTS/etc.  So other parts of the kernel don't
 have to cook up ioctl() calls.

+kmalloc_sizes-fix.patch

 Fix the fix for the cleanup of the kmalloc_sizes array.

+proc-interrupts-kmalloc-size.patch

 Pile more kludges on the last lot.

+setserial-fix.patch

 Return the right thing from uart_set_info()

+objrmap-sort-vma-list.patch

 Broken back out of the objrmap patch.  It is still in a bit of flux.

+objrmap-vma-sorting-fix.patch

 Fixes for the i_mmap and i_mmap_shared list sorting.

+i8042-share-irqs.patch

 Teach the i8042 driver to share irq12

-earlier-keyboard-init.patch

 Seems to be breaking things.

+vmalloc-stats.patch

 Display info about the vmalloc arena state in /proc/meminfo

+meminfo-doc.patch

 Documentation for /proc/meminfo

+gfp_repeat.patch

 Implement __GFP_REPEAT: so we can consolidate lots of alloc-with-retry code.

+alloc_buffer_head-take-gfp.patch

 alloc_buffer_head() should not be assuming the allocation mode.

+pte_alloc_one-use-gfp_repeat.patch

 Use __GFP_REPEAT for pte page allocations.

+pmd_alloc_one-use-gfp_repeat.patch

 Use __GFP_REPEAT for pmd page allocations.

-dynamic-hd_struct-allocation-fixes.patch

 Folded into dynamic-hd_struct-allocation.patch

+dynamic-hd_struct-devfs-fix.patch

 Fix the dymamic hd_struct allocation patch for devfs

+lockmeter.patch

 Spinlock contention metering code.  See

	http://oss.sgi.com/projects/lockmeter/



All 104 patches

linus.patch

mm.patch
  add -mmN to EXTRAVERSION

kgdb-ga.patch
  kgdb stub for ia32 (George Anzinger's one)

kgdb-ga-warning-fix.patch
  -mm traps.c warning

kgdb-ga-up-warning-fix.patch

p4-oprofile-fix.patch
  Fix oprofile on P4's

flush_workqueue-hang-fix.patch
  flush_work_queue() fixes

tty-shutdown-race-fix.patch
  fix tty shutdown race

ppa-null-pointer-fix.patch

dmfe-kfree_skb-fix.patch
  dmfe: don't free skb with local interrupts disabled

config_spinline.patch
  uninline spinlocks for profiling accuracy.

ppc64-reloc_hide.patch

ppc64-pci-patch.patch
  Subject: pci patch

ppc64-aio-32bit-emulation.patch
  32/64bit emulation for aio

ppc64-scruffiness.patch
  Fix some PPC64 compile warnings

ppc64-update.patch
  ppc64 update

ppc64-update-fixes.patch

sym-do-160.patch
  make the SYM driver do 160 MB/sec

config-PAGE_OFFSET.patch
  Configurable kenrel/user memory split

kobject-leak-fix.patch
  kobject hotplug fixes

radix_tree_delete-api-cleanup.patch
  radix_tree_delete API improvement

gen_rtc-compile-fix.patch
  Fix gen_rtc compilation error

buffer-debug.patch
  buffer.c debugging

warn-null-wakeup.patch
  remove the test for null waitqueue in __wake_up()

remove-flush_page_to_ram.patch
  Remove flush_page_to_ram()

ext3-truncate-ordered-pages.patch
  ext3: explicitly free truncated pages

reiserfs_file_write-5.patch

sched_idle-typo-fix.patch
  fix sched_idle typo

rcu-stats.patch
  RCU statistics reporting

ext3-journalled-data-assertion-fix.patch
  Remove incorrect assertion from ext3

nfs-speedup.patch

nfs-oom-fix.patch
  nfs oom fix

sk-allocation.patch
  Subject: Re: nfs oom

nfs-more-oom-fix.patch

rpciod-atomic-allocations.patch
  Make rcpiod use atomic allocations

linux-isp.patch

isp-update-1.patch

ext3-quota-deadlock-fix.patch
  Fix deadlock with ext3+quota

dont-clear-PG_uptodate-on-ENOSPC.patch
  don't clear PG_uptodate on ENOSPC

stack-protection-fix.patch
  Subject: [patch] correct vm_page_prot on stack pages

