On Wed, Apr 30, 2025 at 10:34:59AM +0200, Markus Elfring wrote: > … > >> +++ b/fs/xfs/scrub/repair.c > >> @@ -382,7 +382,7 @@ xrep_calc_ag_resblks( > >> refcbt_sz); > >> xfs_perag_put(pag); > >> > >> - return max(max(bnobt_sz, inobt_sz), max(rmapbt_sz, refcbt_sz)); > >> + return max3(bnobt_sz, inobt_sz, max(rmapbt_sz, refcbt_sz)); > > > > I have nothing against the patch itself, but honestly I don't see how it > > improves anything. It boils down to nesting comparison instructions too, and > > doesn't make the code more clear IMHO. > > So, unless somebody else has a stronger reason to have this change, NAK from my side. > Would you be looking for a wrapper call variant like max4()? I have no preference really, between a max(max(), max()) and a max4(a, b, c, d), the latter is a tad easier to the eyes, if it's worth adding a new max() macro for that, it's another thing. Although a quick search on the source code returned returned several usages of the max(max3(a,b,c), d) patterns, so I think indeed the kernel could benefit of a max4() :) > > Regards, > Markus
