Skip to contentSkip to conversation

Revert "sna: Refresh last detection timestamp on hotplug notifies"

PATCH In review Lyude 4 participants 5 messages Jun 17, 2016 – Jun 27, 2016

lore
Share & tools
lore

Lyude

author
#0 lore raw

PATCH v2 1/1 Revert "sna: Refresh last detection timestamp on hotplug notifies"

From: Lyude Paul <[email protected]>

DRM does not always update the status of each connector during a hotplug event, and it's generally expected that userspace is supposed to handle that by reprobing. This happens in a couple situations:
suspend/resume, MST hotplugs, and probably a few others. As a result, making this assumption actually breaks MST hotplugging.

Changes since v1:
- Fix removal of break statements from switch case

---
 src/sna/sna_display.c | 14 +++++---------
 1 file changed, 5 insertions(+), 9 deletions(-)
src/sna/sna_display.c +5 -9
index 0cf2bdb..79c72cc 100644
@@ -5114,7 +5114,7 @@ void sna_mode_discover(struct sna *sna, bool tell)
 	ScreenPtr screen = xf86ScrnToScreen(sna->scrn);
 	xf86CrtcConfigPtr config = XF86_CRTC_CONFIG_PTR(sna->scrn);
 	struct drm_mode_card_res res;
-	uint32_t connectors[32], now;
+	uint32_t connectors[32];
 	unsigned changed = 0;
 	unsigned serial;
 	int i, j;
@@ -5146,7 +5146,6 @@ void sna_mode_discover(struct sna *sna, bool tell)
 	if (serial == 0)
 		serial = ++sna->mode.serial;
 
-	now = GetTimeInMillis();
 	for (i = 0; i < res.count_connectors; i++) {
 		DBG(("%s: connector[%d] = %d\n", __FUNCTION__, i, connectors[i]));
 		for (j = 0; j < sna->mode.num_real_output; j++) {
@@ -5172,13 +5171,10 @@ void sna_mode_discover(struct sna *sna, bool tell)
 			continue;
 
 		if (sna_output->serial == serial) {
-			if (output_check_status(sna, sna_output)) {
-				DBG(("%s: output %s (id=%d), retained state\n",
-				     __FUNCTION__, output->name, sna_output->id));
-				sna_output->last_detect = now;
-			} else {
-				DBG(("%s: output %s (id=%d), changed state, reprobing\n",
-				     __FUNCTION__, output->name, sna_output->id));
+			if (!output_check_status(sna, sna_output)) {
+				DBG(("%s: output %s (id=%d), changed state, reprobing]\n",
+				     __FUNCTION__, output->name, sna_output->id,
+				     sna_output->serial, serial));
 				sna_output->last_detect = 0;
 				changed |= 4;
 			}
Signature · 6 lines
2.5.5

_______________________________________________
Intel-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/intel-gfx
#1 lore raw
Quoted 7 lines · Lyude

From: Lyude Paul <[email protected]>

DRM does not always update the status of each connector during a hotplug event, and it's generally expected that userspace is supposed to handle that by reprobing. This happens in a couple situations:
suspend/resume, MST hotplugs, and probably a few others. As a result, making this assumption actually breaks MST hotplugging.

Generally expected? So logind reprobes? We force a probe on wakeup? -Chris

Signature · 5 lines
Chris Wilson, Intel Open Source Technology Centre
_______________________________________________
Intel-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/intel-gfx
author
#2 lore raw
Quoted 10 lines · Chris Wilson
Quoted 9 lines · Lyude

From: Lyude Paul <[email protected]>

DRM does not always update the status of each connector during a hotplug event, and it's generally expected that userspace is supposed to
handle that by reprobing. This happens in a couple situations:
suspend/resume, MST hotplugs, and probably a few others. As a result,

This was meant more to be an example of when DRM connector states aren't consistent. This being said; i915 does do a reprobe on resuming so that is still true. Just to make sure; you saw the IRC convo on this airlied/danvet had right?

Quoted 4 lines
Quoted 1 line

making this assumption actually breaks MST hotplugging.

Generally expected? So logind reprobes? We force a probe on wakeup? -Chris

Signature · 7 lines
Cheers,
	Lyude

_______________________________________________
Intel-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/intel-gfx
#3 lore raw
In reply to Chris Wilson
Quoted 10 lines · Chris Wilson
Quoted 7 lines · Lyude

From: Lyude Paul <[email protected]>

DRM does not always update the status of each connector during a hotplug event, and it's generally expected that userspace is supposed to handle that by reprobing. This happens in a couple situations:
suspend/resume, MST hotplugs, and probably a few others. As a result, making this assumption actually breaks MST hotplugging.

Generally expected? So logind reprobes? We force a probe on wakeup?

Yes, expected by the API from day one.

If you want an API to do something else, write a new API. Don't go shoehorning behaviour into a generic API that is driver specific.

So please fix the DDX or I'll push the patch and you can revert it if you don't like but at least
then I know you are being painful and I can justify carrying the patch upstream until the
Intel DDX is no longer something we care about.

Dave.
_______________________________________________
Intel-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/intel-gfx

#4 lore raw

✗ Ro.CI.BAT: failure for Revert "sna: Refresh last detection timestamp on hotplug notifies" (rev2)

In reply to Lyude
== Series Details ==

== Series Details ==

Series: Revert "sna: Refresh last detection timestamp on hotplug notifies" (rev2)
URL : https://patchwork.freedesktop.org/series/7369/
State : failure

== Summary ==

Applying: Revert "sna: Refresh last detection timestamp on hotplug notifies"
fatal: sha1 information is lacking or useless (src/sna/sna_display.c).
error: could not build fake ancestor
Patch failed at 0001 Revert "sna: Refresh last detection timestamp on hotplug notifies" The copy of the patch that failed is found in: .git/rebase-apply/patch When you have resolved this problem, run "git am --continue".
If you prefer to skip this patch, run "git am --skip" instead.
To restore the original branch and stop patching, run "git am --abort".

_______________________________________________
Intel-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/intel-gfx