Skip to contentSkip to conversation

Call4Response: Correct PST table for ATHLONs

DISCUSSION Lars Gemeinhardt 4 participants 7 messages Jul 2, 2003

lore
Share & tools
lore

Lars Gemeinhardt

author
#0 lore raw

Hello cpufreq-user and developer,

can everybody with an correct PST table (diffent FIDs and VIDs) and an ATHLON cpu post the "dmesg" snip (you get this using the dmesg command after the modul is loaded) to this mailing-list... Maybe we can setup with this the correct values to the cpufreq driver (the K7 powernow part)...then otherwise the Acer and Sony notebooks doesn't work perfectly :-(

Thanks in advanced

Ciao Lars

#1 lore raw

Am Mittwoch, 2. Juli 2003 09:52 schrieb Lars Gemeinhardt:

Quoted 8 lines

Hello cpufreq-user and developer,

can everybody with an correct PST table (diffent FIDs and VIDs) and an ATHLON cpu post the "dmesg" snip (you get this using the dmesg command after the modul is loaded) to this mailing-list...
Maybe we can setup with this the correct values to the cpufreq driver (the K7 powernow part)...then otherwise the Acer and Sony notebooks doesn't work perfectly :-(

Your wish is my command:

This is a Vaio PCG-FX601 (Mobile Duron at 1100Mhz) running 2.4.21-rc1 and cpufreq-2.4.21-pre4-3 plus acpi and swsusp patches

Here the is dmesg (hope it helps):

powernow: AMD K7 CPU detected.
powernow: PowerNOW! Technology present. Can scale: frequency and voltage.
powernow: Found PSB header at c00f74d0
powernow: Table version: 0x12
powernow: Flags: 0x0 (Mobile voltage regulator)
powernow: Settling Time: 100 microseconds.
powernow: Has 28 PST tables. (Only dumping ones relevant to this CPU).
powernow: PST:9 (@c00f757a)
powernow:  cpuid: 0x762  fsb: 100        maxFID: 0x0     startvid: 0xc
powernow:    FID: 0x4 (5.0x [500MHz])    VID: 0x13 (1.200V)
powernow:    FID: 0x6 (6.0x [600MHz])    VID: 0x13 (1.200V)
powernow:    FID: 0x8 (7.0x [700MHz])    VID: 0x11 (1.250V)
powernow:    FID: 0xc (9.0x [900MHz])    VID: 0xd (1.350V)
powernow:    FID: 0x0 (11.0x [1100MHz])  VID: 0xc (1.400V)
powernow: Minimum speed 500 MHz. Maximum speed 1100 MHz.

Gregor

Signature · 3 lines
Gregor Hlawacek                                                  Austria
http://www.unileoben.ac.at/~m9327555/        May the source be with you!
 This posting is 100% M$ free. Guaranteed by Tux, the friendly penguin.
#2 lore raw
In reply to Lars Gemeinhardt

Alle 09:52, mercoledì 2 luglio 2003, Lars Gemeinhardt ha scritto:

Quoted 8 lines

Hello cpufreq-user and developer,

can everybody with an correct PST table (diffent FIDs and VIDs) and an ATHLON cpu post the "dmesg" snip (you get this using the dmesg command after the modul is loaded) to this mailing-list...
Maybe we can setup with this the correct values to the cpufreq driver (the K7 powernow part)...then otherwise the Acer and Sony notebooks doesn't work perfectly :-(

This is an Acer Aspire 1300DXV with and Amd Mobile Duron 1200MHz running Linux 2.5.70 vanilla

powernow: AMD K7 CPU detected.
powernow: PowerNOW! Technology present. Can scale: frequency and voltage.
powernow: Found PSB header at c00e5100
powernow: Table version: 0x12
powernow: Flags: 0x0 (Mobile voltage regulator)
powernow: Settling Time: 100 microseconds.
powernow: Has 1 PST tables. (Only dumping ones relevant to this CPU).
powernow: PST:0 (@c00e5110)
powernow:  cpuid: 0x762 fsb: 100        maxFID: 0x2     startvid: 0xb
powernow:    FID: 0x4 (5.0x [500MHz])   VID: 0x13 (1.200V)
powernow:    FID: 0x6 (6.0x [600MHz])   VID: 0x13 (1.200V)
powernow:    FID: 0xa (8.0x [800MHz])   VID: 0xe (1.300V)
powernow:    FID: 0xe (10.0x [1000MHz]) VID: 0xc (1.400V)
powernow:    FID: 0x2 (12.0x [1200MHz]) VID: 0xb (1.450V)

powernow: Minimum speed 500 MHz. Maximum speed 1200 MHz.

