Hi, On 12/23/22 19:12, Linus Torvalds wrote:
On Fri, Dec 23, 2022 at 10:00 AM Ævar Arnfjörð Bjarmason <avarab@xxxxxxxxx> wrote:Whereas the fix here is a fix for a promise we're currently making which hasn't been true since v4.18.Hah. Ack. Did anybody ever actually notice? I wonder if the newer limit of 64 characters for kworkers shouldn't even be mentioned at all, and if the 16-byte truncation for user space should also be just removed. Those limits should never have been some documented API, they were always just implementation details, after all. Linus
I agree. A variable implementation detail like this doesn't provide anything valuable to users; especially since there's no statbility promise at all. I'd rewrite to just remove the (16) implementation detail.
Ævar, would you send an v2 that removes implementation details, rather than fixing the details?
Thanks for the patch set! Cheers, Alex On 12/23/22 18:59, Ævar Arnfjörð Bjarmason wrote: > diff --git a/man5/proc.5 b/man5/proc.5 > index 115c8592e..b23dd1479 100644 > --- a/man5/proc.5 > +++ b/man5/proc.5> @@ -2092,9 +2092,13 @@ The filename of the executable, in parentheses. Tools such as
> may alternatively (or additionally) use > .IR/proc/ pid /cmdline. > .IP > -Strings longer than > +For userspace, strings longer than > .B TASK_COMM_LEN > (16) characters (including the terminating null byte) are silently truncated. > +Since Linux version 4.18.0 a longer limit of 64 (including the > +terminating null byte) has applied to the kernel's own workqueue > +workers (whose names start with "kworker/"). > +.IP > This is visible whether or not the executable is swapped out. > .TP > (3) \fIstate\fP \ %c -- <http://www.alejandro-colomar.es/>
Attachment:
OpenPGP_signature
Description: OpenPGP digital signature
