2.6.7-rc1-mm1, Simplify DMI matching data
A message that is not in the archive (sent off-list or not yet received): [email protected]
Quoted 2 lines · Andrey Panin
simplify DMI blacklist table by removing the need to fill unused slots with
NO_MATCHmacro.
Can you please delay that patch for 2.7?
2.6 is for bug fixes, not for cleanups.
There are large third party patchkits for DMI and "cleaning up" the format will just cause lots of rejects and pain. A stable kernel is supposed to be somewhat stable in internal interfaces too.
Thanks,
-Andi
Quoted 8 lines · Andi Kleen
Quoted 2 lines
simplify DMI blacklist table by removing the need to fill unused slots with
NO_MATCHmacro.Can you please delay that patch for 2.7?
2.6 is for bug fixes, not for cleanups.There are large third party patchkits for DMI and "cleaning up" the format will just cause lots of rejects and pain.
Alternatively, those third parties could get their act together and submit those patches back upstream.
Dave
Quoted 5 lines · Dave Jones
Quoted 3 lines · Andi Kleen
Quoted 2 lines
simplify DMI blacklist table by removing the need to fill unused slots with
NO_MATCHmacro.
Quoted 8 lines
Quoted 5 lines
Can you please delay that patch for 2.7?
2.6 is for bug fixes, not for cleanups.There are large third party patchkits for DMI and "cleaning up" the format will just cause lots of rejects and pain.
Alternatively, those third parties could get their act together and submit those patches back upstream.
Often this is not the best thing to do - e.g. for upstream it is better to track down the bugs and try to fix them, even if that takes a long time or find some other cleaner solution that doesn't involve blacklisting. For a third party there are often time constraints (e.g. for a release) where there is no time to track down everything and blacklisting has to be more extensively used.
My point stays that kernel interfaces should stay stable in the stable series as far as possible (= unless terminally broken, but that's clearly not the case here). If you feel the need to clean up something better wait for the unstable series.
-Andi
Quoted 4 lines · Andi Kleen
My point stays that kernel interfaces should stay stable in the stable series as far as possible (= unless terminally broken, but that's clearly not the case here). If you feel the need to clean up something better wait for the unstable series.
So what interface doews the patch break? I can only see that it adds interfaces.
Quoted 15 lines · Andi Kleen
Quoted 5 lines · Dave Jones
Quoted 3 lines · Andi Kleen
Quoted 2 lines
simplify DMI blacklist table by removing the need to fill unused slots with
NO_MATCHmacro.Quoted 8 lines
Quoted 5 lines
Can you please delay that patch for 2.7?
2.6 is for bug fixes, not for cleanups.There are large third party patchkits for DMI and "cleaning up" the format will just cause lots of rejects and pain.
Alternatively, those third parties could get their act together and submit those patches back upstream.
Quoted 6 lines
Often this is not the best thing to do - e.g. for upstream it is better to track down the bugs and try to fix them, even if that takes a long time or find some other cleaner solution that doesn't involve blacklisting. For a third party there are often time constraints (e.g. for a release) where there is no time to track down everything and blacklisting has to be more extensively used.
See the next patch then, it should make life of third party developers much simpler. Also I can modify the patch to provide NO_MATCH constant, so there will be no visible differencies.
Quoted 4 lines
My point stays that kernel interfaces should stay stable in the stable series as far as possible (= unless terminally broken, but that's clearly not the case here). If you feel the need to clean up something better wait for the unstable series.
I can't call dmi_scan.c a kernel interface, currently it's a crap.
Signature · 2 lines
Andrey Panin | Linux and UNIX system administrator [email protected] | PGP key: wwwkeys.pgp.net
Quoted 6 lines
Quoted 4 lines
My point stays that kernel interfaces should stay stable in the stable series as far as possible (= unless terminally broken, but that's clearly not the case here). If you feel the need to clean up something better wait for the unstable series.
I can't call
dmi_scan.ca kernel interface, currently it's a crap.
Disagree. It works just fine in its current form, your patch doesn't fix any bugs.
-Andi
Quoted 9 lines · Andi Kleen
Quoted 6 lines
Quoted 4 lines
My point stays that kernel interfaces should stay stable in the stable series as far as possible (= unless terminally broken, but that's clearly not the case here). If you feel the need to clean up something better wait for the unstable series.
I can't call
dmi_scan.ca kernel interface, currently it's a crap.Disagree. It works just fine in its current form, your patch doesn't fix any bugs.
Actually it does. The ASUS P4P800 entry is very broken. Andrey's patches inadvertantly fix it by doing away with the necessity for NO_MATCH entries.
Dave
Quoted 3 lines
Actually it does. The ASUS P4P800 entry is very broken. Andrey's patches inadvertantly fix it by doing away with the necessity for
NO_MATCHentries.
Well, several thousand lines of changes for a trivial bug fix that could be done much simpler also doesn't seem like the right approach for bug fixing in a stable kernel.
We can only hope that 2.7 opens soon, so that all those "I really need to clean up something NOW" people leave 2.6 alone, so that it can finally mature and become stable.
-Andi
Quoted 1 line · Andi Kleen
There are large third party patchkits for DMI
I think you just identified the problem...
Quoted 1 line
and "cleaning up" the format will just cause lots of rejects and pain.
And this isn't it. ;)
Signature · 3 lines
"Debugging is twice as hard as writing the code in the first place. Therefore, if you write the code as cleverly as possible, you are, by definition, not smart enough to debug it." - Brian W. Kernighan
Hi!
Quoted 5 lines
Quoted 2 lines
simplify DMI blacklist table by removing the need to fill unused slots with
NO_MATCHmacro.Can you please delay that patch for 2.7?
2.6 is for bug fixes, not for cleanups.
Current DMI setup is terminally broken by keeping whole blacklist in one place. Fixing it very good idea.
Pavel
Signature · 1 line
934a471f20d6580d5aad759bf0d97ddc