From mboxrd@z Thu Jan 1 00:00:00 1970 From: Claude Code Review Bot To: dri-devel-reviews@example.com Subject: Claude review: misc: fastrpc: fix context leak and hang on signal-interrupted invoke Date: Tue, 26 May 2026 07:14:36 +1000 Message-ID: In-Reply-To: <20260525124222.3082420-1-anandu.e@oss.qualcomm.com> References: <20260525124222.3082420-1-anandu.e@oss.qualcomm.com> X-Mailer: Claude Code Patch Reviewer Content-Type: text/plain; charset="utf-8" Content-Transfer-Encoding: quoted-printable MIME-Version: 1.0 Overall Series Review Subject: misc: fastrpc: fix context leak and hang on signal-interrupted inv= oke Author: Anandu Krishnan E Patches: 2 Reviewed: 2026-05-26T07:14:36.687555 --- This is a single-patch fix for a real and well-described bug: when a signal= interrupts `wait_for_completion_interruptible()` during a fastrpc invoke, = the context leaks and retrying causes DSP hangs due to duplicate invocation= s. The approach of saving interrupted contexts on a per-user list and resto= ring them on retry is reasonable and follows the existing code patterns. The patch also replaces the old `invoke_interrupted_mmaps` mechanism (which= only saved DMA buffers at the channel level) with context-level preservati= on at the user level =E2=80=94 a more correct granularity. However, there are several issues ranging from a potential correctness bug = to missing handling for edge cases. **Verdict: Needs revision** =E2=80=94 the matching logic has a correctness = issue with PID vs TID semantics, the `handle` parameter is not checked in t= he restore path, and the mmaps leak concern is not addressed. --- Generated by Claude Code Patch Reviewer