2.6.4 under heavy ioload disables sis5513 DMA

2.6.4 under heavy ioload disables sis5513 DMA

[Michael Frank]

Happens every few hours with heavy io and cpu load:

hda: dma_timer_expiry: dma status == 0x21
hda: DMA timeout error
hda: dma timeout error: status=0xd0 { Busy }

hda: DMA disabled
ide0: reset: success

DMA auto-reenabled by boot time hdparm -k

lspci -vv

00:02.5 IDE interface: Silicon Integrated Systems [SiS] 5513 [IDE] (prog-if 80 [Master])
         Subsystem: Micro-Star International Co., Ltd.: Unknown device 5332
         Control: I/O+ Mem- BusMaster+ SpecCycle- MemWINV- VGASnoop- ParErr- Stepping- SERR- FastB2B-
         Status: Cap+ 66Mhz- UDF- FastB2B- ParErr- DEVSEL=medium >TAbort- <TAbort- <MAbort- >SERR- <PERR-
         Latency: 128
         Interrupt: pin ? routed to IRQ 10
         Region 4: I/O ports at 4000 [size=16]
         Capabilities: [58] Power Management version 2
                 Flags: PMEClk- DSI- D1+ D2+ AuxCurrent=0mA PME(D0-,D1-,D2-,D3hot-,D3cold+)
                 Status: D0 PME-Enable- DSel=0 DScale=0 PME-


hdparm -iI

/dev/hda:

  Model=IC35L090AVV207-0, FwRev=V23OA63A, SerialNo=VNVC00G3CABSMD
  Config={ HardSect NotMFM HdSw>15uSec Fixed DTR>10Mbs }
  RawCHS=16383/16/63, TrkSize=0, SectSize=0, ECCbytes=52
  BuffType=DualPortCache, BuffSize=1821kB, MaxMultSect=16, MultSect=16
  CurCHS=16383/16/63, CurSects=16514064, LBA=yes, LBAsects=160836480
  IORDY=on/off, tPIO={min:240,w/IORDY:120}, tDMA={min:120,rec:120}
  PIO modes:  pio0 pio1 pio2 pio3 pio4
  DMA modes:  mdma0 mdma1 mdma2
  UDMA modes: udma0 udma1 udma2 udma3 udma4 *udma5
  AdvancedPM=yes: disabled (255) WriteCache=enabled
  Drive conforms to: ATA/ATAPI-6 T13 1410D revision 3a:  2 3 4 5 6


ATA device, with non-removable media
powers-up in standby; SET FEATURES subcmd spins-up.
         Model Number:       IC35L090AVV207-0
         Serial Number:      VNVC00G3CABSMD
         Firmware Revision:  V23OA63A
Standards:
         Used: ATA/ATAPI-6 T13 1410D revision 3a
         Supported: 6 5 4 3
Configuration:
         Logical         max     current
         cylinders       16383   65535
         heads           16      1
         sectors/track   63      63
         --
         CHS current addressable sectors:    4128705
         LBA    user addressable sectors:  160836480
         LBA48  user addressable sectors:  160836480
         device size with M = 1024*1024:       78533 MBytes
         device size with M = 1000*1000:       82348 MBytes (82 GB)
Capabilities:
         LBA, IORDY(can be disabled)
         bytes avail on r/w long: 52     Queue depth: 32
         Standby timer values: spec'd by Standard, no device specific minimum
         R/W multiple sector transfer: Max = 16  Current = 16
         Advanced power management level: unknown setting (0x0000)
         Recommended acoustic management value: 128, current value: 254
         DMA: mdma0 mdma1 mdma2 udma0 udma1 udma2 udma3 udma4 *udma5
              Cycle time: min=120ns recommended=120ns
         PIO: pio0 pio1 pio2 pio3 pio4
              Cycle time: no flow control=240ns  IORDY flow control=120ns
Commands/features:
         Enabled Supported:
            *    NOP cmd
            *    READ BUFFER cmd
            *    WRITE BUFFER cmd
            *    Host Protected Area feature set
                 Release interrupt
            *    Look-ahead
            *    Write cache
            *    Power Management feature set
                 Security Mode feature set
            *    SMART feature set
            *    FLUSH CACHE EXT command
            *    Mandatory FLUSH CACHE command
            *    Device Configuration Overlay feature set
            *    48-bit Address feature set
                 Automatic Acoustic Management feature set
                 SET MAX security extension
                 Address Offset Reserved Area Boot
                 SET FEATURES subcommand required to spinup after power up
                 Power-Up In Standby feature set
                 Advanced Power Management feature set
            *    READ/WRITE DMA QUEUED
            *    General Purpose Logging feature set
            *    SMART self-test
            *    SMART error logging
Security:
         Master password revision code = 65534
                 supported
         not     enabled
         not     locked
         not     frozen
         not     expired: security count
         not     supported: enhanced erase
         46min for SECURITY ERASE UNIT.
HW reset results:
         CBLID- above Vih
         Device num = 0 determined by the jumper
Checksum: correct

