2.4.19-pre8-ac5 ide & raid0 bugs

2.4.19-pre8-ac5 ide & raid0 bugs

[Nick Evgeniev]

Hi,

I wrote about ide problems with 2.4.19-pre8 a few days ago (it just trashed
filesystem in a couple hours) & I was told to try 2.4.19-pre8-ac5 it was a
little bit better though every 5-8 hours I've got ide errors in log (at
least it didn't crash my reiserfs volumes yet):
>-----------------------------
May 27 14:38:02 vzhik kernel: hdg: status error: status=0x58 { DriveReady
SeekComplete DataRequest }
May 27 14:38:02 vzhik kernel:
May 27 14:38:02 vzhik kernel: hdg: drive not ready for command
May 27 14:38:02 vzhik kernel: hdg: status error: status=0x58 { DriveReady
SeekComplete DataRequest }
May 27 14:38:02 vzhik kernel:
May 27 14:38:02 vzhik kernel: hdg: drive not ready for command
May 27 17:08:05 vzhik kernel: hdg: drive_cmd: status=0xd0 { Busy }
May 27 17:08:05 vzhik kernel:
May 27 17:08:05 vzhik kernel: hdg: status error: status=0x58 { DriveReady
SeekComplete DataRequest }
>-----------------------------
But now I've got even more bugs in log like:
>-----------------------------
May 29 11:28:06 vzhik kernel: raid0_make_request bug: can't convert block
across chunks or bigger than 16k 37713311 4
May 29 11:28:06 vzhik kernel: raid0_make_request bug: can't convert block
across chunks or bigger than 16k 37713343 4
May 29 11:28:06 vzhik kernel: raid0_make_request bug: can't convert block
across chunks or bigger than 16k 37713375 4
May 29 11:28:06 vzhik kernel: raid0_make_request bug: can't convert block
across chunks or bigger than 16k 37713407 2
May 29 11:28:07 vzhik kernel: raid0_make_request bug: can't convert block
across chunks or bigger than 16k 38161563 4
May 29 11:28:07 vzhik kernel: raid0_make_request bug: can't convert block
across chunks or bigger than 16k 38161595 4
May 29 11:28:07 vzhik kernel: raid0_make_request bug: can't convert block
across chunks or bigger than 16k 38161627 4
May 29 11:28:07 vzhik kernel: raid0_make_request bug: can't convert block
across chunks or bigger than 16k 38161659 4
May 29 11:28:07 vzhik kernel: raid0_make_request bug: can't convert block
across chunks or bigger than 16k 37713308 4
>-----------------------------

I don't even think about trying 2.4.19-pre9 since it doesn't has any ide
related issues in its changelist.
The question is -- What I have to try to get WORKING ide driver under
"STABLE" kernel?

Re: 2.4.19-pre8-ac5 ide & raid0 bugs

[Neil Brown]

On Wednesday May 29, nick@octet.com wrote:
> Hi,
> 
> I wrote about ide problems with 2.4.19-pre8 a few days ago (it just trashed
> filesystem in a couple hours) & I was told to try 2.4.19-pre8-ac5 it was a
> little bit better though every 5-8 hours I've got ide errors in log (at
> least it didn't crash my reiserfs volumes yet):
---snip--
> >-----------------------------
> But now I've got even more bugs in log like:
> >-----------------------------
> May 29 11:28:06 vzhik kernel: raid0_make_request bug: can't convert block
> across chunks or bigger than 16k 37713311 4
> May 29 11:28:06 vzhik kernel: raid0_make_request bug: can't convert block
> across chunks or bigger than 16k 37713343 4

This is a request for a 4K block that is not 4K aligned... this
shouldn't happen.
Are you using LVM or something to partition the raid0 array?
... though I seem to recall that LVM always partitions in multiples of
4K so that shouldn't be a problem.

More detais of configuration please.

> The question is -- What I have to try to get WORKING ide driver under
> "STABLE" kernel?

Use SCSI :-?

NeilBrown

Re: 2.4.19-pre8-ac5 ide & raid0 bugs

[Nick Evgeniev]

Hi,

Yes, I'm using lvm 1.1rc1 on top of raid0. Hardware details:
dual piii box on via Apollo pro133 chipset (via694)
two ATA-100 drives (FUJITSU MPG3204AH) attached to on-board  controller
(PROMISE 20265)
& one drive attached to VIA controller. RAID0 is made on top of first two
drives.

