diff options
| author | Christian Brauner <brauner@kernel.org> | 2026-08-20 01:09:30 +0200 |
|---|---|---|
| committer | Christian Brauner <brauner@kernel.org> | 2026-09-22 13:25:01 +0200 |
| commit | a769e2e7114266e40a62b483360c73d83da82e11 (patch) | |
| tree | 5ba1a1a53d90635ae25393b4b44a9d79c0d9480c /tools/perf/scripts/python | |
| parent | 98a2606e329542625a67d4d980778581148c7e0a (diff) | |
coredump: add COREDUMP_RECORDS to the coredump socket protocol
Currently a coredump sent over a socket is raw data. The kernel knows
things about the data it's sending that are useful for a coredump
server. For example, it knows where the unpopulated parts of a mapping
are. We can't communicate this to userspace currently though.
Add a COREDUMP_RECORDS feature bit and a struct coredump_record_header
so userspace can negotiate that feature. Instead of a byte stream it
gets a header plus data. Reassembling the records yields the same
coredump that would have been sent without them.
The record itself is also versioned and thus extensible with the same
protocol as the ack-req sync.
A record stream ends explicitly. A COREDUMP_RECORD_END record closes it,
carries no data and reports the size of the coredump. It is only sent
once the whole coredump has been written. If it's missing the coredump
should be treated as truncated.
This just adds the infrastructure.
Link: https://patch.msgid.link/20260820-work-coredump-sparse-v2-13-ba32dd718c51@kernel.org
Signed-off-by: Christian Brauner (Amutable) <brauner@kernel.org>
Diffstat (limited to 'tools/perf/scripts/python')
0 files changed, 0 insertions, 0 deletions
