Messages in this thread | | | Date | Fri, 1 Oct 2021 16:12:00 -0700 | From | Mike Rapoport <> | Subject | Re: [PATCH] memblock: Neaten logging |
| |
On Fri, Oct 01, 2021 at 01:34:10PM -0700, Joe Perches wrote: > On Fri, 2021-10-01 at 09:37 -0700, Mike Rapoport wrote: > > On Thu, Sep 30, 2021 at 06:27:15PM -0700, Joe Perches wrote: > > > On Thu, 2021-09-30 at 14:12 -0700, Mike Rapoport wrote: > > > > Hi Joe, > > > > > > > > On Wed, Sep 29, 2021 at 09:43:14PM -0700, Joe Perches wrote: > > > > > Use more typical kernel logging styles. > > > > > > > > > > o Add and use #define pr_fmt KBUILD_MODNAME ": " fmt > > > > > > > > I don't see it as an improvement. On the contrary, the output becomes > > > > somewhat tautological: > > > > > > And rather easier to grep as the prefix is constant. > > > > memblock_ is perfectly greppable > > Of course, but only when it's there, 2 instances out of 9.
I didn't object to the patch as a whole. I just don't like the pr_fmt definition.
> $ git grep '\bpr_' mm/memblock.c > mm/memblock.c: pr_warn("Could not allocate %pap bytes of mirrored memory\n", > mm/memblock.c: pr_err("memblock: Failed to double %s array from %ld to %ld entries !\n", > mm/memblock.c: pr_warn("Could not allocate %pap bytes of mirrored memory\n",
These can have %s for __func__ as a prefix,
> mm/memblock.c: pr_warn("%s: No memory registered yet\n", __func__);
this one already has it,
> mm/memblock.c: pr_info(" %s.cnt = 0x%lx\n", type->name, type->cnt); > mm/memblock.c: pr_info(" %s[%#x]\t[%pa-%pa], %pa bytes%s flags: %#x\n", > mm/memblock.c: pr_info("MEMBLOCK configuration:\n"); > mm/memblock.c: pr_info(" memory size = %pa reserved size = %pa\n",
and these are just fine now even without the prefix.
-- Sincerely yours, Mike.
| |