ALi 15xx and Phoenix NoteBIOS suspend to disk problems?

ALi 15xx and Phoenix NoteBIOS suspend to disk problems?

[Joseph Pingenot]

Greetings.

I have a Toshiba Satellite 1605CDS (laptop).  It has a Phoenix BIOS, and
  an ALi 1541 chipset (aparently with M5229).  Relevant 2.5.14 boot info:
May  8 04:31:10 paulus agpgart: Maximum main memory to use for agp memory: 122M
May  8 04:31:10 paulus agpgart: Detected Ali M1541 chipset
May  8 04:31:10 paulus agpgart: AGP aperture is 64M @ 0xe0000000
May  8 04:31:10 paulus ATA/ATAPI driver v7.0.0
May  8 04:31:10 paulus ATA: system bus speed 33MHz
May  8 04:31:10 paulus ATA: interface: Acer Laboratories Inc. [ALi] M5229 IDE, on PCI slot 00:0f.0
May  8 04:31:10 paulus PCI: No IRQ known for interrupt pin A of device 00:0f.0.
May  8 04:31:10 paulus ATA: chipset rev.: 32
May  8 04:31:10 paulus ATA: non-legacy mode: IRQ probe delayed
May  8 04:31:10 paulus ide0: BM-DMA at 0xfcf0-0xfcf7, BIOS settings: hda:DMA, hdb:pio
May  8 04:31:10 paulus ide1: BM-DMA at 0xfcf8-0xfcff, BIOS settings: hdc:DMA, hdd:pio

Anyhow, ever since I've seen the ALi 15xx chipset option (in the kernel
  IDE chipset-specific config section), that is, since possibly before
  kernel version 2.4.0, whenever I've enabled the ALi chipset support and
  then suspend the laptop, the BIOS halts partway through the suspend-to-disk
  screen (the "Sytem Information" and "Conventional Memory" sections complete
  the save-to-disk operation, but the "Extended Memory" bar never gets past
  zero progress; this is all in the "Phoenix NoteBIOS 4.0 Save To Disk Manager"
  screen, which starts a few screen flickers after calling apm -s.  BIOS is
  set to suspend to disk instead of suspending to RAM).
Any ideas what might be causing these problems?  Thanks!

-Joseph
-- 
Joseph======================================================jap3003@ksu.edu
[While discussing 8 new IIS (Microsoft's webserver) vulnerabilities]
"One workaround we rather like is called Apache, but we digress...."
  Greene, The Register, http://www.theregister.co.uk./content/4/24795.html

[OOPS & PATCH 2.4.18] NULL pointer dereference in ide.c, ide_revalidate_disk()

[Paul Clements]

[-- Attachment #1: Type: TEXT/PLAIN, Size: 4424 bytes --]


Looking back at LKML archives, I think this issue was reported by someone back in
November 01, and I just saw the same issue on a system a few days ago. I think I
have tracked down the cause of the oops. I looked to see if it had been fixed in
2.4.19-pre8 and it had not, so I have attached a patch against 2.4.18. Please apply.

The kernel oops was this:

----

May  3 23:20:07 liono kernel: hda: ATAPI 32X CD-ROM drive, 256kB Cache, UDMA(33)
May  3 23:20:07 liono kernel: Uniform CD-ROM driver Revision: 3.12
May  3 23:20:07 liono kernel: hda: ide_cdrom_setup failed to register device with the cdrom driver.
May  3 23:20:07 liono kernel: Unable to handle kernel NULL pointer dereference at virtual address 00000028
May  3 23:20:07 liono kernel:  printing eip:
May  3 23:20:07 liono kernel: c019b54a
May  3 23:20:07 liono kernel: *pde = 00000000
May  3 23:20:07 liono kernel: Oops: 0000
May  3 23:20:07 liono kernel: CPU:    0
May  3 23:20:07 liono kernel: EIP:    0010:[ide_revalidate_disk+250/304]    Not tainted
May  3 23:20:07 liono kernel: EIP:    0010:[<c019b54a>]    Not tainted
May  3 23:20:07 liono kernel: EFLAGS: 00010212
May  3 23:20:07 liono kernel: EIP is at ide_revalidate_disk [kernel] 0xfa
May  3 23:20:07 liono kernel: eax: 00000000   ebx: 00000300   ecx: 00000000   edx: 00000000
May  3 23:20:07 liono kernel: esi: c03bdba0   edi: 00001100   ebp: 00000040   esp: c90e3ee0
May  3 23:20:07 liono kernel: ds: 0018   es: 0018   ss: 0018
May  3 23:20:07 liono kernel: Process modprobe (pid: 2028, stackpage=c90e3000)
May  3 23:20:07 liono kernel: Stack: 00000300 00000000 00000000 00000000 00000000 c03bde7c c019b5d2 00000300
May  3 23:20:07 liono kernel:        cd3ce200 00000000 cd3ce348 00000002 c019dba5 d0972aa7 d09759e0 d096f000
May  3 23:20:07 liono kernel:        00000001 00000001 00000001 c011c685 d09758dc c903e000 000067d8 c91cd000
May  3 23:20:07 liono kernel: Call Trace: [revalidate_drives+82/112] revalidate_drives [kernel] 0x52
May  3 23:20:07 liono kernel: Call Trace: [<c019b5d2>] revalidate_drives [kernel] 0x52
May  3 23:20:07 liono kernel: [ide_register_module+53/64] ide_register_module [kernel] 0x35
May  3 23:20:07 liono kernel: [<c019dba5>] ide_register_module [kernel] 0x35
May  3 23:20:07 liono kernel: [aic7xxx_mod:aic7xxx_verbose+1171231/206116241] ide_cdrom_init [ide-cd] 0x187
May  3 23:20:07 liono kernel: [<d0972aa7>] ide_cdrom_init [ide-cd] 0x187
May  3 23:20:07 liono kernel: [aic7xxx_mod:aic7xxx_verbose+1183320/206104152] __insmod_ide-cd_S.data_L192 [ide-cd] 0xa0
May  3 23:20:07 liono kernel: [<d09759e0>] __insmod_ide-cd_S.data_L192 [ide-cd] 0xa0
May  3 23:20:07 liono kernel: [sys_init_module+1365/1616] sys_init_module [kernel] 0x555
May  3 23:20:07 liono kernel: [<c011c685>] sys_init_module [kernel] 0x555
May  3 23:20:07 liono kernel: [aic7xxx_mod:aic7xxx_verbose+1183060/206104412] sense_data_texts [ide-cd] 0x107c
May  3 23:20:07 liono kernel: [<d09758dc>] sense_data_texts [ide-cd] 0x107c
May  3 23:20:07 liono kernel: [aic7xxx_mod:aic7xxx_verbose+1156312/206131160] __insmod_ide-cd_O/lib/modules/2.4.9-21smp/kernel/drivers/ide/ide-cd.o_M3C472201_V132105 [ide-cd] 0x60
May  3 23:20:07 liono kernel: [<d096f060>] __insmod_ide-cd_O/lib/modules/2.4.9-21smp/kernel/drivers/ide/ide-cd.o_M3C472201_V132105 [ide-cd] 0x60
May  3 23:20:07 liono kernel: [system_call+51/56] system_call [kernel] 0x33
May  3 23:20:07 liono kernel: [<c010719b>] system_call [kernel] 0x33
May  3 23:20:07 liono kernel: [__put_unused_buffer_head+107/416] __put_unused_buffer_head [kernel] 0x6b
May  3 23:20:07 liono kernel: [<c014002b>] __put_unused_buffer_head [kernel] 0x6b
May  3 23:20:07 liono kernel:
May  3 23:20:07 liono kernel:
May  3 23:20:07 liono kernel: Code: 8b 40 28 85 c0 74 04 56 ff d0 5a 80 a6 b6 00 00 00 fb 8d 86

----

So what this tells me is that (probably) a NULL pointer was dereferenced while looking for a structure 
member with an offset of 40 (0x28) bytes. Looking through the ide_revalidate_disk function (ide.c) I find
that the only structure member being referenced, which also has an offset of 40 is drive->driver->revalidate 
(via the DRIVER macro):

    if (DRIVER(drive)->revalidate)

But there is no check for driver != NULL before it is dereferenced. In many other places throughout ide.c 
these explicit checks are present.