sparc-PTE_FILE_MAX_BITS-fix.patch
  Variable PTE_FILE_MAX_BITS

file_lock-spinlock.patch
  convert file_lock to a spinlock

bootmem-speedup.patch
  bootmem speedup from the IA64 tree

mem_map-init-arch-hooks.patch
  architecture hooks for mem_map initialization

posix-timer-hang-fix-2.patch
  Posix timer hang fix

tty-modem-control-api.patch
  Subject: Re: uart_ioctl OOPS with irtty-sir

kmalloc_sizes-fix.patch
  Fix kmalloc_sizes[] indexing

proc-interrupts-kmalloc-size.patch
  /proc/interrupts allocates too much memory

setserial-fix.patch
  Subject: [PATCH 2.5] Minor fix for driver/serial/core.c

kblockd.patch
  Create `kblockd' workqueue

as-iosched.patch
  anticipatory I/O scheduler

as-np-reads-1.patch
  AS: read-vs-read fixes

as-np-reads-2.patch
  AS: more read-vs-read fixes

as-predict-data-direction.patch
  as: predict direction of next IO

as-remove-frontmerge.patch
  AS: remove frontmerge tunable

as-misc-cleanups.patch
  AS: misc cleanups

as-minor-tweaks.patch
  AS: tuning and tweaks

as-remove-stats.patch
  AS: remove statistics

as-locking-fix.patch
  AS: Fix minor race

as-disable-thinktime.patch

as-use-queue_empty.patch
  AS: Use the queue_empty API

cfq-2.patch
  CFQ scheduler, #2

unplug-use-kblockd.patch
  Use kblockd for running request queues

objrmap.patch
  object-based rmap

objrmap-sort-vma-list.patch
  objrmap: optimise per-mapping vma searches

objrmap-vma-sorting-fix.patch
  fix obj vma sorting

32bit-dev_t-nfs-export-fix.patch
  Fix nfsd exports with big dev_t

sched-2.5.64-D3.patch
  sched-2.5.64-D3, more interactivity changes

scheduler-tunables.patch
  scheduler tunables

show_task-free-stack-fix.patch
  show_task() fix and cleanup

yellowfin-set_bit-fix.patch
  yellowfin driver set_bit fix

htree-nfs-fix.patch
  Fix ext3 htree / NFS compatibility problems

task_prio-fix.patch
  simple task_prio() fix

i8042-share-irqs.patch
  allow i8042 interrupt sharing

vmalloc-stats.patch
  vmalloc stats in /proc/meminfo

meminfo-doc.patch
  /proc/meminfo documentation

gfp_repeat.patch
  implement __GFP_REPEAT

alloc_buffer_head-take-gfp.patch
  make alloc_buffer_head take gfp_flags

pte_alloc_one-use-gfp_repeat.patch
  use __GFP_REPEAT in pte_alloc_one()

pmd_alloc_one-use-gfp_repeat.patch
  use __GFP_REPEAT in pmd_alloc_one()

dynamic-hd_struct-allocation.patch
  Allocate hd_structs dynamically

dynamic-hd_struct-devfs-fix.patch
  Fix dynamic hd_struct allocation for devfs

slab_store_user-large-objects.patch
  slab debug: perform redzoning against larger objects

htree-nfs-fix-2.patch
  htree nfs fix

put_task_struct-debug.patch

percpu_counter.patch
  percpu_counters: approximate but scalable counters

blockgroup_lock.patch
  blockgroup_lock: hashed spinlocks for ext2 and ext3 blockgroup locking

