Hi,
I don't have any IDE machines handy, and since these problems that IDE had
in the last days, I wonder what's become of it. Has anyone been so brave
as to try out 2.5.26 w/the included IDE (IDE 98)? How is it?
Regards,
Thunder
--
(Use http://www.ebb.org/ungeek if you can't decode)
------BEGIN GEEK CODE BLOCK------
Version: 3.12
GCS/E/G/S/AT d- s++:-- a? C++$ ULAVHI++++$ P++$ L++++(+++++)$ E W-$
N--- o? K? w-- O- M V$ PS+ PE- Y- PGP+ t+ 5+ X+ R- !tv b++ DI? !D G
e++++ h* r--- y-
------END GEEK CODE BLOCK------
Em Fri, Jul 19, 2002 at 10:51:26PM -0600, Thunder from the hill escreveu:
> I don't have any IDE machines handy, and since these problems that IDE had
> in the last days, I wonder what's become of it. Has anyone been so brave
> as to try out 2.5.26 w/the included IDE (IDE 98)? How is it?
I've been using the current linus BK tree, which already has more than what
is in 2.5.26 in a dual pentium 100 machine with a buggy CMD640 machine without
problems up to now. 32 MB of memory, 16 MB of swap, old QUANTUM LPS210A:
hda: 412110 sectors w/98KiB Cache, CHS=723/15/38
Did some light stress testing, swapping, etc, no problems up to now, and with
Rik's minimal rmap patch that is already at Linus BK tree. 8)
- Arnaldo
Thunder from the hill wrote:
> I don't have any IDE machines handy, and since these problems that IDE had
> in the last days, I wonder what's become of it. Has anyone been so brave
> as to try out 2.5.26 w/the included IDE (IDE 98)? How is it?
I was able to freeze 2.5.25 regularly (locking issues?) by trying to
move large files between different filesystems on the same drive.
IDE 98 in 2.5.26 handles that fine and hasn't locked up a single
time in normal use.
--
Skip
IDE 98 is just compilation of 94, 95, 96, 97, 98-pre plus some minor
fixes. Runs fine for me.
And once again: vanilla 2.5.25 on IDE is bad for your health.
Regards
--
Bartlomiej
On Fri, 19 Jul 2002, Thunder from the hill wrote:
> Hi,
>
> I don't have any IDE machines handy, and since these problems that IDE had
> in the last days, I wonder what's become of it. Has anyone been so brave
> as to try out 2.5.26 w/the included IDE (IDE 98)? How is it?
>
> Regards,
> Thunder
> --
> (Use http://www.ebb.org/ungeek if you can't decode)
> ------BEGIN GEEK CODE BLOCK------
> Version: 3.12
> GCS/E/G/S/AT d- s++:-- a? C++$ ULAVHI++++$ P++$ L++++(+++++)$ E W-$
> N--- o? K? w-- O- M V$ PS+ PE- Y- PGP+ t+ 5+ X+ R- !tv b++ DI? !D G
> e++++ h* r--- y-
> ------END GEEK CODE BLOCK------
>
> -
> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> the body of a message to [email protected]
> More majordomo info at http://vger.kernel.org/majordomo-info.html
> Please read the FAQ at http://www.tux.org/lkml/
>
Hi,
On Sat, 20 Jul 2002, Bartlomiej Zolnierkiewicz wrote:
> And once again: vanilla 2.5.25 on IDE is bad for your health.
Nobody ever mentioned 2.5.25...
Regards,
Thunder
--
(Use http://www.ebb.org/ungeek if you can't decode)
------BEGIN GEEK CODE BLOCK------
Version: 3.12
GCS/E/G/S/AT d- s++:-- a? C++$ ULAVHI++++$ P++$ L++++(+++++)$ E W-$
N--- o? K? w-- O- M V$ PS+ PE- Y- PGP+ t+ 5+ X+ R- !tv b++ DI? !D G
e++++ h* r--- y-
------END GEEK CODE BLOCK------
> > And once again: vanilla 2.5.25 on IDE is bad for your health.
>
> Nobody ever mentioned 2.5.25...
I'm wondering, why haven't IDE 99 and IDE 100 been posted on lkml?
T.
