__alloc_pages: 4-order allocation failed

__alloc_pages: 4-order allocation failed

[Ho Chak Hung]

Hi,

I got the error __alloc_pages: 4-order allocation failed in a module that uses and frees a lot of pages.
Basically, I am trying implement a page cache for the module. First, I keep allocating pages using page_cache_alloc() until it fails, then I free a whole bunch of pages using freepages((unsigned long)page_address(page))

Would anyone please give me some advice about how to solve this problem?
Thanks a lot.

__________________________________________________________________
Get your own FREE, personal Netscape Webmail account today at http://webmail.netscape.com/

Re: __alloc_pages: 4-order allocation failed

[Jes Sorensen]

>>>>> "Ho" == Ho Chak Hung <hunghochak@netscape.net> writes:

Ho> Hi, I got the error __alloc_pages: 4-order allocation failed in a
Ho> module that uses and frees a lot of pages.  Basically, I am trying
Ho> implement a page cache for the module. First, I keep allocating
Ho> pages using page_cache_alloc() until it fails, then I free a whole
Ho> bunch of pages using freepages((unsigned long)page_address(page))

Ho> Would anyone please give me some advice about how to solve this
Ho> problem?  Thanks a lot.

You ran out of memory, ie. there were no more free blocks of 16
consecutive pages available in the system. This is what happens on a
system with little memory or which is loaded with memory intensive
applications.

Jes

Re: __alloc_pages: 4-order allocation failed

[David Whysong]

Jes Sorensen (jes@sunsite.dk) wrote:

>You ran out of memory, ie. there were no more free blocks of 16
>consecutive pages available in the system. This is what happens on a
>system with little memory or which is loaded with memory intensive
>applications.

I'm seeing the same thing here on a machine with 256 MB RAM and 1.5
gigabytes of swap. There is no chance I am using anywhere near that
much virtual memory.

Something is wrong with the MM in 2.4.6-pre9.

Dave

David Whysong                                       dwhysong@physics.ucsb.edu
Astrophysics graduate student         University of California, Santa Barbara
My public PGP keys are on my web page - http://www.physics.ucsb.edu/~dwhysong
DSS PGP Key 0x903F5BD6  :  FE78 91FE 4508 106F 7C88  1706 B792 6995 903F 5BD6
D-H PGP key 0x5DAB0F91  :  BC33 0F36 FCCD E72C 441F  663A 72ED 7FB7 5DAB 0F91

Re: __alloc_pages: 4-order allocation failed

[Marcelo Tosatti]

On Thu, 5 Jul 2001, David Whysong wrote:

> Jes Sorensen (jes@sunsite.dk) wrote:
> 
> >You ran out of memory, ie. there were no more free blocks of 16
> >consecutive pages available in the system. This is what happens on a
> >system with little memory or which is loaded with memory intensive
> >applications.
> 
> I'm seeing the same thing here on a machine with 256 MB RAM and 1.5
> gigabytes of swap. There is no chance I am using anywhere near that
> much virtual memory.
> 
> Something is wrong with the MM in 2.4.6-pre9.

David, 

The messages are "harmless". The SCSI layer is trying to allocate big
chunks of memory (for the scattergather tables IIRC) just as an
optimization. 

If its not possible to allocate big chunks, the SCSI code will use lower
order allocations. (smaller chunks) 

Re: __alloc_pages: 4-order allocation failed

[Szabolcs Szakacsits]

On Thu, 26 Apr 2001, Jeff V. Merkey wrote:

> I am seeing this as well on 2.4.3 with both _get_free_pages() and
> kmalloc().  In the kmalloc case, the modules hang waiting
> for memory.

One possible source of this hang is due to the change below in
2.4.3, non GPF_ATOMIC and non-recursive allocations (PF_MEMALLOC is set)
will loop until the requested continuous memory is available.

	Szaka

diff -u --recursive --new-file v2.4.2/linux/mm/page_alloc.c
linux/mm/page_alloc.c--- v2.4.2/linux/mm/page_alloc.c        Sat Feb  3
19:51:32 2001
+++ linux/mm/page_alloc.c       Tue Mar 20 15:05:46 2001
@@ -455,8 +455,7 @@
                        memory_pressure++;
                        try_to_free_pages(gfp_mask);
                        wakeup_bdflush(0);
-                       if (!order)
-                               goto try_again;
+                       goto try_again;
                }
        }

Re: __alloc_pages: 4-order allocation failed

[Jeff V. Merkey]

On Thu, Apr 26, 2001 at 08:00:12PM +0200, Szabolcs Szakacsits wrote:
> 
> On Thu, 26 Apr 2001, Jeff V. Merkey wrote:

The request should fail after two or three attempts rather than hang
the entire system waiting for memory.

