[PROBLEM] mtrr's not set, 2.6.13

[PROBLEM] mtrr's not set, 2.6.13

[Jim McCloskey]

I'm not sure who to report this to ....

Somewhere between 2.6.11.3 and 2.6.12 (also under 2.6.13), the
following change occurred on this box.

Under 2.6.11.3, mtrr ranges are automatically set when X is started:

----------------------------------------------------------------------
running 2.6.11.3:

cat /proc/mtrr
reg00: base=0x00000000 (   0MB), size= 512MB: write-back, count=1
reg01: base=0xe8000000 (3712MB), size= 128MB: write-combining, count=2
reg02: base=0xe0000000 (3584MB), size= 128MB: write-combining, count=1
----------------------------------------------------------------------

After installation of 2.6.12/13, mtrr ranges are not set:

----------------------------------------------------------------------
running 2.6.13:

Xorg.0.log:

(WW) RADEON(0): Failed to set up write-combining range (0xe8000000,0x8000000)

/var/log/messages:

Aug 30 17:37:13 localhost kernel: mtrr: type mismatch for e8000000,8000000 old: write-back new: write-combining
Aug 30 17:37:14 localhost kernel: mtrr: type mismatch for e0000000,8000000 old: write-back new: write-combining
Aug 30 17:37:14 localhost kernel: [drm] Initialized radeon 1.16.0 20050311 on minor 0: ATI Technologies Inc RV280 [Radeon 9200]
Aug 30 17:37:14 localhost kernel: mtrr: type mismatch for e8000000,8000000 old: write-back new: write-combining

cat /proc/mtrr
reg00: base=0x00000000 (   0MB), size=983552MB: write-back, count=1
----------------------------------------------------------------------

Under 2.6.13 it's fairly easy to force a crash of the X server (e.g. by
playing Tuxracer badly).

The only change here is in the kernel-version. I haven't tried all the
point releases between 2.6.11.3 and 2.6.12, but the relevant Changelogs
don't suggest that anything relevant changed.

I'll add what details I can below. Please let me know what other
information I can supply that might help resolve this.

Thanks,

Jim

----------------------------------------------------------------------
X server:

Source: xorg-x11
Version: 6.8.2.dfsg.1-2

from Debian's (testing) xserver-xorg package.

----------------------------------------
The graphics card:

0000:01:00.0 VGA compatible controller: ATI Technologies Inc RV280
[Radeon 9200] (rev 01) (prog-if 00 [VGA])
        Subsystem: ATI Technologies Inc: Unknown device 2002
        Flags: bus master, 66MHz, medium devsel, latency 32, IRQ 16
        Memory at e8000000 (32-bit, prefetchable) [size=128M]
        I/O ports at a000 [size=256]
        Memory at f9000000 (32-bit, non-prefetchable) [size=64K]
        Capabilities: [58] AGP version 3.0
        Capabilities: [50] Power Management version 2
----------------------------------------
/proc/cpuinfo:

processor	: 0
vendor_id	: GenuineIntel
cpu family	: 15
model		: 3
model name	: Intel(R) Pentium(R) 4 CPU 2.80GHz
stepping	: 3
cpu MHz		: 2800.074
cache size	: 1024 KB
fdiv_bug	: no
hlt_bug		: no
f00f_bug	: no
coma_bug	: no
fpu		: yes
fpu_exception	: yes
cpuid level	: 5
wp		: yes
flags		: fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush dts acpi mmx fxsr sse sse2 ss ht tm pbe pni monitor ds_cpl cid
bogomips	: 5583.66
----------------------------------------
Output of ver_linux
 
Gnu C                  4.0.1
Gnu make               3.80
binutils               2.16.1
util-linux             2.12p
mount                  2.12p
module-init-tools      3.2-pre1
e2fsprogs              1.37
reiserfsprogs          line
reiser4progs           line
Linux C Library        2.3.5
Dynamic linker (ldd)   2.3.5
Procps                 3.2.1
Net-tools              1.60
Console-tools          0.2.3
Sh-utils               5.2.1
udev                   068
Modules Loaded         radeon drm md5 ipv6 snd_cs46xx snd_rawmidi snd_seq_device snd_intel8x0 snd_ac97_codec snd_pcm_oss snd_mixer_oss snd_pcm snd_timer snd soundcore snd_page_alloc i2c_i801 i2c_core hw_random ehci_hcd uhci_hcd usbcore b44 mii parport_pc lp parport intel_agp agpgart pktcdvd ide_cd cdrom loop rtc evdev