On Sat, 20 Jul 2002, Tomas Szepe wrote:
> I'm wondering, why haven't IDE 99 and IDE 100 been posted on lkml?
>
> T.
Ask Martin, I don't know.
Regards
--
Bartlomiej
Thunder from the hill <[email protected]> writes:
> Hi,
>
> I don't have any IDE machines handy, and since these problems that IDE had
> in the last days, I wonder what's become of it. Has anyone been so brave
> as to try out 2.5.26 w/the included IDE (IDE 98)? How is it?
On my Duron system where LG CD-RW GCE-8320B is connected as hda to
00:07.1 IDE interface: VIA Technologies, Inc. Bus Master IDE (rev 10)
and three IDE discs (hde: IBM-DTLA-307030 [30GB], hdg: ST380021A [80GB],
hdh: ST360021A [60GB]) connected to:
00:0b.0 Unknown mass storage controller: Promise Technology, Inc. 20268 (rev 01)
2.5.26+xfs patch boots, detects VIA IDE controller, finds my cdrw
and stops - no oops just booting stops, HDD LED (connected to promise
controller) is on. Nothing more happens.
On 2.4.18 + Hedrick patches there is no problem.
> Thunder
--
Arkadiusz Mi?kiewicz IPv6 ready PLD Linux at http://www.pld.org.pl
misiek(at)pld.org.pl AM2-6BONE, 1024/3DB19BBD, arekm(at)ircnet, PWr
On 20 Jul 2002, Arkadiusz Miskiewicz wrote:
> Thunder from the hill <[email protected]> writes:
>
> > Hi,
> >
> > I don't have any IDE machines handy, and since these problems that IDE had
> > in the last days, I wonder what's become of it. Has anyone been so brave
> > as to try out 2.5.26 w/the included IDE (IDE 98)? How is it?
> On my Duron system where LG CD-RW GCE-8320B is connected as hda to
> 00:07.1 IDE interface: VIA Technologies, Inc. Bus Master IDE (rev 10)
> and three IDE discs (hde: IBM-DTLA-307030 [30GB], hdg: ST380021A [80GB],
> hdh: ST360021A [60GB]) connected to:
> 00:0b.0 Unknown mass storage controller: Promise Technology, Inc. 20268 (rev 01)
> 2.5.26+xfs patch boots, detects VIA IDE controller, finds my cdrw
> and stops - no oops just booting stops, HDD LED (connected to promise
> controller) is on. Nothing more happens.
>
> On 2.4.18 + Hedrick patches there is no problem.
And on some recent 2.4.19-rc ar -ac?
2.5.27 have my forward port of Hank's fixes for Promise controllers.
It may help... (or not).
Regards
--
Bartlomiej
>
> > Thunder
>
> --
> Arkadiusz Mi?kiewicz IPv6 ready PLD Linux at http://www.pld.org.pl
> misiek(at)pld.org.pl AM2-6BONE, 1024/3DB19BBD, arekm(at)ircnet, PWr
> > I'm wondering, why haven't IDE 99 and IDE 100 been posted on lkml?
> >
> Ask Martin, I don't know.
Right. I'm puzzled here.
How safe does Martin think people will suppose the patches are if they were
not meant to be reviewed in the standard lkml manner? Given the problems
with 2.5.25 (and 'problems' would be an euphemism here) I'd expect to see
something like a 'safer coding' approach, not 'stealth operations.'
T.
Tomas,
Give Bartlomiej a break!
I have been working with and teaching him how to read the document.
I refused to give him the answers, he had to learn it while I guided him.
It took about 3-4 evenings on IRC for Bart to get it, while he still has
much to learn that only time will grant him. I got in a shorter period
than me, but I had to learn much if it on my own and only at the threshold
of getting did somebody in the storage industry give me a swift kick in
the pants.
Somebody submit a patch to make Bart the maintainer of 2.5.
Sincerely,
Andre Hedrick
LAD Storage Consulting Group
On Sun, 21 Jul 2002, Tomas Szepe wrote:
> > > I'm wondering, why haven't IDE 99 and IDE 100 been posted on lkml?
> > >
> > Ask Martin, I don't know.
>
> Right. I'm puzzled here.
