Silicon Image 3112 SerialATA FIX! (kernel patch inside)

Heretic · Posted: Mon Dec 29, 2003 1:57 am Post subject:

No, you were already using inefficient software RAID, and no.

The SI3112 never did hardware RAID, it was always BIOS assisted software RAID, and it was really slow at that. ataraid is gone for that reason.

behd · Apprentice Joined: 11 Feb 2003 Posts: 155

So it would means that people which have installed their OS with SI3112 raid won't be able to switch to a 2.6.n kernel ? (or wil have to re-install every OS hosted on such config, aaaargh... )

> The SI3112 never did hardware RAID, it was always BIOS assisted software RAID,
> and it was really slow at that. ataraid is gone for that reason.
What do you mean by really slow at that ? almost the same as using only one disk ?
(isn't there any boost when dealing with large files ?? NB. I directly setuped my computer with raid0, so I can't say if there's a lot of difference with or without the raid...). btw. do you have any benchmark ?

Thanks anyway for your answer

Heretic · Posted: Tue Dec 30, 2003 12:10 pm Post subject:

taskara · Posted: Fri Jan 02, 2004 12:17 am Post subject:

this has been a great read - thanks Ryan and Bart

_________________
Kororaa install method - have Gentoo up and running quickly and easily, fully automated with an installer!

ktech · Guru Joined: 12 Apr 2003 Posts: 340

hello.

I thinks it's off-topic, but perhaps any of you can help me. I have an nforce-2 motherboard with pata and sata support. The sata is throught the SI 3112 controller.

It's an Athlon 2500+ and I have my Seagate Barracuda V ST380023A (PATA) HD connected to the PATA controller and I get this in hdparm:

/dev/hda:
Timing buffer-cache reads: 1780 MB in 2.00 seconds = 889.25 MB/sec
Timing buffered disk reads: 98 MB in 3.02 seconds = 32.42 MB/sec

I think 32 MB/sec in this drive is very little, don't you think that?

The motherboard cames with a PATA to SATA adapter. Do you think I can gain any speed if I connect my HD to SATA in this way?

Thanks a lot!

taskara · Posted: Fri Jan 02, 2004 3:34 am Post subject:

hi,

the converter will give you ZERO performance gain.

it's probably more an issue of running in the correct dma mode, and having the right driver installed etc.

what kernel are you using

and can u post

Heretic · Posted: Fri Jan 02, 2004 4:27 am Post subject:

ktech · Guru Joined: 12 Apr 2003 Posts: 340

Hello

I was not talking again gaining performance with the converter, but with the "more advanced" controller.

The motherboard is an nforce-2 so I think the "amd and nvidia ide" is the correct.

I have enabled in kernel config the options:
* ATAPI support
* Enhanced IDE....
* Include IDE / ATA2 DISK support
* Use multi-mode by default
* Include IDE/ATAPI cdrom support
* PCI ide chipset support
* Generic PCI Bus-Master DMA support
* Use PCI DMA by default when available
* AMD and nVidia IDE support
* Silicon Image chipset support

bash-2.05b# hdparm -i /dev/hda

/dev/hda:

Model=ST380023A, FwRev=3.53, SerialNo=3KA140JH
Config={ HardSect NotMFM HdSw>15uSec Fixed DTR>10Mbs RotSpdTol>.5% }
RawCHS=16383/16/63, TrkSize=0, SectSize=0, ECCbytes=4
BuffType=unknown, BuffSize=2048kB, MaxMultSect=16, MultSect=16
CurCHS=16383/16/63, CurSects=16514064, LBA=yes, LBAsects=156301488
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=no WriteCache=enabled
Drive conforms to: ATA/ATAPI-6 T13 1410D revision 2:

* signifies the current active mode

ktech · Guru Joined: 12 Apr 2003 Posts: 340

sorry.... I forget this:

bash-2.05b# hdparm /dev/hda

/dev/hda:
multcount = 16 (on)
IO_support = 1 (32-bit)
unmaskirq = 1 (on)
using_dma = 1 (on)
keepsettings = 0 (off)
readonly = 0 (off)
readahead = 256 (on)
geometry = 65535/16/63, sectors = 156301488, start = 0
bash-2.05b#

I think I followed the suggestion by one guy that say setting the Read-ahead to 4K but nothing improved.

Thanks!

ktech · Guru Joined: 12 Apr 2003 Posts: 340

And I gorgot mentioning that I'm running 2.6-mm2 kernel.

I have tried with 2.6.0 test-x in the past.... even 2.6.0.... no improvements...

