Re: svn commit: r277796 - in head/sys: dev/acpica dev/syscons dev/vt sys
On 27/01/2015 19:47, Konstantin Belousov wrote: On Tue, Jan 27, 2015 at 05:33:19PM +, Andriy Gapon wrote: +EVENTHANDLER_INVOKE(power_suspend_early); +stop_all_proc(); It also makes sense to call syncer_suspend() after stop_all_proc(), I suppose. It would ensure that a failed resume still ends with the consistent filesystem metadata. Probably yes. I can make that change if you endorse it. Or you can make it if that'll be faster. Thanks! -- Andriy Gapon ___ svn-src-head@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/svn-src-head To unsubscribe, send any mail to svn-src-head-unsubscr...@freebsd.org
Re: svn commit: r277796 - in head/sys: dev/acpica dev/syscons dev/vt sys
On 28/01/2015 00:36, Ivan Klymenko wrote: acpiconf -s 3 works perfectly, but there is one problem. I use not the main timecounter kern.timecounter.hardware=HPET after turning on the power button does not occur laptop recovery hdac0 and it is likely that the same applies to the timer. Indirect evidence of this - it's a quick video playback, for example in flash on YouTube. dmesg_first : http://pastebin.com/a4gC5PGy dmesg_afterS3 : http://pastebin.com/vyHtrakZ my HW: pciconf -lvbce : https://bz-attachments.freebsd.org/attachment.cgi?id=148947 devinfo -vr : https://bz-attachments.freebsd.org/attachment.cgi?id=148948 Is that a new problem introduced by this commit? I have a hard time connecting the described symptoms to the code affected by this commit. -- Andriy Gapon ___ svn-src-head@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/svn-src-head To unsubscribe, send any mail to svn-src-head-unsubscr...@freebsd.org
Re: svn commit: r277796 - in head/sys: dev/acpica dev/syscons dev/vt sys
On Wed, Jan 28, 2015 at 03:53:34PM +0200, Andriy Gapon wrote: On 27/01/2015 19:47, Konstantin Belousov wrote: On Tue, Jan 27, 2015 at 05:33:19PM +, Andriy Gapon wrote: +EVENTHANDLER_INVOKE(power_suspend_early); +stop_all_proc(); It also makes sense to call syncer_suspend() after stop_all_proc(), I suppose. It would ensure that a failed resume still ends with the consistent filesystem metadata. Probably yes. I can make that change if you endorse it. Or you can make it if that'll be faster. Thanks! I prefer that the change is made by somebody who uses and tests the suspension code, i.e. you. ___ svn-src-head@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/svn-src-head To unsubscribe, send any mail to svn-src-head-unsubscr...@freebsd.org
Re: svn commit: r277796 - in head/sys: dev/acpica dev/syscons dev/vt sys
В Tue, 27 Jan 2015 17:33:19 + (UTC) Andriy Gapon a...@freebsd.org пишет: Author: avg Date: Tue Jan 27 17:33:18 2015 New Revision: 277796 URL: https://svnweb.freebsd.org/changeset/base/277796 Log: hook userland threads suspend + resume into acpi suspend code Also, split power_suspend into power_suspend and power_suspend_early. power_suspend_early is called before the userland is frozen. power_suspend is called after the userland is frozen. Currently only VT switching is hooked to power_suspend_early. This is needed because switching away from X server requires its cooperation, so obviously X server must not be frozen when that happens. Freezing userland during ACPI suspend is useful because not all drivers correctly handle suspension concurrent with other activity. This is especially applicable to drivers ported from other operating systems that suspend all software activity between placing drivers and hardware into suspended state. In particular drm2/radeon (radeonkms) depends on