Athlon: Try this (was: Re: Athlon bug stomping #2)
A message that is not in the archive (sent off-list or not yet received): [email protected]
Hello Petr,
Saturday, September 15, 2001, 10:16:59 AM, you wrote:
PV> Hi,
PV> I'll sleep much better with ...
Quoted 5 lines
Review
Quoted 5 lines
(quoted diff) +5
-0+static void __init pci_fixup_athlon_bug(struct pci_dev *d) +{ + u8 v; + printk("Trying to stomp on Athlon bug...\n"); + pci_read_config_byte(d, 0x55, &v);
PV>+ if (v & 0x80) {
Quoted 2 lines
Review
Quoted 2 lines
(quoted diff) +2
-0+ v &= 0x7f; /* clear bit 55.7 */ + pci_write_config_byte(d, 0x55, v);
PV>+ }
How about this:
+static void __init pci_fixup_athlon_bug(struct pci_dev *d) +{
+ u8 v;
+ pci_read_config_byte(d, 0x55, &v);
+ if(v & 0x80) {
+ printk(KERN_NOTICE "Stomping on Athlon bug.\n");
+ v &= 0x7f; /* clear bit 55.7 */
+ pci_write_config_byte(d, 0x55, v);+ }
+}
Well, these are cosmetic changes anyway...
What is more important now:
1) Do we have people who still see Athlon bug with the patch?
2) Do Alan read these messages? ;-)
Signature · 3 lines
Best regards, VDA mailto:[email protected] http://port.imtp.ilyichevsk.odessa.ua/vda/
Quoted 11 lines
(quoted diff) +10
-0+static void __init pci_fixup_athlon_bug(struct pci_dev *d) +{ + u8 v; + pci_read_config_byte(d, 0x55, &v); + if(v & 0x80) { + printk(KERN_NOTICE "Stomping on Athlon bug.\n"); + v &= 0x7f; /* clear bit 55.7 */ + pci_write_config_byte(d, 0x55, v); + } +}
Quoted 4 lines
Well, these are cosmetic changes anyway...
What is more important now:
1) Do we have people who still see Athlon bug with the patch?
2) Do Alan read these messages? ;-)
Im watching the discussion with interest. If it proves to be the magic bullet I will ask VIA for guidance on the issue
Quoted 11 lines · Alan Cox
Quoted 11 lines
(quoted diff) +10
-0+static void __init pci_fixup_athlon_bug(struct pci_dev *d) +{ + u8 v; + pci_read_config_byte(d, 0x55, &v); + if(v & 0x80) { + printk(KERN_NOTICE "Stomping on Athlon bug.\n"); + v &= 0x7f; /* clear bit 55.7 */ + pci_write_config_byte(d, 0x55, v); + } +}
Quoted 7 lines
Quoted 4 lines
Well, these are cosmetic changes anyway...
What is more important now:
1) Do we have people who still see Athlon bug with the patch?
2) Do Alan read these messages? ;-)Im watching the discussion with interest. If it proves to be the magic bullet I will ask VIA for guidance on the issue
Not being a developer or guru on the internal software workings of the hardware here, I have to ask, does this clear up some bug, or does it disable the optimizations causing the problem? Is this a *fix* or a band-aid?
Quoted 16 lines
Quoted 16 lines
Quoted 15 lines
(quoted diff) +10
-0+static void __init pci_fixup_athlon_bug(struct pci_dev *d) +{ + u8 v; + pci_read_config_byte(d, 0x55, &v); + if(v & 0x80) { + printk(KERN_NOTICE "Stomping on Athlon bug.\n"); + v &= 0x7f; /* clear bit 55.7 */ + pci_write_config_byte(d, 0x55, v); + } +}Well, these are cosmetic changes anyway...
What is more important now:
1) Do we have people who still see Athlon bug with the patch?
2) Do Alan read these messages? ;-)
Quoted 7 lines
Quoted 2 lines
Im watching the discussion with interest. If it proves to be the magic bullet I will ask VIA for guidance on the issue
Not being a developer or guru on the internal software workings of the hardware here, I have to ask, does this clear up some bug, or does it disable the optimizations causing the problem? Is this a *fix* or a band-aid?