ext2-no-lock_super-ng.patch
  use spinlocking in the ext2 block allocator

ext2-ialloc-no-lock_super-ng.patch
  use spinlocking in the ext2 inode allocator

dev_t-32-bit.patch
  [for playing only] change type of dev_t

dev_t-remove-B_FREE.patch
  dev_t: eliminate B_FREE

sg-dev_t-fix.patch
  32-bit dev_t fix for sg

xfs-dev_t-warning-fix.patch
  xfs dev_t printk warning fix

init-sections-in-kallsyms.patch
  Put all functions in kallsyms

aggregated-disk-stats.patch
  Aggregated disk statistics

oops-dump-preceding-code.patch
  i386 oops output: dump preceding code

usb-disconnect-crash-fix.patch
  Subject: Re: [linux-usb-devel] timer hang with current 2.5 BK

lockmeter.patch

ext3-no-bkl.patch

journal_dirty_metadata-speedup.patch

journal_get_write_access-speedup.patch

ext3-concurrent-block-inode-allocation.patch
  Subject: [PATCH] concurrent block/inode allocation for EXT3

ext3-concurrent-block-allocation-fix-1.patch

ext3-concurrent-block-allocation-hashed.patch
  Subject: Re: [PATCH] concurrent block/inode allocation for EXT3

Re: 2.5.67-mm2

[Felipe Alfaro Solana]

On Sun, 2003-04-13 at 03:08, Andrew Morton wrote:
> ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/patches/2.5/2.5.67/2.5.67-mm2/
> 
> . Lots of misc saved-up things.
> 
> . I've changed the 32-bit dev_t patch to provide a 12:20 split rather than
>   16:16.  This patch is starting to drag a bit and unless someone stops me I
>   might just go submit the thing.

Any patches for CardBus/PCMCIA support? It's broken for me since
2.5.66-mm2 (it works with 2.5.66-mm1) probably due to PCI changes or the
new PCMCIA state machine: if I boot my machine with my 3Com CardBus NIC
plugged in, the kernel deadlocks while checking the sockets, but it
works when booting with the card unplugged, and then plugging it back
once the system is stable (for example, init 1).

I have written to Russell King, but had no response from him. Maybe he
is too busy. I'm stuck at 2.5.66-mm1 on my laptop.

-- 
Please AVOID sending me WORD, EXCEL or POWERPOINT attachments.
See http://www.fsf.org/philosophy/no-word-attachments.html
Linux Registered User #287198

Re: 2.5.67-mm2

[Jeremy Hall]

I dunno about that, but mm2 locks in the boot process and doesn't display 
anything to me through gdb even though it is supposed to.  I have gdb 
console=gdb but that doesn't make the messages flow.

_J

In the new year, Felipe Alfaro Solana wrote:
> On Sun, 2003-04-13 at 03:08, Andrew Morton wrote:
> > ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/patches/2.5/2.5.67/2.5.67-mm2/
> > 
> > . Lots of misc saved-up things.
> > 
> > . I've changed the 32-bit dev_t patch to provide a 12:20 split rather than
> >   16:16.  This patch is starting to drag a bit and unless someone stops me I
> >   might just go submit the thing.
> 
> Any patches for CardBus/PCMCIA support? It's broken for me since
> 2.5.66-mm2 (it works with 2.5.66-mm1) probably due to PCI changes or the
> new PCMCIA state machine: if I boot my machine with my 3Com CardBus NIC
> plugged in, the kernel deadlocks while checking the sockets, but it
> works when booting with the card unplugged, and then plugging it back
> once the system is stable (for example, init 1).
> 
> I have written to Russell King, but had no response from him. Maybe he
> is too busy. I'm stuck at 2.5.66-mm1 on my laptop.
> 
> -- 
> Please AVOID sending me WORD, EXCEL or POWERPOINT attachments.
> See http://www.fsf.org/philosophy/no-word-attachments.html
> Linux Registered User #287198
> 
> --
> To unsubscribe, send a message with 'unsubscribe linux-mm' in
> the body to majordomo@kvack.org.  For more info on Linux MM,
> see: http://www.linux-mm.org/ .
> Don't email: <a href=mailto:"aart@kvack.org">aart@kvack.org</a>
> 

Re: 2.5.67-mm2

[Andrew Morton]

Jeremy Hall <jhall@maoz.com> wrote:
>
> I dunno about that, but mm2 locks in the boot process and doesn't display 
> anything to me through gdb even though it is supposed to.  I have gdb 
> console=gdb but that doesn't make the messages flow.
> 

You want "gdb console=gdb".  It changed.

