Hi all,
we're trying to get a Cool Road Runner board by Lippert (see
http://www.emjembedded.com/products/single/coolroadr.html)
to run under Linux (SuSE 6.4, kernel 2.2.14).
The CompactFlash disk (a 32 MB SanDisk) is recognized as /dev/hda,
but the system fails to see the /dev/hdb disk (an IBM DARA-206000
jumpered as slave). When the IDE driver loads, it displays
hda:pio, hdb:DMA - and yes, the BIOS assigns UDMA33 to the slave drive
while the master is detected as Mode1.
The IDE controller is a CS5530.
Is there a chance that a newer kernel will detect the second disk?
If I disconnect the slave drive, I can see "hdb:pio" :-((( but not
the drive, of course B-)
Any ideas?
Steffen
--
Steffen Grunewald | GFZ | PB 2.2 | Telegrafenberg E3 | D-14473 Potsdam
? email: steffen(at)gfz-potsdam.de | fax/fon: +49-331-288-1266/-1245 ?
It has just been discovered that research causes cancer in rats.
On Fri, Mar 30, 2001 at 05:59:00PM +0200, Steffen Grunewald wrote:
> Hi all,
>
> we're trying to get a Cool Road Runner board by Lippert (see
> http://www.emjembedded.com/products/single/coolroadr.html)
> to run under Linux (SuSE 6.4, kernel 2.2.14).
>
well looking at that page, there's a linux link to
http://www.whitedwarflinux.org which looks like a distro specific to
this board... you may want to try this out before continuing on as it is
considered an embedded system and they appear to have their own open
source patches for 2.2.14
--
.oO Gnea [gnea at rochester dot rr dot com] Oo.
.oO url: http://garson.org/~gnea Oo.
"You can tune a filesystem, but you can't tuna fish." -unknown
On Fri 2001-03-30 (12:14), Gnea wrote:
> > http://www.emjembedded.com/products/single/coolroadr.html
>
> well looking at that page, there's a linux link to
> http://www.whitedwarflinux.org which looks like a distro specific to
> this board...
Did look there, but they seem to have stopped at 2.2.14, too.
Sent a mail to them,. let's see...
Thanks, and enjoy the weekend
Steffen
--
Steffen Grunewald | GFZ | PB 2.2 | Telegrafenberg E3 | D-14473 Potsdam
? email: steffen(at)gfz-potsdam.de | fax/fon: +49-331-288-1266/-1245 ?
Swim nude. Sharks hate to peel their food.
Steffen Gruenwald writes:
> The CompactFlash disk (a 32 MB SanDisk) is recognized as /dev/hda,
> but the system fails to see the /dev/hdb disk (an IBM DARA-206000
> jumpered as slave). When the IDE driver loads, it displays
> hda:pio, hdb:DMA - and yes, the BIOS assigns UDMA33 to the slave drive
> while the master is detected as Mode1.
> The IDE controller is a CS5530.
This was just discussed this week by Andre Hedrick. You need to add a
mount option like "hdb=flash" (I wasn't paying much attention). This
is because CF disks do not properly handle detection of slaves. See:
http://marc.theaimsgroup.com/?l=linux-kernel&m=98580536318380&w=4
Cheers, Andreas
--
Andreas Dilger \ "If a man ate a pound of pasta and a pound of antipasto,
\ would they cancel out, leaving him still hungry?"
http://www-mddsp.enel.ucalgary.ca/People/adilger/ -- Dogbert
On Fri 2001-03-30 (10:52), Andreas Dilger wrote:
> Steffen Gruenwald writes:
> > The CompactFlash disk (a 32 MB SanDisk) is recognized as /dev/hda,
> > but the system fails to see the /dev/hdb disk (an IBM DARA-206000
> > jumpered as slave). When the IDE driver loads, it displays
> > hda:pio, hdb:DMA - and yes, the BIOS assigns UDMA33 to the slave drive
> > while the master is detected as Mode1.
> > The IDE controller is a CS5530.