AFAICT, it's a *fix*. The register in question is "for debug use only, never write anything other than zero in it" but generally shows non-zero when a faulty BIOS is installed. We're zeroing the bit that appears to cause the problem. It's highly unlikely to be an extra optimisation - if it is, it's an unsafe one.
Signature · 7 lines
-------------------------------------------------------------- from: Jonathan "Chromatix" Morton mail: [email protected] (not for attachments) website: http://www.chromatix.uklinux.net/vnc/ geekcode: GCS$/E dpu(!) s:- a20 C+++ UL++ P L+++ E W+ N- o? K? w--- O-- M++$ V? PS PE- Y+ PGP++ t- 5- X- R !tv b++ DI+++ D G e+ h+ r++ y+(*) tagline: The key to knowledge is not to rely on people to teach you it.
Quoted 4 lines
Not being a developer or guru on the internal software workings of the hardware here, I have to ask, does this clear up some bug, or does it disable the optimizations causing the problem? Is this a *fix* or a band-aid?
Without documentation - who knows
Quoted 11 lines
(quoted diff) +10
-0+static void __init pci_fixup_athlon_bug(struct pci_dev *d) +{ + u8 v; + pci_read_config_byte(d, 0x55, &v); + if(v & 0x80) { + printk(KERN_NOTICE "Stomping on Athlon bug.\n"); + v &= 0x7f; /* clear bit 55.7 */ + pci_write_config_byte(d, 0x55, v); + } +}
Quoted 3 lines
Well, these are cosmetic changes anyway...
What is more important now:
1) Do we have people who still see Athlon bug with the patch?
Just by any chance - does anybody have KT133 (not KT133A) datasheet? I just noticed at home that my KT133 has reg 55 set to 0x89 and it happilly lives... So maybe some BIOS vendors used KT133 instead of KT133A BIOS image?
Thanks,
Petr Vandrovec [email protected]
Quoted 12 lines · Petr Vandrovec
Quoted 12 lines
(quoted diff) +10
-0+static void __init pci_fixup_athlon_bug(struct pci_dev *d) +{ + u8 v; + pci_read_config_byte(d, 0x55, &v); + if(v & 0x80) { + printk(KERN_NOTICE "Stomping on Athlon bug.\n"); + v &= 0x7f; /* clear bit 55.7 */ + pci_write_config_byte(d, 0x55, v); + } +}
Quoted 8 lines
Quoted 3 lines
Well, these are cosmetic changes anyway...
What is more important now:
1) Do we have people who still see Athlon bug with the patch?Just by any chance - does anybody have KT133 (not KT133A) datasheet? I just noticed at home that my KT133 has reg 55 set to
0x89and it happilly lives... So maybe some BIOS vendors used KT133 instead of KT133A BIOS image?
Hmm, I have a "KT133 Athlon Norbridge, Preliminary Revision 1.0 May 12, 2000" PDF. Register 55 is described like this :
Device 0 Offset 55 Debug .............................................. RW
7-5 Reserved (do not program)...................... default = 0
4 Write Policy for CPU Write to DRAM
0 Issue DRAM write when FIFO holds more
than two requests of DRAM controller idle . def
1 Disable Write Policy
3-0 Reserved (do not program)...................... default = 0
Which doesn't explain things much more since the bits in question (7, 3, 0) is marked as "Reserved".
I also have a question; if "movntq; sfence" type of memory copy can cause data corruption in kernel space, it can in theory also do so in user space right ? So, if I'm right this bug could also be on machines running a 2.2 kernel with userspace programs using 3DNow (or SSE even) instructions.
Regards,
Signature · 4 lines
Steffen Persvold | Scalable Linux Systems | Try out the world's best mailto:[email protected] | http://www.scali.com | performing MPI implementation: Tel: (+47) 2262 8950 | Olaf Helsets vei 6 | - ScaMPI 1.12.2 - Fax: (+47) 2262 8951 | N0621 Oslo, NORWAY | >300MBytes/s and <4uS latency
Quoted 4 lines · Steffen Persvold
I also have a question; if "movntq; sfence" type of memory copy can cause data corruption in kernel space, it can in theory also do so in user space right ? So, if I'm right this bug could also be on machines running a 2.2 kernel with userspace programs using 3DNow (or SSE even) instructions.
