* 2.4.3-p8 pci_fixup_vt8363 + ASUS A7V "Optimal" = IDE disk corruption
@ 2001-03-29 5:29 Wayne Whitney
2001-03-29 16:16 ` Arjan van de Ven
0 siblings, 1 reply; 5+ messages in thread
From: Wayne Whitney @ 2001-03-29 5:29 UTC (permalink / raw)
To: LKML
Hi,
I'm running kernel 2.4.3-pre8 on an ASUS A7V (BIOS 1007) motherboard and
recently noticed that it sometimes corrupts my hard disk, an IBM 75GXP on
the onboard PDC20265 IDE controller. The corruption is detectable with a
simple 'dd if=/dev/urandom of=test bs=16384 count=32768; cp test test2 ;
diff test test2'.
I traced the problem to a combination of choosing "Optimal" for the System
Permorfance Setting in the BIOS and the the new pci_fixup_vt8363 added to
arch/i386/kernel/pci-pc.c in kernel 2.4.3-pre3. So I did a bunch of tests
using no pci_fixup function, the pci_fixup_vt8363 function, and the
following subset of pci_fixup_vt8363:
pci_read_config_byte(d, 0x54, &tmp);
if(tmp & (1<<2)) {
printk("PCI: Bus master Pipeline request disabled\n");
pci_write_config_byte(d, 0x54, tmp & ~(1<<2));
}
pci_read_config_byte(d, 0x70, &tmp);
if(tmp & (1<<2)) {
printk("PCI: Disabled Master Read Caching\n");
pci_write_config_byte(d, 0x70, tmp & ~(1<<2));
}
The results for me:
Normal Optimal
------ -------
no pci_fixup no corruption no corruption
pci_fixup_vt8363 subset corruption corruption
pci_fixup_vt8363 no corruption corruption
At this point my skills and perseverance gave out, but if someone would
like me to do a few more specific tests, I could.
Below is the output of 'lspci -xxx -s 0:0' on this hardware, with no
pci_fixup, for both the Normal and Optimal BIOS settings, in the form of a
unified diff. Hopefully this will shed some light on what the BIOS is
doing, as we don't see to have pci_fixup_vt8363 quite right yet.
Cheers,
Wayne
[root@pizza /mnt]# diff -u --unified=20 normal optimal
--- normal Wed Mar 28 20:37:57 2001
+++ optimal Wed Mar 28 20:35:32 2001
@@ -1,18 +1,18 @@
00:00.0 Host bridge: VIA Technologies, Inc. VT8363/8365 [KT133/KM133] (rev 02)
00: 06 11 05 03 06 00 10 22 02 00 00 06 00 00 00 00
10: 08 00 00 e4 00 00 00 00 00 00 00 00 00 00 00 00
20: 00 00 00 00 00 00 00 00 00 00 00 00 43 10 33 80
30: 00 00 00 00 a0 00 00 00 00 00 00 00 00 00 00 00
40: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
50: 17 a4 6b b4 07 89 20 20 80 80 10 10 10 10 20 20
-60: 33 ff 55 a0 d4 d4 d4 00 40 78 86 2f 08 23 00 00
-70: d8 c0 cc 0d 0e a1 d2 00 01 b4 09 02 00 00 00 03
+60: 33 ff 55 a0 d4 d4 d4 00 44 5c 86 2f 08 23 00 00
+70: de c0 cc 0d 0e a1 d2 00 01 b4 19 02 00 00 00 03
80: 0f 40 00 00 c0 00 00 00 03 00 25 1f 00 00 00 00
90: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
a0: 02 c0 20 00 17 02 00 1f 00 00 00 00 6e 02 14 00
b0: 62 ec 80 e5 32 33 28 00 00 00 00 00 00 00 00 00
c0: 01 00 02 00 00 00 00 00 00 00 00 00 00 00 00 00
d0: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
e0: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
f0: 00 00 00 00 00 00 00 0e 22 00 00 00 00 00 91 06
^ permalink raw reply [flat|nested] 5+ messages in thread* Re: 2.4.3-p8 pci_fixup_vt8363 + ASUS A7V "Optimal" = IDE disk corruption
2001-03-29 5:29 2.4.3-p8 pci_fixup_vt8363 + ASUS A7V "Optimal" = IDE disk corruption Wayne Whitney
@ 2001-03-29 16:16 ` Arjan van de Ven
0 siblings, 0 replies; 5+ messages in thread
From: Arjan van de Ven @ 2001-03-29 16:16 UTC (permalink / raw)
To: Wayne Whitney; +Cc: LKML
On Wed, Mar 28, 2001 at 09:29:46PM -0800, Wayne Whitney wrote:
> I'm running kernel 2.4.3-pre8 on an ASUS A7V (BIOS 1007) motherboard and
> recently noticed that it sometimes corrupts my hard disk, an IBM 75GXP on
> the onboard PDC20265 IDE controller. The corruption is detectable with a
> simple 'dd if=/dev/urandom of=test bs=16384 count=32768; cp test test2 ;
> diff test test2'.
