Thread (63 messages) 63 messages, 7 authors, 2025-09-09

Re: [PATCH v6 5/6] tracing: Show inode and device major:minor in deferred user space stacktrace

From: Linus Torvalds <torvalds@linux-foundation.org>
Date: 2025-08-29 17:10:01
Also in: bpf, lkml

On Fri, 29 Aug 2025 at 09:57, Steven Rostedt [off-list ref] wrote:
The reason is to keep from triggering the event that records the pathname
for every look up.
BUT THAT WAS NEVER THE POINT.

There is only a single 64-bit number. No lookup. No pointer following.
No nothing.

The whole point of hashing was to get an *opaque* thing very quickly.
Not a pathname. No reference counting. No verifying whether you have
seen it before.

Literally just something that you can match up in the trace file much
much later.

(And, honestly, the likely thing is that you never match it up at all
- you can delay the "match it up" until a human actually looks at a
trace, which is presumably going to be a "one in a million" case).

              Linus
Keyboard shortcuts
hback out one level
jnext message in thread
kprevious message in thread
ldrill in
Escclose help / fold thread tree
?toggle this help