Sure. I did crash a computer running an old kernel (not 2.2 but 2.4.0-testX without the optimised fast_copy_page) from a non-privileged user space program containing the same code as the optimised fast_copy_page.
Jan
Quoted 3 lines · Jan Niehusmann
Sure. I did crash a computer running an old kernel (not 2.2 but 2.4.0-testX without the optimised
fast_copy_page) from a non-privileged user space program containing the same code as the optimisedfast_copy_page.
Port this to Windows if you want to raise Hell. Call it something nasty (like "AMD/VIA bug exploit") to get attention.
Quoted 11 lines · Petr Vandrovec
Quoted 11 lines
(quoted diff) +10
-0+static void __init pci_fixup_athlon_bug(struct pci_dev *d) +{ + u8 v; + pci_read_config_byte(d, 0x55, &v); + if(v & 0x80) { + printk(KERN_NOTICE "Stomping on Athlon bug.\n"); + v &= 0x7f; /* clear bit 55.7 */ + pci_write_config_byte(d, 0x55, v); + } +}
Quoted 8 lines
Quoted 3 lines
Well, these are cosmetic changes anyway...
What is more important now:
1) Do we have people who still see Athlon bug with the patch?Just by any chance - does anybody have KT133 (not KT133A) datasheet? I just noticed at home that my KT133 has reg 55 set to
0x89and it happilly lives... So maybe some BIOS vendors used KT133 instead of KT133A BIOS image?
Same here ...
Signature · 2 lines
Vojtech Pavlik SuSE Labs
Quoted 4 lines
Quoted 2 lines
to
0x89and it happilly lives... So maybe some BIOS vendors used KT133 instead of KT133A BIOS image?Same here ...
One way to test this hypthesis maybe to run dmidecode on the machines and see if they report KT133 or KT133A. Its also possible some BIOS code does blindly program bit 7 even tho its reserved and should have been kept unchanged.
Alan
Quoted 9 lines · Alan Cox
Quoted 4 lines
Quoted 2 lines
to
0x89and it happilly lives... So maybe some BIOS vendors used KT133 instead of KT133A BIOS image?Same here ...
One way to test this hypthesis maybe to run dmidecode on the machines and see if they report KT133 or KT133A. Its also possible some BIOS code does blindly program bit 7 even tho its reserved and should have been kept unchanged.
I'm not sure if this is the output you are talking about (I see no KT133 strings on it) but here it goes (this motherboard uses KT133A and exhibits the bug):
SMBIOS 2.3 present.
DMI 2.3 present.
40 structures occupying 1181 bytes.
DMI table at 0x000F0800.
Handle 0x0000
DMI type 0, 20 bytes. BIOS Information Block Vendor: Award Software International, Inc. Version: 6.00 PG Release: 07/05/2001 BIOS base: 0xE0000 ROM size: 192K Capabilities: Flags: 0x000000007FCBDE90
Handle 0x0001
DMI type 1, 25 bytes.
System Information Block
Vendor: VIA Technologies, Inc.
Product: VT8363
Version:
Serial Number:
Handle 0x0002
DMI type 2, 8 bytes. Board Information Block Vendor: <http://www.abit.com.tw> Product: 8363-686A(KT7,KT7A,KT7A-RAID,KT7E) Version: Serial Number:
Handle 0x0003
DMI type 3, 17 bytes.
Chassis Information Block
Vendor:
Chassis Type: Desktop
Version:
Serial Number:
Asset Tag:
Handle 0x0004
DMI type 4, 35 bytes.
Processor
Socket Designation: Socket A
Processor Type: Central Processor
Processor Family: M1
Processor Manufacturer: AMD
Processor Version: AMD Duron(tm)
Serial Number:
Asset Tag:
Vendor Part Number:
Handle 0x0006
DMI type 6, 12 bytes.