>
> This was just discussed this week by Andre Hedrick. You need to add a
> mount option like "hdb=flash" (I wasn't paying much attention). This
> is because CF disks do not properly handle detection of slaves. See:
>
> http://marc.theaimsgroup.com/?l=linux-kernel&m=98580536318380&w=4
OK, that makes things clearer. In addition to that, I also found that
2.2.15 fixes some things wrt the CS5530 controller chip. So I'll try both
things: a newer kernel (2.2.16 is in the SuSE upgrades) _and_ the append
parameter.
And BTW: Isn't Andre's writing style very lovely ?
Steffen, starting with fresh hope into a new week
--
Steffen Grunewald | GFZ | PB 2.2 | Telegrafenberg E3 | D-14473 Potsdam
? email: steffen(at)gfz-potsdam.de | fax/fon: +49-331-288-1266/-1245 ?
If at first you don't succeed, then you're in good company. - Lincoln
On Mon, 2 Apr 2001, Steffen Grunewald wrote:
> And BTW: Isn't Andre's writing style very lovely ?
DGD, and who made you a grammer teacher of english on LKML?
Never claimed to be an "English Major" soooooo WTF is the price of eggs in
Germany today? This is no different to trying an orthoginal rotation to
one's person.
> Steffen, starting with fresh hope into a new week
Later..
Andre Hedrick
Linux ATA Development
On Sun 2001-04-01 (23:55), Andre Hedrick wrote:
> On Mon, 2 Apr 2001, Steffen Grunewald wrote:
>
> > And BTW: Isn't Andre's writing style very lovely ?
>
> DGD, and who made you a grammer teacher of english on LKML?
Hey cool down dude.
That's exactly the attitude I was speaking about. Not your or my grammar.
And WRT your kernel code - that's no matter of English majors, but of
the compiler and the users (which seems to work fine BTW, no pun intended)
> Never claimed to be an "English Major" soooooo WTF is the price of eggs in
> Germany today?
I think 1,39 for 10 would be a bargain. But who cares?
Steffen, who prefers chocolate
--
Steffen Grunewald | GFZ | PB 2.2 | Telegrafenberg E3 | D-14473 Potsdam
? email: steffen(at)gfz-potsdam.de | fax/fon: +49-331-288-1266/-1245 ?
Hi! I'm a .signature virus. Copy me to your sig and help me spread!
On Mon, 2 Apr 2001, Steffen Grunewald wrote:
> On Sun 2001-04-01 (23:55), Andre Hedrick wrote:
> > On Mon, 2 Apr 2001, Steffen Grunewald wrote:
> >
> > > And BTW: Isn't Andre's writing style very lovely ?
> >
> > DGD, and who made you a grammer teacher of english on LKML?
>
> Hey cool down dude.
I have been and I am actively try, but it is hard when I try to be nice
and get hammered.
> That's exactly the attitude I was speaking about. Not your or my grammar.
No you got the attitude line because of the direct personal slide against
me "And BTW: Isn't Andre's writing style very lovely ?"
> And WRT your kernel code - that's no matter of English majors, but of
> the compiler and the users (which seems to work fine BTW, no pun intended)
Also, several messages earlier I pointed out that I had not documented the
feature because it was only attempted once, and only with 2 CFA's in a
bazar ata-bridge.
I then explained why the detection was failing and pointed where to verify.
After 3-5 attempts and I can not get the point across because the other
party keeps going off in different directions to do "what about this",
I finally pointed out facts that distrub people, and gave up on trying to
show/present/give the answer and offered to then enforce their beliefs of
reality.
So I state a few facts very pointed to get their attention again and that
is additude??
> > Never claimed to be an "English Major" soooooo WTF is the price of eggs in
> > Germany today?
>
> I think 1,39 for 10 would be a bargain. But who cares?
Exactly, however....
In the greater context of the your first message, what was your point, but
to inflame me with "And BTW: Isn't Andre's writing style very lovely ?".
If you and never made that comment, you would not have pissed me off.
Personal commentary is not desired.
Everyone know that I am not good at "documentation", and to slap/slam me
with mockery makes it all the less likely I'll consider trying again.
Regards,
Andre Hedrick
Linux ATA Development
OK can we just have a technical discussion?