What CPU type?

Try just 2.5.67 plus 
ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/patches/2.5/2.5.67/2.5.67-mm2/broken-out/linus.patch

try disabling kgdb in config.

etcetera.

Re: 2.5.67-mm2

[Jeremy Hall]

mm1 worked.

to be clear, my append line looks like

append="gdb console=gdb pci=biosirq" and maybe some other stuff on there

to get my machine to boot in single processor mode, I have to add nosmp 
and acpi=off

if I don't add acpi=off I get apic errors on my cpus, as well as if I say 
noapic like the boot process says to, it says

amd errata #22 may be present.  in the event of instability boot with the 
noapic option

dual athlon 1900 palamino's on a tyan 2462ung

_J

In the new year, Andrew Morton wrote:
> Jeremy Hall <jhall@maoz.com> wrote:
> >
> > I dunno about that, but mm2 locks in the boot process and doesn't display 
> > anything to me through gdb even though it is supposed to.  I have gdb 
> > console=gdb but that doesn't make the messages flow.
> > 
> 
> You want "gdb console=gdb".  It changed.
> 
> What CPU type?
> 
> Try just 2.5.67 plus 
> ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/patches/2.5/2.5.67/2.5.67-mm2/broken-out/linus.patch
> 
> try disabling kgdb in config.
> 
> etcetera.
> 
> --
> To unsubscribe, send a message with 'unsubscribe linux-mm' in
> the body to majordomo@kvack.org.  For more info on Linux MM,
> see: http://www.linux-mm.org/ .
> Don't email: <a href=mailto:"aart@kvack.org">aart@kvack.org</a>
> 

Re: 2.5.67-mm2

[Jeremy Hall]

ah, here we go

BUG(); line 907 of mm/slab.c

name == "size-32"

in_interrupt() is an undefined symbol in this context, so I don't know 
whether we are in interrupt context or not, if indeed that is what this 
is

no symbol BYTES_PER_WORD, but size == 32

MAX_OBJ_ORDER and PAGE_SIZE are unknown symbols in this context

dtor appears to be 0, and it looks like it should be set to something

offset == 0 and it looks like it should be < 0 or > size

in kmem_cache_create from kmem_cache_sizes_init at mm/slab.c:616

_J

In the new year, Jeremy Hall wrote:
> mm1 worked.
> 
> to be clear, my append line looks like
> 
> append="gdb console=gdb pci=biosirq" and maybe some other stuff on there
> 
> to get my machine to boot in single processor mode, I have to add nosmp 
> and acpi=off
> 
> if I don't add acpi=off I get apic errors on my cpus, as well as if I say 
> noapic like the boot process says to, it says
> 
> amd errata #22 may be present.  in the event of instability boot with the 
> noapic option
> 
> dual athlon 1900 palamino's on a tyan 2462ung
> 
> _J
> 
> In the new year, Andrew Morton wrote:
> > Jeremy Hall <jhall@maoz.com> wrote:
> > >
> > > I dunno about that, but mm2 locks in the boot process and doesn't display 
> > > anything to me through gdb even though it is supposed to.  I have gdb 
> > > console=gdb but that doesn't make the messages flow.
> > > 
> > 
> > You want "gdb console=gdb".  It changed.
> > 
> > What CPU type?
> > 
> > Try just 2.5.67 plus 
> > ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/patches/2.5/2.5.67/2.5.67-mm2/broken-out/linus.patch
> > 
> > try disabling kgdb in config.
> > 
> > etcetera.
> > 
> > --
> > To unsubscribe, send a message with 'unsubscribe linux-mm' in
> > the body to majordomo@kvack.org.  For more info on Linux MM,
> > see: http://www.linux-mm.org/ .
> > Don't email: <a href=mailto:"aart@kvack.org">aart@kvack.org</a>
> > 
> 
> --
> To unsubscribe, send a message with 'unsubscribe linux-mm' in
> the body to majordomo@kvack.org.  For more info on Linux MM,
> see: http://www.linux-mm.org/ .
> Don't email: <a href=mailto:"aart@kvack.org">aart@kvack.org</a>
> 

Re: 2.5.67-mm2

[Andrew Morton]

Jeremy Hall <jhall@maoz.com> wrote:
>
> ah, here we go
> 
> BUG(); line 907 of mm/slab.c
> 

Yup, it looks like the lockmeter patch has borked the preempt_count when
CONFIG_LOCKMETER=n.  Sorry, I didn't test it with preempt enabled.