/me throws arms in the air in despiration
It seems that the exact same settings corrupt a lot on one board and not on
another. But if you remove the quirk, the corruption is the
other way around........
Greetings,
Arjan van de Ven
^ permalink raw reply [flat|nested] 5+ messages in thread
* re: 2.4.3-p8 pci_fixup_vt8363 + ASUS A7V "Optimal" = IDE disk corruption
@ 2001-03-29 12:05 Todd M. Roy
0 siblings, 0 replies; 5+ messages in thread
From: Todd M. Roy @ 2001-03-29 12:05 UTC (permalink / raw)
To: Wayne Whitney, linux-kernel
Wayne,
I have also been seeing disk corruption with my ASUS A7V with both
2.4.3-pre7 and pre8.
-- todd --
Hi,
I'm running kernel 2.4.3-pre8 on an ASUS A7V (BIOS 1007) motherboard and
recently noticed that it sometimes corrupts my hard disk, an IBM 75GXP
on
the onboard PDC20265 IDE controller. The corruption is detectable with
a
simple 'dd if=/dev/urandom of=test bs=16384 count=32768; cp test test2 ;
diff test test2'.
I traced the problem to a combination of choosing "Optimal" for the
System
Permorfance Setting in the BIOS and the the new pci_fixup_vt8363 added
to
arch/i386/kernel/pci-pc.c in kernel 2.4.3-pre3. So I did a bunch of
tests
using no pci_fixup function, the pci_fixup_vt8363 function, and the
following subset of pci_fixup_vt8363:
pci_read_config_byte(d, 0x54, &tmp);
if(tmp & (1<<2)) {
printk("PCI: Bus master Pipeline request disabled\n");
pci_write_config_byte(d, 0x54, tmp & ~(1<<2));
}
pci_read_config_byte(d, 0x70, &tmp);
if(tmp & (1<<2)) {
printk("PCI: Disabled Master Read Caching\n");
pci_write_config_byte(d, 0x70, tmp & ~(1<<2));
}
The results for me:
Normal Optimal
------ -------
no pci_fixup no corruption no corruption
pci_fixup_vt8363 subset corruption corruption
pci_fixup_vt8363 no corruption corruption
At this point my skills and perseverance gave out, but if someone would
like me to do a few more specific tests, I could.
Below is the output of 'lspci -xxx -s 0:0' on this hardware, with no
pci_fixup, for both the Normal and Optimal BIOS settings, in the form of
a
unified diff. Hopefully this will shed some light on what the BIOS is
doing, as we don't see to have pci_fixup_vt8363 quite right yet.
Cheers,
Wayne
--
.~. Todd Roy, Senior Database Administrator .~.
/V\ Holstein Association, U.S.A. Inc. /V\
// \\ troy@holstein.com // \\
/( )\ 1-802-254-4551x4230 /( )\
^^-^^ ^^-^^
^ permalink raw reply [flat|nested] 5+ messages in thread[parent not found: <m2puf0hft9.fsf@trasno.mitica>]
* Re: 2.4.3-p8 pci_fixup_vt8363 + ASUS A7V "Optimal" = IDE disk corruption
[not found] <m2puf0hft9.fsf@trasno.mitica>
@ 2001-03-29 18:09 ` Wayne Whitney
2001-05-09 17:31 ` Juan Quintela
0 siblings, 1 reply; 5+ messages in thread
From: Wayne Whitney @ 2001-03-29 18:09 UTC (permalink / raw)
To: LKML; +Cc: toddroy, Juan Quintela
On 29 Mar 2001, Juan Quintela wrote:
> Hi I have the same motherboard and BIOS version. I was having
> filesystem corruption. There is a bugfix (from Arjan van der Ven) in
> the ac tree (around ac20 I think), could you test the last ac patch
> and test if the filesystem corruption persist??