Memory Bank
Socket: BANK_0
Banks: 0 1
Speed: 60nS
Type:
Installed Size: 64Mbyte
Enabled Size: 64Mbyte
Handle 0x0007
DMI type 6, 12 bytes.
Memory Bank
Socket: BANK_1
Banks: 2 3
Speed: 60nS
Type:
Installed Size: 64Mbyte
Enabled Size: 64Mbyte
Handle 0x0008
DMI type 6, 12 bytes.
Memory Bank
Socket: BANK_2
Banks: 4 5
Speed: 60nS
Type: UNKNOWN
Installed Size: Not Installed
Enabled Size: Not Installed
Handle 0x000A
DMI type 7, 19 bytes.
Cache
Socket: Internal Cache
L1 Internal Cache: write-back
L1 Cache Size: 128K
L1 Cache Maximum: 128K
L1 Cache Type: Synchronous
Handle 0x000B
DMI type 7, 19 bytes.
Cache
Socket: External Cache
L2 External Cache: write-back
L2 Cache Size: 64K
L2 Cache Maximum: 64K
L2 Cache Type: Synchronous
Handle 0x000C
DMI type 8, 9 bytes.
Port Connector
Internal Designator: PRIMARY IDE
Internal Connector Type: On Board IDE External Designator:
External Connector Type: None
Port Type: Other
Handle 0x000D
DMI type 8, 9 bytes.
Port Connector
Internal Designator: SECONDARY IDE
Internal Connector Type: On Board IDE External Designator:
External Connector Type: None
Port Type: Other
Handle 0x000E
DMI type 8, 9 bytes.
Port Connector
Internal Designator: FDD
Internal Connector Type: On Board Floppy External Designator:
External Connector Type: None
Port Type: „(û
Handle 0x000F
DMI type 8, 9 bytes.
Port Connector
Internal Designator: COM1
Internal Connector Type: 9 Pin Dual Inline (pin 10 cut) External Designator:
External Connector Type: DB-9 pin male
Port Type: Serial Port 16450 Compatible
Handle 0x0010
DMI type 8, 9 bytes.
Port Connector
Internal Designator: COM2
Internal Connector Type: 9 Pin Dual Inline (pin 10 cut) External Designator:
External Connector Type: DB-9 pin male
Port Type: Serial Port 16450 Compatible
Handle 0x0011
DMI type 8, 9 bytes.
Port Connector
Internal Designator: LPT1
Internal Connector Type: DB-25 pin female External Designator:
External Connector Type: DB-25 pin female Port Type: Parallel Port ECP/EPP
Handle 0x0012
DMI type 8, 9 bytes.
Port Connector
Internal Designator: Keyboard
Internal Connector Type: Other
External Designator:
External Connector Type: PS/2
Port Type: Keyboard Port
Handle 0x0013
DMI type 8, 9 bytes.
Port Connector
Internal Designator: PS/2 Mouse
Internal Connector Type: PS/2
External Designator: No Detected
External Connector Type: PS/2
Port Type: Mouse Port
Handle 0x0014
DMI type 9, 13 bytes.
Card Slot
Slot: ISA
Type: 16bit ISA
Slot Features: 5v
Handle 0x0015
DMI type 9, 13 bytes.
Card Slot
Slot: PCI
Type: 32bit PCI
Status: In use.
Slot Features: 5v
Handle 0x0016
DMI type 9, 13 bytes.
Card Slot
Slot: PCI
Type: 32bit PCI
Status: In use.
Slot Features: 5v
Handle 0x0017
DMI type 9, 13 bytes.
Card Slot
Slot: PCI
Type: 32bit PCI
Status: In use.
Slot Features: 5v
Handle 0x0018
DMI type 9, 13 bytes.
Card Slot
Slot: PCI
Type: 32bit PCI
Status: In use.
Slot Features: 5v
Handle 0x0019
DMI type 9, 13 bytes.
Card Slot
Slot: PCI
Type: 32bit PCI
Status: Available.
Slot Features: 5v
Handle 0x001A
DMI type 9, 13 bytes.
Card Slot
Slot: PCI
Type: 32bit PCI
Status: Available.
Slot Features: 5v
Handle 0x001B
DMI type 10, 8 bytes.