Jeff

> 
> > I am seeing this as well on 2.4.3 with both _get_free_pages() and
> > kmalloc().  In the kmalloc case, the modules hang waiting
> > for memory.
> 
> One possible source of this hang is due to the change below in
> 2.4.3, non GPF_ATOMIC and non-recursive allocations (PF_MEMALLOC is set)
> will loop until the requested continuous memory is available.
> 
> 	Szaka
> 
> diff -u --recursive --new-file v2.4.2/linux/mm/page_alloc.c
> linux/mm/page_alloc.c--- v2.4.2/linux/mm/page_alloc.c        Sat Feb  3
> 19:51:32 2001
> +++ linux/mm/page_alloc.c       Tue Mar 20 15:05:46 2001
> @@ -455,8 +455,7 @@
>                         memory_pressure++;
>                         try_to_free_pages(gfp_mask);
>                         wakeup_bdflush(0);
> -                       if (!order)
> -                               goto try_again;
> +                       goto try_again;
>                 }
>         }
> 
> -
> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
> Please read the FAQ at  http://www.tux.org/lkml/

Re: __alloc_pages: 4-order allocation failed

[Feng Xian]

The problem is I didn't see those error message on 2.4.2 or 2.4.0, only on
2.4.3. That's the reason I posted the question here. Maybe I will try
2.4.4

Thanks all for you guys!

Alex

On Thu, 26 Apr 2001, Feng Xian wrote:

>
> Yes I am running nvidia module. i tried nv, X use less memory but nv
> doesn't give me the NV_GLX extension, xlock will crash for some 3d mode.
>
> Alex
>
> On Thu, 26 Apr 2001, Andi Kleen wrote:
>
> > On Thu, Apr 26, 2001 at 08:09:06AM -0400, Feng Xian wrote:
> > > It looks like the X consumes most of the memory (almost used up all the
> > > physical memory, more than 100M), it uses NVidia driver. I was also
> > > running pppoe but that took less memory.
> >
> > You're probably using the NVidia provided driver module, right?
> > Try it without it and the "nv" driver.
> >
> >
> > -Andi
> >
>
>

-- 
 Feng Xian
     .-.
     /v\    L   I   N   U   X
    // \\  >Phear the Penguin<
   /(   )\
    ^^-^^

Re: __alloc_pages: 4-order allocation failed

[Andi Kleen]

On Thu, Apr 26, 2001 at 09:50:15AM -0400, Feng Xian wrote:
> 
> Yes I am running nvidia module. i tried nv, X use less memory but nv
> doesn't give me the NV_GLX extension, xlock will crash for some 3d mode.

In this case you should report any kernel problems you see to NVidia
first, except if you can also reproduce them without the nvidia module.

-Andi

Re: __alloc_pages: 4-order allocation failed

[Feng Xian]

Yes I am running nvidia module. i tried nv, X use less memory but nv
doesn't give me the NV_GLX extension, xlock will crash for some 3d mode.

Alex

On Thu, 26 Apr 2001, Andi Kleen wrote:

> On Thu, Apr 26, 2001 at 08:09:06AM -0400, Feng Xian wrote:
> > It looks like the X consumes most of the memory (almost used up all the
> > physical memory, more than 100M), it uses NVidia driver. I was also
> > running pppoe but that took less memory.
>
> You're probably using the NVidia provided driver module, right?
> Try it without it and the "nv" driver.
>
>
> -Andi
>

-- 
 Feng Xian
     .-.
     /v\    L   I   N   U   X
    // \\  >Phear the Penguin<
   /(   )\
    ^^-^^

Re: __alloc_pages: 4-order allocation failed

[Andi Kleen]

On Thu, Apr 26, 2001 at 08:09:06AM -0400, Feng Xian wrote:
> It looks like the X consumes most of the memory (almost used up all the
> physical memory, more than 100M), it uses NVidia driver. I was also
> running pppoe but that took less memory.

You're probably using the NVidia provided driver module, right?
Try it without it and the "nv" driver.


-Andi 

Re: __alloc_pages: 4-order allocation failed

[Feng Xian]

Thanks for the suggestion. but where to get pre-2.4.4 kernel? when I
looked into the kernel traffic mail list, peoples are talking about 2.4.4,
but i checked kernel.org, the lastest one i found is 2.4.3

regards,

Alex


On Wed, 25 Apr 2001, Mark Hahn wrote:

> > I am running linux-2.4.3 on a Dell dual PIII machine with 128M memory.
>
> 2.4.3 has many known flaws; why not try a pre-2.4.4 kernel?
>
> > __alloc_pages: 4-order allocation failed.
>
> these happen when someone tries to allocate large contiguous blocks.
>
> > and sometime the system will crash. I looked into the memory info,
> > there still has some free physical memory (20M) left and swap space is
> > almost not in use. (250M swap)
>
> your ram is fragmented; it's the contiguity that is causing the failure.
>
> > I didn't have this problem when I ran 2.4.0 (I even didn't see it on
> > 2.4.2) could anybody tell me what's wrong or where should I look into this
> > problem?
>
> you can simply perturb ram (flush, etc), or use a newer kernel.
>

