Re: [PATCH net-next v2 1/2] selftests: net-drv: exercise queue stats when the device is down

[Stanislav Fomichev <sdf@fomichev.me>]

On 08/01, Petr Machata wrote:
> 
> Stanislav Fomichev <sdf@fomichev.me> writes:
> 
> > On 07/31, Jakub Kicinski wrote:
> >> On Wed, 31 Jul 2024 13:34:58 +0200 Petr Machata wrote:
> >> > > +        qstat = netfam.qstats_get({"ifindex": cfg.ifindex}, dump=True)
> >> > > +    except NlError as e:
> >> > > +        if e.error == 95:  
> >> > 
> >> > Could you do this as if e.error == errno.ENOTSUP?
> >> 
> >> just to be clear EOPNOTSUPP ..
> >
> > That might be the reason it's coded explicitly as 95? :-D
> 
> Both exist, I just didn't notice the latter.
> 
> >>> import errno
> >>> errno.ENOTSUP
> 95
> >>> errno.EOPNOTSUPP
> 95

I believe Jakub was talking about kernel's ENOTSUPP (524) vs EOPNOTSUPP (95):

$ grep ENOTSUPP include/linux/errno.h
#define ENOTSUPP        524     /* Operation is not supported */

$ grep EOPNOTSUPP include/uapi/asm-generic/errno.h
#define EOPNOTSUPP      95      /* Operation not supported on transport endpoint */

These two are frequently confused.

OTOH, ENOTSUP looks like a userspace/libc invention:

$ grep -w ENOTSUP /usr/include/bits/errno.h
# ifndef ENOTSUP
#  define ENOTSUP               EOPNOTSUPP

I'm gonna stick to kernel's EOPNOTSUPP to make it look similar to what
we have on the kernel side.