Date: Thu, 15 Dec 2016 13:38:11 -0800 From: John Baldwin <jhb@freebsd.org> To: Roger Pau =?ISO-8859-1?Q?Monn=E9?= <roger.pau@citrix.com> Cc: freebsd-arch@freebsd.org Subject: Re: Order of device suspend/resume Message-ID: <7469755.xT5lfhErkd@ralph.baldwin.cx> In-Reply-To: <20161215114033.r33nt3fqhnfi7hqw@dhcp-3-221.uk.xensource.com> References: <20161215114033.r33nt3fqhnfi7hqw@dhcp-3-221.uk.xensource.com>
next in thread | previous in thread | raw e-mail | index | archive | help
On Thursday, December 15, 2016 11:40:33 AM Roger Pau Monn=E9 wrote: > Hello, >=20 > I'm currently dealing with a bug in the Xen suspend/resume sequence, = and I've=20 > found that lacking a way to order device priority during suspend/resu= me is=20 > proving quite harmful for Xen (and maybe other systems too). The curr= ent=20 > suspend/resume code simply scans the root bus, and suspends/resumes e= very device=20 > based on the order they are attached to their parents. The problem he= re is that=20 > there's no way to tell that some devices should be resumed before oth= ers, for=20 > example the event timers/time counters/uarts should definitely be res= ume before=20 > other devices, but that's seems to happens mostly out of chance. >=20 > Currently most time related devices are attached directly to the nexu= s, which=20 > means they will get resumed first, but for example the uart is curren= tly=20 > attached to the pci bus IIRC, which means it gets resumed quite late.= On Xen=20 > systems, this is even worse. The Xen PV bus (that contains all Xen-re= lated=20 > devices) is attached the last one (because it tends to pick up unused= memory=20 > regions for it's own usage) and this bus also contains the PV timecou= nter which=20 > should be resumed _before_ other devices, or else timecounting will b= e=20 > completely screwed and things can get stuck in indefinitely long loop= s (due to=20 > the fact that the timecounter is implemented based on the uptime of t= he host,=20 > and that changes from host-to-host). >=20 > In order to solve this I could add a hack to the Xen resume process (= which is=20 > already different from the ACPI one), but this looks gross. I could a= lso attach=20 > the Xen PV timer to the nexus directly (as it was done before), but I= also=20 > prefer to keep all Xen-related devices in the same bus for coherency.= Last=20 > option would be to add some kind of suspend/resume priorities to the = devices,=20 > and do more than one suspend/resume pass. This is more complex and re= quires more=20 > changes, so I would like to know if it would be helpful for other sys= tems, or if=20 > someone has already attempted to do it. I think Justin Hibbits had some patches to make use of the boot-time ne= w-bus passes for suspend and resume which I think would help with this. You = suspend things in the reverse order of boot and resume operates in the same ord= er as boot. --=20 John Baldwin
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?7469755.xT5lfhErkd>