On Tue, 1 Apr 2008, Arnd Bergmann wrote:
> On Tuesday 01 April 2008, stephane eranian wrote:
> > > it into 2.6.26, now would be the time to post the patches.
> > > I can certainly offer some help getting it into shape for inclusion
> > > if you're interested.
> > >
> > I don't think the code is yet ready to start submitting patches. We need
> > some more work to isolate things more. As I said the area of fous right
> > now should be the generic to arch-specific interface. There is also another
> > set of issues with the PMu description modules and how they interface with
> > the arch-specific code. I would like to see more code pushed into the PMU
> > description modules.
>
> Ok, however I think it would be good to post the patches for review anyway,
> as soon as they are reasonably split up. It's important to let people know
> that someone is still working on them, and give them a chance to complain
> now, rather than waiting for the time of submission for people to give
> comments.
I second this: the sooner you post the patches for review, the sooner you will
receive valuable review comments, and the less rework is needed if (hopefully
none or not to many) deep changes have to be made.
> > > * include/linux/perfmon.h contains an #ifdef CONFIG_PERFMON, which is
> > > not valid for a user space header. Config symbols are never set in
> > > user space, so this doesn't work.
> > >
> > User code is never supposed to include kernel header files directly. Perfmon
> > applications instead use a "user-level" version of perfmon.h which is
> > provided
> > by libpfm. This is file is totally different from
> > include/linux/perfmon.h. But it
> > may be that the CONFIG_PERFMON is superfluous now that I have separated
> > the kernel-only piece. Will take a look.
>
> Since the perfmon.h file defines the ABI between kernel and user space, it
> would be better to have them identical, even if libpfm uses a copy instead
> of the version from /usr/include/linux.
>
> > > * You need to list in include/*/Kbuild the files that should be
> > > exported to user space, i.e. perfmon.h. Otherwise, "make headers_install"
> > > doesn't pick them up
> > >
> > Okay, so now I understand your earlier point about perfmon.h I did not know
> > about the headers_install option. Are those headers supposed to be used by
> > applications or just kernel modules?
>
> These exported headers are only meant for user space, not for modules. In
> fact, all parts hidden in #ifdef __KERNEL__ get removed in the headers_install
> stage, so you can't use them.
The goal is to make the exported headers the ones to be used by all userspace.
Currently several C libraries ships their own sanitized kernel headers, which
makes keeping everything in sync more complicated.
With kind regards,
Geert Uytterhoeven
Software Architect
Sony Network and Software Technology Center Europe
The Corporate Village · Da Vincilaan 7-D1 · B-1935 Zaventem · Belgium
Phone: +32 (0)2 700 8453
Fax: +32 (0)2 700 8622
E-mail: [EMAIL PROTECTED]
Internet: http://www.sony-europe.com/
Sony Network and Software Technology Center Europe
A division of Sony Service Centre (Europe) N.V.
Registered office: Technologielaan 7 · B-1840 Londerzeel · Belgium
VAT BE 0413.825.160 · RPR Brussels
Fortis Bank Zaventem · Swift GEBABEBB08A · IBAN BE39001382358619
-------------------------------------------------------------------------
Check out the new SourceForge.net Marketplace.
It's the best place to buy or sell services for
just about anything Open Source.
http://ad.doubleclick.net/clk;164216239;13503038;w?http://sf.net/marketplace
_______________________________________________
perfmon2-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/perfmon2-devel
