siI3112 crash on enabling dma

siI3112 crash on enabling dma

[Adriaan Peeters]

Hello,

I'm using 2.4.20-ac2, when I boot the system, SATA disks on a sil3112
controller (IWill IS150-R) are detected, but dma is not enabled.
The disks work perfectly in raid 1 mode, and I can enable dma using 
hdparm -X66 -d1 /dev/hda
hdparm -X66 -d1 /dev/hdc

But sometimes this fails and the entire system hangs. I suspect it
happens when I enable dma mode while there is some harddisk activity.

I tried 2.4.21-rc7-ac1 too, but the dma isn't enabled by default either.

If I can test some more, please let me know,

Regards,
Adriaan Peeters

PS: Please cc me when replying, I'm not subscribed to the list.

Interesting parts from dmesg:

Uniform Multi-Platform E-IDE driver Revision: 7.00beta-2.4
ide: Assuming 33MHz system bus speed for PIO modes; override with
idebus=xx
SiI3112 Serial ATA: IDE controller at PCI slot 00:0c.0
PCI: Found IRQ 5 for device 00:0c.0
SiI3112 Serial ATA: chipset revision 2
SiI3112 Serial ATA: not 100% native mode: will probe irqs later
    ide0: MMIO-DMA at 0xd081c000-0xd081c007, BIOS settings: hda:pio,
hdb:pio
    ide1: MMIO-DMA at 0xd081c008-0xd081c00f, BIOS settings: hdc:pio,
hdd:pio
VP_IDE: IDE controller at PCI slot 00:11.1
PCI: No IRQ known for interrupt pin A of device 00:11.1. Please try
using pci=biosirq.
VP_IDE: chipset revision 6
VP_IDE: not 100% native mode: will probe irqs later
ide: Assuming 33MHz system bus speed for PIO modes; override with
idebus=xx
VP_IDE: VIA vt8235 (rev 00) IDE UDMA133 controller on pci00:11.1
    ide2: BM-DMA at 0x8400-0x8407, BIOS settings: hde:DMA, hdf:pio
    ide3: BM-DMA at 0x8408-0x840f, BIOS settings: hdg:DMA, hdh:pio
hda: Maxtor 6Y080M0, ATA DISK drive
hda: DMA disabled
hdc: Maxtor 6Y080M0, ATA DISK drive
hdc: DMA disabled
hde: CD-ROM 40X/AKU, ATAPI CD/DVD-ROM drive
hde: DMA disabled
hdg: QUANTUM BIGFOOT2550A, ATA DISK drive
hdg: DMA disabled
blk: queue c047d8c4, I/O limit 4095Mb (mask 0xffffffff)
ide0 at 0xd081c080-0xd081c087,0xd081c08a on irq 5
ide1 at 0xd081c0c0-0xd081c0c7,0xd081c0ca on irq 5
ide2 at 0x1f0-0x1f7,0x3f6 on irq 14
ide3 at 0x170-0x177,0x376 on irq 15
hda: host protected area => 1
hda: 160086528 sectors (81964 MB) w/7936KiB Cache, CHS=158816/16/63
hdc: host protected area => 1
hdc: 160086528 sectors (81964 MB) w/7936KiB Cache, CHS=158816/16/63
hdg: task_no_data_intr: status=0x51 { DriveReady SeekComplete Error }
hdg: task_no_data_intr: error=0x04 { DriveStatusError }
hdg: 5033952 sectors (2577 MB) w/87KiB Cache, CHS=4994/16/63, DMA
hde: ATAPI 40X CD-ROM drive, 128kB Cache, UDMA(33)
Uniform CD-ROM driver Revision: 3.12
ide-floppy driver 0.99.newide
Partition check:
 hda:<6> [PTBL] [9964/255/63] hda1 hda2 hda3
 hdc:<6> [PTBL] [9964/255/63] hdc1 hdc2 hdc3
 hdg:<6> [PTBL] [624/128/63] hdg1 hdg2
ide-floppy driver 0.99.newide
 ataraid/d0: ataraid/d0p1 ataraid/d0p2 ataraid/d0p3
Drive 0 is 78167 Mb (22 / 0) 
Drive 1 is 78167 Mb (3 / 0) 
Raid1 array consists of 2 drives. 
-- 
Adriaan Peeters <apeeters@lashout.net>

Re: siI3112 crash on enabling dma

[Alan Cox]

On Gwe, 2003-06-06 at 20:52, Adriaan Peeters wrote:
> I tried 2.4.21-rc7-ac1 too, but the dma isn't enabled by default either.

