Skip to contentSkip to conversation

Incorrect 80 wire detection with amd 760mpx & 2.4.21-pre4-ac7

DISCUSSION Nicholas Wourms 3 participants 3 messages Mar 2, 2003 – Mar 3, 2003

lore
Share & tools
lore
#0 lore raw

FYI:
I suspect that: http://marc.theaimsgroup.com/?l=linux-kernel&m=104619727013220&w=2
is related to my problem.

Anyhow, I'm using a UDMA5 WesternDigital drive on a ASUS K7M266-D motherboard. With a plain, stock 2.4.20 kernel, the viper driver properly recognizes which channel has the 80 wire cable (in my case ide0). The hard disk is the primary master, a cd-r drive is the primary slave, and a zip drive is the secondary slave. I can successfully set UDMA5 with hdparm without any problems. However, after upgrading to 2.4.21-pre4-ac7, I noticed that the drive was stuck at UDMA2. Checking /proc/ide/amd74XX yeilds some unexpected results:

----------AMD BusMastering IDE Configuration----------------

Driver Version:                     2.9
South Bridge:                       Advanced Micro Devices [AMD] AMD-768 [Opus] IDE
Revision:                           IDE 0x4
Highest DMA rate:                   UDMA100
BM-DMA base:                        0xd800
PCI clock:                          33.3MHz
-----------------------Primary IDE-------Secondary IDE------
Prefetch Buffer:              yes                 yes
Post Write Buffer:            yes                 yes
Enabled:                      yes                 yes
Simplex only:                  no                  no
Cable Type:                   40w                 80w
-------------------drive0----drive1----drive2----drive3-----
Transfer Mode:       UDMA      UDMA      UDMA       PIO
Address Setup:       30ns      30ns      30ns     120ns
Cmd Active:          90ns      90ns      90ns      90ns
Cmd Recovery:        90ns      90ns      30ns      30ns
Data Active:         90ns      90ns      90ns     330ns
Data Recovery:       30ns      90ns      30ns     270ns
Cycle Time:          60ns      60ns      60ns     600ns
Transfer Rate:   33.3MB/s  33.3MB/s  33.3MB/s   3.3MB/s

It appears that the driver has got it backwards, identifying my 80 wire cable as a 40 wire cable and visa-versa. As I mentioned, this is completely opposite to the behavior of 2.4.20. I've poked around the source, but I can't come up with anything new to what the other person discovered. Trying to pass ide0=ata66 doesn't seem to have any effect on the situation. I can provide further information upon request, but I don't think it will be necessary at this point.

Cheers,
Nicholas

#1 lore raw
Quoted 8 lines · Nicholas Wourms

FYI:
I suspect that:
http://marc.theaimsgroup.com/?l=linux-kernel&m=104619727013220&w=2
is related to my problem.

Anyhow, I'm using a UDMA5 WesternDigital drive on a ASUS K7M266-D motherboard. With a plain, stock 2.4.20 kernel, the viper driver properly recognizes which channel has the

Yep. I'll apply the obvious fix if Vojtech doesn't. Its on the known list

#2 lore raw
Quoted 6 lines · Alan Cox
Quoted 5 lines · Nicholas Wourms

FYI:
I suspect that:
http://marc.theaimsgroup.com/?l=linux-kernel&m=104619727013220&w=2
is related to my problem.

Quoted 6 lines
Quoted 3 lines

Anyhow, I'm using a UDMA5 WesternDigital drive on a ASUS K7M266-D motherboard. With a plain, stock 2.4.20 kernel, the viper driver properly recognizes which channel has the

Yep. I'll apply the obvious fix if Vojtech doesn't. Its on the known list

Obvious fix attached.

Signature · 2 lines
Vojtech Pavlik
SuSE Labs
drivers/ide/pci/amd74xx.c +4 -3
@@ -1,5 +1,5 @@
 /*
- * Version 2.9
+ * Version 2.10
  *
  * AMD 755/756/766/8111 and nVidia nForce IDE driver for Linux.
  *
@@ -103,7 +103,7 @@
 
 	amd_print("----------AMD BusMastering IDE Configuration----------------");
 
-	amd_print("Driver Version:                     2.9");
+	amd_print("Driver Version:                     2.10");
 	amd_print("South Bridge:                       %s", bmide_dev->name);
 
 	pci_read_config_byte(dev, PCI_REVISION_ID, &t);
@@ -309,7 +309,8 @@
 
 		case AMD_UDMA_100:
 			pci_read_config_byte(dev, AMD_CABLE_DETECT, &t);
-			amd_80w = ((u & 0x3) ? 1 : 0) | ((u & 0xc) ? 2 : 0);
+			pci_read_config_dword(dev, AMD_UDMA_TIMING, &u);
+			amd_80w = ((t & 0x3) ? 1 : 0) | ((t & 0xc) ? 2 : 0);
 			for (i = 24; i >= 0; i -= 8)
 				if (((u >> i) & 4) && !(amd_80w & (1 << (1 - (i >> 4))))) {
 					printk(KERN_WARNING "AMD_IDE: Bios didn't set cable bits corectly. Enabling workaround.\n");