[gentoo-user] Re: [gentoo-security] AMD64 + Hard Drive weirdness...

[gentoo-user] Re: [gentoo-security] AMD64 + Hard Drive weirdness...

[Jeff]

Moving my thread over to the proper list... first...

Now then - thanks to everyone on the list for your help. I've had barely
any sleep lately, so I must apologize first, for putting the original
thread onto the security mailing list by mistake.

For anyone who's wondering - I have an AMD64 box, with a new Gentoo
AMD64 install. The hard drive read times are obnoxiously slow - so, I'm
going to attribute this to the wrong driver being loaded for the controller.

See here:

hdparm -tT /dev/hda

/dev/hda:
 Timing cached reads:   3016 MB in  2.00 seconds = 1507.91 MB/sec
 Timing buffered disk reads:    4 MB in  3.68 seconds =   1.09 MB/sec

Horribly slow! This machine should be blazing fast, with the 7200 rpm
200 GB hard drive, AMD64 3500+ processor, 1.5 MB RAM, and very modern
motherboard to compliment.

So, in the meantime, I'm trying to track down the culprit that's making
my drive run so slow.

Here's my lspci:

00:00.0 Host bridge: ATI Technologies Inc RS480 Host Bridge (rev 10)
00:02.0 PCI bridge: ATI Technologies Inc RS480 PCI-X Root Port
00:12.0 IDE interface: ATI Technologies Inc ATI 4379 Serial ATA Controller
00:13.0 USB Controller: ATI Technologies Inc IXP SB400 USB Host Controller
00:13.1 USB Controller: ATI Technologies Inc IXP SB400 USB Host Controller
00:13.2 USB Controller: ATI Technologies Inc IXP SB400 USB2 Host Controller
00:14.0 SMBus: ATI Technologies Inc IXP SB400 SMBus Controller (rev 11)
00:14.1 IDE interface: ATI Technologies Inc Standard Dual Channel PCI
IDE Controller ATI
00:14.3 ISA bridge: ATI Technologies Inc IXP SB400 PCI-ISA Bridge
00:14.4 PCI bridge: ATI Technologies Inc IXP SB400 PCI-PCI Bridge
00:18.0 Host bridge: Advanced Micro Devices [AMD] K8 [Athlon64/Opteron]
HyperTransport Technology Configuration
00:18.1 Host bridge: Advanced Micro Devices [AMD] K8 [Athlon64/Opteron]
Address Map
00:18.2 Host bridge: Advanced Micro Devices [AMD] K8 [Athlon64/Opteron]
DRAM Controller
00:18.3 Host bridge: Advanced Micro Devices [AMD] K8 [Athlon64/Opteron]
Miscellaneous Control
01:00.0 VGA compatible controller: nVidia Corporation NV41.0 (rev a2)
02:05.0 FireWire (IEEE 1394): VIA Technologies, Inc. IEEE 1394 Host
Controller (rev 80)
02:09.0 Ethernet controller: 3Com Corporation 3c905C-TX/TX-M [Tornado]
(rev 78)
02:0a.0 Multimedia audio controller: Creative Labs SB0400 Audigy2 Value

And modules:

Module                  Size  Used by
nvidia               4057916  12
snd_pcm_oss            56224  0
snd_mixer_oss          19392  1 snd_pcm_oss
eth1394                22608  0
snd_emu10k1           122180  1
snd_rawmidi            30112  1 snd_emu10k1
snd_seq_device         10576  2 snd_emu10k1,snd_rawmidi
snd_ac97_codec        108120  1 snd_emu10k1
snd_pcm               100936  3 snd_pcm_oss,snd_emu10k1,snd_ac97_codec
snd_timer              27336  2 snd_emu10k1,snd_pcm
snd_ac97_bus            3392  1 snd_ac97_codec
snd_page_alloc         12560  2 snd_emu10k1,snd_pcm
snd_util_mem            6016  1 snd_emu10k1
snd_hwdep              11936  1 snd_emu10k1
3c59x                  50420  0
mii                     7040  1 3c59x
ata_piix               12548  0
sata_vsc                9988  0
sata_sis                9796  0
sata_sx4               15812  0
sata_nv                11652  0
sata_via               10436  0
sata_svw                9540  0
sata_sil               11588  0
sata_promise           14148  0
libata                 65296  9
ata_piix,sata_vsc,sata_sis,sata_sx4,sata_nv,sata_via,sata_svw,sata_sil,sata_promise
sbp2                   27076  0
ohci1394               35532  0
ieee1394              109752  3 eth1394,sbp2,ohci1394
ohci_hcd               22340  0
uhci_hcd               34848  0
usb_storage            71360  0
usbhid                 41056  0
ehci_hcd               35336  0

I'll be spending the rest of the day trying to figure out what's going
on here. Of course, if anyone has some insight, that would ultimately be
most helpful! Off to work I go...