Regards
Michael

SiS APIC, hacker looking for docs/help, was : Re: 2.6.4 under heavy ioload disables sis5513 DMA

[Lionel Bouton]

Michael Frank wrote the following on 03/18/2004 11:52 AM :
> Happens every few hours with heavy io and cpu load:
> 
> hda: dma_timer_expiry: dma status == 0x21
> hda: DMA timeout error
> hda: dma timeout error: status=0xd0 { Busy }
> 
> hda: DMA disabled
> ide0: reset: success
> 
> DMA auto-reenabled by boot time hdparm -k
> 

Hum, I'm wondering if -k is fully functionnal (hdparm man page hints 
that this isn't supported by all drives and I don't remember any 
success/failure stories here).

> lspci -vv
> 
> 00:02.5 IDE interface: Silicon Integrated Systems [SiS] 5513 [IDE] 

SiS chipset : is APIC functionnal ? (cat /proc/interrupts)

If not, I believe this might be the problem and the solutions still 
eludes me (I don't think the problem lies in the IDE driver but in APIC 
support).

I've 2 SiS based mainboards forced to use XT-PIC (SiS735 and SiS645 
based) here but without this kind of problems (everything works until I 
start to add to many PCI cards in one system...). I'm willing to start 
hacking around (mostly on the 645 as the 735 is an always-on system).

Is reading the arch/i386/kernel/*pic* files (and probably others) enough 
to start or is there somewhere else to look for information ?

Regards,
-- 
Lionel Bouton - inet6
---------------------------------------------------------------------
    o              Siege social: 51, rue de Verdun - 92158 Suresnes
   /      _ __ _   Acces Bureaux: 33 rue Benoit Malon - 92150 Suresnes
  / /\  /_  / /_   France
  \/  \/_  / /_/   Tel. +33 (0) 1 41 44 85 36
   Inetsys S.A.    Fax  +33 (0) 1 46 97 20 10

Re: SiS APIC, hacker looking for docs/help, was : Re: 2.6.4 under heavy ioload disables sis5513 DMA

[Michael Frank]

On Thu, 18 Mar 2004 12:30:17 +0100, Lionel Bouton <Lionel.Bouton@inet6.fr> wrote:

> Michael Frank wrote the following on 03/18/2004 11:52 AM :
>> Happens every few hours with heavy io and cpu load:
>>
>> hda: dma_timer_expiry: dma status == 0x21
>> hda: DMA timeout error
>> hda: dma timeout error: status=0xd0 { Busy }
>>
>> hda: DMA disabled
>> ide0: reset: success
>>
>> DMA auto-reenabled by boot time hdparm -k
>>
>
> Hum, I'm wondering if -k is fully functionnal (hdparm man page hints
> that this isn't supported by all drives and I don't remember any
> success/failure stories here).

It is OK here as otherwhise io-performance would collapse and hdparm -iI would show PIO.

>
>> lspci -vv
>>
>> 00:02.5 IDE interface: Silicon Integrated Systems [SiS] 5513 [IDE]
>
> SiS chipset : is APIC functionnal ? (cat /proc/interrupts)

APIC disabled in BIOS.
APIC not compiled in kernel. This board did not work well with APIC,
lots of spurious interrupts and delayloop-cal hangs during boot also
with 2.4 but have not tried in 6 months.

>
> If not, I believe this might be the problem and the solutions still
> eludes me (I don't think the problem lies in the IDE driver but in APIC
> support).
>
> I've 2 SiS based mainboards forced to use XT-PIC (SiS735 and SiS645
> based) here but without this kind of problems (everything works until I
> start to add to many PCI cards in one system...). I'm willing to start
> hacking around (mostly on the 645 as the 735 is an always-on system).

