Re: [PATCH] spi: remove return value check of debugfs_create_dir()

[Greg Kroah-Hartman <gregkh@xxxxxxxxxxxxxxxxxxx>]

On Mon, Apr 24, 2023 at 03:54:13PM +0200, Peter Enderborg wrote:
> On 4/24/23 15:22, Greg Kroah-Hartman wrote:
> > On Mon, Apr 24, 2023 at 03:17:09PM +0200, Peter Enderborg wrote:
> >> On 4/24/23 14:53, Greg Kroah-Hartman wrote:
> >>>>> We can do things with the debug information without filesystem enabled.
> >>> What exactly do you mean by this?
> >>>
> >>>
> >> We can read out data from kernel with a ramdumper and analyse with crash.
> >>
> >> See https://access.redhat.com/documentation/en-us/red_hat_enterprise_linux/7/html/kernel_administration_guide/kernel_crash_dump_guide 
> >>
> >> If it is useful or not I can not say, but the dws->regset. is lost and can not be read with a post mortem debugger.
> > What is "dws"?  What is "regset"?
> 
> That is from the patch. It is used as an example.
> 
> 
> > What is the root problem here?
> 
> That it is a mater of taste.  It should not be a mater of taste.
> 
>       if (!dws->debugfs)
>                 return -ENOMEM;

Right here, you abort the normal operation of the driver if something
went wrong with debugfs, which is not a good idea.  That's the goal
here, nothing else.

thanks,

greg k-h