That suprises me somewhat. 7-ac1 should force DMA on

> hda: Maxtor 6Y080M0, ATA DISK drive
> hda: DMA disabled
> hdc: Maxtor 6Y080M0, ATA DISK drive
> hdc: DMA disabled

These are Maxtor drives with SATA convertors ? If so make sure the
convertor is set for UDMA100 not UDMA133 mode. (See www.siimage.com
support pages)

Can you send me a dmesg from 7ac1 please

Re: siI3112 crash on enabling dma

[Adriaan Peeters]

On Sat, 2003-06-07 at 03:09, Alan Cox wrote:
> On Gwe, 2003-06-06 at 20:52, Adriaan Peeters wrote:
> > I tried 2.4.21-rc7-ac1 too, but the dma isn't enabled by default either.
> 
> That suprises me somewhat. 7-ac1 should force DMA on
> 
> > hda: Maxtor 6Y080M0, ATA DISK drive
> > hda: DMA disabled
> > hdc: Maxtor 6Y080M0, ATA DISK drive
> > hdc: DMA disabled
> 
> These are Maxtor drives with SATA convertors ? If so make sure the
> convertor is set for UDMA100 not UDMA133 mode. (See www.siimage.com
> support pages)

No, these are native SATA drives. The only jumper settings are the usual
master/slave jumpers, and slave mode is enabled (no jumpers). But that
shouldn't matter.

> 
> Can you send me a dmesg from 7ac1 please

Here you go:

Uniform Multi-Platform E-IDE driver Revision: 7.00beta4-2.4
ide: Assuming 33MHz system bus speed for PIO modes; override with
idebus=xx
SiI3112 Serial ATA: IDE controller at PCI slot 00:0c.0
PCI: Found IRQ 5 for device 00:0c.0
SiI3112 Serial ATA: chipset revision 2
SiI3112 Serial ATA: not 100% native mode: will probe irqs later
    ide0: MMIO-DMA , BIOS settings: hda:pio, hdb:pio
    ide1: MMIO-DMA , BIOS settings: hdc:pio, hdd:pio
VP_IDE: IDE controller at PCI slot 00:11.1
VP_IDE: chipset revision 6
VP_IDE: not 100% native mode: will probe irqs later
ide: Assuming 33MHz system bus speed for PIO modes; override with
idebus=xx
VP_IDE: VIA vt8235 (rev 00) IDE UDMA133 controller on pci00:11.1
    ide2: BM-DMA at 0x8400-0x8407, BIOS settings: hde:DMA, hdf:pio
    ide3: BM-DMA at 0x8408-0x840f, BIOS settings: hdg:DMA, hdh:pio
hda: Maxtor 6Y080M0, ATA DISK drive
hdc: Maxtor 6Y080M0, ATA DISK drive
hde: CD-ROM 40X/AKU, ATAPI CD/DVD-ROM drive
hdg: QUANTUM BIGFOOT2550A, ATA DISK drive
blk: queue c0475ef4, I/O limit 4095Mb (mask 0xffffffff)
ide0 at 0xe081c080-0xe081c087,0xe081c08a on irq 5
ide1 at 0xe081c0c0-0xe081c0c7,0xe081c0ca on irq 5
ide2 at 0x1f0-0x1f7,0x3f6 on irq 14
ide3 at 0x170-0x177,0x376 on irq 15
hda: attached ide-disk driver.
hda: host protected area => 1
hda: 160086528 sectors (81964 MB) w/7936KiB Cache, CHS=158816/16/63
hdc: attached ide-disk driver.
hdc: host protected area => 1
hdc: 160086528 sectors (81964 MB) w/7936KiB Cache, CHS=158816/16/63
hdg: attached ide-disk driver.
hdg: task_no_data_intr: status=0x51 { DriveReady SeekComplete Error }
hdg: task_no_data_intr: error=0x04 { DriveStatusError }
hdg: 5033952 sectors (2577 MB) w/87KiB Cache, CHS=4994/16/63, DMA
hde: attached ide-cdrom driver.
hde: ATAPI 40X CD-ROM drive, 128kB Cache, UDMA(33)
Uniform CD-ROM driver Revision: 3.12
ide-floppy driver 0.99.newide
Partition check:
 hda: [PTBL] [9964/255/63] hda1 hda2 hda3
 hdc: [PTBL] [9964/255/63] hdc1 hdc2 hdc3
 hdg: [PTBL] [624/128/63] hdg1 hdg2
ide-floppy driver 0.99.newide
 ataraid/d0: ataraid/d0p1 ataraid/d0p2 ataraid/d0p3
Drive 0 is 78167 Mb (22 / 0) 
Drive 1 is 78167 Mb (3 / 0) 
Raid1 array consists of 2 drives. 

bari:~# hdparm -d /dev/hda

/dev/hda:
 using_dma    =  0 (off)
bari:~# hdparm -d /dev/hdc

/dev/hdc:
 using_dma    =  0 (off)
bari:~#

Thanks,

-- 
Adriaan Peeters <apeeters@lashout.net>

Re: siI3112 crash on enabling dma

[Adriaan Peeters]

On Fri, 2003-06-06 at 21:52, Adriaan Peeters wrote:
> Hello,
> 
> I'm using 2.4.20-ac2, when I boot the system, SATA disks on a sil3112
> controller (IWill IS150-R) are detected, but dma is not enabled.
> The disks work perfectly in raid 1 mode, and I can enable dma using 
> hdparm -X66 -d1 /dev/hda
> hdparm -X66 -d1 /dev/hdc
> 
> But sometimes this fails and the entire system hangs. I suspect it
> happens when I enable dma mode while there is some harddisk activity.