I'll fix that up.  Meanwhile you can revert the lockmeter patch or disable
preemption.

Re: 2.5.67-mm2

[William Lee Irwin III]

On Sat, Apr 12, 2003 at 11:03:46PM -0400, Jeremy Hall wrote:
> I dunno about that, but mm2 locks in the boot process and doesn't display 
> anything to me through gdb even though it is supposed to.  I have gdb 
> console=gdb but that doesn't make the messages flow.

An early printk patch (any of the several going around) may give you an
idea of where it's barfing.


-- wli

Re: 2.5.67-mm2

[Jeremy Hall]

well I guess I could step through one thing at a time, n, because I DO get 
an initial trap it comes as soon as cpus are brought up

but that would take a long time and I'm sure there's LOTS of code.

_J

In the new year, William Lee Irwin III wrote:
> On Sat, Apr 12, 2003 at 11:03:46PM -0400, Jeremy Hall wrote:
> > I dunno about that, but mm2 locks in the boot process and doesn't display 
> > anything to me through gdb even though it is supposed to.  I have gdb 
> > console=gdb but that doesn't make the messages flow.
> 
> An early printk patch (any of the several going around) may give you an
> idea of where it's barfing.
> 
> 
> -- wli
> 

Re: 2.5.67-mm2

[Andrew Morton]

This should fix it up.

 include/linux/spinlock.h |   14 +++++++-------
 1 files changed, 7 insertions(+), 7 deletions(-)

diff -puN include/linux/spinlock.h~lockmeter-fixes include/linux/spinlock.h
--- 25/include/linux/spinlock.h~lockmeter-fixes	2003-04-12 21:35:49.000000000 -0700
+++ 25-akpm/include/linux/spinlock.h	2003-04-12 21:35:57.000000000 -0700
@@ -328,20 +328,20 @@ do { \
 
 #define spin_unlock_irqrestore(lock, flags) \
 do { \
-	spin_unlock(lock); \
+	_raw_spin_unlock(lock); \
 	local_irq_restore(flags); \
 	preempt_enable(); \
 } while (0)
 
 #define _raw_spin_unlock_irqrestore(lock, flags) \
 do { \
-	spin_unlock(lock); \
+	_raw_spin_unlock(lock); \
 	local_irq_restore(flags); \
 } while (0)
 
 #define spin_unlock_irq(lock) \
 do { \
-	spin_unlock(lock); \
+	_raw_spin_unlock(lock); \
 	local_irq_enable(); \
 	preempt_enable(); \
 } while (0)
@@ -355,14 +355,14 @@ do { \
 
 #define read_unlock_irqrestore(lock, flags) \
 do { \
-	read_unlock(lock); \
+	_raw_read_unlock(lock); \
 	local_irq_restore(flags); \
 	preempt_enable(); \
 } while (0)
 
 #define read_unlock_irq(lock) \
 do { \
-	read_unlock(lock); \
+	_raw_read_unlock(lock); \
 	local_irq_enable(); \
 	preempt_enable(); \
 } while (0)
@@ -376,14 +376,14 @@ do { \
 
 #define write_unlock_irqrestore(lock, flags) \
 do { \
-	write_unlock(lock); \
+	_raw_write_unlock(lock); \
 	local_irq_restore(flags); \
 	preempt_enable(); \
 } while (0)
 
 #define write_unlock_irq(lock) \
 do { \
-	write_unlock(lock); \
+	_raw_write_unlock(lock); \
 	local_irq_enable(); \
 	preempt_enable(); \
 } while (0)

_

Re: 2.5.67-mm2

[Joshua Kwan]

