Messages in this thread |  | | Date | Sun, 13 Mar 2022 21:09:58 +0300 | From | Alexey Dobriyan <> | Subject | Re: [PATCH 1/5] x86/alternative: simplify DUMP_BYTES macro |
| |
On Sat, Mar 12, 2022 at 08:36:11AM -0800, Joe Perches wrote: > On Fri, 2022-03-11 at 17:43 +0300, Alexey Dobriyan wrote: > > Avoid zero length check with clever whitespace placement in the format > > string. > [] > > diff --git a/arch/x86/kernel/alternative.c b/arch/x86/kernel/alternative.c > [] > > @@ -66,13 +66,10 @@ do { \ > > if (unlikely(debug_alternative)) { \ > > int j; \ > > \ > > - if (!(len)) \ > > - break; \ > > - \ > > printk(KERN_DEBUG pr_fmt(fmt), ##args); \ > > - for (j = 0; j < (len) - 1; j++) \ > > - printk(KERN_CONT "%02hhx ", buf[j]); \ > > - printk(KERN_CONT "%02hhx\n", buf[j]); \ > > + for (j = 0; j < (len); j++) \ > > + printk(KERN_CONT " %02hhx", buf[j]); \ > > + printk(KERN_CONT "\n"); \ > > } \ > > This could also use %02x and not %02hhx
I doubt as there is funky stuff possible with 255 and such values. Format specifiers aren't the purpose of the patch anyway.
> And MAX_PATCH_LEN is 255 but is that really possible?
Yes if you try hard enough.
> Maybe if the actual patch length is always <= 64 this could use > printk(KERN_CONT "%*ph\n", (int)len, buf); > instead and avoid all possible interleaving?
It is for debugging feature nobody uses (because it works).
|  |