Matteo

Signature · 3 lines
'Detectoring is like gambling,' said Vimes, putting down the clove. 'The 
secret is to know the winner in advance.'
(Jingo)
#3 lore raw
In reply to Lars Gemeinhardt
Quoted 9 lines · Lars Gemeinhardt

Hello cpufreq-user and developer,

can everybody with an correct PST table (diffent FIDs and VIDs) and an ATHLON cpu post the "dmesg" snip (you get this using the dmesg command after the modul is loaded) to this mailing-list...
Maybe we can setup with this the correct values to the cpufreq driver (the K7 powernow part)...then otherwise the Acer and Sony notebooks doesn't work perfectly :-(

You may want perhaps to get x86info instead.

Just do something like that:

cvs -z3 -d:pserver:[email protected]:/cvsroot/x86info co x86info

then edit AMD/dumppsb.c
at a line which look like

if ((etuple(cpu) == pst->cpuid) && (maxfid==pst->maxfid) && (startvid==pst->startvid))

replace with
if (1)

Then all PST tables should be dumped for different cpuid.

Signature · 4 lines
Ducrot Bruno

--  Which is worse:  ignorance or apathy?
--  Don't know.  Don't care.
author
#4 lore raw
Quoted 21 lines · Ducrot Bruno
Quoted 7 lines · Lars Gemeinhardt

Hello cpufreq-user and developer,

can everybody with an correct PST table (diffent FIDs and VIDs) and an ATHLON cpu post the "dmesg" snip (you get this using the dmesg command after the modul is loaded) to this mailing-list... Maybe we can setup with this the correct values to the cpufreq driver (the K7 powernow part)...then otherwise the Acer and Sony notebooks doesn't work perfectly :-(

You may want perhaps to get x86info instead.

Just do something like that:

cvs -z3 -d:pserver:[email protected]:/cvsroot/x86info co x86info

then edit AMD/dumppsb.c
at a line which look like

Quoted 10 lines

if ((etuple(cpu) == pst->cpuid) && (maxfid==pst->maxfid) && (startvid==pst->startvid))

replace with
if (1)

Then all PST tables should be dumped for different cpuid.

I think this is not what I expect ... I tested x86info and here is my output-snip:

FID changes will happen
VID changes will happen
Current VID multiplier code: 1.300
Current FSB multiplier code: 13.0
Voltage ID codes: Maximum=1.300V Startup=1.300V Currently=1.300V Frequency ID codes: Maximum=13.0x Startup=5.0x Currently=13.0x Decoding BIOS PST tables (maxfid=14, startvid=e)
Found PSB header at 0x4018f4e0
Table version: 0x12
Flags: 0x0 (Mobile voltage regulator)
Settling Time: 100 microseconds.
Has 28 PST tables. (Only dumping ones relevant to this CPU). PST:26 (@0x4018f6bc)
cpuid: 0x780 fsb: 100 maxFID: 0x14 startvid: 0xe
num of p states in this table: 5
FID: 0x4 (5.0x [500MHz]) VID: 0xe (1.300V)
FID: 0x8 (7.0x [700MHz]) VID: 0xe (1.300V)
FID: 0xc (9.0x [900MHz]) VID: 0xe (1.300V)
FID: 0x0 (11.0x [1100MHz]) VID: 0xe (1.300V)
FID: 0x14 (13.0x [1300MHz]) VID: 0xe (1.300V)

Connector type: Socket A (462 Pin PGA)

I think the FIDs are OK, but the VIDs are all equal (but "VID changes will happen") - this means the cpu consumes to much power :-(
My idea was that people with a correct PST table for the ATHLON 1500+ (and other CPUs) can give the right values, because in my opinion this values are CPU specific and not PST/BIOS specific....