[-- Attachment #1: Type: text/plain, Size: 1890 bytes --]

On Sat, Apr 12, 2003 at 09:42:14PM -0700, Andrew Morton wrote:
> 
> This should fix it up.
> 
>  include/linux/spinlock.h |   14 +++++++-------
>  1 files changed, 7 insertions(+), 7 deletions(-)
> 

Okay, so now it boots fine, but this happens [abridged]:

Unable to handle kernel NULL pointer dereference at virtual address 00000040
 printing eip:
c0223cf4
*pde = 00000000
Oops: 0002 [#1]
CPU:    0
EIP:    0060:[<c0223cf4>]    Not tainted VLI
EFLAGS: 00010282
EIP is at devclass_add_driver+0x34/0x8a
eax: 00000040   ebx: fffffff8   ecx: fffffff0   edx: ffff0001
esi: 00000040   edi: 00000000   ebp: c03af298   esp: c1297f50
ds: 0007b   es: 007b   ss: 0068
Process swapper: (pid: 1, threadinfo=c1296000 task=c1294080)
Stack: c03af500 c03af2b4 c03ae7c4 00000000 c03ae780 c0223620 c03af298 c03af280
       00000001 00000000 00000000 c0223a72 c03af298 c0223a72 c03af048 fffffffc
       00000034 c026d15d c03af298 c01afc02 c03af048 c0402ebe c03af020 0000003c
Call Trace:
 [<c0223620>] bus_add_driver+0x82/0xd2
 [<c0223a72>] driver_register+0x38/0x3c
 [<c0223a72>] driver_register+0x38/0x3c
 [<c026d15d>] usb_register+0x73/0xae
 [<c01afc02>] pci_register_driver+0x47/0x57
 [<c0402ebe>] uhci_hcd_init+0xd9/0x137
 [<c0402f2b>] hid_init+0xf/0x21
 [<c03ec6e0>] do_initcalls+0x28/0x94
 [<c0129ce8>] init_workqueues+0xf/0x27
 [<c0105099>] init+0x36/0x195
 [<c0105063>] init+0x0/0x195
 [<c0106ffd>] kernel_thread_helper+0x5/0xb

Code: 24 0c 89 6c 24 10 89 74 24 08 8b 6c 24 18 31 ff 8b 45 08 89 04 24 e8 f7 02
00 00 85 c0 89 c3 74 4a 8d 70 48 ba 01 00 ff ff 98 f0 <0f> c1 10 85 d2 0f 85 f
03 00 00 89 2c 24 e8 c0 fe ff ff 85 c0
 <0>Kernel panic: Attempted to kill init!

This doesn't happen with 2.5.67{-bk4}. I looked at the USB changes for 2.5.67-mm2
and they are just about the same as in -bk4. I didn't look at every line, though.

Regards,
Josh

[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]

Re: 2.5.67-mm2

[J Sloan]

Andrew Morton wrote:

>This should fix it up.
>
> include/linux/spinlock.h |   14 +++++++-------
> 1 files changed, 7 insertions(+), 7 deletions(-)
>
Yes with that patch it works a charm.

Pounding the box now with no problems,

Joe

Re: 2.5.67-mm2

[Valdis.Kletnieks]

[-- Attachment #1: Type: text/plain, Size: 834 bytes --]

On Sun, 13 Apr 2003 03:55:29 +0200, Felipe Alfaro Solana said:

> Any patches for CardBus/PCMCIA support? It's broken for me since
> 2.5.66-mm2 (it works with 2.5.66-mm1) probably due to PCI changes or the
> new PCMCIA state machine: if I boot my machine with my 3Com CardBus NIC
> plugged in, the kernel deadlocks while checking the sockets, but it
> works when booting with the card unplugged, and then plugging it back
> once the system is stable (for example, init 1).

Also seeing this with a Xircom card under vanilla 2.5.67.

lspci reports this card as:

03:00.0 Ethernet controller: Xircom Cardbus Ethernet 10/100 (rev 03)
03:00.1 Serial controller: Xircom Cardbus Ethernet + 56k Modem (rev 03)

Russel King posted an analysis back on April 1, which indicated he knew
about the problem, understood it, and was working on it.


