Re: [PATCH 2/2] x86: modify show_shared_cpu_map in intel_cacheinfo

[Mike Travis <travis@sgi.com>]

Bert Wesarg wrote:
> On Fri, Mar 28, 2008 at 3:40 PM, Mike Travis <travis@sgi.com> wrote:
>> Bert Wesarg wrote:
>>  > On Fri, Mar 28, 2008 at 12:16 AM, Mike Travis <travis@sgi.com> wrote:
>>  >> Used cpulist_scnprintf to print cpus on a leaf instead of requiring
>>  >>  a new "cpumask_scnprintf_len" function to determine the size of
>>  >>  the temporary buffer.  cpulist_scnprintf can be used to print directly
>>  >>  to the output buffer, eliminating the need for the temporary buffer.
>>  >>
>>  >>  Based on:
>>  >>         git://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux-2.6.git
>>  >>         git://git.kernel.org/pub/scm/linux/kernel/git/x86/linux-2.6-x86.git
>>  >>
>>  >>  Signed-off-by: Mike Travis <travis@sgi.com>
>>  >>  ---
>>  >>   arch/x86/kernel/cpu/intel_cacheinfo.c |   25 +++++++++++++++++--------
>>  >>   1 file changed, 17 insertions(+), 8 deletions(-)
>>  >>
>>  >>  --- linux.trees.git.orig/arch/x86/kernel/cpu/intel_cacheinfo.c
>>  >>  +++ linux.trees.git/arch/x86/kernel/cpu/intel_cacheinfo.c
>>  >>  @@ -593,14 +593,23 @@ static ssize_t show_size(struct _cpuid4_
>>  >>
>>  >>   static ssize_t show_shared_cpu_map(struct _cpuid4_info *this_leaf, char *buf)
>>  >>   {
>>  >>  -       int n = 0;
>>  >>  -       int len = cpumask_scnprintf_len(nr_cpu_ids);
>>  >>  -       char *mask_str = kmalloc(len, GFP_KERNEL);
>>  >>  -
>>  >>  -       if (mask_str) {
>>  >>  -               cpumask_scnprintf(mask_str, len, this_leaf->shared_cpu_map);
>>  >>  -               n = sprintf(buf, "%s\n", mask_str);
>>  >>  -               kfree(mask_str);
>>  >>  +       /*
>>  >>  +        * cpulist_scnprintf() has the advantage of compressing
>>  >>  +        * consecutive cpu numbers into a single range which seems
>>  >>  +        * appropriate for cpus on a leaf.  This will change what is
>>  >>  +        * output so scripts that process the output will have to change.
>>  > So this breaks user space?
>>  >
>>  > Bert
>>
>>  Potentially, yes.  But I suspect with 4096 cpus, user scripts will have
>>  to change anyways.  Currently it is represented as sets of 32 bit mask
>>  outputs with comma separators, so 1152 characters would be output.
> But you can declare it as a programming error on user space side. If
> you change the format, the brown-paper-bag is yours.
> 
>>  Is there a special notice I should provide to announce this change?
> I hope so. At least lwn.net has an API changes site:
> 
> http://lwn.net/Articles/2.6-kernel-api/

I did look at that site.   Besides being "kind of" out of date (last mod: 10/19/07),
it didn't appear to track changes in information displayed by proc/sysfs functions.

> 
> I also looked into MAINTAINERS, but it seems there is no official API
> 'maintainer'.
> 
>>  (And this output does conform with other syntax for printing and parsing
>>  strings of bits.)
> Aren't the most cpumaps (like cpu/cpu*/topology/*_siblings or
> node/node*/cpumap) bitmasks?

I did an informal survey and you are right, the majority of references do use
cpumask_scnprintf instead of cpulist_scnprintf.  Maybe the later function was
added later?

To me though, it would seem that:

240-255

is more readable than:

00000000,00000000,00000000,00000000,00000000,00000000,00000000,0000ffff

And as I mentioned, bitmask_parselist() [libbitmask(3)] does parse the output.

Thanks,
Mike