Date: Sat, 24 Dec 2022 11:09:54 +0000 From: "Alexander V. Chernikov" <melifaro@ipfw.ru> To: Mateusz Guzik <mjguzik@gmail.com> Cc: Zhenlei Huang <zlei.huang@gmail.com>, freebsd-jail@freebsd.org Subject: Re: Is it possible to employ epoch to simplify managing prison lifecycle Message-ID: <E608BF71-9B4E-49EF-86DE-AFBDF180C2CE@ipfw.ru> In-Reply-To: <CAGudoHEncQfMXuG1UHrGRbaHchVfp2pC79mZvHNBUDh-XNhDFA@mail.gmail.com> References: <9BD54A54-A809-4D3E-BCBA-639E6C61FE37@FreeBSD.org> <CAGudoHHe8zo%2B7x4Myhotj60BJsASuV109Aj_Rkhg95RNVsTeaw@mail.gmail.com> <4E70D6D2-4E80-4AAD-BB3C-9295F586D1FF@ipfw.ru> <CAGudoHEncQfMXuG1UHrGRbaHchVfp2pC79mZvHNBUDh-XNhDFA@mail.gmail.com>
next in thread | previous in thread | raw e-mail | index | archive | help
> On 23 Dec 2022, at 15:27, Mateusz Guzik <mjguzik@gmail.com> wrote: >=20 > On 12/23/22, Alexander V. Chernikov <melifaro@ipfw.ru> wrote: >>=20 >>=20 >>> On 16 Dec 2022, at 16:29, Mateusz Guzik <mjguzik@gmail.com> wrote: >>>=20 >>> On 12/16/22, Zhenlei Huang <zlei.huang@gmail.com> wrote: >>>> Hi, >>>>=20 >>>> While hacking `sys/kern/kern_jail.c` I got lost. >>>>=20 >>>> There're lots of ref / unref and flags to prevent visit invalid = prison >>>> while >>>> concurrent modification is possible and some refs looks weird. >>>>=20 >>>> Is it possible to employ epoch(9) to simplify managing of prison >>>> lifecycle >>>> ? >>>>=20 >>>=20 >>> Some of the ref/unref cycles are probably avoidable to begin with, = but >>> ultimately the thing to do here is to employ per-cpu reference >>> counting, if at all needed. >>>=20 >>> I have a wip patch to provide such a mechanism, it may or may not = land >>> this month. >> That would be nice. I=E2=80=99d love to convert nextops refcounting = to that one. >> Do you envision similar semantics as Linux percpu_ref? I mean, does = one need >> to explicitly mark =E2=80=9Cnot in active use=E2=80=9D stage? >=20 > There *something* needed to disable per-cpu operation, otherwise how > can you ever know if the count is 0, apart from going over all cpus > every time, which defeats the point. Ack, sounds reasonable. Happy to test the KPI once it=E2=80=99s = available. >=20 > More specifically, I have a on/off switch for said per-cpu op. This is > modeled after what I did for counters in vfs, see vfs_ref et al. >=20 > --=20 > Mateusz Guzik <mjguzik gmail.com> >=20
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?E608BF71-9B4E-49EF-86DE-AFBDF180C2CE>