From mboxrd@z Thu Jan 1 00:00:00 1970 From: Claude Code Review Bot To: dri-devel-reviews@example.com Subject: Claude review: gpu: nova-core: gsp: shuffle boot code a bit to keep chipset-specific parts close Date: Sat, 16 May 2026 09:54:08 +1000 Message-ID: In-Reply-To: <20260515-nova-unload-v5-5-c4d6250ad160@nvidia.com> References: <20260515-nova-unload-v5-0-c4d6250ad160@nvidia.com> <20260515-nova-unload-v5-5-c4d6250ad160@nvidia.com> X-Mailer: Claude Code Patch Reviewer Content-Type: text/plain; charset="utf-8" Content-Transfer-Encoding: 7bit MIME-Version: 1.0 Patch Review This patch reorders two things: 1. Moves `wpr_meta` allocation before `run_fwsec_frts()` -- benign, just memory allocation. 2. Moves `SetSystemInfo` and `SetRegistry` commands from **before** GSP falcon boot to **after** RISC-V active confirmation: ```rust - self.cmdq - .send_command_no_wait(bar, commands::SetSystemInfo::new(pdev))?; - self.cmdq - .send_command_no_wait(bar, commands::SetRegistry::new())?; - gsp_falcon.reset(bar)?; ... + self.cmdq + .send_command_no_wait(bar, commands::SetSystemInfo::new(pdev))?; + self.cmdq + .send_command_no_wait(bar, commands::SetRegistry::new())?; ``` The commit message says "This has no effect on the GSP boot process" which is plausible since these are `send_command_no_wait` calls that queue messages for the GSP to process later (before `wait_gsp_init_done`). However, the original placement (before falcon boot) means the commands are in the queue when the GSP starts up and begins reading. The new placement (after falcon boot) means the GSP starts, begins reading the queue, and these commands arrive while it's already running. The functional result should be the same since the GSP waits for these before signaling init done, but this is a subtle ordering change worth explicit confirmation in review that it's safe. --- --- Generated by Claude Code Patch Reviewer