Messages in this thread | | | Date | Mon, 27 Apr 2015 17:26:16 +0100 | From | Sudeep Holla <> | Subject | Re: [PATCH] bitmap: remove explicit newline handling using scnprintf format string |
| |
On 27/04/15 17:14, Tejun Heo wrote: > Hello, Sudeep. > > On Mon, Apr 27, 2015 at 10:46:58AM +0100, Sudeep Holla wrote: >> int bitmap_print_to_pagebuf(bool list, char *buf, const unsigned long *maskp, >> int nmaskbits) >> { >> - ptrdiff_t len = PTR_ALIGN(buf + PAGE_SIZE - 1, PAGE_SIZE) - buf - 2; >> + ptrdiff_t len = PTR_ALIGN(buf + PAGE_SIZE - 1, PAGE_SIZE) - buf; >> int n = 0; >> >> - if (len > 1) { >> - n = list ? scnprintf(buf, len, "%*pbl", nmaskbits, maskp) : >> - scnprintf(buf, len, "%*pb", nmaskbits, maskp); >> - buf[n++] = '\n'; >> - buf[n] = '\0'; >> - } >> + if (len > 1) >> + n = list ? scnprintf(buf, len, "%*pbl\n", nmaskbits, maskp) : >> + scnprintf(buf, len, "%*pb\n", nmaskbits, maskp); >> return n; >> } >> EXPORT_SYMBOL(bitmap_print_to_pagebuf); > > So, there's one behavior difference stemming from this. When the > buffer is too small, the original code would still output '\n' at the > end while the new code would just continue to print the formatted > string.
Completely agree and in-fact we did discuss that internally too. But since this function deals only with page size buffers, we thought it's highly unlikely to hit that corner case.
> Given that bitmap outputs can be pretty long, this behavior > difference has a minute but still non-zero chance of causing something > surprising. There are multiple copies of the above function in arch > codes too.
I assumed that I had consolidated most of them in commit 5aaba36318e5 ("cpumask: factor out show_cpumap into separate helper function"). I might have missed, will have a look at it again.
> We prolly want to audit the usages to verify that the > passed in buffer is always big enough at which point the above > function and its copies can simply be replaced with direct scnprintf() > calls. This function doesn't actually add anything. >
Ah, right that would be much simpler.
Regards, Sudeep
| |