Andre Hedrick wrote:
> On Mon, 2 Apr 2001, Steffen Grunewald wrote:
>
[hostilities snipped]
> Also, several messages earlier I pointed out that I had not documented the
> feature because it was only attempted once, and only with 2 CFA's in a
> bazar ata-bridge.
1. All compact flash have an inbuilt ATA controller. I.E. they can be
used exactly like a harddisk, directly off the IDE controller of a
motherboard.
I.E. no need for PCMCIA or any of that. I understood from your
responses
that you didn't realise this?
2. Compact Flash in this application (I.E. solid state hard disk) is
getting very
popular as prices are tumbling.
3. Having a config parameter (uneeded kludge in my opinion), like hdx=flash
even if hdx is not a compact flash is confusing. Can we call it
hdx=probe
which fits nicely with the noprobe option.
> I then explained why the detection was failing and pointed where to verify.
No you didn't. You mentioned a 30 second timeout, but not why it
was caused. Have you seen this yourself or can you point us at who
reported this to you?
> After 3-5 attempts and I can not get the point across because the other
> party keeps going off in different directions to do "what about this",
Emm, I think *you* were going off describing your application with
a "bazar ata-bridge", not the simple use of a compact flash as a
hard disk.
> I finally pointed out facts that distrub people, and gave up on trying to
> show/present/give the answer and offered to then enforce their beliefs of
> reality.
>
> So I state a few facts very pointed to get their attention again and that
> is additude??
Actually I thought the final email was a little more concise/informative, thanks.
[more hostility snipped]
Padraig.
Andre Hedrick writes:
> On Mon, 2 Apr 2001, Steffen Grunewald wrote:
>
> > And BTW: Isn't Andre's writing style very lovely ?
>
> DGD, and who made you a grammer teacher of english on LKML?
> Never claimed to be an "English Major" soooooo WTF is the price of
> eggs in Germany today? This is no different to trying an orthoginal
> rotation to one's person.
Well, I for one think people should make the effort to write better
English. It makes it easier to parse. I realise it's hard, more so for
some than others. I also realise that there's the temptation to just
blurt out quickly and move on. But overall it's still better to make
it easier for people to read. As I say in the FAQ "write for maximum
efficiency of reading". More time is wasted if thousands of brains
stall for a minute while parsing, than if one brain stalls for 5
minutes in generating. Respect the time of the readers. Their time is
important too.
However, flaming about it isn't the right answer either. Nor are snide
comments. The point can be raised politely (i.e. "I have difficulty
parsing your messages because of your writing style/grammar/spelling").
Regards,
Richard....
Permanent: [email protected]
Current: [email protected]
Andre Hedrick wrote:
> On Mon, 2 Apr 2001, Padraig Brady wrote:
>
>> OK can we just have a technical discussion?
>
> Please, lets do, I am tired of the battles
>
>> I.E. no need for PCMCIA or any of that. I understood from your
>> responses that you didn't realise this?
>
> This valid that I do not know everything and that CFA does interesting
> things more than what was specified in the past.
cool.
>> 2. Compact Flash in this application (I.E. solid state hard disk) is
>> getting very popular as prices are tumbling.
>>
>> 3. Having a config parameter (uneeded kludge in my opinion), like hdx=flash
>> even if hdx is not a compact flash is confusing. Can we call it hdx=probe
>> which fits nicely with the noprobe option.
>>
>>> I then explained why the detection was failing and pointed where to verify.
>>
>> No you didn't. You mentioned a 30 second timeout, but not why it
>> was caused. Have you seen this yourself or can you point us at who
>> reported this to you?
>
> Sorry phone call and email got mixed togather.
> But I did explain that there could be a failure to detect if PDIAG/DASP
> if one or the other devices was held to long and the wrong device reported
> a signature in the task register. Also that the if you reversed the two
> device it would correctly report always.
>
Hmm, OK.
>>
>>> After 3-5 attempts and I can not get the point across because the other
>>> party keeps going off in different directions to do "what about this",
>>
>> Emm, I think *you* were going off describing your application with
>> a "bazar ata-bridge", not the simple use of a compact flash as a
>> hard disk.