The attached patch (against 2.4.18) adds this explicit check.

--
Paul Clements
SteelEye Technology
Paul.Clements@SteelEye.com

[-- Attachment #2: Type: TEXT/PLAIN, Size: 348 bytes --]

--- linux-2.4.18.PRISTINE/drivers/ide/ide.c	Mon Apr 29 10:44:12 2002
+++ linux-2.4.18.alt/drivers/ide/ide.c	Tue May  7 17:16:53 2002
@@ -1892,7 +1892,7 @@
 		drive->part[p].nr_sects   = 0;
 	};
 
-	if (DRIVER(drive)->revalidate)
+	if (DRIVER(drive) && DRIVER(drive)->revalidate)
 		DRIVER(drive)->revalidate(drive);
 
 	drive->busy = 0;

[RESEND] [OOPS & PATCH 2.4.18] NULL pointer dereference in ide.c, ide_revalidate_disk()

[Paul Clements]

[-- Attachment #1: Type: TEXT/PLAIN, Size: 4801 bytes --]


Resending this patch. Looks like it did not make 2.4.19-pre9. Is there some problem in the patch 
or in my analysis, or has the problem been fixed in some other (better) way?

Thanks,
Paul


On Wed, 8 May 2002, Paul Clements wrote:

> 
> Looking back at LKML archives, I think this issue was reported by someone back in
> November 01, and I just saw the same issue on a system a few days ago. I think I
> have tracked down the cause of the oops. I looked to see if it had been fixed in
> 2.4.19-pre8 and it had not, so I have attached a patch against 2.4.18. Please apply.
> 
> The kernel oops was this:
> 
> ----
> 
> May  3 23:20:07 liono kernel: hda: ATAPI 32X CD-ROM drive, 256kB Cache, UDMA(33)
> May  3 23:20:07 liono kernel: Uniform CD-ROM driver Revision: 3.12
> May  3 23:20:07 liono kernel: hda: ide_cdrom_setup failed to register device with the cdrom driver.
> May  3 23:20:07 liono kernel: Unable to handle kernel NULL pointer dereference at virtual address 00000028
> May  3 23:20:07 liono kernel:  printing eip:
> May  3 23:20:07 liono kernel: c019b54a
> May  3 23:20:07 liono kernel: *pde = 00000000
> May  3 23:20:07 liono kernel: Oops: 0000
> May  3 23:20:07 liono kernel: CPU:    0
> May  3 23:20:07 liono kernel: EIP:    0010:[ide_revalidate_disk+250/304]    Not tainted
> May  3 23:20:07 liono kernel: EIP:    0010:[<c019b54a>]    Not tainted
> May  3 23:20:07 liono kernel: EFLAGS: 00010212
> May  3 23:20:07 liono kernel: EIP is at ide_revalidate_disk [kernel] 0xfa
> May  3 23:20:07 liono kernel: eax: 00000000   ebx: 00000300   ecx: 00000000   edx: 00000000
> May  3 23:20:07 liono kernel: esi: c03bdba0   edi: 00001100   ebp: 00000040   esp: c90e3ee0
> May  3 23:20:07 liono kernel: ds: 0018   es: 0018   ss: 0018
> May  3 23:20:07 liono kernel: Process modprobe (pid: 2028, stackpage=c90e3000)
> May  3 23:20:07 liono kernel: Stack: 00000300 00000000 00000000 00000000 00000000 c03bde7c c019b5d2 00000300
> May  3 23:20:07 liono kernel:        cd3ce200 00000000 cd3ce348 00000002 c019dba5 d0972aa7 d09759e0 d096f000
> May  3 23:20:07 liono kernel:        00000001 00000001 00000001 c011c685 d09758dc c903e000 000067d8 c91cd000
> May  3 23:20:07 liono kernel: Call Trace: [revalidate_drives+82/112] revalidate_drives [kernel] 0x52
> May  3 23:20:07 liono kernel: Call Trace: [<c019b5d2>] revalidate_drives [kernel] 0x52
