Gustavo A. R. Silva wrote: > > > On 25/06/25 15:08, [email protected] wrote: > > dan.j.williams@ wrote: > >> Gustavo A. R. Silva wrote: > >> [..] > >>>> I think it would be a pretty small and direct replacement: > >>>> > >>>> TRAILING_OVERLAP(struct nd_cmd_pkg, pkg, nd_payload, > >>>> struct nd_intel_get_security_state cmd; > >>>> ) nd_cmd = { > >>>> ... > >>> > >>> Yes, this works. Hopefully, maintainers will comment on this and let us > >>> know what they prefer. :) > >> > >> Hey Gustavo, apologies for the latency here. I think TRAILING_OVERLAP() > >> looks lovely for this if only because I can read that and have an idea > >> what it means vs wondering what this _offset_to_fam is about and needing > >> to read the comment. > >> > >> If you can get me that patch on top of the TRAILING_OVERLAP() branch I > >> can test it out and ack it to let it do in through the KSPP tree. > > > > Just to move this along, I gave this conversion a try and all looks good > > here. So feel free to fold this in and add: > > > > Acked-by: Dan Williams <[email protected]> > > Tested-by: Dan Williams <[email protected]> > > > > ...and take it through the KSPP tree with the TRAILING_OVERLAP() merge. > > Thank you, Dan! :)
Uh oh, I spoke too soon. I saw most of the tests pass when I sent that mail, but the firmware-update.sh regresses. It passes on vanilla v6.16-rc3 and fails with both the original open-coded _offset_to_fam[] approach and TRAILING_OVERLAP() conversion. Let me try to get some more debug info. The test is: meson test -C build firmware-update.sh ...from the ndctl project: https://github.com/pmem/ndctl