----------------------------------------
lspci -v under 2.6.11.3

0000:00:00.0 Host bridge: Intel Corp. 82865G/PE/P DRAM Controller/Host-Hub Interface (rev 02)
	Subsystem: ABIT Computer Corp.: Unknown device 101e
	Flags: bus master, fast devsel, latency 0
	Memory at e8000000 (32-bit, prefetchable) [size=128M]
	Capabilities: [e4] #09 [0106]

0000:00:02.0 VGA compatible controller: Intel Corp. 82865G Integrated Graphics Device (rev 02) (prog-if 00 [VGA])
	Subsystem: ABIT Computer Corp.: Unknown device 101e
	Flags: bus master, fast devsel, latency 0, IRQ 16
	Memory at f0000000 (32-bit, prefetchable) [size=128M]
	Memory at fa000000 (32-bit, non-prefetchable) [size=512K]
	I/O ports at 9000 [size=8]
	Capabilities: [d0] Power Management version 1

0000:00:1e.0 PCI bridge: Intel Corp. 82801 PCI Bridge (rev c2) (prog-if 00 [Normal decode])
	Flags: bus master, fast devsel, latency 0
	Bus: primary=00, secondary=01, subordinate=01, sec-latency=32
	Memory behind bridge: f8000000-f9ffffff

0000:00:1f.0 ISA bridge: Intel Corp. 82801EB/ER (ICH5/ICH5R) LPC Bridge (rev 02)
	Flags: bus master, medium devsel, latency 0
----------------------------------------------------------------------
lspci -v under 2.6.12

0000:00:00.0 Host bridge: Intel Corp. 82865G/PE/P DRAM Controller/Host-Hub Interface (rev 02)
	Subsystem: ABIT Computer Corp.: Unknown device 101e
	Flags: bus master, fast devsel, latency 0
	Memory at e0000000 (32-bit, prefetchable) [size=128M]
	Capabilities: [e4] #09 [0106]
	Capabilities: [a0] AGP version 3.0

0000:00:01.0 PCI bridge: Intel Corp. 82865G/PE/P PCI to AGP Controller (rev 02) (prog-if 00 [Normal decode])
	Flags: bus master, 66MHz, fast devsel, latency 64
	Bus: primary=00, secondary=01, subordinate=01, sec-latency=32
	I/O behind bridge: 0000a000-0000afff
	Memory behind bridge: f8000000-f9ffffff
	Prefetchable memory behind bridge: e8000000-f7ffffff

0000:00:1e.0 PCI bridge: Intel Corp. 82801 PCI Bridge (rev c2) (prog-if 00 [Normal decode])
	Flags: bus master, fast devsel, latency 0
	Bus: primary=00, secondary=02, subordinate=02, sec-latency=32
	Memory behind bridge: fa000000-fbffffff

0000:00:1f.0 ISA bridge: Intel Corp. 82801EB/ER (ICH5/ICH5R) LPC Bridge (rev 02)
	Flags: bus master, medium devsel, latency 0

0000:01:00.0 VGA compatible controller: ATI Technologies Inc RV280 [Radeon 9200] (rev 01) (prog-if 00 [VGA])
	Subsystem: ATI Technologies Inc: Unknown device 2002
	Flags: bus master, 66MHz, medium devsel, latency 32, IRQ 16
	Memory at e8000000 (32-bit, prefetchable) [size=128M]
	I/O ports at a000 [size=256]
	Memory at f9000000 (32-bit, non-prefetchable) [size=64K]
	Capabilities: [58] AGP version 3.0
	Capabilities: [50] Power Management version 2