>
> Not quite, the electronic differences and flash in native mode is
> incompatable, if you put it in to a mode that is 5V compatable then it
> does seem possible and reasonable to work. Your imperical data points
> verify this issue.
cool
> What really needs to happen is that all the devices that are CFA-like
> which require name parsing for detecting should have the "flash" rule
> imposed. Whereas the ones that correctly report 0x848A for word 0 of the
> identify page may be exempt.
sounds good if we can easily differentiate between buggy & non-buggy flash.
> This seems like a reasonable step given that you are pointing out you
> a have modern CFA's that are more than just CFA's.
I'm not sure I have. They seem to following the latest spec I
downloaded from http://www.compactflash.org
> Would that work for you?
>
>>> I finally pointed out facts that distrub people, and gave up on trying to
>>> show/present/give the answer and offered to then enforce their beliefs of
>>> reality.
>>>
>>> So I state a few facts very pointed to get their attention again and that
>>> is additude??
>>
>> Actually I thought the final email was a little more concise/informative, thanks.
>
> Well I am glad that somebody gleened some information and providing
> feedback so that forward progress is possible, and not the classic battles.
cool,
Padraig.
Andre Hedrick wrote:
> On Mon, 2 Apr 2001, Padraig Brady wrote:
>
>> I'm not sure I have. They seem to following the latest spec I
>> downloaded from http://www.compactflash.org
>
> I am not paying $2500-$5000 annual for membership sorry.
> It is bad enough that I burn $800 for T13 plus about $1000 per meeting.
> $7000 is my personal financial limit.
>
> If you want to give me the SPEC to review and no NDA cool, but CFA and SDA
> I have not interest in the legal action that will happen I expose SDA for
> what it is ...
I just filled a form on the WWW site and downloaded:
"CF+ and CompactFlash Specificiation revision 1.4"
Padraig.
On Mon, 23 Apr 2001 21:57:12 -0700 (PDT), Andre Hedrick wrote:
>> OK can we just have a technical discussion?
>Please, lets do, I am tired of the battles
Me three... I am much better at tech than flames although I seemed to have missed all the
flambe.
BTW... why is this thread called Road Runner?
>> I.E. no need for PCMCIA or any of that. I understood from your
>> responses that you didn't realise this?
>This valid that I do not know everything and that CFA does interesting
>things more than what was specified in the past.
The converse is true as well.. Having only used the CFA devices as solid state IDE drives I
was unaware that there were any that didn't do the True IDE mode.
>Sorry phone call and email got mixed togather.
>But I did explain that there could be a failure to detect if PDIAG/DASP
>if one or the other devices was held to long and the wrong device reported
>a signature in the task register. Also that the if you reversed the two
>device it would correctly report always.
Yes... Now that I am up on the docs I see exactly how this can happen. Mishandled signals
would cause device 0 to go 30 seconds before it gave up on Device 1.
Andre, do you think it's reasonable to assume that if most PC BIOS's will detect the CFA then
it probally is handling the POR sequence and EXECUTE DEVICE DIAGNOSIC correctly?
If so then I feel pretty confident that it should work ok.. Over the 2 years we have been
using CFAs as solid state HD's I have never seen a case where the BIOS didn't autodetect it
properly as master or slave or have I had a problem with a long timeout. I have plugged it
up to all manner of machine old and new.. An the usual configuration was that the CFA was
Device 0 and some random IDE drive we have laying around was device 1.
Now I am not saying it won't happen.. just that I haven't seen it.
>Not quite, the electronic differences and flash in native mode is
>incompatible, if you put it in to a mode that is 5V compatable then it
>does seem possible and reasonable to work. Your imperical data points
>verify this issue.
Not to mention that the datasheet for the device indicates that this is possible and intended
not just happenstance. *grin*
>What really needs to happen is that all the devices that are CFA-like
>which require name parsing for detecting should have the "flash" rule
>imposed. Whereas the ones that correctly report 0x848A for word 0 of the
>identify page may be exempt.
Yes I think so... Although I haven't done a rigorous reading of the SanDisk datasheet I
haven't found anything so far that would lead me to believe that operates any different than
an ATA-2 device when operating in True IDE mode.