1. I don't have any problem with drive attached to via controller (in DMA33
mode)
2. The only problem that I have is with UDMA100 drives attached to promise
controller... Also I have to notice that log messages about ide errors are
related to drive that has no errors (according to smartctl -l, while another
drive has a few records in its log)
3. I had no problems with ide with 2.4.7 kernel for a months...

> > >-----------------------------
> > But now I've got even more bugs in log like:
> > >-----------------------------
> > May 29 11:28:06 vzhik kernel: raid0_make_request bug: can't convert
block
> > across chunks or bigger than 16k 37713311 4
> > May 29 11:28:06 vzhik kernel: raid0_make_request bug: can't convert
block
> > across chunks or bigger than 16k 37713343 4
>
> This is a request for a 4K block that is not 4K aligned... this
> shouldn't happen.
> Are you using LVM or something to partition the raid0 array?
> ... though I seem to recall that LVM always partitions in multiples of
> 4K so that shouldn't be a problem.

Re: 2.4.19-pre8-ac5 ide & raid0 bugs

[Bill Davidsen]

On Wed, 29 May 2002, Nick Evgeniev wrote:

> Hi,
> 
> I wrote about ide problems with 2.4.19-pre8 a few days ago (it just trashed
> filesystem in a couple hours) & I was told to try 2.4.19-pre8-ac5 it was a
> little bit better though every 5-8 hours I've got ide errors in log (at
> least it didn't crash my reiserfs volumes yet):

I see a lot of the 0x58 with taskfile enabled, are you doing that? I even
see it mounting an "IDE" compact flash! I ran out of time to try w/o
taskfile_io.

> >-----------------------------
> May 27 14:38:02 vzhik kernel: hdg: status error: status=0x58 { DriveReady
> SeekComplete DataRequest }
> May 27 14:38:02 vzhik kernel:
> May 27 14:38:02 vzhik kernel: hdg: drive not ready for command
> May 27 14:38:02 vzhik kernel: hdg: status error: status=0x58 { DriveReady
> SeekComplete DataRequest }
> May 27 14:38:02 vzhik kernel:
> May 27 14:38:02 vzhik kernel: hdg: drive not ready for command
> May 27 17:08:05 vzhik kernel: hdg: drive_cmd: status=0xd0 { Busy }
> May 27 17:08:05 vzhik kernel:
> May 27 17:08:05 vzhik kernel: hdg: status error: status=0x58 { DriveReady
> SeekComplete DataRequest }
> >-----------------------------
> But now I've got even more bugs in log like:
> >-----------------------------
> May 29 11:28:06 vzhik kernel: raid0_make_request bug: can't convert block
> across chunks or bigger than 16k 37713311 4
> May 29 11:28:06 vzhik kernel: raid0_make_request bug: can't convert block
> across chunks or bigger than 16k 37713343 4
> May 29 11:28:06 vzhik kernel: raid0_make_request bug: can't convert block
> across chunks or bigger than 16k 37713375 4
> May 29 11:28:06 vzhik kernel: raid0_make_request bug: can't convert block
> across chunks or bigger than 16k 37713407 2
> May 29 11:28:07 vzhik kernel: raid0_make_request bug: can't convert block
> across chunks or bigger than 16k 38161563 4
> May 29 11:28:07 vzhik kernel: raid0_make_request bug: can't convert block
> across chunks or bigger than 16k 38161595 4
> May 29 11:28:07 vzhik kernel: raid0_make_request bug: can't convert block
> across chunks or bigger than 16k 38161627 4
> May 29 11:28:07 vzhik kernel: raid0_make_request bug: can't convert block
> across chunks or bigger than 16k 38161659 4
> May 29 11:28:07 vzhik kernel: raid0_make_request bug: can't convert block
> across chunks or bigger than 16k 37713308 4
> >-----------------------------
> 
> I don't even think about trying 2.4.19-pre9 since it doesn't has any ide
> related issues in its changelist.
> The question is -- What I have to try to get WORKING ide driver under
> "STABLE" kernel?
> 
> 
> -
> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
> Please read the FAQ at  http://www.tux.org/lkml/
> 

-- 
bill davidsen <davidsen@tmr.com>
  CTO, TMR Associates, Inc
Doing interesting things with little computers since 1979.

