Looks like Linux does not have (dropped?) _DIR, just `STATX_ATTR_AUTOMOUNT`

this is stx_attributes_mask

for what it's worth memset(&x, 0, sizeof(x)) is slightly more common in kernel now than bzero(&x, sizeof(x))

Thanks! Somehow it didn't occur to me to compare the struct layout.

for what it's worth memset(&x, 0, sizeof(x)) is slightly more common in kernel now than bzero(&x, sizeof(x))

In this file it's mostly bzero(), though. I think we should leave it as it is, for consistency.

small optimisatio: may be mask alredy known flags? to see which is bit is unsupported?

indeed a minor point, but the message reads as if it implies the printed flags are all unsupported so I think it makes sense to either mask the known flags or make the message something like "statx flags 0x%x includes unsupported flags"

I like the idea of masking off the recognized flags, leaving just the unrecognized ones. However, I'd prefer to do it in a followup commit, as we have at least a dozen places which do that, and I think it's better to fix them together.

agree, Im asking as glibc-2.32 statx test fails, unsupported flags are: 0x2000 and 0x4000

For reference, those are AT_STATX_FORCE_SYNC (0x2000) and AT_STATX_DONT_SYNC (0x4000).

AT_STATX_FORCE_SYNC
       Force the attributes to be synchronized with the server.
       This may require that a network filesystem perform a data
       writeback to get the timestamps correct.

AT_STATX_DONT_SYNC
       Don't synchronize anything, but rather just take whatever
       the system has cached if possible.  This may mean that the
       information returned is approximate, but, on a network
       filesystem, it may not involve a round trip to the server
       - even if no lease is held.