Messages in this thread | | | Date | Fri, 26 Aug 2011 16:27:21 +0200 | From | Tejun Heo <> | Subject | Re: [RFC] fs, proc: Introduce the /proc/<pid>/map_files/ directory v2 |
| |
Hello,
On Fri, Aug 26, 2011 at 05:23:34PM +0300, Kirill A. Shutemov wrote: > On Fri, Aug 26, 2011 at 04:06:20PM +0200, Tejun Heo wrote: > > > +struct map_files_info { > > > + struct file *file; > > > + unsigned char name[16+16+2]; /* max: %016lx-%016lx\0 */ > > > + unsigned long len; > > > +}; > > > > That's slightly above 50 bytes. > > We can use 4 * sizeof(unsigned long) + 2 for name to save a bit on 32 bit > architectures.
Ooh, I wasn't trying to say it was too big. I was trying to say it would be okay to have that on stack. :)
> BTW, show_map_vma() uses "%08lx-%08lx" to print ->start, ->end. I think > it's good idea to use this here as well to be consistent with > /proc/$PID/{s,}maps
I'm not too sure about this one. First of all, I think %08lx is more of an accident than intention (just forgot to update it while transitioning to 64bit) and when ls'd the output can be a bit messy without leading 0's, but again my opinion here isn't very strong.
Thanks.
-- tejun
| |