-- 
 Feng Xian
     .-.
     /v\    L   I   N   U   X
    // \\
   /(   )\
    ^^-^^

Re: __alloc_pages: 4-order allocation failed

[Feng Xian]

On Wed, 25 Apr 2001, Marcelo Tosatti wrote:

>
>
> On Wed, 25 Apr 2001, Feng Xian wrote:
>
> > Hi,
> >
> > I am running linux-2.4.3 on a Dell dual PIII machine with 128M memory.
> > After the machine runs a while, dmesg shows,
> >
> > __alloc_pages: 4-order allocation failed.
> > __alloc_pages: 3-order allocation failed.
> > __alloc_pages: 4-order allocation failed.
> > __alloc_pages: 4-order allocation failed.
> > __alloc_pages: 4-order allocation failed.
> > __alloc_pages: 4-order allocation failed.
> >
> >
> > and sometime the system will crash. I looked into the memory info,
> > there still has some free physical memory (20M) left and swap space is
> > almost not in use. (250M swap)
> >
> > I didn't have this problem when I ran 2.4.0 (I even didn't see it on
> > 2.4.2) could anybody tell me what's wrong or where should I look into this
> > problem?
>
> Feng,
>
> Which apps are you running when this happens ?

It looks like the X consumes most of the memory (almost used up all the
physical memory, more than 100M), it uses NVidia driver. I was also
running pppoe but that took less memory.

>
> Thanks
>
>

-- 
 Feng Xian

Re: __alloc_pages: 4-order allocation failed

[Mike Galbraith]

On Thu, 26 Apr 2001, Jeff V. Merkey wrote:

> I am seeing this as well on 2.4.3 with both _get_free_pages() and
> kmalloc().  In the kmalloc case, the modules hang waiting
> for memory.

Would adding __builtin_return_address(0) to the warning help locate?

	-Mike

Re: __alloc_pages: 4-order allocation failed

[Jeff V. Merkey]

I am seeing this as well on 2.4.3 with both _get_free_pages() and 
kmalloc().  In the kmalloc case, the modules hang waiting
for memory.

Jeff

On Wed, Apr 25, 2001 at 09:09:57PM -0400, Feng Xian wrote:
> Hi,
> 
> I am running linux-2.4.3 on a Dell dual PIII machine with 128M memory.
> After the machine runs a while, dmesg shows,
> 
> __alloc_pages: 4-order allocation failed.
> __alloc_pages: 3-order allocation failed.
> __alloc_pages: 4-order allocation failed.
> __alloc_pages: 4-order allocation failed.
> __alloc_pages: 4-order allocation failed.
> __alloc_pages: 4-order allocation failed.
> 
> 
> and sometime the system will crash. I looked into the memory info,
> there still has some free physical memory (20M) left and swap space is
> almost not in use. (250M swap)
> 
> I didn't have this problem when I ran 2.4.0 (I even didn't see it on
> 2.4.2) could anybody tell me what's wrong or where should I look into this
> problem?
> 
> Thanks,
> 
> Alex
> 
> -- 
> Feng Xian
> 
> -
> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
> Please read the FAQ at  http://www.tux.org/lkml/

Re: __alloc_pages: 4-order allocation failed

[Marcelo Tosatti]

On Wed, 25 Apr 2001, Feng Xian wrote:

> Hi,
> 
> I am running linux-2.4.3 on a Dell dual PIII machine with 128M memory.
> After the machine runs a while, dmesg shows,
> 
> __alloc_pages: 4-order allocation failed.
> __alloc_pages: 3-order allocation failed.
> __alloc_pages: 4-order allocation failed.
> __alloc_pages: 4-order allocation failed.
> __alloc_pages: 4-order allocation failed.
> __alloc_pages: 4-order allocation failed.
> 
> 
> and sometime the system will crash. I looked into the memory info,
> there still has some free physical memory (20M) left and swap space is
> almost not in use. (250M swap)
> 
> I didn't have this problem when I ran 2.4.0 (I even didn't see it on
> 2.4.2) could anybody tell me what's wrong or where should I look into this
> problem?

Feng,

Which apps are you running when this happens ?

Thanks 

__alloc_pages: 4-order allocation failed

[Feng Xian]

Hi,

I am running linux-2.4.3 on a Dell dual PIII machine with 128M memory.
After the machine runs a while, dmesg shows,

__alloc_pages: 4-order allocation failed.
__alloc_pages: 3-order allocation failed.
__alloc_pages: 4-order allocation failed.
__alloc_pages: 4-order allocation failed.
__alloc_pages: 4-order allocation failed.
__alloc_pages: 4-order allocation failed.


and sometime the system will crash. I looked into the memory info,
there still has some free physical memory (20M) left and swap space is
almost not in use. (250M swap)

I didn't have this problem when I ran 2.4.0 (I even didn't see it on
2.4.2) could anybody tell me what's wrong or where should I look into this
problem?

Thanks,

Alex

-- 
Feng Xian