On 1/6/26 6:21 PM, Timur Tabi wrote:
> On Tue, 2026-01-06 at 18:10 -0800, John Hubbard wrote:
>> nova-core (dev_*!) [after patch] 39 32 55%
>> nova-core (dev_*!) [before patch] 32 39 45%
>
> This can't be right, because I checked thoroughly to make sure that my patch
> fixed ALL dev_xx
> format strings.
Did you? haha :)
drivers/gpu/nova-core/fb.rs:81-85
dev_warn!(
&self.device,
"failed to unregister sysmem flush page: {:?}",
e
)
drivers/gpu/nova-core/gsp/boot.rs:83-87
dev_err!(
dev,
"FWSEC-FRTS returned with error code {:#x}",
frts_status
);
drivers/gpu/nova-core/gsp/cmdq.rs:616-620
dev_err!(
self.dev,
"GSP RPC: receive: Call {} - bad checksum",
header.sequence()
);
drivers/gpu/nova-core/gsp/sequencer.rs:391-395
dev_err!(
sequencer.dev,
"Error running command at index {}",
sequencer.seq_info.cmd_index
);
drivers/gpu/nova-core/gsp/sequencer.rs:401-404
dev_dbg!(
sequencer.dev,
"CPU Sequencer commands completed successfully"
);
However, you *are* correct about the "this can't be right". I had a bug
in my search script. Updated numbers do actually lean much more
toward using the \n on screen:
Codebase WITH \n WITHOUT \n % with \n
----------------------------------------------------------------
samples/rust (dev_*!) 39 1 98%
samples/rust (pr_*!) 29 6 83%
nova-core (dev_*!) 68 5 93%
nova (dev_*!) 0 0 --
tyr (dev_*!) 4 3 57%
pwm (dev_*!) 7 2 78%
binder (pr_*!) 22 57 28%
thanks,
--
John Hubbard
