sata_via oops on copying files?

sata_via oops on copying files?

[Albert Astals Cid]

Please CC me, i'm not in the list.

Not sure if this or bugzilla is preferred, so doing both, shout me if i
should wait on bugzilla.

Description of the hardware:
 * Soekris net 4526 motherboard [0]
 * MiniPci to Compact Flash adapter that uses Via VT6421A chipset [1] 

When booting up the sata_via module says

sata_via 0000:00:11.0: version 2.4                                      
       
sata_via 0000:00:11.0: routed to hard irq line 11                       
       
scsi0 : sata_via                                                        
       
scsi1 : sata_via                                                        
       
scsi2 : sata_via                                                        
       
ata1: SATA max UDMA/133 port i16@0xe000 bmdma 0xe040 irq 11             
       
ata2: SATA max UDMA/133 port i16@0xe010 bmdma 0xe048 irq 11             
       
ata3: PATA max UDMA/133 port i16@0xe020 bmdma 0xe050 irq 11             
       
ata1: SATA link down (SStatus 0 SControl 310)                           
       
ata2: SATA link down (SStatus 0 SControl 310)                           
       
irq 11: nobody cared (try booting with the "irqpoll" option)            
       
Pid: 787, comm: scsi_eh_2 Not tainted 2.6.29.3 #1                       
       
Call Trace:                                                             
       