I'm using 2.6.0-mm2 because someone says me it improves performance in nforce2 motherboards because in this kernel I can enable APIC / LAPCI.

Thanks and sorry for all of this posts...

taskara · Posted: Fri Jan 02, 2004 1:58 pm Post subject:

ahh ok.. well I think that's prob about right.

2.4 kernel is faster than 2.6 for amd ide.

under 2.4 you'd prob get 40-45 or so

also your drive only has 2mb cache.. and people with 7200rpm with 8mb cache get about 40-45..

everything is set correctly...

oh one final thought, do u have something on hdb? if so, move it and try again..
_________________
Kororaa install method - have Gentoo up and running quickly and easily, fully automated with an installer!

ktech · Guru Joined: 12 Apr 2003 Posts: 340

No, I have nothing in hdb... only hdc...

Then... do you suggest me something more to try?

Have you heard something about _why_ is 2.4 fastest that 2.6? Or perhaps any plan to improve the driver?

Do you think I can get faster speeds with SATA controller throught PATA-SATA adapter?

Thanks!

taskara · Posted: Sat Jan 03, 2004 12:21 am Post subject:

no idea.. I remember switching from 2.4 to 2.6 and got slower speeds on the ide controller - it may be because the old driver (at faster speeds) was causing some problems, no idea.

and I don't think you will get any performance increase from the converter, like someone else said it will prob run slower.
_________________
Kororaa install method - have Gentoo up and running quickly and easily, fully automated with an installer!

sneakerski · Apprentice Joined: 14 Oct 2003 Posts: 168

well, i'm using the same kernel and this patch is included, i've got 2 maxtor 160GB plus 9s in RAID 0. performance is around 110MB/s on the sil 3112 controller, it's atleast worth giving a shot removing the converter. incidentally i've got the nf7-s too (v2)

ktech · Guru Joined: 12 Apr 2003 Posts: 340

What set of patches do you recommends me in order to run a Seagate drive in the Si 3112 chip?

it's not that I don't believe you, but I must try this because my transfer speeds suck so much.

Thanks!

P.S.: I'm using now 2.6.0-mm2

taskara · Posted: Sat Jan 03, 2004 10:35 am Post subject:

PrakashP · Posted: Sun Jan 04, 2004 10:28 pm Post subject:

@Heretic

Nice work you did, I just saw this thread now. Checked siimage 1.09 and so far the best driver I had. Nevertheless a few issues:

I have a Samsung pata drive connected to siimage with bios 4.2.43 via pata->sata converter (has some benefits for me).

With all older drivers I could not do hdparm -d1 /dev/hde. (Nevertheless the drive was in fast dma mode from boot onwards, so only issue for me was unability to use swsusp.) Then I would get controller reset and crapping up, so that I would have press the reset button. Now this doesn't happen, but my drive (or controller) doesn't get into DMA mode. hdparm -t just gives me a few miserable mb/sec. dmesg (tried -d1 twice):

hde: dma_intr: status=0xd0 { Busy }

hde: DMA disabled
ide2: reset phy, status=0x00000113, siimage_reset
ide2: reset: success
hde: sata_error = 0x00000000, watchdog = 0, siimage_mmio_ide_dma_test_irq
hde: dma_intr: status=0xd0 { Busy }

hde: DMA disabled
ide2: reset phy, status=0x00000113, siimage_reset
ide2: reset: success

But using -d1 -X69 gives me back my speed. So far so good.

But nevertheless I am constantly getting this error/warning on drive access:

hde: sata_error = 0x00000000, watchdog = 0, siimage_mmio_ide_dma_test_irq

Any ideas how to fix this? This one is really nasty as I have to comment it out in the driver.

Below a few infos which might be useful. Would be great if you could help me or at least advise me how to fix it.

dmesg:
SiI3112 Serial ATA: IDE controller at PCI slot 0000:01:0b.0
SiI3112 Serial ATA: chipset revision 2
SiI3112 Serial ATA: 100% native mode on irq 18
ide2: MMIO-DMA , BIOS settings: hde:DMA, hdf:DMA
ide3: MMIO-DMA , BIOS settings: hdg:pio, hdh:pio
hde: SAMSUNG SP1614N, ATA DISK drive
ide2 at 0xf981c080-0xf981c087,0xf981c08a on irq 18
hde: max request size: 64KiB
hde: 312581808 sectors (160041 MB) w/8192KiB Cache, CHS=19457/255/63, UDMA(100)
/dev/ide/host2/bus0/target0/lun0: p1 p2 p3 < p5 p6 p7 >

bash-2.05b# hdparm -iI /dev/hde

/dev/hde:

Model=SAMSUNG SP1614N, FwRev=TM100-24, SerialNo=0735J1FW702444
Config={ HardSect NotMFM HdSw>15uSec Fixed DTR>10Mbs }
RawCHS=16383/16/63, TrkSize=34902, SectSize=554, ECCbytes=4
BuffType=DualPortCache, BuffSize=8192kB, MaxMultSect=16, MultSect=16
CurCHS=16383/16/63, CurSects=16514064, LBA=yes, LBAsects=268435455
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
AdvancedPM=no WriteCache=enabled
Drive conforms to: (null):

* signifies the current active mode

ATA device, with non-removable media
Model Number: SAMSUNG SP1614N
Serial Number: 0735J1FW702444
Firmware Revision: TM100-24
Standards:
Supported: 7 6 5 4
Likely used: 7
Configuration:
Logical max current
cylinders 16383 65535
heads 16 1
sectors/track 63 63
--
CHS current addressable sectors: 4128705
LBA user addressable sectors: 268435455
LBA48 user addressable sectors: 312581808
device size with M = 1024*1024: 152627 MBytes
device size with M = 1000*1000: 160041 MBytes (160 GB)
Capabilities:
LBA, IORDY(can be disabled)
Queue depth: 1
Standby timer values: spec'd by Standard, no device specific minimum
R/W multiple sector transfer: Max = 16 Current = 16
Recommended acoustic management value: 254, current value: 0
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:
* READ BUFFER cmd
* WRITE BUFFER cmd
* Host Protected Area feature set
* 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
* DOWNLOAD MICROCODE cmd
* SMART self-test
* SMART error logging
Security:
Master password revision code = 65534
supported
not enabled
not locked
not frozen
not expired: security count
supported: enhanced erase
96min for SECURITY ERASE UNIT. 96min for ENHANCED SECURITY ERASE UNIT.
HW reset results:
CBLID- below Vih
Device num = 0 determined by the jumper
Checksum: correct

bash-2.05b# hdparm -tT /dev/hde

/dev/hde:
Timing buffer-cache reads: 2040 MB in 2.00 seconds = 1019.14 MB/sec
Timing buffered disk reads: 184 MB in 3.03 seconds = 60.74 MB/sec

ballyn · n00b Joined: 14 Aug 2003 Posts: 73

GTVincent · Posted: Sat Jan 10, 2004 5:10 am Post subject:

I have tried the patch, but with it my computer hangs during read/write mounting of the file systems.

/dev/hda: 120 GB WD SATA
/dev/hdc: 36GB Raptor SATA
A7N8X Deluxe/SiI3112
kernel 2.6.1

without the patch:
/dev/hda:
Timing buffer-cache reads: 1420 MB in 2.00 seconds = 708.34 MB/sec
Timing buffered disk reads: 60 MB in 3.08 seconds = 19.50 MB/sec

/dev/hdc:
Timing buffer-cache reads: 1380 MB in 2.00 seconds = 689.42 MB/sec
Timing buffered disk reads: 86 MB in 3.02 seconds = 28.49 MB/sec

I'm not unhappy, but faster would be nicer...

taskara · Posted: Sat Jan 10, 2004 5:35 am Post subject:

PrakashP · Posted: Sat Jan 10, 2004 5:53 am Post subject:

@ballyn

2.6.1-love1

Heretic · Posted: Sat Jan 10, 2004 7:40 am Post subject:

PrakashP · Posted: Sat Jan 10, 2004 8:00 am Post subject:

BIOS is latest (and AFAIK with older ones it was the same).

I put the drive to SATA, as I want my DVD-ROM and CD-RW have a channel for their own, so all devices are independant. SO I have maximum performance. I think my cable is OK (all firmly inserted). I am sure it is a driver issue. Only this hde: sata_error = 0x00000000, watchdog = 0, siimage_mmio_ide_dma_test_irq message is really annoying. Do you have an idea why it *constantly* appears? (If it wasn't clear: Every time data is transfered from hd, this message appears, thus dmesg gets flooded...)

Using libata siimage I have no annoying messages and everything seems fine (except that swsusp doesn't currently support scsi layer).

What is really strange: I cannot (well I don't want to;)) make the kernel to init my hd in pio mode, ie even if I don't select DMA on in the kernel compile options, the HD on siimage will be running in DMA mode, so I don't understand, which path the driver takes in my case.