 |
|
| Linux Forum Index » Linux Hardware » sata_sil and UDMA 6... |
|
Page 1 of 1 |
|
| Author |
Message |
| Hans-Joachim Zierke... |
 Posted: Wed Oct 14, 2009 9:38 am |
|
|
|
Guest
|
My SATA drive supports UDMA 6 and is recognized by the controller as
UDMA 6 during boot. (The controller can do 150).
But Linux (Debian Lenny) and its sata_sil driver do the following:
[ 1.790976] libata version 3.00 loaded.
[ 1.792392] sata_sil 0000:00:0d.0: version 2.3
[ 1.792507] ACPI: PCI Interrupt 0000:00:0d.0[A] -> GSI 16 (level, low) -> IRQ 16
[ 1.794742] sata_sil 0000:00:0d.0: Applying R_ERR on DMA activate FIS errata fix
[ 1.808790] scsi1 : sata_sil
[ 1.827902] scsi2 : sata_sil
[ 1.847935] scsi3 : sata_sil
[ 1.871953] scsi4 : sata_sil
[ 1.874332] ata1: SATA max UDMA/100 mmio m1024 at (no spam) 0xdb000000 tf 0xdb000080 irq 16
[ 1.877695] ata2: SATA max UDMA/100 mmio m1024 at (no spam) 0xdb000000 tf 0xdb0000c0 irq 16
[ 1.880072] ata3: SATA max UDMA/100 mmio m1024 at (no spam) 0xdb000000 tf 0xdb000280 irq 16
[ 1.880953] ata4: SATA max UDMA/100 mmio m1024 at (no spam) 0xdb000000 tf 0xdb0002c0 irq 16
<...>
[ 2.208076] ata1: SATA link up 1.5 Gbps (SStatus 113 SControl 310)
[ 2.216636] ata1.00: ATA-8: WDC WD10EADS-00M2B0, 01.00A01, max UDMA/133
[ 2.219357] ata1.00: 1953525168 sectors, multi 16: LBA48 NCQ (depth 0/32)
[ 2.230730] ata1.00: configured for UDMA/100
[ 2.548014] ata2: SATA link down (SStatus 0 SControl 310)
[ 2.868014] ata3: SATA link down (SStatus 0 SControl 310)
[ 3.188014] ata4: SATA link down (SStatus 0 SControl 310)
Question: The current drive is slower than UDMA/100 anyway, so it doesn't
hurt at the moment. Will this change, if ata2 gets a faster drive?
Hans-Joachim
--
It sometimes seems to me that people have a need for the externalization
of evil. They have the need to think that there is, somewhere, an enemy
boundlessly evil, because this makes them feel boundlessly good.
George Kennan |
|
|
| Back to top |
|
|
|
| Pascal Hambourg... |
| Posted: Wed Oct 14, 2009 9:48 am |
|
|
|
Guest
|
Hello,
Hans-Joachim Zierke a écrit :
Quote: My SATA drive supports UDMA 6 and is recognized by the controller as
UDMA 6 during boot. (The controller can do 150).
...
[ 2.208076] ata1: SATA link up 1.5 Gbps (SStatus 113 SControl 310)
[ 2.216636] ata1.00: ATA-8: WDC WD10EADS-00M2B0, 01.00A01, max UDMA/133
[ 2.219357] ata1.00: 1953525168 sectors, multi 16: LBA48 NCQ (depth 0/32)
[ 2.230730] ata1.00: configured for UDMA/100
...
Question: The current drive is slower than UDMA/100 anyway, so it doesn't
hurt at the moment. Will this change, if ata2 gets a faster drive?
Question : the UDMA mode defines signal timings for the parallel
transport (PATA), so does it actually matter for serial transport (SATA)
devices or is it just for software compatibility and ignored ? |
|
|
| Back to top |
|
|
|
| Hans-Joachim Zierke... |
| Posted: Sat Oct 24, 2009 2:53 pm |
|
|
|
Guest
|
Pascal Hambourg schrieb:
Quote: [ 2.208076] ata1: SATA link up 1.5 Gbps (SStatus 113 SControl 310)
[ 2.216636] ata1.00: ATA-8: WDC WD10EADS-00M2B0, 01.00A01, max UDMA/133
[ 2.219357] ata1.00: 1953525168 sectors, multi 16: LBA48 NCQ (depth 0/32)
[ 2.230730] ata1.00: configured for UDMA/100
...
Question: The current drive is slower than UDMA/100 anyway, so it doesn't
hurt at the moment. Will this change, if ata2 gets a faster drive?
Question : the UDMA mode defines signal timings for the parallel
transport (PATA), so does it actually matter for serial transport (SATA)
devices or is it just for software compatibility and ignored ?
If it's a fake for software compatibility, it certainly won't get reported
to human readers.
Hans-Joachim
--
It sometimes seems to me that people have a need for the externalization
of evil. They have the need to think that there is, somewhere, an enemy
boundlessly evil, because this makes them feel boundlessly good.
George Kennan |
|
|
| Back to top |
|
|
|
| Pascal Hambourg... |
| Posted: Sun Oct 25, 2009 8:41 am |
|
|
|
Guest
|
Hans-Joachim Zierke a écrit :
Quote: Pascal Hambourg schrieb:
Question: The current drive is slower than UDMA/100 anyway, so it doesn't
hurt at the moment. Will this change, if ata2 gets a faster drive?
Question : the UDMA mode defines signal timings for the parallel
transport (PATA), so does it actually matter for serial transport (SATA)
devices or is it just for software compatibility and ignored ?
If it's a fake for software compatibility, it certainly won't get reported
to human readers.
The software doesn't know it's fake, and reports it to human readers.
Have you tested the sustained throughput in different UDMA modes ? If it
does not change, and more importantly, if it is higher that the
specified maximum throughput for a given mode (e.g. > 33 MB/s in UDMA
mode 2), then it's obviously fake. |
|
|
| Back to top |
|
|
|
|
|
All times are GMT - 5 Hours
The time now is Fri Nov 27, 2009 1:11 pm
|
|