0000:01:00.1 Display controller: ATI Technologies Inc RV280 [Radeon 9200] (Secondary) (rev 01)
	Subsystem: ATI Technologies Inc: Unknown device 2003
	Flags: 66MHz, medium devsel
	Memory at f0000000 (32-bit, prefetchable) [disabled] [size=128M]
	Memory at f9010000 (32-bit, non-prefetchable) [disabled] [size=64K]
	Capabilities: [50] Power Management version 2

----------------------------------------------------------------------
lspci -v under 2.6.13:

0000:00:00.0 Host bridge: Intel Corp. 82865G/PE/P DRAM Controller/Host-Hub Interface (rev 02)
	Subsystem: ABIT Computer Corp.: Unknown device 101e
	Flags: bus master, fast devsel, latency 0
	Memory at e0000000 (32-bit, prefetchable) [size=128M]
	Capabilities: [e4] #09 [0106]
	Capabilities: [a0] AGP version 3.0

0000:00:01.0 PCI bridge: Intel Corp. 82865G/PE/P PCI to AGP Controller (rev 02) (prog-if 00 [Normal decode])
	Flags: bus master, 66MHz, fast devsel, latency 64
	Bus: primary=00, secondary=01, subordinate=01, sec-latency=32
	I/O behind bridge: 0000a000-0000afff
	Memory behind bridge: f8000000-f9ffffff
	Prefetchable memory behind bridge: e8000000-f7ffffff

0000:00:1e.0 PCI bridge: Intel Corp. 82801 PCI Bridge (rev c2) (prog-if 00 [Normal decode])
	Flags: bus master, fast devsel, latency 0
	Bus: primary=00, secondary=02, subordinate=02, sec-latency=32
	Memory behind bridge: fa000000-fbffffff
	Prefetchable memory behind bridge: 20000000-200fffff

0000:00:1f.0 ISA bridge: Intel Corp. 82801EB/ER (ICH5/ICH5R) LPC Bridge (rev 02)
	Flags: bus master, medium devsel, latency 0

0000:01:00.0 VGA compatible controller: ATI Technologies Inc RV280 [Radeon 9200] (rev 01) (prog-if 00 [VGA])
	Subsystem: ATI Technologies Inc: Unknown device 2002
	Flags: bus master, 66MHz, medium devsel, latency 32, IRQ 17
	Memory at e8000000 (32-bit, prefetchable) [size=128M]
	I/O ports at a000 [size=256]
	Memory at f9000000 (32-bit, non-prefetchable) [size=64K]
	Expansion ROM at f8000000 [disabled] [size=128K]
	Capabilities: [58] AGP version 3.0
	Capabilities: [50] Power Management version 2

0000:01:00.1 Display controller: ATI Technologies Inc RV280 [Radeon 9200] (Secondary) (rev 01)
	Subsystem: ATI Technologies Inc: Unknown device 2003
	Flags: 66MHz, medium devsel
	Memory at f0000000 (32-bit, prefetchable) [disabled] [size=128M]
	Memory at f9010000 (32-bit, non-prefetchable) [disabled] [size=64K]
	Capabilities: [50] Power Management version 2

----------------------------------------------------------------------

Re: [PROBLEM] mtrr's not set, 2.6.13

[Andrew Morton]

Jim McCloskey <mcclosk@ucsc.edu> wrote:
>
>  I'm not sure who to report this to ....

This works.