 [<c0134c16>] __report_bad_irq+0x2e/0x6f                                
       
 [<c0134d44>] note_interrupt+0xed/0x149                                 
       
 [<c013537c>] handle_level_irq+0x8d/0xcf                                
       
 [<c01043d8>] do_IRQ+0x50/0x68                                          
       
 [<c0103227>] common_interrupt+0x27/0x30                                
       
 [<c0119484>] ? __do_softirq+0x31/0xbb                                  
       
 [<c0119533>] do_softirq+0x25/0x2a                                      
       
 [<c011965e>] irq_exit+0x28/0x5c                                        
       
 [<c01043dd>] do_IRQ+0x55/0x68                                          
       
 [<c0103227>] common_interrupt+0x27/0x30                                
       
 [<c01f007b>] ? cfq_insert_request+0x21b/0x383                          
       
 [<c0250000>] ? sata_link_init_spd+0xc1/0x185                           
       
 [<c01f8cbc>] ? iowrite8+0x20/0x30                                      
       
 [<c025ba7c>] ata_sff_dev_select+0x1c/0x26                              
       
 [<c025d7a1>] ata_sff_softreset+0x53/0x106                              
       
 [<c0257447>] ata_do_reset+0x49/0x4f                                    
       
 [<c025d74e>] ? ata_sff_softreset+0x0/0x106                             
       
 [<c02587e1>] ata_eh_reset+0x431/0xa10                                  
       
 [<c025d854>] ? ata_sff_prereset+0x0/0x98                               
       
 [<c025d74e>] ? ata_sff_softreset+0x0/0x106                             
       
 [<c0259c9e>] ata_eh_recover+0x270/0xa79                                
       
 [<c025d74e>] ? ata_sff_softreset+0x0/0x106                             
       
 [<c025bc7c>] ? ata_sff_postreset+0x0/0x5b                              
       
 [<c025d74e>] ? ata_sff_softreset+0x0/0x106                             
       
 [<c025d854>] ? ata_sff_prereset+0x0/0x98                               
       
 [<c025d74e>] ? ata_sff_softreset+0x0/0x106                             
       
 [<c025d854>] ? ata_sff_prereset+0x0/0x98                               
       
 [<c025a62b>] ata_do_eh+0x2b/0x6d                                       
       
 [<c025bc7c>] ? ata_sff_postreset+0x0/0x5b                              
       
 [<c025bc72>] ata_sff_error_handler+0x13d/0x147                         
       
 [<c025bc7c>] ? ata_sff_postreset+0x0/0x5b                              
       
 [<c025d74e>] ? ata_sff_softreset+0x0/0x106                             
       
 [<c025ad6c>] ata_scsi_error+0x1fb/0x46b                                
       
 [<c02456dd>] scsi_error_handler+0x6a/0x298                             
       
 [<c031e075>] ? schedule+0x290/0x2eb                                    
       
 [<c011327f>] ? complete+0x2a/0x48                                      
       
 [<c0245673>] ? scsi_error_handler+0x0/0x298                            
       
 [<c012472f>] kthread+0x39/0x5e                                         
       
 [<c01246f6>] ? kthread+0x0/0x5e                                        
       
 [<c0103397>] kernel_thread_helper+0x7/0x10                             
       
handlers:                                                               
       
[<c025d04c>] (ata_sff_interrupt+0x0/0x205)                              
       
Disabling IRQ #11                                                       
       
ata3.01: ATA-0: CF CARD 4GB, Ver2.19K, max UDMA/100                     
       
ata3.01: 7880544 sectors, multi 0: LBA
ata3.01: limited to UDMA/33 due to 40-wire cable
ata3.01: configured for UDMA/33
ata3.01: configured for UDMA/33
ata3: EH complete
scsi 2:0:1:0: Direct-Access     ATA      CF CARD 4GB      Ver2 PQ: 0
ANSI: 5
sd 2:0:1:0: [sda] 7880544 512-byte hardware sectors: (4.03 GB/3.75 GiB)
sd 2:0:1:0: [sda] Write Protect is off
sd 2:0:1:0: [sda] Mode Sense: 00 3a 00 00
sd 2:0:1:0: [sda] Write cache: disabled, read cache: enabled, doesn't
support
DPO or FUA
sd 2:0:1:0: [sda] 7880544 512-byte hardware sectors: (4.03 GB/3.75 GiB)
sd 2:0:1:0: [sda] Write Protect is off
sd 2:0:1:0: [sda] Mode Sense: 00 3a 00 00
sd 2:0:1:0: [sda] Write cache: disabled, read cache: enabled, doesn't
support
DPO or FUA
 sda: sda1 sda2
sd 2:0:1:0: [sda] Attached SCSI disk


When copying files with a simple script like
i=1; while [ true ]; do echo "Round $i"; cp blabla blabla2; sync; rm
blabla2;
sync; i=$(($i + 1)); done

I get crashes like

BUG: unable to handle kernel paging request at 6545a000
IP: [<c012d8eb>] tick_nohz_stop_idle+0x20/0x49
*pde = 00000000
Oops: 0002 [#1] PREEMPT
last sysfs file: /sys/class/net/lo/operstate
Modules linked in:

Pid: 0, comm: swapper Not tainted (2.6.29.3 #1)
EIP: 0060:[<c012d8eb>] EFLAGS: 00010002 CPU: 0
EIP is at tick_nohz_stop_idle+0x20/0x49
EAX: cdd635da EBX: 00000000 ECX: 0000016c EDX: 00000054
ESI: ffffffcf EDI: 00000000 EBP: c03f7f58 ESP: c03f7f58
 DS: 007b ES: 007b FS: 0000 GS: 0000 SS: 0068
Process swapper (pid: 0, ti=c03f6000 task=c03c5388 task.ti=c03f6000)
Stack:
 c03f7f64 c012d926 c041a8e0 c03f7f6c c01193bc c03f7f80 c01043a1 c041a8e0
 00049800 c03f0000 c03f7fc0 c0103227 c041a8e0 c0459e0c 00000054 00049800
 c03f0000 c03f7fc0 c03f6000 c3a4007b c03c007b 00000000 ffffffcf c0107662
Call Trace:
 [<c012d926>] ? tick_check_idle+0x12/0x3f
 [<c01193bc>] ? irq_enter+0x2f/0x41
 [<c01043a1>] ? do_IRQ+0x19/0x68
 [<c0103227>] ? common_interrupt+0x27/0x30
 [<c0107662>] ? default_idle+0x28/0x3d
 [<c0101dde>] ? cpu_idle+0x30/0x5e
 [<c0318c9a>] ? rest_init+0x66/0x68
Code: 02 00 00 00 8d 65 f8 5b 5e 5d c3 83 3d d8 9f 45 c0 00 55 89 e5 74
3b e8
20 9f ff ff c7 05 d8 9f 45 c0 00 00 00 00 a3 fc 9f 45 c0 <89> 15 00 a0
45 c0 2b
05 dc 9f 45 c0 1b 15
EIP: [<c012d8eb>] tick_nohz_stop_idle+0x20/0x49 SS:ESP 0068:c03f7f58