Thanks all on the list(s).

Robert Larson wrote:
> Hello Jeff,
> 
> 
> I've had 3 machines exhibit this kind of behaviour in the last few months.
> 
> On the first machine, it was an intermitten IDE controller failure (probably 
> related to heat and expansion of motherboard compoenents).  I was able to 
> bypass it by installing a PCI SATA controller.  The way that I was able to 
> figure this out was by running knoppix on it (I tried windows too, just in 
> case).  When running knoppix (and, that OTHER os), the problems still 
> occured.
> 
> The second and third machines were having problems because the wrong drivers 
> were loaded for the motherboard IDE controller.  On the first of these 
> machines, I ran knoppix and it correctly loaded the drivers (I used lsmod to 
> find them ;-).  On the second of these machines, it was a production machine, 
> and it took a lot of time because I couldn't just bring it down.  I was 
> getting "operation not permitted" when trying to enable DMA.  Eventually, I 
> had performed lspci, and saw the controller, then noticed that it was 
> compiled into the kernel as a different controller.
> 
> As far as it goes for your situation, I would recommend running knoppix to see 
> if the autodetection can resolve it.  If that doesn't work, it may be that 
> it's simply getting confused between those two similar controllers.  Does 
> "hdparm /dev/hda" show any useful info?  How about "hdparm -i /dev/hda"?  If 
> you try to make settings (such as set DMA "hdparm -d1 /dev/hda") does it spit 
> out errors?  I think hdparm may tell you more in this situation because the 
> disc reads are insanely slow (1MB/sec should be more like 50MB/sec).  It 
> might be worth walking through this just to see if it gives you any errors:
> http://gentoo-wiki.com/HOWTO_Use_hdparm_to_improve_IDE_device_performance
> 
> I hope this gives you enough to go on...
> 
> 
> Regards,
> 
> 
> Robert Larson
> 
> 
> 
> On Thursday 16 February 2006 08:58 am, Jeff wrote:
> 
>>Hey all.
>>
>>Gentoo Linux AMD64 - running pretty sweet - except, I've noticed that
>>even under minimal loads, my system seems to have mini-lockups
>>frequently. For instance, when downloading a mere 10 emails, my system
>>seems to choke - to the point where I can't even move my mouse for a
>>good 10 seconds or so. Opening a gnome-terminal is painful - sometimes,
>>apps make the desktop freeze for a good 20-30 seconds. I did a large
>>emerge last night using the gnome-terminal, and it rendered my Gnome
>>desktop almost completely frozen.
>>
>>Any idea what might be causing this, and what steps I should take to
>>troubleshoot this? The best I can tell, the problem seems to be hard
>>drive related, as it does a lot of chewing before the app finally let's
>>rip. The hard drive is an IDE, not SATA as you might expect from the
>>info below. Other than these strange mini-lockups, my system is buzzing
>>right along at a good clip!
>>
>>Thanks!
>>
> 
> [snip]

-- 
Jabba the Hutt:
	This bounty hunter is my kind of scum: fearless and
	inventive.

-- 
gentoo-user@gentoo.org mailing list

Re: [gentoo-user] Re: [gentoo-security] AMD64 + Hard Drive weirdness...

[darren kirby]

[-- Attachment #1: Type: text/plain, Size: 794 bytes --]

quoth the Jeff:
>
> See here:
>
> hdparm -tT /dev/hda
>
> /dev/hda:
>  Timing cached reads:   3016 MB in  2.00 seconds = 1507.91 MB/sec
>  Timing buffered disk reads:    4 MB in  3.68 seconds =   1.09 MB/sec
>
> Horribly slow! This machine should be blazing fast, with the 7200 rpm
> 200 GB hard drive, AMD64 3500+ processor, 1.5 MB RAM, and very modern
> motherboard to compliment.
>
Can we see output of "hdparm -i /dev/hda".  I am trying to saee if you have 
DMA enabled. Your timings are horribly slow.

