Date: Sat, 18 Jul 2020 22:20:10 +0200 From: "Kristof Provost" <kp@FreeBSD.org> To: "Konstantin Belousov" <kostikbel@gmail.com> Cc: src-committers@freebsd.org, svn-src-all@freebsd.org, svn-src-head@freebsd.org Subject: Re: svn commit: r363308 - head/sys/net Message-ID: <1E8C2818-9388-4722-A3A1-1F6B54A597A8@FreeBSD.org> In-Reply-To: <20200718184750.GY44314@kib.kiev.ua> References: <202007181243.06IChBTR073126@repo.freebsd.org> <20200718184750.GY44314@kib.kiev.ua>
next in thread | previous in thread | raw e-mail | index | archive | help
On 18 Jul 2020, at 20:47, Konstantin Belousov wrote: > On Sat, Jul 18, 2020 at 12:43:11PM +0000, Kristof Provost wrote: >> Author: kp >> Date: Sat Jul 18 12:43:11 2020 >> New Revision: 363308 >> URL: https://svnweb.freebsd.org/changeset/base/363308 >> >> Log: >> bridge: Don't sleep during epoch >> >> While it doesn't trigger INVARIANTS or WITNESS on head it does in >> stable/12. >> There's also no reason for it, as we can easily report the out of >> memory error >> to the caller (i.e. userspace). All of these can already fail. > This makes syscalls (ioctl) fail randomly. Can you pre-allocate the > buffers > before entering epoch, instead ? > Not easily, no. The bridge ioctl handling is all done via bridge_ioctl(), which enters epoch and dispatches to the bridge_control_table. We’d have to modify every single ioctl function. These are also not the only ioctl functions that can return ENOMEM (or EINVAL). bridge_ioctl_add() already did, for example. Best regards, Kristof
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?1E8C2818-9388-4722-A3A1-1F6B54A597A8>