IIRC SanDisk was the original people to come out with IDE CFA and everyone else just copied
them. I have the SanDisk datasheets that I can send you if you need them to verify stuff. I
believe that if you verify it with the SanDisk then all the other MFG's should work as well.
--
Richard A. Smith Bitworks, Inc.
[email protected] 501.846.5777
Sr. Design Engineer http://www.bitworks.com
Richard A. Smith wrote:
> IIRC SanDisk was the original people to come out with IDE CFA and everyone else just copied
> them. I have the SanDisk datasheets that I can send you if you need them to verify stuff. I
> believe that if you verify it with the SanDisk then all the other MFG's should work as well.
If only. In my limited experience SanDisk cards have been the most
tolerant. I suspect that Sandisk actually implement the full range of
timings documented in the spec, and nobody else bothers.
This isn't normally a problem on PC hardware, but if you try to
implement an interface to talk to a CF card in an embedded system you
find this out.
- Adrian Cox
On Mon, 02 Apr 2001 21:50:59 +0100, Adrian Cox wrote:
>> IIRC SanDisk was the original people to come out with IDE CFA and everyone
>> else just copied them. I have the SanDisk datasheets that I can send you
>> if you need them to verify stuff. I believe that if you verify it with
>> the SanDisk then all the other MFG's should work as well.
>
>If only. In my limited experience SanDisk cards have been the most
>tolerant. I suspect that Sandisk actually implement the full range of
>timings documented in the spec, and nobody else bothers.
>
>This isn't normally a problem on PC hardware, but if you try to
>implement an interface to talk to a CF card in an embedded system you
>find this out.
Hmmm... most of our embedded systems are based on a PC somehow either via a processor card or
an actual PC system that we design so perhaps I have't stressed the limits yet.
We do actually use SST (Silcon Storage Technolog) CF's as well and they seem to function just
identical to the SanDisk but not quite as robust... I have had several of the SST's develope
a problem in the partition table and as thus the just error when you try to mount them.
Several people on the liunx-embedded list also have similar experiences.
That seems to follow your observations...
Will it be worth while for you if I break out the scope and examine how our CF's handle the
PDIAG signal or can we just go on faith that they do indeed work as expected?
--
Richard A. Smith Bitworks, Inc.
[email protected] 501.846.5777
Sr. Design Engineer http://www.bitworks.com
Richard A. Smith wrote:
> On Mon, 02 Apr 2001 21:50:59 +0100, Adrian Cox wrote:
>> If only. In my limited experience SanDisk cards have been the most
>> tolerant. I suspect that Sandisk actually implement the full range of
>> timings documented in the spec, and nobody else bothers.
[...]
> We do actually use SST (Silcon Storage Technolog) CF's as well and they seem to function just
> identical to the SanDisk but not quite as robust... I have had several of the SST's develope
> a problem in the partition table and as thus the just error when you try to mount them.
> Several people on the liunx-embedded list also have similar experiences.
>
> That seems to follow your observations...
>
> Will it be worth while for you if I break out the scope and examine how our CF's handle the
> PDIAG signal or can we just go on faith that they do indeed work as expected?
Might be interesting to see. The worst trouble I've had has been with
noname parts from high street computer stores, and I don't know who the
original manufacturer was. If I knew what subset of functionality
digital cameras used I'd know what could be relied on.
- Adrian
Andre Hedrick writes:
> On Mon, 2 Apr 2001, Richard Gooch wrote:
>
> > However, flaming about it isn't the right answer either. Nor are snide
> > comments. The point can be raised politely (i.e. "I have difficulty
> > parsing your messages because of your writing style/grammar/spelling").
>
> Hi Richard,
>
> You are correct and I freely admit that I have problems in
> communication on paper, but do better in person as you found out, I
> hope.
Sure. It was good to finally meet you. I understood you fine, despite
the funny American accent ;-)
> I think I will do what most every one else does is drop from lkml
> and only deal with direct mail.
Well, I hope you'll reconsider that decision. Whether or not I have
trouble parsing your emails, you still have a contribution to make.
All I can do is hope people pay more attention to generating easier to
parse output :-)
> Also it was good to meet you this weekend.
Indeed. Putting a face to a name is always helpful. Makes a person
seem more real (no big surprise, since us monkeys are wired like
that).