>
> How safe does Martin think people will suppose the patches are if they were
> not meant to be reviewed in the standard lkml manner? Given the problems
> with 2.5.25 (and 'problems' would be an euphemism here) I'd expect to see
> something like a 'safer coding' approach, not 'stealth operations.'
>
> T.
> -
> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> the body of a message to [email protected]
> More majordomo info at http://vger.kernel.org/majordomo-info.html
> Please read the FAQ at http://www.tux.org/lkml/
>
Andre,
> Give Bartlomiej a break!
Hmm? I was pointing my finger at Martin, not Bart.
> I have been working with and teaching him how to read the document.
> I refused to give him the answers, he had to learn it while I guided him.
> It took about 3-4 evenings on IRC for Bart to get it, while he still has
> much to learn that only time will grant him. I got in a shorter period
> than me, but I had to learn much if it on my own and only at the threshold
> of getting did somebody in the storage industry give me a swift kick in
> the pants.
>
> Somebody submit a patch to make Bart the maintainer of 2.5.
Like I said yesterday, I'd love to see that happen, on the grounds that:
o He has exhibited understanding of the subtleties of the IDE code --
unlike MD.
o He doesn't have a conflicting/macho personality -- unlike MD.
o He actually knows what has to be done and has planned out a long-term
schedule -- unlike MD.
T.
Hi,
On Sun, 21 Jul 2002, Tomas Szepe wrote:
> > Give Bartlomiej a break!
>
> Hmm? I was pointing my finger at Martin, not Bart.
I guess he meant "a start".
> o He has exhibited understanding of the subtleties of the IDE code --
> unlike MD.
> o He doesn't have a conflicting/macho personality -- unlike MD.
> o He actually knows what has to be done and has planned out a long-term
> schedule -- unlike MD.
Whatever, I think there are some problems with the current IDE
development. Martin is possibly a good developer, but he doesn't always
seem all that experienced regarding IDE. Also, he releases too many testal
patches which simply aren't going to work. Unfortunately, Linus ate them
at once, so we sometimes had deadly situations.
These facts make me think that Martin is probably not the best maintainer
for it. He might be a good developer - whatever -, but what is worth, he's
certainly not the best one when it comes to planning. Thus, I think the
question is: will it be better for IDE if Martin has the lead, current
situation, or if Martin can unfold all his powerfulness in testing while
someone else takes up the pieces and makes IDE of it, planning what comes
next?
I'm not proposing a Martin Dalecki frontend, I'm just propagating an
election.
Regards,
Thunder
--
(Use http://www.ebb.org/ungeek if you can't decode)
------BEGIN GEEK CODE BLOCK------
Version: 3.12
GCS/E/G/S/AT d- s++:-- a? C++$ ULAVHI++++$ P++$ L++++(+++++)$ E W-$
N--- o? K? w-- O- M V$ PS+ PE- Y- PGP+ t+ 5+ X+ R- !tv b++ DI? !D G
e++++ h* r--- y-
------END GEEK CODE BLOCK------
At 12:14 AM 7/21/2002 -0600, Thunder from the hill wrote:
>Hi,
>
>On Sun, 21 Jul 2002, Tomas Szepe wrote:
> > > Give Bartlomiej a break!
> >
> > Hmm? I was pointing my finger at Martin, not Bart.
>
>I guess he meant "a start".
>
> > o He has exhibited understanding of the subtleties of the IDE code --
> > unlike MD.
> > o He doesn't have a conflicting/macho personality -- unlike MD.
> > o He actually knows what has to be done and has planned out a long-term
> > schedule -- unlike MD.
>
>Whatever, I think there are some problems with the current IDE
>development. Martin is possibly a good developer, but he doesn't always
>seem all that experienced regarding IDE. Also, he releases too many testal
>patches which simply aren't going to work. Unfortunately, Linus ate them
>at once, so we sometimes had deadly situations.
>
>These facts make me think that Martin is probably not the best maintainer
>for it. He might be a good developer - whatever -, but what is worth, he's
>certainly not the best one when it comes to planning. Thus, I think the
>question is: will it be better for IDE if Martin has the lead, current
>situation, or if Martin can unfold all his powerfulness in testing while
>someone else takes up the pieces and makes IDE of it, planning what comes
>next?