> May  3 23:20:07 liono kernel: [ide_register_module+53/64] ide_register_module [kernel] 0x35
> May  3 23:20:07 liono kernel: [<c019dba5>] ide_register_module [kernel] 0x35
> May  3 23:20:07 liono kernel: [aic7xxx_mod:aic7xxx_verbose+1171231/206116241] ide_cdrom_init [ide-cd] 0x187
> May  3 23:20:07 liono kernel: [<d0972aa7>] ide_cdrom_init [ide-cd] 0x187
> May  3 23:20:07 liono kernel: [aic7xxx_mod:aic7xxx_verbose+1183320/206104152] __insmod_ide-cd_S.data_L192 [ide-cd] 0xa0
> May  3 23:20:07 liono kernel: [<d09759e0>] __insmod_ide-cd_S.data_L192 [ide-cd] 0xa0
> May  3 23:20:07 liono kernel: [sys_init_module+1365/1616] sys_init_module [kernel] 0x555
> May  3 23:20:07 liono kernel: [<c011c685>] sys_init_module [kernel] 0x555
> May  3 23:20:07 liono kernel: [aic7xxx_mod:aic7xxx_verbose+1183060/206104412] sense_data_texts [ide-cd] 0x107c
> May  3 23:20:07 liono kernel: [<d09758dc>] sense_data_texts [ide-cd] 0x107c
> May  3 23:20:07 liono kernel: [aic7xxx_mod:aic7xxx_verbose+1156312/206131160] __insmod_ide-cd_O/lib/modules/2.4.9-21smp/kernel/drivers/ide/ide-cd.o_M3C472201_V132105 [ide-cd] 0x60
> May  3 23:20:07 liono kernel: [<d096f060>] __insmod_ide-cd_O/lib/modules/2.4.9-21smp/kernel/drivers/ide/ide-cd.o_M3C472201_V132105 [ide-cd] 0x60
> May  3 23:20:07 liono kernel: [system_call+51/56] system_call [kernel] 0x33
> May  3 23:20:07 liono kernel: [<c010719b>] system_call [kernel] 0x33
> May  3 23:20:07 liono kernel: [__put_unused_buffer_head+107/416] __put_unused_buffer_head [kernel] 0x6b
> May  3 23:20:07 liono kernel: [<c014002b>] __put_unused_buffer_head [kernel] 0x6b
> May  3 23:20:07 liono kernel:
> May  3 23:20:07 liono kernel:
> May  3 23:20:07 liono kernel: Code: 8b 40 28 85 c0 74 04 56 ff d0 5a 80 a6 b6 00 00 00 fb 8d 86
> 
> ----
> 
> So what this tells me is that (probably) a NULL pointer was dereferenced while looking for a structure 
> member with an offset of 40 (0x28) bytes. Looking through the ide_revalidate_disk function (ide.c) I find
> that the only structure member being referenced, which also has an offset of 40 is drive->driver->revalidate 
> (via the DRIVER macro):
> 
>     if (DRIVER(drive)->revalidate)
> 
> But there is no check for driver != NULL before it is dereferenced. In many other places throughout ide.c 
> these explicit checks are present.
> 
> The attached patch (against 2.4.18) adds this explicit check.
> 
> --
> Paul Clements
> SteelEye Technology
> Paul.Clements@SteelEye.com
> 

[-- Attachment #2: Type: TEXT/PLAIN, Size: 348 bytes --]

--- linux-2.4.18.PRISTINE/drivers/ide/ide.c	Mon Apr 29 10:44:12 2002
+++ linux-2.4.18.alt/drivers/ide/ide.c	Tue May  7 17:16:53 2002
@@ -1892,7 +1892,7 @@
 		drive->part[p].nr_sects   = 0;
 	};
 
-	if (DRIVER(drive)->revalidate)
+	if (DRIVER(drive) && DRIVER(drive)->revalidate)
 		DRIVER(drive)->revalidate(drive);
 
 	drive->busy = 0;

[RESEND] [OOPS & TRIVIAL PATCH 2.4.20] NULL pointer dereference in ide.c, ide_revalidate_disk()

[Paul Clements]

[-- Attachment #1: Type: TEXT/PLAIN, Size: 5116 bytes --]

Problem still exists in 2.4.20. Please apply if this has not been fixed in some other way.