The system seemed to be stable for a few days, but then the following
happened:
- First I got read-only filesystem messages
- a few seconds later I/O errors
- After rebooting, the root filesystem (ext3) could not be mounted
- After fscking and a lot of deleted inodes, large portions of the
filesystem were gone to lost+found (/etc, /bin, ...)

I can only relate this to the driver, it seems very unstable :(

Some info: 2.4.20-ac2, asus a7v8x motherboard (onboard promise raid not
used), iwill IS150-R raid controler, 2 Maxtor 6Y080M0 SATA disks in
raid1.

-- 
Adriaan Peeters <apeeters@lashout.net>

Re: siI3112 crash on enabling dma

[Justin Cormack]

On Tue, 2003-06-10 at 17:05, Adriaan Peeters wrote:
> On Fri, 2003-06-06 at 21:52, Adriaan Peeters wrote:
> > Hello,
> > 
> > I'm using 2.4.20-ac2, when I boot the system, SATA disks on a sil3112
> > controller (IWill IS150-R) are detected, but dma is not enabled.
> > The disks work perfectly in raid 1 mode, and I can enable dma using 
> > hdparm -X66 -d1 /dev/hda
> > hdparm -X66 -d1 /dev/hdc
> > 
> > But sometimes this fails and the entire system hangs. I suspect it
> > happens when I enable dma mode while there is some harddisk activity.
> 
> The system seemed to be stable for a few days, but then the following
> happened:
> - First I got read-only filesystem messages
> - a few seconds later I/O errors
> - After rebooting, the root filesystem (ext3) could not be mounted
> - After fscking and a lot of deleted inodes, large portions of the
> filesystem were gone to lost+found (/etc, /bin, ...)
> 
> I can only relate this to the driver, it seems very unstable :(
> 
> Some info: 2.4.20-ac2, asus a7v8x motherboard (onboard promise raid not
> used), iwill IS150-R raid controler, 2 Maxtor 6Y080M0 SATA disks in
> raid1.

I got severe filesystem corruption on an Sii3112 system (Supermicro
E7505 with 2 onboard SiI3112) but I later memtested the machine and
found bad ram, so I dont think it is a valid example of corruption.
However I am still slightly suspicious. Did you have 
echo "max_kb_per_request:15" > /proc/ide/hdXX/settings
set for the relevant drives? This is a known bug (I didnt have this
either, another reason why this is not a valid datapoint...)


Justin

Re: siI3112 crash on enabling dma

[Adriaan Peeters]

On Tue, 2003-06-10 at 22:29, Justin Cormack wrote:

> I got severe filesystem corruption on an Sii3112 system (Supermicro
> E7505 with 2 onboard SiI3112) but I later memtested the machine and
> found bad ram, so I dont think it is a valid example of corruption.
> However I am still slightly suspicious. Did you have 

The memory is stable, no problems in memtest.

> echo "max_kb_per_request:15" > /proc/ide/hdXX/settings
> set for the relevant drives? This is a known bug (I didnt have this
> either, another reason why this is not a valid datapoint...)

These were not set when the error occured. I found this option after the
crash somewhere on the net, but I can't find any references to it on
google at the moment :( Can it cause that severe filesystem corruption ?

Was DMA enabled on your system, and what drives did you use ?

I tested several firmware revisions, disabled mmio in the driver,
nothing helped. No DMA at boot, and the crash is easy to reproduce, for
example:

console1: dd if=/dev/hda of=/dev/null
console2: hdparm -X66 -d1 /dev/hda

result without max_kb_per_request:
---
/dev/hda:
 setting using_dma to 1 (on)
blk: queue c0468b00, I/O limit 4095Mb (mask 0xffffffff)
 setting xfermode to 66 (UltraDMA mode2)
hda: dma_intr: status=0x58 { DriveReady SeekComplete DataRequest }

hda: dma_intr: status=0xd0 { Busy }

hda: DMA disabled
--

result with max_kb_per_request:
---
/dev/hda:
 setting using_dma to 1 (on)
blk: queue c0468b00, I/O limit 4095Mb (mask 0xffffffff)
 setting xfermode to 66 (UltraDMA mode2)
hda: dma_timer_expiry: dma status = 0x21
hda: timeout waiting for DMA
hda: timeout waiting for DMA
hda: status error: status=0x58 { DriveReady SeekComplete DataRequest }
--

The system hangs and nothing is written to syslog.

I don't know where to look next.

-- 
Adriaan Peeters <apeeters@lashout.net>