>
>I'm not proposing a Martin Dalecki frontend, I'm just propagating an
>election.
Since I know spit about IDE/ATA/ATAPI/SCSI, I'll keep my mouth shut and
leave judgement/"voting" to those who fully understand the technical issues.
-Mike
Hi,
On Sun, 21 Jul 2002, Mike Galbraith wrote:
> Since I know spit about IDE/ATA/ATAPI/SCSI, I'll keep my mouth shut and
> leave judgement/"voting" to those who fully understand the technical issues.
You probably shouldn't. Technical decisions should be made by technicians,
but decisions about the technicians should be made by the human resources
dept., and since we claim to be a constitutional monarchy, we might try
out a democratic decision...
Regards,
Thunder
--
(Use http://www.ebb.org/ungeek if you can't decode)
------BEGIN GEEK CODE BLOCK------
Version: 3.12
GCS/E/G/S/AT d- s++:-- a? C++$ ULAVHI++++$ P++$ L++++(+++++)$ E W-$
N--- o? K? w-- O- M V$ PS+ PE- Y- PGP+ t+ 5+ X+ R- !tv b++ DI? !D G
e++++ h* r--- y-
------END GEEK CODE BLOCK------
At 01:31 AM 7/21/2002 -0600, Thunder from the hill wrote:
>Hi,
>
>On Sun, 21 Jul 2002, Mike Galbraith wrote:
> > Since I know spit about IDE/ATA/ATAPI/SCSI, I'll keep my mouth shut and
> > leave judgement/"voting" to those who fully understand the technical
> issues.
>
>You probably shouldn't. Technical decisions should be made by technicians,
>but decisions about the technicians should be made by the human resources
>dept., and since we claim to be a constitutional monarchy, we might try
>out a democratic decision...
No, I'm absolutely sure I'm doing the right thing.
Regarding constitutional monarchy, that's high-grade horse shit. LT has always
been the benevolent dictator [1] of his tree.
-Mike
1. aka heartless bastard. (for the humor impaired, those are his words;)
> >On Sun, 21 Jul 2002, Mike Galbraith wrote:
> >> Since I know spit about IDE/ATA/ATAPI/SCSI, I'll keep my mouth shut and
> >> leave judgement/"voting" to those who fully understand the technical
> >issues.
> >
> >You probably shouldn't. Technical decisions should be made by technicians,
> >but decisions about the technicians should be made by the human resources
> >dept., and since we claim to be a constitutional monarchy, we might try
> >out a democratic decision...
>
> No, I'm absolutely sure I'm doing the right thing.
Well you don't necessarily have to be an IDE guru to realize something's
wrong when you see a bloke constantly breaking the subsystem, practically
never fixing it up himself, disappearing for a month w/o saying a word
after having fried 2.5.25 completely and not really caring about what
others have to say about the code.
And you've noticed the IDE 2.4 forward-port, right?
T.
Hi,
On Sun, 21 Jul 2002, Mike Galbraith wrote:
> Regarding constitutional monarchy, that's high-grade horse shit. LT has
^^
>From Jargon File (4.3.0, 30 APR 2001) [jargon]:
Linus /leen'us'/ or /lin'us'/, not /li:'nus/ Linus Torvalds, the author
of {Linux}. Nobody in the hacker culture has been as readily recognized
by first name alone since Ken (Thompson).
So I think calling him LT is way, way wrong if you think he's worth the
comparison to ken.
> always been the benevolent dictator [1] of his tree.
That's nothing that's speaking against the monarchy. However, I think he's
still bound to do things that he just has to do, because otherwise he'd be
hung ;-)
Regards,
Thunder
--
(Use http://www.ebb.org/ungeek if you can't decode)
------BEGIN GEEK CODE BLOCK------
Version: 3.12
GCS/E/G/S/AT d- s++:-- a? C++$ ULAVHI++++$ P++$ L++++(+++++)$ E W-$
N--- o? K? w-- O- M V$ PS+ PE- Y- PGP+ t+ 5+ X+ R- !tv b++ DI? !D G
e++++ h* r--- y-
------END GEEK CODE BLOCK------
At 10:23 AM 7/21/2002 +0200, Tomas Szepe wrote:
> > >On Sun, 21 Jul 2002, Mike Galbraith wrote:
> > >> Since I know spit about IDE/ATA/ATAPI/SCSI, I'll keep my mouth shut and
> > >> leave judgement/"voting" to those who fully understand the technical
> > >issues.
> > >
> > >You probably shouldn't. Technical decisions should be made by technicians,
> > >but decisions about the technicians should be made by the human resources
> > >dept., and since we claim to be a constitutional monarchy, we might try
> > >out a democratic decision...
> >
> > No, I'm absolutely sure I'm doing the right thing.
>
>Well you don't necessarily have to be an IDE guru to realize something's
>wrong when you see a bloke constantly breaking the subsystem, practically
>never fixing it up himself, disappearing for a month w/o saying a word
>after having fried 2.5.25 completely and not really caring about what
>others have to say about the code.
No, you don't have to be a guru to notice that the rewrite is proving
difficult.
>And you've noticed the IDE 2.4 forward-port, right?
Yes. I also read Jens stated reasons for doing the port.. made perfect sense.
-Mike
On Sun, 21 Jul 2002, Mike Galbraith wrote:
> >Well you don't necessarily have to be an IDE guru to realize something's
> >wrong when you see a bloke constantly breaking the subsystem, practically
> >never fixing it up himself, disappearing for a month w/o saying a word
> >after having fried 2.5.25 completely and not really caring about what
> >others have to say about the code.
>
> No, you don't have to be a guru to notice that the rewrite is proving
> difficult.
You are right there, but anyone could have told you that even before the
rewrite started. A rewrite of the IDE drivers should mean one out of two
things:
1) You are aware of the fact IDE disks contain user data, thus if you hack
the ide drivers, you better do / test it thoroughly. Of course, you are
not perfect, mistakes are possible, but know that people won't like them.
2) You assume EVERYONE that is working on the development kernel (even on
parts that have nothing to do with IDE) has a seperate computer for kernel
development or SCSI disks. (You don't want your IDE driver to destroy that
data partition / disk, do you ?).
Well... IMHO, 2) is not an option for a good maintainer, and at the
moment, 1) seems not to hold.
I was willing to test the 2.5 tree, but the way IDE development is going
at the moment makes me keeping my hands off any 2.5 kernel. I just don't
have a second system that is fast enough to recomile a kernel every few
days.
Well... I already gave my opinion about the IDE maintenance in the early
2.5 days, glad to see I was right back then.
Guys, see you again after 2.6.0, I probably won't compile any single
kernel before then.
Jos
On Sun, 21 Jul 2002, Jos Hulzink wrote:
> I was willing to test the 2.5 tree, but the way IDE development is going
> at the moment makes me keeping my hands off any 2.5 kernel. I just don't
> have a second system that is fast enough to recomile a kernel every few
> days.
Compile on a faster machine and transfer kernel images across, using NFS
can simplify this quite a lot.
> Well... I already gave my opinion about the IDE maintenance in the early
> 2.5 days, glad to see I was right back then.
Well i'm sure lotsa folks can say 'i told you so' but that doesn't help a
bit, how much you put in is directly related to what you'll get out...
> Guys, see you again after 2.6.0, I probably won't compile any single
> kernel before then.
... in 2.6
Cheers,
Zwane Mwaikambo
--
function.linuxpower.ca
Hi,
On Mon, 22 Jul 2002, Mike Galbraith wrote:
> Oh, I have thought about it. I'm in no way defending Martin (he's a big
> boy, can do that by himself) or saying everything's fine. I merely
> objected to the content and presentation of the posts I replied to.
Thing is: if the people think that it actually _is_ good that Martin is
the "Linux 2.5 IDE guy", he's got nothing to fear. All I wanted is to
summon a bit of a discussion about several issues that are not okay in any
way. It's like with the code itself: either bugs get resolved, or the code
must be pulled. Of course I guess it'd be cool if the bugs could get
fixed.