I took a look at 2.4.2-ac28; rather than test the kernel itself, I used
setpci to duplicate what the fixup function does. The upshot is that
2.4.2-ac28 does not cause any corruption with the ASUS A7V 1007 "Optimal".
Its pci_fixup_vt8363 function has a subset of the tests from 2.4.3-pre8,
namely it omits:
pci_read_config_byte(d, 0x54, &tmp);
if(tmp & (1)) {
printk("PCI: Fast Write to Read turnaround disabled\n");
pci_write_config_byte(d, 0x54, tmp & ~(1));
}
pci_read_config_byte(d, 0x70, &tmp);
if(tmp & (1<<2)) {
printk("PCI: Disabled Master Read Caching\n");
pci_write_config_byte(d, 0x70, tmp & ~(1<<2));
}
This second test was part of the 2.4.3-pre8 pci_fixup_vt8363 subset from
my previous email which causes corruption on the ASUS A7V 1007 for both
Optimal and Normal settings. I verified that if I add it back in, I do
get corruption with ASUS A7V 1007 "Optimal". By omitting it, I guess
2.4.2-ac28 avoids the corruption of 2.4.3-pre8.
Perhaps 2.4.3-pre9 should adopt the pci_fixup_vt8363 from 2.4.2-ac28?
Below is a patch.
Cheers,
Wayne
--- linux-2.4.3-pre8/arch/i386/kernel/pci-pc.c Wed Mar 28 22:56:04 2001
+++ linux-2.4.2-ac28/arch/i386/kernel/pci-pc.c Wed Mar 28 22:51:00 2001
@@ -968,23 +971,13 @@
printk("PCI: Bus master Pipeline request disabled\n");
pci_write_config_byte(d, 0x54, tmp & ~(1<<2));
}
- pci_read_config_byte(d, 0x54, &tmp);
- if(tmp & (1)) {
- printk("PCI: Fast Write to Read turnaround disabled\n");
- pci_write_config_byte(d, 0x54, tmp & ~(1));
- }
pci_read_config_byte(d, 0x70, &tmp);
if(tmp & (1<<3)) {
printk("PCI: Disabled enhanced CPU to PCI writes\n");
pci_write_config_byte(d, 0x70, tmp & ~(1<<3));
}
- pci_read_config_byte(d, 0x70, &tmp);
- if(tmp & (1<<2)) {
- printk("PCI: Disabled Master Read Caching\n");
- pci_write_config_byte(d, 0x70, tmp & ~(1<<2));
- }
pci_read_config_byte(d, 0x71, &tmp);
- if ((tmp & (1<<3))==0) {
+ if((tmp & (1<<3)) == 0) {
printk("PCI: Bursting cornercase bug worked around\n");
pci_write_config_byte(d, 0x71, tmp | (1<<3));
}
^ permalink raw reply [flat|nested] 5+ messages in thread* Re: 2.4.3-p8 pci_fixup_vt8363 + ASUS A7V "Optimal" = IDE disk corruption
2001-03-29 18:09 ` Wayne Whitney
@ 2001-05-09 17:31 ` Juan Quintela
0 siblings, 0 replies; 5+ messages in thread
From: Juan Quintela @ 2001-05-09 17:31 UTC (permalink / raw)
To: whitney; +Cc: LKML, toddroy
Hi
sorry for the delay, it is working for your motherboard the
lastest kernels of Mandrake? I think that all the problems
have been solved?
Sorry for the delay, as I was finishing more things.
Later, Juan.