[-- Attachment #2: Type: application/pgp-signature, Size: 226 bytes --]

Re: 2.5.67-mm2

[Felipe Alfaro Solana]

On Sun, 2003-04-13 at 05:17, Valdis.Kletnieks@vt.edu wrote:
> On Sun, 13 Apr 2003 03:55:29 +0200, Felipe Alfaro Solana said:
> 
> > Any patches for CardBus/PCMCIA support? It's broken for me since
> > 2.5.66-mm2 (it works with 2.5.66-mm1) probably due to PCI changes or the
> > new PCMCIA state machine: if I boot my machine with my 3Com CardBus NIC
> > plugged in, the kernel deadlocks while checking the sockets, but it
> > works when booting with the card unplugged, and then plugging it back
> > once the system is stable (for example, init 1).
> 
> Also seeing this with a Xircom card under vanilla 2.5.67.
> 
> lspci reports this card as:
> 
> 03:00.0 Ethernet controller: Xircom Cardbus Ethernet 10/100 (rev 03)
> 03:00.1 Serial controller: Xircom Cardbus Ethernet + 56k Modem (rev 03)
> 
> Russel King posted an analysis back on April 1, which indicated he knew
> about the problem, understood it, and was working on it.

Yeah! I know, but I wrote him and didn't get a response, so I'm a little
bit worried. I assume he'll be too busy.

-- 
Please AVOID sending me WORD, EXCEL or POWERPOINT attachments.
See http://www.fsf.org/philosophy/no-word-attachments.html
Linux Registered User #287198

Re: 2.5.67-mm2

[Russell King]

On Sun, Apr 13, 2003 at 03:55:29AM +0200, Felipe Alfaro Solana wrote:
> I have written to Russell King, but had no response from him. Maybe he
> is too busy. I'm stuck at 2.5.66-mm1 on my laptop.

I've been on holiday for the past week, and am currently playing the
catch-up game.  People will have to wait a while longer until I can
get back to PCMCIA - I also look after the ARM kernel, which is also
requiring my attention at present.

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

Re: 2.5.67-mm2

[Ingo Oeser]

Hi Andrew,
hi lists readers,

On Sat, Apr 12, 2003 at 06:08:52PM -0700, Andrew Morton wrote:
> +gfp_repeat.patch
> 
>  Implement __GFP_REPEAT: so we can consolidate lots of alloc-with-retry code.

What about reworking the semantics of kmalloc()?

Many users of kmalloc get the flags and size reversed (major
source of hard to find bugs), so wouldn't it be simpler to have:

__kmalloc() /* The old kmalloc()*/

kmalloc()               /* kmalloc(, GFP_KERNEL) */
kmalloc_user()          /* kmalloc(, GFP_USER) */
kmalloc_dma()           /* kmalloc(, GFP_KERNEL | GFP_DMA) */
kmalloc_dma_repeat()    /* kmalloc(, GFP_KERNEL | GFP_DMA | __GFP_REPEAT) */
kmalloc_repeat()        /* kmalloc(, GFP_KERNEL | __GFP_REPEAT) */
kmalloc_atomic()        /* kmalloc(, GFP_ATOMIC) */
kmalloc_atomic_dma()    /* kmalloc(, GFP_ATOMIC | GFP_DMA) */

an so on? These functions will of course just be static inline
wrappers for __kmalloc().

These functions above would just take a size and not confuse
programmers anymore (as prototypes with compatible arguments
usally do).

If it's just a matter of "nobody had the time do do it, yet",
than this is doable, if only slowly.

If this is considered nonsense, then I will shut-up.

What do you think?

