On Wed Mar 4, 2026 at 8:25 PM JST, Gary Guo wrote: > On Wed Mar 4, 2026 at 2:46 AM GMT, Eliot Courtney wrote: >> Fix some inaccuracies / old doc comments. >> >> Reviewed-by: Zhi Wang <[email protected]> >> Tested-by: Zhi Wang <[email protected]> >> Signed-off-by: Eliot Courtney <[email protected]> >> --- >> drivers/gpu/nova-core/gsp/cmdq.rs | 17 ++++++++--------- >> 1 file changed, 8 insertions(+), 9 deletions(-) >> >> diff --git a/drivers/gpu/nova-core/gsp/cmdq.rs >> b/drivers/gpu/nova-core/gsp/cmdq.rs >> index 492e9489e808..4829830b6921 100644 >> --- a/drivers/gpu/nova-core/gsp/cmdq.rs >> +++ b/drivers/gpu/nova-core/gsp/cmdq.rs >> @@ -531,6 +531,7 @@ fn notify_gsp(bar: &Bar0) { >> /// >> /// # Errors >> /// >> + /// - `EMSGSIZE` if the command exceeds the maximum queue element size. >> /// - `ETIMEDOUT` if space does not become available within the timeout. >> /// - `EIO` if the variable payload requested by the command has not >> been entirely > > What's the benefit of enumerating all the error codes like this? Unless all > the > mentioned error code here is supposed to be handled, then it doesn't gain much > for them to be mentioned, no? > > For the errors that do need special handling, we probably want to use enums to > force their handling.
Agreed about using enums, that's a change we will want to do driver-wide at some point. Preferably once we have an equivalent of `thiserror` to make the transition easier. Meanwhile I think it is expected to document the returned error codes, as (1) they provide an easy way to lookup the reason for runtime errors instead of grepping the code, and (2) they can eventually be converted into the doccomments of the enum error types once we switch to them.