and

BUG: unable to handle kernel NULL pointer dereference at (null)
IP: [<c0129f24>] update_wall_time+0x1a/0x6f2
*pde = 00000000
Oops: 0000 [#1] PREEMPT
last sysfs file: /sys/class/net/lo/operstate
Modules linked in:

Pid: 2416, comm: bash Not tainted (2.6.29.3 #1)
EIP: 0060:[<c0129f24>] EFLAGS: 00010202 CPU: 0
EIP is at update_wall_time+0x1a/0x6f2
EAX: 00000001 EBX: 00000001 ECX: c03c9100 EDX: 00000000
ESI: 00000000 EDI: ffffff9c EBP: c281ff30 ESP: c281fed4
 DS: 007b ES: 007b FS: 0000 GS: 0033 SS: 0068
Process bash (pid: 2416, ti=c281e000 task=c3996670 task.ti=c281e000)
Stack:
 c380c320 c39a254c 00067442 00000031 3ba096ed 00000001 00000001 00000000
 c399669c aa88243a 0000000d c3996670 00000000 c399669c c281ff20 c0111777
 c03235d4 00000000 00000008 c281ff3c 00000001 00000000 c03c98e0 c281ff4c
Call Trace:
 [<c0111777>] ? task_tick_fair+0x1b/0x90
 [<c011cb87>] ? do_timer+0x1e/0xab
 [<c01676bd>] ? mntput_no_expire+0x19/0xf9
 [<c012ca9e>] ? tick_periodic+0x3c/0x72
 [<c012cae3>] ? tick_handle_periodic+0xf/0x68
 [<c0104754>] ? timer_interrupt+0x3c/0x44
 [<c01342dc>] ? handle_IRQ_event+0x21/0x48
 [<c0135368>] ? handle_level_irq+0x79/0xcf
 [<c01043d8>] ? do_IRQ+0x50/0x68
 [<c0103227>] ? common_interrupt+0x27/0x30
Code: 9e 45 c0 55 89 e5 f7 d9 f7 da e8 e9 eb fe ff 5d c3 55 89 e5 57 56
53 83
ec 54 83 3d 30 2c 3f c0 00 0f 85 d4 06 00 00 a1 68 9e 45 <c0> ff 50 10
8b 1d 68
9e 45 c0 8b 4b 14 2b
EIP: [<c0129f24>] update_wall_time+0x1a/0x6f2 SS:ESP 0068:c281fed4



I've tried passing the irqpoll option to the kernel boot line, the
sata_via
boot warning disappears but the crashes are still present.

Can it be workaround? Like e.g. making it go at PIO mode 0 i don't need
huge speeds here.

Any idea of what can be causing this? Is it a problem of the
motherboard? Of the adapter? Of the software?

Albert

[0] http://www.soekris.com/net4526.htm
[1]
http://www.globalamericaninc.com/Mini_PCI_Modules-Storage_Controller/c55_429/p1801120/1801120_-_Mini-PCI_Type_IIIB_124-pin_Storage_Interface_Module_Supports_1x_44-pin_EIDE,_2x_SATA-150,_1x_Compact_Flash/product_info.html

Re: sata_via oops on copying files?