the described procedure. The driver does not have any internal synchronization between suspension activities and processing of userland requests. Many thanks to kib for the code that allows to freeze and thaw all userland threads. Note that ideally we also need to park / inhibit (non-special) kernel threads as well to ensure that they do not call into drivers. MFC after: 17 days Thank you for your work! acpiconf -s 3 works perfectly, but there is one problem. I use not the main timecounter kern.timecounter.hardware=HPET after turning on the power button does not occur laptop recovery hdac0 and it is likely that the same applies to the timer. Indirect evidence of this - it's a quick video playback, for example in flash on YouTube. dmesg_first : http://pastebin.com/a4gC5PGy dmesg_afterS3 : http://pastebin.com/vyHtrakZ my HW: pciconf -lvbce : https://bz-attachments.freebsd.org/attachment.cgi?id=148947 devinfo -vr : https://bz-attachments.freebsd.org/attachment.cgi?id=148948 Thanks. ___ svn-src-head@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/svn-src-head To unsubscribe, send any mail to svn-src-head-unsubscr...@freebsd.org
svn commit: r277796 - in head/sys: dev/acpica dev/syscons dev/vt sys
Author: avg
Date: Tue Jan 27 17:33:18 2015
New Revision: 277796
URL: https://svnweb.freebsd.org/changeset/base/277796
Log:
hook userland threads suspend + resume into acpi suspend code
Also, split power_suspend into power_suspend and power_suspend_early.
power_suspend_early is called before the userland is frozen.
power_suspend is called after the userland is frozen.
Currently only VT switching is hooked to power_suspend_early.
This is needed because switching away from X server requires its
cooperation, so obviously X server must not be frozen when that happens.
Freezing userland during ACPI suspend is useful because not all drivers
correctly handle suspension concurrent with other activity. This is
especially applicable to drivers ported from other operating systems
that suspend all software activity between placing drivers and hardware
into suspended state.
In particular drm2/radeon (radeonkms) depends on the described
procedure. The driver does not have any internal synchronization
between suspension activities and processing of userland requests.
Many thanks to kib for the code that allows to freeze and thaw all
userland threads.
Note that ideally we also need to park / inhibit (non-special) kernel
threads as well to ensure that they do not call into drivers.
MFC after:17 days
Modified:
head/sys/dev/acpica/acpi.c
head/sys/dev/syscons/syscons.c
head/sys/dev/vt/vt_core.c
head/sys/sys/eventhandler.h
Modified: head/sys/dev/acpica/acpi.c
==
--- head/sys/dev/acpica/acpi.c Tue Jan 27 15:28:46 2015(r277795)
+++ head/sys/dev/acpica/acpi.c Tue Jan 27 17:33:18 2015(r277796)
@@ -2749,6 +2749,8 @@ acpi_EnterSleepState(struct acpi_softc *
return_ACPI_STATUS (AE_OK);
}
+EVENTHANDLER_INVOKE(power_suspend_early);
+stop_all_proc();
EVENTHANDLER_INVOKE(power_suspend);
if (smp_started) {
@@ -2892,6 +2894,8 @@ backout:
thread_unlock(curthread);
}
+resume_all_proc();
+
EVENTHANDLER_INVOKE(power_resume);
/* Allow another sleep request after a while. */
Modified: head/sys/dev/syscons/syscons.c
==
--- head/sys/dev/syscons/syscons.c Tue Jan 27 15:28:46 2015
(r277795)
+++ head/sys/dev/syscons/syscons.c Tue Jan 27 17:33:18 2015
(r277796)
@@ -549,7 +549,7 @@ sc_attach_unit(int unit, int flags)
/* Register suspend/resume/shutdown callbacks for the kernel console. */
if (sc_console_unit == unit) {
- EVENTHANDLER_REGISTER(power_suspend, scsuspend, NULL,