No (PCI,AGP) cards plugged in.

>
> Is reading the arch/i386/kernel/*pic* files (and probably others) enough
> to start or is there somewhere else to look for information ?
>

Dunno how DMA timeout is related to interrupts or are you suggesting it
is loosing dma-complete interrupts?

Same board runs same and higher loads with 2.4.2[345] flawlessly. Also
8 hours OK with 2.4.26-pre4 last night + 430 cycles of swsusp2.

By now, it happened with 2.6.4 four (4) times in 2.7 hours

IO is still at 20MB/s average btw so it is still using DMA.

Regards
Michael

Re: SiS APIC, hacker looking for docs/help, was : Re: 2.6.4 under heavy ioload disables sis5513 DMA

[Lionel Bouton]

Michael Frank wrote the following on 03/18/2004 12:52 PM :

>
>>
>> Is reading the arch/i386/kernel/*pic* files (and probably others) enough
>> to start or is there somewhere else to look for information ?
>>
>
> Dunno how DMA timeout is related to interrupts or are you suggesting it
> is loosing dma-complete interrupts?
>

I don't know the details (yet I hope), but I'm quite sure that interrupt 
handling in XT-PIC mode leads to problems on several SiS configurations 
(when you reorganize PCI cards in a system and the behaviour changes or 
when you disable the VGA IRQ and some things start to work, the suspect 
becomes obvious). One user reported that putting 2 disks on one channel 
instead of one on each (so 1 IRQ used instead of 2) solve instability 
issues too... I ruled out IDE driver problems several times by 
repeatedly checking the code and the run-time register values against 
known-good values. My lack of knowledge on the interrupt handling 
details is what prevents me from being 100% sure that the problem lies 
here. This is why I'm willing to work on this subject.

> Same board runs same and higher loads with 2.4.2[345] flawlessly. Also
> 8 hours OK with 2.4.26-pre4 last night + 430 cycles of swsusp2.
>

Now I remember why your name ringed a bell ! Thanks for your testing 
work on swsusp, my amount of free time went up thanks to 2.4's swsusp.

Regards,

-- 
Lionel Bouton - inet6
---------------------------------------------------------------------
   o              Siege social: 51, rue de Verdun - 92158 Suresnes
  /      _ __ _   Acces Bureaux: 33 rue Benoit Malon - 92150 Suresnes
 / /\  /_  / /_   France
 \/  \/_  / /_/   Tel. +33 (0) 1 41 44 85 36
  Inetsys S.A.    Fax  +33 (0) 1 46 97 20 10
 

Re: SiS APIC, hacker looking for docs/help, was : Re: 2.6.4 under heavy ioload disables sis5513 DMA

[Ross Dickson]

<snip>
> > Dunno how DMA timeout is related to interrupts or are you suggesting it 
> > is loosing dma-complete interrupts? 
> > 
> > I don't know the details (yet I hope), but I'm quite sure that interrupt 
> > handling in XT-PIC mode leads to problems on several SiS configurations 
> > (when you reorganize PCI cards in a system and the behaviour changes or 
> > when you disable the VGA IRQ and some things start to work, the suspect 
> > becomes obvious). One user reported that putting 2 disks on one channel 
> > instead of one on each (so 1 IRQ used instead of 2) solve instability 
> > issues too... I ruled out IDE driver problems several times by 
> > repeatedly checking the code and the run-time register values against 
> > known-good values. My lack of knowledge on the interrupt handling 
> > details is what prevents me from being 100% sure that the problem lies 
> > here. This is why I'm willing to work on this subject. 
 
I had a report that my patch helped an sis740 board run in apic ioapic mode.
Don't know if it would help your situation. Here is relevant link.
http://linux.derkeiler.com/Mailing-Lists/Kernel/2004-03/4278.html

There is a problem with apm mode with my patch - small fix here if reqd.
http://linux.derkeiler.com/Mailing-Lists/Kernel/2004-03/4410.html

Hope it helps.
Ross.

> > Same board runs same and higher loads with 2.4.2[345] flawlessly. Also 
> > 8 hours OK with 2.4.26-pre4 last night + 430 cycles of swsusp2. 
> >