Regards

Ingo Oeser

Re: 2.5.67-mm2

[Arjan van de Ven]

[-- Attachment #1: Type: text/plain, Size: 613 bytes --]

On Sun, 2003-04-13 at 15:12, Ingo Oeser wrote:
> Hi Andrew,
> hi lists readers,
> 
> On Sat, Apr 12, 2003 at 06:08:52PM -0700, Andrew Morton wrote:
> > +gfp_repeat.patch
> > 
> >  Implement __GFP_REPEAT: so we can consolidate lots of alloc-with-retry code.
> 
> What about reworking the semantics of kmalloc()?
> 
> Many users of kmalloc get the flags and size reversed (major
> source of hard to find bugs), so wouldn't it be simpler to have:

that in itself is easy to find btw; just give every GFP_* an extra
__GFP_REQUIRED bit and then check inside kmalloc for that bit (MSB?) to
be set.....

[-- Attachment #2: This is a digitally signed message part --]
[-- Type: application/pgp-signature, Size: 189 bytes --]

Re: 2.5.67-mm2

[Denis Vlasenko]

On 13 April 2003 17:54, Arjan van de Ven wrote:
> On Sun, 2003-04-13 at 15:12, Ingo Oeser wrote:
> > Hi Andrew,
> > hi lists readers,
> >
> > On Sat, Apr 12, 2003 at 06:08:52PM -0700, Andrew Morton wrote:
> > > +gfp_repeat.patch
> > >
> > >  Implement __GFP_REPEAT: so we can consolidate lots of
> > > alloc-with-retry code.
> >
> > What about reworking the semantics of kmalloc()?
> >
> > Many users of kmalloc get the flags and size reversed (major
> > source of hard to find bugs), so wouldn't it be simpler to have:
>
> that in itself is easy to find btw; just give every GFP_* an extra
> __GFP_REQUIRED bit and then check inside kmalloc for that bit (MSB?)
> to be set.....

This will incur runtime penalty
--
vda

Re: 2.5.67-mm2

[Arjan van de Ven]

On Mon, Apr 14, 2003 at 09:24:26AM +0300, Denis Vlasenko wrote:
> >
> > that in itself is easy to find btw; just give every GFP_* an extra
> > __GFP_REQUIRED bit and then check inside kmalloc for that bit (MSB?)
> > to be set.....
> 
> This will incur runtime penalty

but only for CONFIG_DEBUG_KMALLOC or whatever

Re: 2.5.67-mm2

[Joel Becker]

On Sat, Apr 12, 2003 at 06:08:52PM -0700, Andrew Morton wrote:
> . I've changed the 32-bit dev_t patch to provide a 12:20 split rather than
>   16:16.  This patch is starting to drag a bit and unless someone stops me I
>   might just go submit the thing.

	Cool, but before you go off and push, maybe kick the appropriate
folks about making the 32/64 decision?

Joel

-- 

"When choosing between two evils, I always like to try the one
 I've never tried before."
        - Mae West

Joel Becker
Senior Member of Technical Staff
Oracle Corporation
E-mail: joel.becker@oracle.com
Phone: (650) 506-8127

Re: 2.5.67-mm2

[Andrew Morton]

Joel Becker <Joel.Becker@oracle.com> wrote:
>
> On Sat, Apr 12, 2003 at 06:08:52PM -0700, Andrew Morton wrote:
> > . I've changed the 32-bit dev_t patch to provide a 12:20 split rather than
> >   16:16.  This patch is starting to drag a bit and unless someone stops me I
> >   might just go submit the thing.
> 
> 	Cool, but before you go off and push, maybe kick the appropriate
> folks about making the 32/64 decision?
> 

It'll be 32+32.  I was just trolling.

Re: 2.5.67-mm2

[Randy.Dunlap]

On Mon, 14 Apr 2003 14:31:41 -0700 Andrew Morton <akpm@digeo.com> wrote:

| Joel Becker <Joel.Becker@oracle.com> wrote:
| >
| > On Sat, Apr 12, 2003 at 06:08:52PM -0700, Andrew Morton wrote:
| > > . I've changed the 32-bit dev_t patch to provide a 12:20 split rather than
| > >   16:16.  This patch is starting to drag a bit and unless someone stops me I
| > >   might just go submit the thing.
| > 
| > 	Cool, but before you go off and push, maybe kick the appropriate
| > folks about making the 32/64 decision?
| > 
| 
| It'll be 32+32.  I was just trolling.

Good.  and good.

--
~Randy