>  Somewhere between 2.6.11.3 and 2.6.12 (also under 2.6.13), the
>  following change occurred on this box.
> 
>  Under 2.6.11.3, mtrr ranges are automatically set when X is started:
> 
>  ----------------------------------------------------------------------
>  running 2.6.11.3:
> 
>  cat /proc/mtrr
>  reg00: base=0x00000000 (   0MB), size= 512MB: write-back, count=1
>  reg01: base=0xe8000000 (3712MB), size= 128MB: write-combining, count=2
>  reg02: base=0xe0000000 (3584MB), size= 128MB: write-combining, count=1
>  ----------------------------------------------------------------------
> 
>  After installation of 2.6.12/13, mtrr ranges are not set:
> 
>  ----------------------------------------------------------------------
>  running 2.6.13:
> 
>  Xorg.0.log:
> 
>  (WW) RADEON(0): Failed to set up write-combining range (0xe8000000,0x8000000)
> 
>  /var/log/messages:
> 
>  Aug 30 17:37:13 localhost kernel: mtrr: type mismatch for e8000000,8000000 old: write-back new: write-combining
>  Aug 30 17:37:14 localhost kernel: mtrr: type mismatch for e0000000,8000000 old: write-back new: write-combining
>  Aug 30 17:37:14 localhost kernel: [drm] Initialized radeon 1.16.0 20050311 on minor 0: ATI Technologies Inc RV280 [Radeon 9200]
>  Aug 30 17:37:14 localhost kernel: mtrr: type mismatch for e8000000,8000000 old: write-back new: write-combining
> 
>  cat /proc/mtrr
>  reg00: base=0x00000000 (   0MB), size=983552MB: write-back, count=1
>  ----------------------------------------------------------------------
> 
>  Under 2.6.13 it's fairly easy to force a crash of the X server (e.g. by
>  playing Tuxracer badly).
> 
>  The only change here is in the kernel-version. I haven't tried all the
>  point releases between 2.6.11.3 and 2.6.12, but the relevant Changelogs
>  don't suggest that anything relevant changed.

In a 2.6.13 tree could you please do

wget ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/patches/2.6/2.6.13-rc2/2.6.13-rc2-mm1/broken-out/mtrr-suspend-resume-cleanup.patch
patch -p1 -R < mtrr-suspend-resume-cleanup.patch

and retest?

Thanks.

Re: [PROBLEM] mtrr's not set, 2.6.13

[Jim McCloskey]

* Andrew Morton (akpm@osdl.org) wrote:

  |>  In a 2.6.13 tree could you please do
  |>  
  |>  wget ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/patches/2.6/2.6.13-rc2/2.6.13-rc2-mm1/broken-out/mtrr-suspend-resume-cleanup.patch
  |>  patch -p1 -R < mtrr-suspend-resume-cleanup.patch
  |>  
  |>  and retest?

Thank you very much. Unfortunately, this doesn't seem to have had any
effect, as far as I can see:

----------------------------------------------------------------------
/proc/mtrr:

reg00: base=0x00000000 (   0MB), size=983552MB: write-back, count=1

/var/log/syslog:

Sep 12 10:09:10 localhost kernel: mtrr: type mismatch for
e8000000,8000000 old: write-back new: write-combining

/var/log/Xorg.0.log

(WW) RADEON(0): Failed to set up write-combining range (0xe8000000,0x8000000)
----------------------------------------------------------------------

Jim

Re: [PROBLEM] mtrr's not set, 2.6.13

[Dave Jones]

On Mon, Sep 12, 2005 at 10:24:26AM -0700, Jim McCloskey wrote:

 > /proc/mtrr:
 > 
 > reg00: base=0x00000000 (   0MB), size=983552MB: write-back, count=1

That's an incredibly huge amount of system ram :)
Have you done a BIOS update between the kernel upgrades by any chance ?
Or altered any options in the BIOS ?

Does booting the older kernel definitly still work ?

AFAIR, we don't touch the first MTRR, that's typically set up
by the BIOS before we even boot.

		Dave

Re: [PROBLEM] mtrr's not set, 2.6.13

[Shaohua Li]

On Mon, 2005-09-12 at 10:24 -0700, Jim McCloskey wrote:
> * Andrew Morton (akpm@osdl.org) wrote:
> 
>   |>  In a 2.6.13 tree could you please do
>   |>  
>   |>  wget ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/patches/2.6/2.6.13-rc2/2.6.13-rc2-mm1/broken-out/mtrr-suspend-resume-cleanup.patch
>   |>  patch -p1 -R < mtrr-suspend-resume-cleanup.patch
>   |>  
>   |>  and retest?
> 
> Thank you very much. Unfortunately, this doesn't seem to have had any
> effect, as far as I can see:
> 
> ----------------------------------------------------------------------
> /proc/mtrr:
> 
> reg00: base=0x00000000 (   0MB), size=983552MB: write-back, count=1
The size isn't sane. Looks like the upper bits of the size mask isn't 1.
Can you track it down what's the value of variable 'size_of_mask' in
mtrr/main.c?

