Re: [gentoo-server] pdc 20267 and 2.6.x = corruption

[Chris Smart]

hmmm thanks for your info.

I could always stick with 2.4 I guess, but the way it stands my external 
firewire backup drive only works properly under 2.6, so it's a catch 22 ;)

I'll see what else comes up.

ta,
-c

Matt Hallmark wrote:

> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
>
> Well I was having that problem w/ an older ATA100 only promise card in
> 2.4 too.  I thought it due to irq conflicts w/ the other Highpoint
> 372/374 onboard ide ports, the mobo ports and the scsi hba all moving
> data at the same time.  This was back in 2.4.20-22ac days.
>
> My problems were limited to trucking data across hba's.  A copy from
> disk to disk on the same controller would succeed, or even a single copy
> across controllers would be slow, w/ regular pauses for whatever reason,
> but stable.  When I ran more than one copy I'd get freezing, dma
> timeouts, controller reseets, dma turned off etc.  You obviously know
> that drill. :)
>
> FYI, my 20269 is in the same mobo as the earlier card, an Abit w/ i845PE
> chipset.
>
> Matt
>
> Chris Smart wrote:
> | Thanks Matt,
> |
> | You would be using the "new" promise driver then, as it starts at
> | chipset 20268 and upwards.
> |
> | Perhaps it's just the "old" driver that is borked? or perhaps it's my
> | new seagate hdd (which I doubt because 2.4 kernel works a treat).
> |
> | I can't find anyone else with this problem though.. so maybe I'm just
> | crazy.
> |
> | I can copy files for a while or start a kernel compile then "bang" -
> the
> | kernel can't talk to the drive.
> |
> | Cheers
> | -c
> |
> | Matt Hallmark wrote:
> |
> | Well, I'm running:
> | Linux hax 2.6.3-gentoo-r2 #3 Wed May 5 02:22:32 PDT 2004 i686 Intel(R)
> | Pentium(R) 4 CPU 2.40GHz GenuineIntel GNU/Linux
> |
> | WIth a Promise:
> | 20269:0000:02:04.0 Unknown mass storage controller: Promise Technology,
> | Inc. 20269 (rev 02)
> |
> | I've got 2 Maxtor 7200rpm 120's as primary on each Promise channel with
> | these results:
> |
> | 2mb cache:
> | hax root # hdparm -tT /dev/hdf
> |
> | /dev/hdf:
> | ~ Timing buffer-cache reads:   1412 MB in  2.00 seconds = 705.40 MB/sec
> | ~ Timing buffered disk reads:  142 MB in  3.02 seconds =  47.01 MB/sec
> |
> | 8mb cache:
> | hax root # hdparm -tT /dev/hdg
> |
> | /dev/hdg:
> | ~ Timing buffer-cache reads:   1372 MB in  2.00 seconds = 686.10 MB/sec
> | ~ Timing buffered disk reads:  168 MB in  3.03 seconds =  55.53 MB/sec
> |
> | I've also got 4 other ide disks on the mobo controllers and a u160 hba
> | and single 15K boot drive.
> |
> | No ide tweaks were made, this system is less than a week old atm.
> |
> | Matt Hallmark
> |
> | Chris Smart wrote:
> | | sorry.. I meant "..with promise ultra100 controller (namely 20267
> | | chipset) and *2.6.x* kernels?"
> | |
> | | that's 2.6, not 2.5.. thanks
> | |
> | | Chris Smart wrote:
> | |
> | |> Howdy,
> | |> Anyone had any problems (or know of any fixes) with promise ultra100
> | |> controller (namely 20267 chipset) and 2.5.x kernels?
> | |>
> | |> 2.4.x seems rock stable, and fast (55mb/sec)
> | |>
> | |> 2.6.x is slow (30mb/sec) and causes hdd corruptions (using "old"
> | |> promise driver for up to 20267).
> | |>
> | |> I've searched online but can't find any answers, so I turn to my
> | |> trusty friends on the gentoo-server list.
> | |>
> | |> Cheers,
> | |>
> | |> Chris
> | |
> | |
> -----BEGIN PGP SIGNATURE-----
> Version: GnuPG v1.2.4 (GNU/Linux)
> Comment: Using GnuPG with Thunderbird - http://enigmail.mozdev.org
>
> iD8DBQFAnuo3aVzErf2usf4RAuaDAKDOG8+pWCnfNKWygckQo59/leC83ACg9EgU
> qkvp5Se9lQySjVRbUgkYlqw=
> =RU1A
> -----END PGP SIGNATURE-----