On Wed, Feb 12, 2025 at 05:10:42PM -0800, Khazhy Kumykov wrote:
> Heya, I'm backporting 752863bddaca ("block: propagate partition
> scanning errors to the BLKRRPART ioctl") to LTS and noticed it
> conflicts in the blkdev.h header, where in upstream we had already
> introduced another blk_mode_t for bit 5, and this new STRICT_SCAN uses
> bit 6...
>
> In this scenario, would we prefer keep the bit used consistent (so
> have a gap with an unused bit 5 - what I would typically go with), or
> renumber to avoid the gap?
It doesn't really matter as there is no in-kernel ABI But just keeping
the upstream value is probably going to create less confusion going
ahead.