Ciao Lars

#5 lore raw
Quoted 63 lines · Lars Gemeinhardt
Quoted 26 lines · Ducrot Bruno
Quoted 11 lines · Lars Gemeinhardt

Hello cpufreq-user and developer,

can everybody with an correct PST table (diffent FIDs and VIDs) and an ATHLON cpu post the "dmesg" snip (you get this using the dmesg command after the modul is loaded) to this mailing-list...
Maybe we can setup with this the correct values to the cpufreq driver (the K7 powernow part)...then otherwise the Acer and Sony notebooks doesn't work perfectly :-(

You may want perhaps to get x86info instead.

Just do something like that:

cvs -z3 -d:pserver:[email protected]:/cvsroot/x86info co x86info

then edit AMD/dumppsb.c
at a line which look like

Quoted 11 lines

if ((etuple(cpu) == pst->cpuid) && (maxfid==pst->maxfid) && (startvid==pst->startvid))

replace with
if (1)

Then all PST tables should be dumped for different cpuid.

I think this is not what I expect ... I tested x86info and here is my output-snip:

FID changes will happen
VID changes will happen
Current VID multiplier code: 1.300
Current FSB multiplier code: 13.0
Voltage ID codes: Maximum=1.300V Startup=1.300V Currently=1.300V Frequency ID codes: Maximum=13.0x Startup=5.0x Currently=13.0x Decoding BIOS PST tables (maxfid=14, startvid=e)
Found PSB header at 0x4018f4e0
Table version: 0x12
Flags: 0x0 (Mobile voltage regulator)
Settling Time: 100 microseconds.
Has 28 PST tables. (Only dumping ones relevant to this CPU). PST:26 (@0x4018f6bc)
cpuid: 0x780 fsb: 100 maxFID: 0x14 startvid: 0xe
num of p states in this table: 5
FID: 0x4 (5.0x [500MHz]) VID: 0xe (1.300V)
FID: 0x8 (7.0x [700MHz]) VID: 0xe (1.300V)
FID: 0xc (9.0x [900MHz]) VID: 0xe (1.300V)
FID: 0x0 (11.0x [1100MHz]) VID: 0xe (1.300V)
FID: 0x14 (13.0x [1300MHz]) VID: 0xe (1.300V)

Quoted 8 lines

Connector type: Socket A (462 Pin PGA)

I think the FIDs are OK, but the VIDs are all equal (but "VID changes will happen") - this means the cpu consumes to much power :-( My idea was that people with a correct PST table for the ATHLON 1500+ (and other CPUs) can give the right values, because in my opinion this values are CPU specific and not PST/BIOS specific....

well, actually I meant for peoples with 'good' PST.

If you look the output from x86info:

Quoted 1 line

Has 28 PST tables. (Only dumping ones relevant to this CPU).

      ^^              ^^^^

Therefore, by replacing in AMD/dumppst.c
if ((etuple(cpu) == pst->cpuid) && (maxfid==pst->maxfid) && (startvid==pst->startvid))

by
if(1)

you may have hopefully more 'good' PST tables from peoples which have good bios perhaps.

Signature · 4 lines
Ducrot Bruno

--  Which is worse:  ignorance or apathy?
--  Don't know.  Don't care.
author
#6 lore raw
Quoted 7 lines · Ducrot Bruno
Quoted 3 lines

well, actually I meant for peoples with 'good' PST.

OK - fine ...

Quoted 20 lines

If you look the output from x86info:

Quoted 3 lines

Has 28 PST tables. (Only dumping ones relevant to this CPU).

     ^^              ^^^^

Therefore, by replacing in AMD/dumppst.c
if ((etuple(cpu) == pst->cpuid) && (maxfid==pst->maxfid) && (startvid==pst->startvid))

by
if(1)

you may have hopefully more 'good' PST tables from peoples which have good bios perhaps.