Thanks,
Paul


On Wed, 29 May 2002, Paul Clements wrote:

> 
> Resending this patch. Looks like it did not make 2.4.19-pre9. Is there some problem in the patch 
> or in my analysis, or has the problem been fixed in some other (better) way?
> 
> Thanks,
> Paul
> 
> 
> On Wed, 8 May 2002, Paul Clements wrote:
> 
> > 
> > Looking back at LKML archives, I think this issue was reported by someone back in
> > November 01, and I just saw the same issue on a system a few days ago. I think I
> > have tracked down the cause of the oops. I looked to see if it had been fixed in
> > 2.4.19-pre8 and it had not, so I have attached a patch against 2.4.18. Please apply.
> > 
> > The kernel oops was this:
> > 
> > ----
> > 
> > May  3 23:20:07 liono kernel: hda: ATAPI 32X CD-ROM drive, 256kB Cache, UDMA(33)
> > May  3 23:20:07 liono kernel: Uniform CD-ROM driver Revision: 3.12
> > May  3 23:20:07 liono kernel: hda: ide_cdrom_setup failed to register device with the cdrom driver.
> > May  3 23:20:07 liono kernel: Unable to handle kernel NULL pointer dereference at virtual address 00000028
> > May  3 23:20:07 liono kernel:  printing eip:
> > May  3 23:20:07 liono kernel: c019b54a
> > May  3 23:20:07 liono kernel: *pde = 00000000
> > May  3 23:20:07 liono kernel: Oops: 0000
> > May  3 23:20:07 liono kernel: CPU:    0
> > May  3 23:20:07 liono kernel: EIP:    0010:[ide_revalidate_disk+250/304]    Not tainted
> > May  3 23:20:07 liono kernel: EIP:    0010:[<c019b54a>]    Not tainted
> > May  3 23:20:07 liono kernel: EFLAGS: 00010212
> > May  3 23:20:07 liono kernel: EIP is at ide_revalidate_disk [kernel] 0xfa
> > May  3 23:20:07 liono kernel: eax: 00000000   ebx: 00000300   ecx: 00000000   edx: 00000000
> > May  3 23:20:07 liono kernel: esi: c03bdba0   edi: 00001100   ebp: 00000040   esp: c90e3ee0
> > May  3 23:20:07 liono kernel: ds: 0018   es: 0018   ss: 0018
> > May  3 23:20:07 liono kernel: Process modprobe (pid: 2028, stackpage=c90e3000)
> > May  3 23:20:07 liono kernel: Stack: 00000300 00000000 00000000 00000000 00000000 c03bde7c c019b5d2 00000300
> > May  3 23:20:07 liono kernel:        cd3ce200 00000000 cd3ce348 00000002 c019dba5 d0972aa7 d09759e0 d096f000
> > May  3 23:20:07 liono kernel:        00000001 00000001 00000001 c011c685 d09758dc c903e000 000067d8 c91cd000
> > May  3 23:20:07 liono kernel: Call Trace: [revalidate_drives+82/112] revalidate_drives [kernel] 0x52
> > May  3 23:20:07 liono kernel: Call Trace: [<c019b5d2>] revalidate_drives [kernel] 0x52
> > May  3 23:20:07 liono kernel: [ide_register_module+53/64] ide_register_module [kernel] 0x35
> > May  3 23:20:07 liono kernel: [<c019dba5>] ide_register_module [kernel] 0x35
> > May  3 23:20:07 liono kernel: [aic7xxx_mod:aic7xxx_verbose+1171231/206116241] ide_cdrom_init [ide-cd] 0x187
> > May  3 23:20:07 liono kernel: [<d0972aa7>] ide_cdrom_init [ide-cd] 0x187
