Skip to contentSkip to conversation

pata_hpt3x2n: fix overclocked MWDMA0 timing

PATCH In review Bartlomiej Zolnierkiewicz 3 participants 5 messages Nov 27, 2009 – Dec 3, 2009 libata

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 pata_hpt3x2n: fix overclocked MWDMA0 timing

Remove superfluous timings table entry while at it.

---
Sergei, XFER_UDMA_5 timing also looks suspicious, please take a look when you have a minute, thanks.

 drivers/ata/pata_hpt3x2n.c |    5 ++---
 1 file changed, 2 insertions(+), 3 deletions(-)
drivers/ata/pata_hpt3x2n.c +1 -2
Index: b/drivers/ata/pata_hpt3x2n.c
===================================================================
@@ -80,14 +80,13 @@ static struct hpt_clock hpt3x2n_clocks[]
 
 	{	XFER_MW_DMA_2,	0x2c829c62	},
 	{	XFER_MW_DMA_1,	0x2c829c66	},
-	{	XFER_MW_DMA_0,	0x2c829d2c	},
+	{	XFER_MW_DMA_0,	0x2c829d2e	},
 
 	{	XFER_PIO_4,	0x0c829c62	},
 	{	XFER_PIO_3,	0x0c829c84	},
 	{	XFER_PIO_2,	0x0c829ca6	},
 	{	XFER_PIO_1,	0x0d029d26	},
 	{	XFER_PIO_0,	0x0d029d5e	},
-	{	0,		0x0d029d5e	}
 };
 
 /**
#1 lore raw
Quoted 6 lines · Bartlomiej Zolnierkiewicz

Remove superfluous timings table entry while at it.

---
Sergei, XFER_UDMA_5 timing also looks suspicious, please take a look when you have a minute, thanks.

Yeah, but it's the same as XFER_UDMA_4, so actually underclocked...
However, it matches what the HPT371N datasheet and the vendor drivers have. The 'hpt366' driver uses more speedy mode, with 22.5 ns cycle. ;-)

MBR, Sergei

#2 lore raw
Quoted 7 lines · Hello, I wrote:
Quoted 4 lines

---
Sergei, XFER_UDMA_5 timing also looks suspicious, please take a look when you have a minute, thanks.

Yeah, but it's the same as XFER_UDMA_4, so actually underclocked...
However, it matches what the HPT371N datasheet and the vendor drivers have. The 'hpt366' driver uses more speedy mode, with 22.5 ns cycle. ;-)

I have just verified: this driver has always used this timing
historically, at least for HPT372+ chips.

MBR, Sergei

#3 lore raw
Quoted 7 lines · Sergei Shtylyov
Quoted 4 lines · Hello, I wrote:
Quoted 4 lines

---
Sergei, XFER_UDMA_5 timing also looks suspicious, please take a look when you have a minute, thanks.

Quoted 6 lines
Quoted 3 lines

Yeah, but it's the same as XFER_UDMA_4, so actually underclocked... However, it matches what the HPT371N datasheet and the vendor drivers have. The 'hpt366' driver uses more speedy mode, with 22.5 ns cycle. ;-)

I have just verified: this driver has always used this timing historically, at least for HPT372+ chips.

Could you clarify which "this timing" you are referring to? :) I never saw an Acked-by on this one.

Jeff

#4 lore raw

Hello.

Quoted 5 lines · Jeff Garzik
Quoted 5 lines
Quoted 4 lines
Quoted 4 lines

---
Sergei, XFER_UDMA_5 timing also looks suspicious, please take a look when you have a minute, thanks.

Quoted 8 lines
Quoted 6 lines
Quoted 3 lines

Yeah, but it's the same as XFER_UDMA_4, so actually underclocked... However, it matches what the HPT371N datasheet and the vendor drivers have. The 'hpt366' driver uses more speedy mode, with 22.5 ns cycle. ;-)

I have just verified: this driver has always used this timing historically, at least for HPT372+ chips.

Could you clarify which "this timing" you are referring to? :)

UDMA5 timing which looked suspicious to Bart: 'hpt366' driver uses
faster timing at 66 MHz clock than the vendor drivers and HPT371N datasheet have (they have it the same as UDMA4).

Quoted 1 line

I never saw an Acked-by on this one.

I usually don't ACK libata patches, but this one can be an exception:

Though I guess you're waiting for Alan's ACK... :-)

Quoted 1 line

Jeff

MBR, Sergei