Date: Wed, 9 Oct 2019 10:47:04 -0400 From: Mark Johnston <markj@freebsd.org> To: Hans Petter Selasky <hps@selasky.org> Cc: Yuri Pankov <yuripv@yuripv.net>, freebsd-net <freebsd-net@freebsd.org>, glebius@freebsd.org Subject: Re: panic: sleeping in an epoch section Message-ID: <20191009144704.GD66126@raichu> In-Reply-To: <a2075acc-243c-da14-180e-686eaf59cfd6@selasky.org> References: <86cc5d82-50d0-93eb-5900-54e8b0032a08@yuripv.net> <050ba95e-d0d5-dd1a-db6f-9a5c07142efe@selasky.org> <20191009135616.GC66126@raichu> <a2075acc-243c-da14-180e-686eaf59cfd6@selasky.org>
next in thread | previous in thread | raw e-mail | index | archive | help
On Wed, Oct 09, 2019 at 04:18:34PM +0200, Hans Petter Selasky wrote: > On 2019-10-09 15:56, Mark Johnston wrote: > > On Wed, Oct 09, 2019 at 10:40:04AM +0200, Hans Petter Selasky wrote: > >> On 2019-10-09 06:36, Yuri Pankov wrote: > >>> Tried updating from r353072 to r353334 and getting the following panic > >>> reproducibly on boot (starting dhclient?): > >>> > >>> panic: sleeping in an epoch section > >>> cpuid = 5 > >>> time = 1570591558 > >>> KDB: stack backtrace: > >>> db_trace_self_wrapper() at db_trace_self_wrapper+0x2b/frame > >>> 0xfffffe00af780140 > >>> vpanic() at vpanic+0x19d/frame 0xfffffe00af780190 > >>> panic() at panic+0x43/frame 0xfffffe00af7801f0 > >>> _sleep() at _sleep+0x463/frame 0xfffffe00af780290 > >>> pause_sbt() at pause_sbt+0x10f/frame 0xfffffe00af7802d0 > >>> e1000_write_phy_reg_mdic() at e1000_write_phy_reg_mdic+0xee/frame > >>> 0xfffffe00af780310 > >>> e1000_enable_phy_wakeup_reg_access_bm() at > >>> e1000_enable_phy_wakeup_reg_access_bm+0x2b/frame 0xfffffe00af780330 > >>> e1000_update_mc_addr_list_pch2lan() at > >>> e1000_update_mc_addr_list_pch2lan+0x3a/frame 0xfffffe00af780370 > >>> em_if_multi_set() at em_if_multi_set+0x1d4/frame 0xfffffe00af7803c0 > >>> iflib_if_ioctl() at iflib_if_ioctl+0x100/frame 0xfffffe00af780430 > >>> if_addmulti() at if_addmulti+0x2af/frame 0xfffffe00af7804d0 > >>> in_joingroup_locked() at in_joingroup_locked+0x235/frame 0xfffffe00af780570 > >>> in_joingroup() at in_joingroup+0x5c/frame 0xfffffe00af7805d0 > >>> in_control() at in_control+0xadf/frame 0xfffffe00af780680 > >>> ifioctl() at ifioctl+0x40f/frame 0xfffffe00af780750 > >>> kern_ioctl() at kern_ioctl+0x295/frame 0xfffffe00af7807b0 > >>> sys_ioctl() at sys_ioctl+0x15d/frame 0xfffffe00af780880 > >>> amd64_syscall() at amd64_syscall+0x2b9/frame 0xfffffe00af7809b0 > >>> fast_syscall_common() at fast_syscall_common+0x101/frame 0xfffffe00af7809b0 > >>> --- syscall (54, FreeBSD ELF64, sys_ioctl), rip = 0x80048051a, rsp = > >>> 0x7fffffffe3e8, rbp = 0x7fffffffe430 --- > >> > >> The SIOCADDMULTI if_ioctl() is not allowed to sleep, because it can be > >> called from the fast-path, so this is a bug in e1000 driver. Does the > >> attached patch workaround the issue? > > > > What fast path are you referring to? The locking protocol used by the > > multicast code was changed specifically to allow for sleeps in driver > > ioctl handlers. > > I recall a long time ago seeing that input packet processing may end up > calling if_ioctl's . Things may have changed since then though. That may be true in general, but I can't see any instances of that for SIOCADDMULTI or SIOCDELMULTI. I think we should always permit ioctl handlers to sleep. In particular, the panic reported above is a bug in r353292.
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20191009144704.GD66126>