Do you have  CONFIG_BLK_DEV_ATIIXP enabled in your kernel?

-d
-- 
darren kirby :: Part of the problem since 1976 :: http://badcomputer.org
"...the number of UNIX installations has grown to 10, with more expected..."
- Dennis Ritchie and Ken Thompson, June 1972

[-- Attachment #2: Type: application/pgp-signature, Size: 191 bytes --]

Re: [gentoo-user] Re: [gentoo-security] AMD64 + Hard Drive weirdness...

[Hemmann, Volker Armin]

On Thursday 16 February 2006 17:32, Jeff wrote:
> Moving my thread over to the proper list... first...
>
> Now then - thanks to everyone on the list for your help. I've had barely
> any sleep lately, so I must apologize first, for putting the original
> thread onto the security mailing list by mistake.
>
> For anyone who's wondering - I have an AMD64 box, with a new Gentoo
> AMD64 install. The hard drive read times are obnoxiously slow - so, I'm
> going to attribute this to the wrong driver being loaded for the
> controller.
>
> See here:
>
> hdparm -tT /dev/hda

do a hdparm /dev/hda followed by hdparm -iI /dev/hda and check, if dma is 
turned on and if yes, which mode.

Oh, and I hope you compiled the correct driver for your chipset into the 
kernel?
-- 
gentoo-user@gentoo.org mailing list

[gentoo-user] Re: [gentoo-security] AMD64 + Hard Drive weirdness... [SOLVED]

[Jeff]

Ok, so with some quick tips from the nice peoples here in the list, I
was almost instantly able to track down the culprit to my hard draive
whackiness. It's my ATI IDE controller:

00:14.1 IDE interface: ATI Technologies Inc Standard Dual Channel PCI
IDE Controller ATI

In the kernel, I found:

Device Drivers  --->
  ATA/ATAPI/MFM/RLL support  --->
    <*>         ATI IXP chipset IDE support

I built this into my kernel, and now I get:

hdparm -tT /dev/hda

/dev/hda:
 Timing cached reads:   2860 MB in  2.00 seconds = 1429.91 MB/sec
 Timing buffered disk reads:  174 MB in  3.02 seconds =  57.54 MB/sec

That looks a HECK of a lot better!

I tried turning on DMA, but that didn't seem to make a difference.

Thanks to all for your help!

Igor V. Rafienko wrote:
> on Feb 16, 2006, 09:58, Jeff wrote:
> 
>> Gentoo Linux AMD64 - running pretty sweet - except, I've noticed that
>> even under minimal loads, my system seems to have mini-lockups
>> frequently.
> 
> 
> 
> That has *never* happened to me. Last lockup was due to running tvtime
> and bzflag at the same time (some xorg driver crap).
> 
> 
>> I did a large emerge last night using the gnome-terminal, and it
>> rendered my Gnome desktop almost completely frozen.
> 
> 
> 
> That vaguely suggests DMA issues.
> 
> 
>> hdparm -tT /dev/hda
>>
>> /dev/hda:
>> Timing cached reads:   2992 MB in  2.00 seconds = 1495.91 MB/sec
>> Timing buffered disk reads:    4 MB in  3.69 seconds =   1.08 MB/sec
> 
> 
> 
> This is unacceptably low. A relatively recent hdd should perform at
> least in the 30-40MBps range. Did you turn DMA on? (hdparm -d1 /dev/hda)
> 
> 
> 
> 
> 
> ivr

-- 
Han Solo:
	Well Princess, it looks like you managed to keep me
	here a while longer.
Princess Leia:
	I had nothing to do with it. General Rieekan thinks
	it's dangerous for anyone to leave the system until
	they've activated the energy shield.
Han Solo:
	That's a good story. I think you just can't bear to
	let a gorgeous guy like me out of your sight.
Princess Leia:
	I don't know where you get you delusions, laser brain!
Chewbacca laughs
Han Solo:
	Laugh it up, fuzzball!

-- 
gentoo-user@gentoo.org mailing list