> > May  3 23:20:07 liono kernel: [aic7xxx_mod:aic7xxx_verbose+1183320/206104152] __insmod_ide-cd_S.data_L192 [ide-cd] 0xa0
> > May  3 23:20:07 liono kernel: [<d09759e0>] __insmod_ide-cd_S.data_L192 [ide-cd] 0xa0
> > May  3 23:20:07 liono kernel: [sys_init_module+1365/1616] sys_init_module [kernel] 0x555
> > May  3 23:20:07 liono kernel: [<c011c685>] sys_init_module [kernel] 0x555
> > May  3 23:20:07 liono kernel: [aic7xxx_mod:aic7xxx_verbose+1183060/206104412] sense_data_texts [ide-cd] 0x107c
> > May  3 23:20:07 liono kernel: [<d09758dc>] sense_data_texts [ide-cd] 0x107c
> > May  3 23:20:07 liono kernel: [aic7xxx_mod:aic7xxx_verbose+1156312/206131160] __insmod_ide-cd_O/lib/modules/2.4.9-21smp/kernel/drivers/ide/ide-cd.o_M3C472201_V132105 [ide-cd] 0x60
> > May  3 23:20:07 liono kernel: [<d096f060>] __insmod_ide-cd_O/lib/modules/2.4.9-21smp/kernel/drivers/ide/ide-cd.o_M3C472201_V132105 [ide-cd] 0x60
> > May  3 23:20:07 liono kernel: [system_call+51/56] system_call [kernel] 0x33
> > May  3 23:20:07 liono kernel: [<c010719b>] system_call [kernel] 0x33
> > May  3 23:20:07 liono kernel: [__put_unused_buffer_head+107/416] __put_unused_buffer_head [kernel] 0x6b
> > May  3 23:20:07 liono kernel: [<c014002b>] __put_unused_buffer_head [kernel] 0x6b
> > May  3 23:20:07 liono kernel:
> > May  3 23:20:07 liono kernel:
> > May  3 23:20:07 liono kernel: Code: 8b 40 28 85 c0 74 04 56 ff d0 5a 80 a6 b6 00 00 00 fb 8d 86
> > 
> > ----
> > 
> > So what this tells me is that (probably) a NULL pointer was dereferenced while looking for a structure 
> > member with an offset of 40 (0x28) bytes. Looking through the ide_revalidate_disk function (ide.c) I find
> > that the only structure member being referenced, which also has an offset of 40 is drive->driver->revalidate 
> > (via the DRIVER macro):
> > 
> >     if (DRIVER(drive)->revalidate)
> > 
> > But there is no check for driver != NULL before it is dereferenced. In many other places throughout ide.c 
> > these explicit checks are present.
> > 
> > The attached patch (against 2.4.18) adds this explicit check.
> > 
> > --
> > Paul Clements
> > SteelEye Technology
> > Paul.Clements@SteelEye.com
> > 
> 

[-- Attachment #2: patch file --]
[-- Type: TEXT/PLAIN, Size: 348 bytes --]

--- linux-2.4.18.PRISTINE/drivers/ide/ide.c	Mon Apr 29 10:44:12 2002
+++ linux-2.4.18.alt/drivers/ide/ide.c	Tue May  7 17:16:53 2002
@@ -1892,7 +1892,7 @@
 		drive->part[p].nr_sects   = 0;
 	};
 
-	if (DRIVER(drive)->revalidate)
+	if (DRIVER(drive) && DRIVER(drive)->revalidate)
 		DRIVER(drive)->revalidate(drive);
 
 	drive->busy = 0;

Re: [RESEND] [OOPS & TRIVIAL PATCH 2.4.20] NULL pointer dereference in ide.c, ide_revalidate_disk()

[Marcelo Tosatti]

On Wed, 11 Dec 2002, Paul Clements wrote:

> Problem still exists in 2.4.20. Please apply if this has not been fixed in some other way.

Paul,

Patch applied.

Thanks.