>>>>> "wayne" == Wayne Whitney <whitney@math.berkeley.edu> writes:
wayne> On 29 Mar 2001, Juan Quintela wrote:
>> Hi I have the same motherboard and BIOS version. I was having
>> filesystem corruption. There is a bugfix (from Arjan van der Ven) in
>> the ac tree (around ac20 I think), could you test the last ac patch
>> and test if the filesystem corruption persist??
wayne> I took a look at 2.4.2-ac28; rather than test the kernel itself, I used
wayne> setpci to duplicate what the fixup function does. The upshot is that
wayne> 2.4.2-ac28 does not cause any corruption with the ASUS A7V 1007 "Optimal".
wayne> Its pci_fixup_vt8363 function has a subset of the tests from 2.4.3-pre8,
wayne> namely it omits:
wayne> pci_read_config_byte(d, 0x54, &tmp);
wayne> if(tmp & (1)) {
wayne> printk("PCI: Fast Write to Read turnaround disabled\n");
wayne> pci_write_config_byte(d, 0x54, tmp & ~(1));
wayne> }
wayne> pci_read_config_byte(d, 0x70, &tmp);
wayne> if(tmp & (1<<2)) {
wayne> printk("PCI: Disabled Master Read Caching\n");
wayne> pci_write_config_byte(d, 0x70, tmp & ~(1<<2));
wayne> }
wayne> This second test was part of the 2.4.3-pre8 pci_fixup_vt8363 subset from
wayne> my previous email which causes corruption on the ASUS A7V 1007 for both
wayne> Optimal and Normal settings. I verified that if I add it back in, I do
wayne> get corruption with ASUS A7V 1007 "Optimal". By omitting it, I guess
wayne> 2.4.2-ac28 avoids the corruption of 2.4.3-pre8.
wayne> Perhaps 2.4.3-pre9 should adopt the pci_fixup_vt8363 from 2.4.2-ac28?
wayne> Below is a patch.
wayne> Cheers,
wayne> Wayne
wayne> --- linux-2.4.3-pre8/arch/i386/kernel/pci-pc.c Wed Mar 28 22:56:04 2001
wayne> +++ linux-2.4.2-ac28/arch/i386/kernel/pci-pc.c Wed Mar 28 22:51:00 2001
wayne> @@ -968,23 +971,13 @@
wayne> printk("PCI: Bus master Pipeline request disabled\n");
wayne> pci_write_config_byte(d, 0x54, tmp & ~(1<<2));
wayne> }
wayne> - pci_read_config_byte(d, 0x54, &tmp);
wayne> - if(tmp & (1)) {
wayne> - printk("PCI: Fast Write to Read turnaround disabled\n");
wayne> - pci_write_config_byte(d, 0x54, tmp & ~(1));
wayne> - }
wayne> pci_read_config_byte(d, 0x70, &tmp);
wayne> if(tmp & (1<<3)) {
wayne> printk("PCI: Disabled enhanced CPU to PCI writes\n");
wayne> pci_write_config_byte(d, 0x70, tmp & ~(1<<3));
wayne> }
wayne> - pci_read_config_byte(d, 0x70, &tmp);
wayne> - if(tmp & (1<<2)) {
wayne> - printk("PCI: Disabled Master Read Caching\n");
wayne> - pci_write_config_byte(d, 0x70, tmp & ~(1<<2));
wayne> - }
wayne> pci_read_config_byte(d, 0x71, &tmp);
wayne> - if ((tmp & (1<<3))==0) {
wayne> + if((tmp & (1<<3)) == 0) {
wayne> printk("PCI: Bursting cornercase bug worked around\n");
wayne> pci_write_config_byte(d, 0x71, tmp | (1<<3));
wayne> }
--
In theory, practice and theory are the same, but in practice they
are different -- Larry McVoy
^ permalink raw reply [flat|nested] 5+ messages in thread
end of thread, other threads:[~2001-05-09 17:33 UTC | newest]
Thread overview: 5+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2001-03-29 5:29 2.4.3-p8 pci_fixup_vt8363 + ASUS A7V "Optimal" = IDE disk corruption Wayne Whitney
2001-03-29 16:16 ` Arjan van de Ven
-- strict thread matches above, loose matches on Subject: below --
2001-03-29 12:05 Todd M. Roy
[not found] <m2puf0hft9.fsf@trasno.mitica>
2001-03-29 18:09 ` Wayne Whitney
2001-05-09 17:31 ` Juan Quintela
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox