2.4.22 timeline was RE: 2.4.21-rc7 ACPI broken
Quoted 5 lines · Marcelo Tosati
The main reason I didnt want to merge it was due to its size. Its just >too big.
Quoted 1 line
Please stop leading me along. Will you EVER merge it?
Yes, I want to, and will merge it. In 2.4.23-pre.
This kind of mails and sentence makes you lost your credibility :
1) You said that ACPI will be merged in 2.4.22-pre,
2) For many people ACPI (and aic7xxx) is top priority for 2.4 kernel
(see various post including alan). The reason being that most laptop are unusable nowadays without ACPI,
3) You do not explicitely says what you plan for 2.4.22...
So, for stupid people like me, could you take a little time to explains the dummy's what are your views about what is top priority for kernel and for what reasons?
I would personnally suggest that you classify the things using the following filter :
a) Server (SMP, SCSI, RAID, journaling filesystems, ...),
b) laptop (ACPI, CPUFREQ, Software suspend, IDE power save,...),
c) desktop (File system efficiency, new hardware support,...),
d) all systems
Or maybe you simply think that software engineering is a kind of black art that is too hard to understand for average people...
Signature · 7 lines
__ / ` Eric Valette /-- __ o _. 6 rue Paul Le Flem (___, / (_(_(__ 35740 Pace Tel: +33 (0)2 99 85 26 76 Fax: +33 (0)2 99 85 26 76 E-mail: [email protected]
Quoted 8 lines · Eric Valette
...
I would personnally suggest that you classify the things using the following filter :
a) Server (SMP, SCSI, RAID, journaling filesystems, ...),
b) laptop (ACPI, CPUFREQ, Software suspend, IDE power save,...),
c) desktop (File system efficiency, new hardware support,...),
d) all systems
...
Why are journaling filesystems only for servers? Is file system efficiency not relevant on servers?
The important sections are more likely (ordered by priority):
- bug fixes (e.g. aic7xxx)
- support for additional hardware (e.g. ACPI update)
- new features (e.g. XFS)
These groups are not mutual exclusive, e.g. the ACPI update also includes new features.
The important thing is that this is inside a stable kernel series and an update that makes things better for 100 people but makes things worse for one person is IMHO bad since it's a regression for one person.
cu
Adrian
Signature · 4 lines
"Is there not promise of rain?" Ling Tan asked suddenly out
of the darkness. There had been need of rain for many days.
"Only a promise," Lao Er said.
Pearl S. Buck - Dragon SeedQuoted 11 lines · Adrian Bunk
Quoted 7 lines · Eric Valette
...
I would personnally suggest that you classify the things using the following filter :
a) Server (SMP, SCSI, RAID, journaling filesystems, ...),
b) laptop (ACPI, CPUFREQ, Software suspend, IDE power save,...),
c) desktop (File system efficiency, new hardware support,...),
d) all systems
...
Quoted 2 lines
Why are journaling filesystems only for servers? Is file system efficiency not relevant on servers?
I was just making suggestions after a 30s thinking. Side comments, readding this mailling list, I had the impression that journaling and filesystem performance do not seem to mix well. Also on server, you have probably extra backup hardware and means (e.g RAID, DAT, DLT, ...)
Quoted 4 lines
The important sections are more likely (ordered by priority):
- bug fixes (e.g. aic7xxx)
- support for additional hardware (e.g. ACPI update)
- new features (e.g. XFS)
Personnaly, I dislike this approach as it as resulted in 2.4 being non usable for servers (SMP deadlocks, IO stalls, unresponsiveness for several seconds, ...) and laptop (ACPI)...
Quoted 1 line
The important thing is that this is inside a stable kernel series and an update that makes things better for 100 people but makes things worse for one person is IMHO bad since it's a regression for one person.
If 2.4 kernel is not usable without patching, It is far worse for me...
Signature · 7 lines
__ / ` Eric Valette /-- __ o _. 6 rue Paul Le Flem (___, / (_(_(__ 35740 Pace Tel: +33 (0)2 99 85 26 76 Fax: +33 (0)2 99 85 26 76 E-mail: [email protected]
Quoted 4 lines · On Wed, 11 Jun 2003 23:15:06 +0200 Adrian Bunk <[email protected]> wrote:
[...]
The important thing is that this is inside a stable kernel series and an update that makes things better for 100 people but makes things worse for one person is IMHO bad since it's a regression for one person.
You cannot fulfill that in reality. Looking at the broad variety of software out there you simply cannot know all the implications a simple bug fix may have. There may well be boxes that rely on a broken code you just fixed. Only god knows. So you sometimes simply have to do "the right thing"(tm) knowing there will always be people who shoot you for it.
Regards,
Stephan
Quoted 8 lines · Stephan von Krawczynski
Quoted 2 lines · On Wed, 11 Jun 2003 23:15:06 +0200 Adrian Bunk <[email protected]> wrote:
[...]
The important thing is that this is inside a stable kernel series and an update that makes things better for 100 people but makes things worse for one person is IMHO bad since it's a regression for one person.
Quoted 5 lines
You cannot fulfill that in reality. Looking at the broad variety of software out there you simply cannot know all the implications a simple bug fix may have. There may well be boxes that rely on a broken code you just fixed. Only god knows. So you sometimes simply have to do "the right thing"(tm) knowing there will always be people who shoot you for it.
Just to go a little bit more in that direction, I already have seen an obvious one-line patch that caused a deadlock in another OS because due to code location change and probably an additionnal cache miss on specific part of the code, an existing synchronisation bug that was never triggered started to effectively happen...
Besides, if you please 1000 users and cause problem to 2 of them because they have broken hardware, I think you are going into the right direction.
Quoted 4 lines
The important sections are more likely (ordered by priority):
- bug fixes (e.g. aic7xxx)
- support for additional hardware (e.g. ACPI update)
- new features (e.g. XFS)
Bug fixes are meant to make the 2.4 kernel more useable right? So I do not reallly see the utimate difference with other things in your category. What I was asking is a rationnal way of chosing the priorities. You gave me yours without real explanation.
The purpose of the original mail was to ask for discussion/clarification on 2.4 development priorities (e.g something like the 2.6 todo list) and a proposal to set up the priorities using generic targetted hardware (server, laptop, desktop) as hint for requirement selection.
Signature · 7 lines
__ / ` Eric Valette /-- __ o _. 6 rue Paul Le Flem (___, / (_(_(__ 35740 Pace Tel: +33 (0)2 99 85 26 76 Fax: +33 (0)2 99 85 26 76 E-mail: [email protected]
Quoted 7 lines · Eric Valette
Quoted 4 lines · Marcelo Tosati
The main reason I didnt want to merge it was due to its size. Its just too big.
Quoted 1 line
Please stop leading me along. Will you EVER merge it?
Quoted 8 lines
Quoted 1 line
Yes, I want to, and will merge it. In 2.4.23-pre.
This kind of mails and sentence makes you lost your credibility :
1) You said that ACPI will be merged in 2.4.22-pre,
2) For many people ACPI (and aic7xxx) is top priority for 2.4 kernel
(see various post including alan). The reason being that most laptop are unusable nowadays without ACPI,
3) You do not explicitely says what you plan for 2.4.22...
My plan for 2.4.22 is:
- Include the new aic7xxx driver.
- Include ACPI. (I now realized its importance). Already discussing with Andrew the best way to do it.
- Fix the latency/interactivity problems (Chris, Nick and Andrea working
on that)
- Merge obviously correct -aa VM patches.
Those are the most important things that are needed now, I think.
Quoted 22 lines · Marcelo Tosatti
Quoted 7 lines · Eric Valette
Quoted 4 lines · Marcelo Tosati
The main reason I didnt want to merge it was due to its size. Its just too big.
Quoted 1 line
Please stop leading me along. Will you EVER merge it?
Quoted 8 lines
Quoted 1 line
Yes, I want to, and will merge it. In 2.4.23-pre.
This kind of mails and sentence makes you lost your credibility :
1) You said that ACPI will be merged in 2.4.22-pre,
2) For many people ACPI (and aic7xxx) is top priority for 2.4 kernel
(see various post including alan). The reason being that most laptop are unusable nowadays without ACPI,
3) You do not explicitely says what you plan for 2.4.22...My plan for 2.4.22 is:
Quoted 9 lines
- Include the new aic7xxx driver.
- Include ACPI. (I now realized its importance). Already discussing with Andrew the best way to do it.
- Fix the latency/interactivity problems (Chris, Nick and Andrea working
on that)
- Merge obviously correct -aa VM patches.Those are the most important things that are needed now, I think.
Would you accept small canges like -march (gcc_check) for x86 ?
Signature · 4 lines
J.A. Magallon <[email protected]> \ Software is like sex: werewolf.able.es \ It's better when it's free Mandrake Linux release 9.2 (Cooker) for i586 Linux 2.4.21-jam1 (gcc 3.3 (Mandrake Linux 9.2 3.3-1mdk))
Quoted 10 lines · Marcelo Tosatti
My plan for 2.4.22 is:
- Include the new aic7xxx driver.
- Include ACPI. (I now realized its importance). Already discussing with Andrew the best way to do it.
- Fix the latency/interactivity problems (Chris, Nick and Andrea working
on that)
- Merge obviously correct -aa VM patches.Those are the most important things that are needed now, I think.
How about the backported cpufreq patch from Bill Nottingham? Can that one go in as well please? It is not a big patch as such and it is working as far as I can tell. I have been running it since it was posted on the list last week.
Regards,
Signature · 2 lines
Anders Karlsson <[email protected]> Trudheim Technology Limited
Quoted 10 lines · Marcelo Tosatti
My plan for 2.4.22 is:
- Include the new aic7xxx driver.
- Include ACPI. (I now realized its importance). Already discussing with Andrew the best way to do it.
- Fix the latency/interactivity problems (Chris, Nick and Andrea working
on that)
- Merge obviously correct -aa VM patches.Those are the most important things that are needed now, I think.
I think many people on this list will be delighted to read this. I'm looking for the 22-pre to help testing them (UP)... As alan said, increasing the user test base may cause to find some remaining bugs but as many people are already forced to patch their 2.4 kernel with aic7xxx and ACPI just to make it useable most severe/obvious bugs have probably already been found...
I think SMP server users will also be happy with the rest of proposed fixes.
Side comments just to try to make 2.4 even better : I think polling the list to see the most annoying problems or most requested feature people have/want is sometime a good idea. Of course you will always get mail from people that what a patch to have support for their baroque hardware but I think you may reach a consensus rapidly as in the previous (rude) mail exchanges.
Hope you will manage to spend the needed time to make this in a timely fashion.
Signature · 7 lines
__ / ` Eric Valette /-- __ o _. 6 rue Paul Le Flem (___, / (_(_(__ 35740 Pace Tel: +33 (0)2 99 85 26 76 Fax: +33 (0)2 99 85 26 76 E-mail: [email protected]
Quoted 4 lines · Anders Karlsson
How about the backported cpufreq patch from Bill Nottingham? Can that one go in as well please? It is not a big patch as such and it is working as far as I can tell. I have been running it since it was posted on the list last week.
It requires the more substantial generic cpufreq backport first. :)
Bill
Quoted 31 lines · J.A. Magallon
Quoted 29 lines · Marcelo Tosatti
Quoted 7 lines · Eric Valette
Quoted 4 lines · Marcelo Tosati
The main reason I didnt want to merge it was due to its size. Its just too big.
Quoted 1 line
Please stop leading me along. Will you EVER merge it?
Quoted 8 lines
Quoted 1 line
Yes, I want to, and will merge it. In 2.4.23-pre.
This kind of mails and sentence makes you lost your credibility :
1) You said that ACPI will be merged in 2.4.22-pre,
2) For many people ACPI (and aic7xxx) is top priority for 2.4 kernel
(see various post including alan). The reason being that most laptop are unusable nowadays without ACPI,
3) You do not explicitely says what you plan for 2.4.22...My plan for 2.4.22 is:
- Include the new aic7xxx driver.
- Include ACPI. (I now realized its importance). Already discussing with Andrew the best way to do it.
- Fix the latency/interactivity problems (Chris, Nick and Andrea working
on that)
- Merge obviously correct -aa VM patches.
Quoted 4 lines
Quoted 2 lines
Those are the most important things that are needed now, I think.
Would you accept small canges like -march (
gcc_check) for x86 ?
I'm not aware of this patch. I might well accecpt it.
Please submit it with a good explanation and I will consider it.