Re: 2.4.19-pre8-ac5 ide & raid0 bugs

[Nick Evgeniev]

Hi,

No, I don't have task_file enabled.

> > I wrote about ide problems with 2.4.19-pre8 a few days ago (it just
trashed
> > filesystem in a couple hours) & I was told to try 2.4.19-pre8-ac5 it was
a
> > little bit better though every 5-8 hours I've got ide errors in log (at
> > least it didn't crash my reiserfs volumes yet):
>
> I see a lot of the 0x58 with taskfile enabled, are you doing that? I even
> see it mounting an "IDE" compact flash! I ran out of time to try w/o
> taskfile_io.

Re: 2.4.19-pre8-ac5 ide & raid0 bugs

[Andre Hedrick]

Because there is an entire set of new calls to deal with cfa or flash.
It really takes a new subdriver.

On Tue, 4 Jun 2002, Bill Davidsen wrote:

> On Wed, 29 May 2002, Nick Evgeniev wrote:
> 
> > Hi,
> > 
> > I wrote about ide problems with 2.4.19-pre8 a few days ago (it just trashed
> > filesystem in a couple hours) & I was told to try 2.4.19-pre8-ac5 it was a
> > little bit better though every 5-8 hours I've got ide errors in log (at
> > least it didn't crash my reiserfs volumes yet):
> 
> I see a lot of the 0x58 with taskfile enabled, are you doing that? I even
> see it mounting an "IDE" compact flash! I ran out of time to try w/o
> taskfile_io.
> 
> > >-----------------------------
> > May 27 14:38:02 vzhik kernel: hdg: status error: status=0x58 { DriveReady
> > SeekComplete DataRequest }
> > May 27 14:38:02 vzhik kernel:
> > May 27 14:38:02 vzhik kernel: hdg: drive not ready for command
> > May 27 14:38:02 vzhik kernel: hdg: status error: status=0x58 { DriveReady
> > SeekComplete DataRequest }
> > May 27 14:38:02 vzhik kernel:
> > May 27 14:38:02 vzhik kernel: hdg: drive not ready for command
> > May 27 17:08:05 vzhik kernel: hdg: drive_cmd: status=0xd0 { Busy }
> > May 27 17:08:05 vzhik kernel:
> > May 27 17:08:05 vzhik kernel: hdg: status error: status=0x58 { DriveReady
> > SeekComplete DataRequest }
> > >-----------------------------
> > But now I've got even more bugs in log like:
> > >-----------------------------
> > May 29 11:28:06 vzhik kernel: raid0_make_request bug: can't convert block
> > across chunks or bigger than 16k 37713311 4
> > May 29 11:28:06 vzhik kernel: raid0_make_request bug: can't convert block
> > across chunks or bigger than 16k 37713343 4
> > May 29 11:28:06 vzhik kernel: raid0_make_request bug: can't convert block
> > across chunks or bigger than 16k 37713375 4
> > May 29 11:28:06 vzhik kernel: raid0_make_request bug: can't convert block
> > across chunks or bigger than 16k 37713407 2
> > May 29 11:28:07 vzhik kernel: raid0_make_request bug: can't convert block
> > across chunks or bigger than 16k 38161563 4
> > May 29 11:28:07 vzhik kernel: raid0_make_request bug: can't convert block
> > across chunks or bigger than 16k 38161595 4
> > May 29 11:28:07 vzhik kernel: raid0_make_request bug: can't convert block
> > across chunks or bigger than 16k 38161627 4
> > May 29 11:28:07 vzhik kernel: raid0_make_request bug: can't convert block
> > across chunks or bigger than 16k 38161659 4
> > May 29 11:28:07 vzhik kernel: raid0_make_request bug: can't convert block
> > across chunks or bigger than 16k 37713308 4
> > >-----------------------------
> > 
> > I don't even think about trying 2.4.19-pre9 since it doesn't has any ide
> > related issues in its changelist.
> > The question is -- What I have to try to get WORKING ide driver under
> > "STABLE" kernel?
> > 
> > 
> > -
> > To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> > the body of a message to majordomo@vger.kernel.org
> > More majordomo info at  http://vger.kernel.org/majordomo-info.html
> > Please read the FAQ at  http://www.tux.org/lkml/
> > 
> 
> -- 
> bill davidsen <davidsen@tmr.com>
>   CTO, TMR Associates, Inc
> Doing interesting things with little computers since 1979.
> 
> -
> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
> Please read the FAQ at  http://www.tux.org/lkml/
> 