On Board Devices Information
Description: : Enabled
Type:
Description: : Enabled
Type:
Handle 0x001C
DMI type 9, 13 bytes.
Card Slot
Slot: AGP
Type: 32bit AGP 2x
Status: Available.
Slot Features: 5v
Handle 0x001D
DMI type 8, 9 bytes.
Port Connector
Internal Designator: USB
Internal Connector Type: None
External Designator:
External Connector Type: Other
Port Type: USB
Handle 0x001E
DMI type 13, 22 bytes. BIOS Language Information Handle 0x001F DMI type 16, 15 bytes. Physical Memory Array Handle 0x0020 DMI type 17, 27 bytes. Memory Device Handle 0x0403 DMI type 0, 0 bytes. BIOS Information Block Vendor: NoneNoneNoneNone Version: " Release: BIOS base: 0x41420 ROM size: 4800K Capabilities: Flags: 0x326B6E614200315F
Handle 0x0605
DMI type 3, 4 bytes.
Chassis Information Block
Vendor:
Chassis Type:
Version:
Serial Number:
Asset Tag:
Handle 0x0201
DMI type 9, 0 bytes.
Card Slot
Slot:
Type:
Slot Features: 3.3v
Handle 0x0403
DMI type 0, 0 bytes. BIOS Information Block Vendor: NoneNoneNoneNone Version: $ Release: BIOS base: 0x41420 ROM size: 4800K Capabilities: Flags: 0x346B6E614200325F
Handle 0x0605
DMI type 3, 4 bytes.
Chassis Information Block
Vendor:
Chassis Type:
Version:
Serial Number:
Asset Tag:
Handle 0xFF00
DMI type 0, 0 bytes.
BIOS Information Block
Vendor:
Version:
Release:
BIOS base: 0x1F000
ROM size: 64K
Capabilities:
Flags: 0x0000251314000020
Handle 0x01FF
DMI type 0, 255 bytes. BIOS Information Block Vendor: Version: Release: BIOS base: 0x01000 ROM size: 0K Capabilities: Flags: 0x0000000025131400
Handle 0x0000
DMI type 0, 0 bytes.
BIOS Information Block
Vendor:
Version:
Release:
BIOS base: 0x00000
ROM size: 0K
Capabilities:
Flags: 0x0000000000000000
Handle 0x0000
DMI type 0, 0 bytes.
BIOS Information Block
Vendor:
Version:
Release:
BIOS base: 0x00000
ROM size: 0K
Capabilities:
Flags: 0x0000000000000000
--
Roberto Jung Drebes <[email protected]>
Porto Alegre, RS - Brasil
http://www.inf.ufrgs.br/~drebes/
Quoted 10 lines · Roberto Jung Drebes
Quoted 4 lines · Alan Cox
One way to test this hypthesis maybe to run dmidecode on the machines and see if they report KT133 or KT133A. Its also possible some BIOS code does blindly program bit 7 even tho its reserved and should have been kept unchanged.
I'm not sure if this is the output you are talking about (I see no KT133 strings on it) but here it goes (this motherboard uses KT133A and exhibits the bug):
Quoted 7 lines
Handle
0x0001
DMI type 1, 25 bytes.
System Information Block
Vendor: VIA Technologies, Inc.
Product: VT8363
Version:
Serial Number:
VT8363 is basic KT133. KT133A == VT8363A, at least according to VT8363A datasheet...
Quoted 7 lines
Handle
0x0002DMI type 2, 8 bytes. Board Information Block Vendor: <http://www.abit.com.tw> Product: 8363-686A(KT7,KT7A,KT7A-RAID,KT7E) Version: Serial Number:
... but there are listed both KT133 and KT133A based motherboards, so maybe it is intentional?
Best regards,
Petr Vandrovec [email protected]
Quoted 10 lines · Petr Vandrovec
Quoted 7 lines · Roberto Jung Drebes
Handle
0x0002DMI type 2, 8 bytes. Board Information Block Vendor: <http://www.abit.com.tw> Product: 8363-686A(KT7,KT7A,KT7A-RAID,KT7E) Version: Serial Number:
Quoted 2 lines
... but there are listed both KT133 and KT133A based motherboards, so maybe it is intentional?