Regards,
Thunder
--
(Use http://www.ebb.org/ungeek if you can't decode)
------BEGIN GEEK CODE BLOCK------
Version: 3.12
GCS/E/G/S/AT d- s++:-- a? C++$ ULAVHI++++$ P++$ L++++(+++++)$ E W-$
N--- o? K? w-- O- M V$ PS+ PE- Y- PGP+ t+ 5+ X+ R- !tv b++ DI? !D G
e++++ h* r--- y-
------END GEEK CODE BLOCK------
At 01:04 AM 7/22/2002 +0200, Bartlomiej Zolnierkiewicz wrote:
>On Sun, 21 Jul 2002, Mike Galbraith wrote:
>
> >> Well you don't necessarily have to be an IDE guru to realize something's
> >> wrong when you see a bloke constantly breaking the subsystem, practically
> >> never fixing it up himself, disappearing for a month w/o saying a word
> >> after having fried 2.5.25 completely and not really caring about what
> >> others have to say about the code.
>
> > No, you don't have to be a guru to notice that the rewrite is proving
> > difficult.
>
>You also don't have to be a guru to notice that recently most of the
>content of the rewrite is moving code here and there, unfolding functions,
>renaming them and changing intendation. Check yourself.
>Also imagine how hard is now to track changes from 2.4 to 2.5
>now and fix bugs.
>
>Yup, please give me a break from having to track this changes. :-)
>
>If you go through all the ide-clean patches you will see that much
>of the cruft has been removed, some things fixed but there is still
>plenty of work to do.
>
>Next problem is that Martin seems to not care that his style of
>development (pushing stuff immediately to Linus instead of lkml -> some
>reasonable delay -> Linus) _constantly_ interferes other people doing
>kernel hacking.
>
>I don't want next flamewar or personal bashing here,
>please only _think_ for a while about issues raised.
Oh, I have thought about it. I'm in no way defending Martin (he's a big boy,
can do that by himself) or saying everything's fine. I merely objected to the
content and presentation of the posts I replied to.
-Mike
On Sun, Jul 21, 2002 at 01:31:40AM -0600, Thunder from the hill wrote:
> On Sun, 21 Jul 2002, Mike Galbraith wrote:
> > Since I know spit about IDE/ATA/ATAPI/SCSI, I'll keep my mouth shut and
> > leave judgement/"voting" to those who fully understand the technical issues.
>
> You probably shouldn't. Technical decisions should be made by technicians,
> but decisions about the technicians should be made by the human resources
> dept., and since we claim to be a constitutional monarchy, we might try
> out a democratic decision...
I usually don't use words like this, but can you just shut up?
I mean - if Bartek, or me, or whoever, wanted to be the 2.5 IDE king
now, I'm sure Martin would pass the thorny crown quite happily.
When noone else says that, I'm very impressed about where the IDE code
got from before Martin started. And after talking to him at the Kernel
Workshop I must say he has a very good vision where he's heading to.
The only problem that pisses of a lot of people is that he just spews
patches at his own rate, not synchronized with Linus, so that Martin's
latest 'stable' doesn't usually match Linus's releases.
Now that the most aggresive changes are behind us, I believe things will
go a tad smoother ...
Anyway, I guess I should just shut up and go back to fixing that SiS IDE
driver, too ...
--
Vojtech Pavlik
SuSE Labs
On Sun, 21 Jul 2002, Mike Galbraith wrote:
>> Well you don't necessarily have to be an IDE guru to realize something's
>> wrong when you see a bloke constantly breaking the subsystem, practically
>> never fixing it up himself, disappearing for a month w/o saying a word
>> after having fried 2.5.25 completely and not really caring about what
>> others have to say about the code.
> No, you don't have to be a guru to notice that the rewrite is proving
> difficult.
You also don't have to be a guru to notice that recently most of the
content of the rewrite is moving code here and there, unfolding functions,
renaming them and changing intendation. Check yourself.
Also imagine how hard is now to track changes from 2.4 to 2.5
now and fix bugs.
Yup, please give me a break from having to track this changes. :-)
If you go through all the ide-clean patches you will see that much
of the cruft has been removed, some things fixed but there is still
plenty of work to do.
Next problem is that Martin seems to not care that his style of
development (pushing stuff immediately to Linus instead of lkml -> some
reasonable delay -> Linus) _constantly_ interferes other people doing
kernel hacking.
I don't want next flamewar or personal bashing here,
please only _think_ for a while about issues raised.
Regards
--
Bartlomiej