Messages in this thread | | | Date | Tue, 05 Jul 2011 12:16:16 +0100 | From | "Jan Beulich" <> | Subject | Re: [PATCH] x86/EFI: properly pre-initialize table pointers |
| |
>>> On 05.07.11 at 11:33, Ingo Molnar <mingo@elte.hu> wrote:
> * Jan Beulich <JBeulich@novell.com> wrote: > >> Consumers of the table pointers in struct efi check for >> EFI_INVALID_TABLE_ADDR to determine validity, hence these pointers >> should all be pre-initialized to this value (rather than zero). >> >> Signed-off-by: Jan Beulich <jbeulich@novell.com> >> >> --- >> arch/x86/platform/efi/efi.c | 12 +++++++++++- >> 1 file changed, 11 insertions(+), 1 deletion(-) > > This changelog is missing some key information: > > - how did you find the bug (by chance via code review or did you see > some actual badness?)
I can certainly add that information and re-send, but would certainly have stated any really bad effect if I had observed such.
> - what practical effect (if any) did you see from this patch? > > - what practical effect (if any) do you expect others to see from this > patch?
Hmm, I can't really see why this kind of information (especially the last) would belong into a changeset comment, the more a pretty obvious one like this.
> This kind of information helps us prioritize bugfixes.
Bug fixes are bug fixes - unless they fix really critical issues (and here that's unlikely to be the case, as the code had been wrong for ages), I don't follow why you require more information to be added than a description of what gets changed (the "why" should really only matter if it's debatable whether the original code was wrong).
Furthermore, this being not the first time you ask (from my pov) to be overly verbose when it comes to describing changes and their effects - is it unreasonable to expect you (not always you personally, but you as being one of the three x86 maintainers; you hopefully agree that from my position it doesn't really matter who answers as long as someone assumes that responsibility) to be a little less lax in responding to submitted patches? I would roughly estimate that 50% of the patches I send get dropped on the floor with no response whatsoever (there are even cases where you welcomed patches, but never applied them or responded with a revised opinion on a re- submission). I clearly realize that the volume you need to deal with must be pretty high, yet I don't think that ought to be an excuse for other than occasional cases of this happening.
And yes, there are also other times when things get integrated *very* quickly, and I definitely appreciate that.
Jan
> Thanks, > > Ingo
| |