Regards,
Richard....
Permanent: [email protected]
Current: [email protected]
After upgrading from 2.4.2 to 2.4.3 with jfs 0.2.1 my machine died last
night.
The last messages in syslog were:
Apr 3 01:35:01 hades kernel: Unable to handle kernel paging request at
virtual
address 8d0800c0
Apr 3 01:35:01 hades kernel: *pde = 00000000
Apr 3 01:35:01 hades kernel: Unable to handle kernel paging request at
virtual address 8d0800b8
Apr 3 01:35:01 hades kernel: *pde = 00000000
And the last application message was:
Apr 3 01:35:01 hades gnome-name-server[225]: input condition is: 0x11,
exiting
Does anyone have a clue what this means?
Regards,
Andreas
On Mon, 2 Apr 2001, Padraig Brady wrote:
> OK can we just have a technical discussion?
Please, lets do, I am tired of the battles
> I.E. no need for PCMCIA or any of that. I understood from your
> responses
> that you didn't realise this?
This valid that I do not know everything and that CFA does interesting
things more than what was specified in the past.
> 2. Compact Flash in this application (I.E. solid state hard disk) is
> getting very
> popular as prices are tumbling.
>
> 3. Having a config parameter (uneeded kludge in my opinion), like hdx=flash
> even if hdx is not a compact flash is confusing. Can we call it
> hdx=probe
> which fits nicely with the noprobe option.
>
>
> > I then explained why the detection was failing and pointed where to verify.
>
> No you didn't. You mentioned a 30 second timeout, but not why it
> was caused. Have you seen this yourself or can you point us at who
> reported this to you?
Sorry phone call and email got mixed togather.
But I did explain that there could be a failure to detect if PDIAG/DASP
if one or the other devices was held to long and the wrong device reported
a signature in the task register. Also that the if you reversed the two
device it would correctly report always.
>
> > After 3-5 attempts and I can not get the point across because the other
> > party keeps going off in different directions to do "what about this",
>
> Emm, I think *you* were going off describing your application with
> a "bazar ata-bridge", not the simple use of a compact flash as a
> hard disk.
Not quite, the electronic differences and flash in native mode is
incompatable, if you put it in to a mode that is 5V compatable then it
does seem possible and reasonable to work. Your imperical data points
verify this issue.
What really needs to happen is that all the devices that are CFA-like
which require name parsing for detecting should have the "flash" rule
imposed. Whereas the ones that correctly report 0x848A for word 0 of the
identify page may be exempt.
This seems like a reasonable step given that you are pointing out you
a have modern CFA's taht are more than just CFA's.
Would that work for you?
>
> > I finally pointed out facts that distrub people, and gave up on trying to
> > show/present/give the answer and offered to then enforce their beliefs of
> > reality.
> >
> > So I state a few facts very pointed to get their attention again and that
> > is additude??
>
> Actually I thought the final email was a little more concise/informative, thanks.
Well I am glad that somebody gleened some information and providing
feedback so that forward progress is possible, and not the classic
battles.
Cheers,
Andre Hedrick
Linux ATA Development
On Mon, 2 Apr 2001, Richard Gooch wrote:
> However, flaming about it isn't the right answer either. Nor are snide
> comments. The point can be raised politely (i.e. "I have difficulty
> parsing your messages because of your writing style/grammar/spelling").
Hi Richard,
You are correct and I freely admit that I have problems in communication
on paper, but do better in person as you found out, I hope. I think I
will do what most every one else does is drop from lkml and only deal with
direct mail.
Also it was good to meet you this weekend.
Cheers,
Andre Hedrick
Linux ATA Development
On Mon, 2 Apr 2001, Padraig Brady wrote:
> I'm not sure I have. They seem to following the latest spec I
> downloaded from http://www.compactflash.org
I am not paying $2500-$5000 annual for membership sorry.
It is bad enough that I burn $800 for T13 plus about $1000 per meeting.
$7000 is my personal financial limit.
If you want to give me the SPEC to review and no NDA cool, but CFA and SDA
I have not interest in the legal action that will happen I expose SDA for
what it is ...
Andre Hedrick
Linux ATA Development