[Alan Cox]

On Tue, 19 May 2009 09:16:55 -0700
"Albert Astals Cid" <albert@miraveo.com> wrote:

> 
> Please CC me, i'm not in the list.
> 
> Not sure if this or bugzilla is preferred, so doing both, shout me if i
> should wait on bugzilla.

None of this looks IDE related - the first looks like an IRQ routing
problem which is usually but not always a buggy BIOS. The random crashes
look like faulty hardware but its always hard to tell hardware and
software bugs.

Does the system pass memtest86 for 24 hours ?

Re: sata_via oops on copying files?

[Albert Astals Cid]

A Dimarts, 19 de maig de 2009, Alan Cox va escriure:
> On Tue, 19 May 2009 09:16:55 -0700
>
> "Albert Astals Cid" <albert@miraveo.com> wrote:
> > Please CC me, i'm not in the list.
> >
> > Not sure if this or bugzilla is preferred, so doing both, shout me if i
> > should wait on bugzilla.
>
> None of this looks IDE related - the first looks like an IRQ routing
> problem which is usually but not always a buggy BIOS. The random crashes
> look like faulty hardware but its always hard to tell hardware and
> software bugs.
>
> Does the system pass memtest86 for 24 hours ?

Hi Alan, 

If i don't use the Compact Flash attached to the miniPCI->IDE->CF adapter it 
works like a charm, it's only when i start writting (reading seems to work 
fine) from the CF card that the kernel oopses, that's why i thought it might 
be IDE/sata_via related.

Is it possible to work around IRQ problems? If so I guess this is not the 
correct mailing list for that question anymore, so can you suggest a better 
one? And if it's not possible i guess all i have left is bugging my hardware 
manufacturer.

Thanks,
  Albert

Re: sata_via oops on copying files?

[Alan Cox]

> If i don't use the Compact Flash attached to the miniPCI->IDE->CF adapter it 
> works like a charm, it's only when i start writting (reading seems to work 
> fine) from the CF card that the kernel oopses, that's why i thought it might 
> be IDE/sata_via related.

Not unreasonable as an assumption - and it may be. It doesn't look like a
typical ATA bug, but its often hard to tell some hardware and software
bugs apart. Do you have any other DMA using miniPCI devices you could try
with the board (eg network devices) to see if it stays stable, or other
cards you could test the CF with ?
 
> Is it possible to work around IRQ problems? If so I guess this is not the 

irqpoll is there for that reason

> correct mailing list for that question anymore, so can you suggest a better 
> one? And if it's not possible i guess all i have left is bugging my hardware 
> manufacturer.

RE: sata_via oops on copying files?

[Albert Astals Cid]

> > If i don't use the Compact Flash attached to the miniPCI->IDE->CF adapter it 
> > works like a charm, it's only when i start writting (reading seems to work 
> > fine) from the CF card that the kernel oopses, that's why i thought it might 
> > be IDE/sata_via related.
> 
> Not unreasonable as an assumption - and it may be. It doesn't look like a
> typical ATA bug, but its often hard to tell some hardware and software
> bugs apart. Do you have any other DMA using miniPCI devices you could try
> with the board (eg network devices) to see if it stays stable, or other
> cards you could test the CF with ?

I have a miniPCI wireless card and it works ok, in the sense that i can
run a looped wget someUrl -O /dev/null for hours and no crash happens.

I've tested the CF card on a USB CF card reader and works too (other
computer though as the crashing one doesn't have USB ports).

Sadly i don't have more CF cards to test in the miniPCI->CF adapter.


>  
> > Is it possible to work around IRQ problems? If so I guess this is not the 
> 
> irqpoll is there for that reason

oh :-/

> 
> > correct mailing list for that question anymore, so can you suggest a better 
> > one? And if it's not possible i guess all i have left is bugging my hardware 
> > manufacturer.
 

Thanks,
  Albert