XPost: linux.debian.bugs.dist   
   From: 10dmar10@gmail.com   
      
   Am 04.02.26 um 18:10 schrieb Ben Hutchings:   
   > On Sat, 2026-01-31 at 12:26 +0100, 10dmar10 wrote:   
   >> Am 31.01.26 um 11:02 schrieb Salvatore Bonaccorso:   
   >>   > Hi,   
   >>   >   
   >>   > On Sat, Jan 31, 2026 at 10:08:25AM +0100, 10dmar10 wrote:   
   > [...]   
   >>   >> while no longer used by default, XFS V4 support is still useful for   
   accessing   
   >>   >> legacy XFS V4 file systems as long as kernel supports it.   
   >>   >   
   >>   > This is correct, the default was changed in upstream with f69260511c69   
   >>   > ("xfs: disable deprecated features by default in Kconfig") and I   
   >>   > believe we should follow that within the forky release cycle.   
   >>   >   
   >>   > Thus for forky I do not think we should diverge here from the default   
   >>   > and re-enable deprecated features, but probably we should document it   
   >>   > at least in the release notes so that people upgrading from trixie to   
   >>   > forky are awaere that V4 filesystems will not be supported anymore.   
   >>   >   
   >>   > A NEWS entry in the packaging itself might be a good idea as well.   
   >>   >   
   >>   > Regards,   
   >>   > Salvatore   
   >>   
   >> Hi,   
   >>   
   >> I wrote this bug report because I assumed the release interval for forky as   
   2027   
   >> -2030, which is within the kernel removal date for XFS V4 feature in 2030.   
   >   
   > Debian stable releases are supported for 5 years on the most popular   
   > architectures.  So for forky this should be 2027-2032.   
   >   
   >> Enabling CONFIG_XFS_SUPPORT_V4 would allow to continue running otherwise   
   >> perfectly fine existing legacy XFS V4 filesystems without unnecessary   
   breaking   
   >> anything within forky's release cycle...   
   >   
   > Those filesystems must be quite old at this point, since v5 appears to   
   > have been the default for new filesystems since xfsprogs 3.2.3 and   
   > Debian 9 "stretch".  And they do need to be upgraded at some point due   
   > to the upstream EOL (and later, Y2038).  How much would we gain by   
   > putting it off another 2 years?   
   >   
   > I agree with Salvatore that this just needs to be documented.   
   >   
   > Ben.   
   >   
      
   Hi,   
      
   my point is not about "How much would we gain by putting it off another 2   
   years?", but about not loosing an existing reliably working capability for   
   another 4 years. AFAIK the only significant change between V4 and V5 XFS is crc   
   for metadata scrubbing...   
      
   --- SoupGate-Win32 v1.05   
    * Origin: you cannot sedate... all the things you hate (1:229/2)