Thanks,
Shaohua

Re: [PROBLEM] mtrr's not set, 2.6.13

[Kenneth Parrish]

-=> In 13 Sep 05  15:30:16 article, Dave Jones wrote to All and Jim <=-

[...]

 JM> /proc/mtrr:
 JM>
 JM> reg00: base=0x00000000 (   0MB), size=983552MB: write-back, count=1

 DJ> That's an incredibly huge amount of system ram :)
 DJ> Have you done a BIOS update between the kernel upgrades by any
 chance ?
 DJ> Or altered any options in the BIOS ?

 DJ> Does booting the older kernel definitly still work ?

 DJ> AFAIR, we don't touch the first MTRR, that's typically set up by
 DJ> the BIOS before we even boot.

not sure when this changed, but this computer, a 99 e-machines cyrix m
ii and via mvp3 unit, is showing more than the 512 megas pc 100 ram it
actually has, apparently, in the first line of /proc/mtrr quoted below

 3 Wed Sep 14 03:44:18 0 ~/build/kernel/linux-2.6.14-rc1 $ cat /proc/mtrr
reg00: base=0xfd000000 (4048MB), size=   4MB: write-combining, count=1
reg01: base=0x000c0000 (   0MB), size= 256KB: uncachable, count=1
reg07: base=0x00000000 (   0MB), size= 512MB: write-back, count=1

X shows the video with four MB, lspci shows 16MB -- been that
way for awhile..

01:00.0 VGA compatible controller: ATI Technologies Inc 3D Rage IIC AGP (rev
7a) (prog-if 00 [VGA])
        Subsystem: ATI Technologies Inc Rage IIC AGP
        Control: I/O+ Mem+ BusMaster+ SpecCycle- MemWINV- VGASnoop- ParErr- Ste
SERR- FastB2B-
        Status: Cap+ 66Mhz- UDF- FastB2B+ ParErr- DEVSEL=medium >TAbort- <TAbor
<MAbort- >SERR- <PERR-
        Latency: 64 (2000ns min), cache line size 08
        Region 0: Memory at fd000000 (32-bit, non-prefetchable) [size=16M]
        Region 1: I/O ports at d800 [size=256]
        Region 2: Memory at feaff000 (32-bit, non-prefetchable) [size=4K]
        Expansion ROM at feac0000 [disabled] [size=128K]
        Capabilities: [5c] Power Management version 1
                Flags: PMEClk- DSI- D1- D2- AuxCurrent=0mA PME(D0-,D1-,D2-,D3ho
                Status: D0 PME-Enable- DSel=0 DScale=0 PME-

Linux fret 2.6.14-rc1 #6 Wed Sep 14 01:05:11 CDT 2005 i686 unknown unknown
GNU/Linux

p.s. this 2.6.14-rc1 kernel shows four lines plus one or two
pixels of the next line, at the bottom of each text console, filled with
what looks like earlier buffer.  and it scrolls, with new output. ??

--- MultiMail/Linux v0.46

Re: [PROBLEM] mtrr's not set, 2.6.13

[Jouni Karvo]

hi,

I have the same problem of MTRR:s and kernel 2.6.13.2

With 2.6.9, the system says:

kex@vdr:~$ cat /proc/mtrr
reg00: base=0x00000000 (   0MB), size= 512MB: write-back, count=1
reg01: base=0xd0000000 (3328MB), size= 128MB: write-combining, count=1

but with 2.6.13.2:

kex@vdr:/var/log$ cat /proc/mtrr
reg00: base=0x00000000 (   0MB), size=983552MB: write-back, count=1

Processor type: (I added the mask printouts to void __init
mtrr_bp_init(void) in mtrr/main.c)

eck reporting enabled on CPU#0.
CPU0: Intel P4/Xeon Extended MCE MSRs (12) available
CPU0: Thermal monitoring enabled
mtrr: v2.0 (20020519)
MTRR: size_or_mask = 0xf0000000, size_and_mask = 0xff00000
CPU: Intel(R) Pentium(R) 4 CPU 2.40GHz stepping 03

Please let me know what info would help tracking this down...

yours,
		Jouni

Btw. sorry I did not have the references for the earlier postings.
-- 
http://www.tkk.fi/%7ekex