+ EVENTHANDLER_REGISTER(power_suspend_early, scsuspend, NULL,
EVENTHANDLER_PRI_ANY);
EVENTHANDLER_REGISTER(power_resume, scresume, NULL,
EVENTHANDLER_PRI_ANY);
Modified: head/sys/dev/vt/vt_core.c
==
--- head/sys/dev/vt/vt_core.c Tue Jan 27 15:28:46 2015(r277795)
+++ head/sys/dev/vt/vt_core.c Tue Jan 27 17:33:18 2015(r277796)
@@ -2556,8 +2556,8 @@ vt_upgrade(struct vt_device *vd)
vd-vd_timer_armed = 1;
/* Register suspend/resume handlers. */
- EVENTHANDLER_REGISTER(power_suspend, vt_suspend_handler, vd,
- EVENTHANDLER_PRI_ANY);
+ EVENTHANDLER_REGISTER(power_suspend_early, vt_suspend_handler,
+ vd, EVENTHANDLER_PRI_ANY);
EVENTHANDLER_REGISTER(power_resume, vt_resume_handler, vd,
EVENTHANDLER_PRI_ANY);
}
Modified: head/sys/sys/eventhandler.h
==
--- head/sys/sys/eventhandler.h Tue Jan 27 15:28:46 2015(r277795)
+++ head/sys/sys/eventhandler.h Tue Jan 27 17:33:18 2015(r277796)
@@ -182,6 +182,7 @@ EVENTHANDLER_DECLARE(shutdown_final, shu
typedef void (*power_change_fn)(void *);
EVENTHANDLER_DECLARE(power_resume, power_change_fn);
EVENTHANDLER_DECLARE(power_suspend, power_change_fn);
+EVENTHANDLER_DECLARE(power_suspend_early, power_change_fn);
/* Low memory event */
typedef void (*vm_lowmem_handler_t)(void *, int);
___
svn-src-head@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to svn-src-head-unsubscr...@freebsd.org
Re: svn commit: r277796 - in head/sys: dev/acpica dev/syscons dev/vt sys
On Tue, Jan 27, 2015 at 05:33:19PM +, Andriy Gapon wrote: Author: avg Date: Tue Jan 27 17:33:18 2015 New Revision: 277796 URL: https://svnweb.freebsd.org/changeset/base/277796 Log: hook userland threads suspend + resume into acpi suspend code Also, split power_suspend into power_suspend and power_suspend_early. power_suspend_early is called before the userland is frozen. power_suspend is called after the userland is frozen. Currently only VT switching is hooked to power_suspend_early. This is needed because switching away from X server requires its cooperation, so obviously X server must not be frozen when that happens. Freezing userland during ACPI suspend is useful because not all drivers correctly handle suspension concurrent with other activity. This is especially applicable to drivers ported from other operating systems that suspend all software activity between placing drivers and hardware into suspended state. In particular drm2/radeon (radeonkms) depends on the described procedure. The driver does not have any internal synchronization between suspension activities and processing of userland requests. Many thanks to kib for the code that allows to freeze and thaw all userland threads. Note that ideally we also need to park / inhibit (non-special) kernel threads as well to ensure that they do not call into drivers. MFC after: 17 days Modified: head/sys/dev/acpica/acpi.c head/sys/dev/syscons/syscons.c head/sys/dev/vt/vt_core.c head/sys/sys/eventhandler.h Thank you for fixing this. Modified: head/sys/dev/acpica/acpi.c == --- head/sys/dev/acpica/acpi.cTue Jan 27 15:28:46 2015 (r277795) +++ head/sys/dev/acpica/acpi.cTue Jan 27 17:33:18 2015 (r277796) @@ -2749,6 +2749,8 @@ acpi_EnterSleepState(struct acpi_softc * return_ACPI_STATUS (AE_OK); } +EVENTHANDLER_INVOKE(power_suspend_early); +stop_all_proc(); It also makes sense to call syncer_suspend() after stop_all_proc(), I suppose. It would ensure that a failed resume still ends with the consistent filesystem metadata. ___ svn-src-head@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/svn-src-head To unsubscribe, send any mail to svn-src-head-unsubscr...@freebsd.org
