Re: [PATCH] scsi: Let scsi_execute_cmd() mark args->sshdr as invalid

On 11.05.23 14:49, Bart Van Assche wrote:

On 5/11/23 05:34, Juergen Gross wrote:

...

Some callers of scsi_execute_cmd() (like e.g. sd_spinup_disk()) are passing an uninitialized struct sshdr and don't look at the return value of scsi_execute_cmd() before looking at the contents of that struct.

Shouldn't the scsi_execute_cmd() callers be fixed instead of modifying scsi_execute_cmd(), e.g. by zero-initializing the sshdr structure?

This would be possible, yes, but introducing new buggy callers could happen.

Additionally the amount of code churn would be much larger.

Juergen