Andre Hedrick
LAD Storage Consulting Group

Re: 2.4.19-pre8-ac5 ide & raid0 bugs

[Bill Davidsen]

On Sat, 8 Jun 2002, Andre Hedrick wrote:

> 
> Because there is an entire set of new calls to deal with cfa or flash.
> It really takes a new subdriver.

Could you clarifiy that to either (a) "the kernel is now broken
completely for CF," (b) "use the non-kernel 3.1.33 pcmcia," or (c) "use
XXX." It's not clear to me if you are stating that CF no longer works at
all, not longer works with taskfile, no longer works but pcmcia separate
modules do, or that I need some other software "new subdriver" not in the
kernel or pcmcia package.

I'm not matching any of those to uni laptops working with kernel plus OLD
pcmcia support for pcmcia in kernel, and SMP not working with any
combination of recent kernel (plain, -jam, -ac, -aa, with patches, etc)
and any pcmcia.
 
> On Tue, 4 Jun 2002, Bill Davidsen wrote:
> 
> > On Wed, 29 May 2002, Nick Evgeniev wrote:

> > > I wrote about ide problems with 2.4.19-pre8 a few days ago (it just trashed
> > > filesystem in a couple hours) & I was told to try 2.4.19-pre8-ac5 it was a
> > > little bit better though every 5-8 hours I've got ide errors in log (at
> > > least it didn't crash my reiserfs volumes yet):
> > 
> > I see a lot of the 0x58 with taskfile enabled, are you doing that? I even
> > see it mounting an "IDE" compact flash! I ran out of time to try w/o
> > taskfile_io.

-- 
bill davidsen <davidsen@tmr.com>
  CTO, TMR Associates, Inc
Doing interesting things with little computers since 1979.

Re: 2.4.19-pre8-ac5 ide & raid0 bugs

[Andre Hedrick]

Basically the driver has never properly supported CFA/Flash.
There are translation calls to decode various events.
It is kind of like ide-scsi or ide-cd for DVD-RAM, both will work but
clearly it is best operate in native interfaces.

In short I need to get around to creating an ide-cfa.c subdriver.

On Mon, 10 Jun 2002, Bill Davidsen wrote:

> On Sat, 8 Jun 2002, Andre Hedrick wrote:
> 
> > 
> > Because there is an entire set of new calls to deal with cfa or flash.
> > It really takes a new subdriver.
> 
> Could you clarifiy that to either (a) "the kernel is now broken
> completely for CF," (b) "use the non-kernel 3.1.33 pcmcia," or (c) "use
> XXX." It's not clear to me if you are stating that CF no longer works at
> all, not longer works with taskfile, no longer works but pcmcia separate
> modules do, or that I need some other software "new subdriver" not in the
> kernel or pcmcia package.
> 
> I'm not matching any of those to uni laptops working with kernel plus OLD
> pcmcia support for pcmcia in kernel, and SMP not working with any
> combination of recent kernel (plain, -jam, -ac, -aa, with patches, etc)
> and any pcmcia.
>  
> > On Tue, 4 Jun 2002, Bill Davidsen wrote:
> > 
> > > On Wed, 29 May 2002, Nick Evgeniev wrote:
> 
> > > > I wrote about ide problems with 2.4.19-pre8 a few days ago (it just trashed
> > > > filesystem in a couple hours) & I was told to try 2.4.19-pre8-ac5 it was a
> > > > little bit better though every 5-8 hours I've got ide errors in log (at
> > > > least it didn't crash my reiserfs volumes yet):
> > > 
> > > I see a lot of the 0x58 with taskfile enabled, are you doing that? I even
> > > see it mounting an "IDE" compact flash! I ran out of time to try w/o
> > > taskfile_io.
> 
> -- 
> bill davidsen <davidsen@tmr.com>
>   CTO, TMR Associates, Inc
> Doing interesting things with little computers since 1979.
> 
> 

Andre Hedrick
LAD Storage Consulting Group