Ah right make sense indeed. On Sat, 8 Jan 2022 at 09:42, Илья Шипицин <[email protected]> wrote: > > David, there are minor failures after this patch > BUG/MEDIUM: mworker: don't use _getsocks in wait mode · > haproxy/haproxy@f82afbb (github.com) > > пт, 31 дек. 2021 г. в 09:33, David CARLIER <[email protected]>: >> >> Hi, >> >> all CPU macros which were incompatible with Linux (ie CPU_AND*, >> CPU_*OR...) had been changed and there is no backward compatibility >> with the old BSD api (at least not for now, they might introduce >> detection like solaris does for ex, FreeBSD 14 is not yet released) >> ultimately once FreeBSD 14 will become the LTS version the HAProxy >> cpuset code could be greatly simplified. >> I would suggest to backport up to the 2.4, that should be fine. >> >> Thanks. >> >> On Fri, 31 Dec 2021 at 06:18, Willy TARREAU <[email protected]> wrote: >> > >> > Hi David, >> > >> > On Fri, Dec 31, 2021 at 05:02:37AM +0000, David CARLIER wrote: >> > > Here a simpler version if that s fine with you. >> > >> > Indeed, even simpler and I also prefer this one. I've now merged it, >> > thank you. Just to be sure, was this a breaking change in FreeBSD or >> > is the old API still supported ? I'm trying to figure if we have to >> > backport this one to stable releases in fact. >> > >> > BTW, I still hadn't had a look at your BPF patchset by lack of time, >> > but I will do soon. Rest assured they're not lost :-) >> > >> > Thanks, >> > Willy >>
0001-BUILD-MINOR-haproxy-build-fix-introduced-by-previous.patch
Description: Binary data