The same south bridge is used for both the KT133 and the KT133A.
Signature · 1 line
Ignacio Vazquez-Abrams <[email protected]>
Quoted 4 lines
Quoted 2 lines
... but there are listed both KT133 and KT133A based motherboards,
so maybe it is intentional?The same south bridge is used for both the KT133 and the KT133A.
And also for other northbridges. I've seen the 82c686a used on an i810 board, which is currently in my friend's games machine.
Signature · 7 lines
-------------------------------------------------------------- from: Jonathan "Chromatix" Morton mail: [email protected] (not for attachments) website: http://www.chromatix.uklinux.net/vnc/ geekcode: GCS$/E dpu(!) s:- a20 C+++ UL++ P L+++ E W+ N- o? K? w--- O-- M++$ V? PS PE- Y+ PGP++ t- 5- X- R !tv b++ DI+++ D G e+ h+ r++ y+(*) tagline: The key to knowledge is not to rely on people to teach you it.
Quoted 9 lines · Alan Cox
Quoted 4 lines
Quoted 2 lines
to
0x89and it happilly lives... So maybe some BIOS vendors used KT133 instead of KT133A BIOS image?Same here ...
One way to test this hypthesis maybe to run dmidecode on the machines and see if they report KT133 or KT133A. Its also possible some BIOS code does blindly program bit 7 even tho its reserved and should have been kept unchanged.
I think it's possible to decide whether a chipset is KT133 or KT133A based on the hostbridge revision. Mine is KT133 and is rev 03.
Signature · 2 lines
Vojtech Pavlik SuSE Labs
Quoted 7 lines
Quoted 4 lines
One way to test this hypthesis maybe to run dmidecode on the machines and see if they report KT133 or KT133A. Its also possible some BIOS code does blindly program bit 7 even tho its reserved and should have been kept unchanged.
I think it's possible to decide whether a chipset is KT133 or KT133A based on the hostbridge revision. Mine is KT133 and is rev 03.
That tells you the chipset of the bridge. dmidecode dumps bios strings which may tell you the chipset the bios was actually for..
Quoted 10 lines · Alan Cox
Quoted 7 lines
Quoted 4 lines
One way to test this hypthesis maybe to run dmidecode on the machines and see if they report KT133 or KT133A. Its also possible some BIOS code does blindly program bit 7 even tho its reserved and should have been kept unchanged.
I think it's possible to decide whether a chipset is KT133 or KT133A based on the hostbridge revision. Mine is KT133 and is rev 03.
That tells you the chipset of the bridge. dmidecode dumps bios strings which may tell you the chipset the bios was actually for..
Ahh, yes.
Signature · 2 lines
Vojtech Pavlik SuSE Labs
Quoted 11 lines · Petr Vandrovec
Quoted 11 lines
(quoted diff) +10
-0+static void __init pci_fixup_athlon_bug(struct pci_dev *d) +{ + u8 v; + pci_read_config_byte(d, 0x55, &v); + if(v & 0x80) { + printk(KERN_NOTICE "Stomping on Athlon bug.\n"); + v &= 0x7f; /* clear bit 55.7 */ + pci_write_config_byte(d, 0x55, v); + } +}
Quoted 10 lines
Quoted 4 lines
Well, these are cosmetic changes anyway...
What is more important now:
1) Do we have people who still see Athlon bug with the patch?Just by any chance - does anybody have KT133 (not KT133A) datasheet? I just noticed at home that my KT133 has reg 55 set to
0x89and it happilly lives... So maybe some BIOS vendors used KT133 instead of KT133A BIOS image?
Same here ... with a board from gigabyte
I had to replace a 128MB PC100 RAM module (maybe related, maybe not) but now it works.
Gerold Jury
In article <[email protected]> you write:
| + printk(KERN_NOTICE "Stomping on Athlon bug.\n");
Okay, I'm pedantic. It's not an Athlon bug, is it? It's a chipset bug? No? If so should say so!
Signature · 4 lines
bill davidsen ([email protected]) Prodigy Internet Server Group Project Leader, USENET news 914-448-1241