On Thu Oct 15, 2026 at 12:45 AM JST, Vladislav Zaharov wrote: > The GSP-RM log buffers are exposed through debugfs, but the Scope that > owns them lives in Gsp, inside GspResources, inside the Gpu built by > probe(). They are DMA allocations of the device and cannot outlive it, > so the entries go away as soon as the GPU is unbound - and, more to the > point, as soon as probe() fails, which is exactly when the log of a GSP > that did not come up is the thing one wants to read. > > Add a gsp_keep_logs module parameter. When it is set, dropping the log > buffers copies whatever the GSP wrote into memory owned by the module > and exposes the copies until the module is unloaded. A buffer whose > "put" pointer is still zero was never written to and is skipped. > > The GSP has normally been stopped by the time the buffers are dropped, > but a boot that timed out can leave it still appending, so a DMA read > barrier orders the read of the "put" pointer before the copy, and that > value, not the one the copy reads again, is what the copy keeps.
That paragraph is nice as a code comment, but that's an implementation detail that doesn't belong to a commit log. The commit log should tell us *why* you are doing this, not go into excruciating details of *how* you did it - only the big lines to make review easier. I'd appreciate if you could do a pass on the commit logs and code comments to deslop them and make them shorter and to the point (similarly to what I tried to do on patch 1). Right now this patch is pretty heavy to read. > > The copies belong to the module data next to the debugfs root, and live > in a "retained" directory created during module init rather than on > first use, which keeps the teardown path of a device from having to > create anything. Keeping them out of the directory used by bound GPUs > also means a device coming back does not find its debugfs name taken by > its own history; nouveau, which recreates the entries under the name of > the GPU that just went away, has that problem. What Nouveau does is irrevelant - that's the kind of unneeeded information I'd like us to avoid. > > While at it, move the log buffer code out of gsp.rs into gsp/logbuffer.rs. "While at it" means this should be two patches: a first one that moves the code into `gsp/logbuffer.rs`, a second one that performs the changes you need there. Right now this moves things while changing them, making it more difficult to review than it could be.
