trace.pl fixes and improvements
cover trace.pl fixes and improvements
From: Tvrtko Ursulin <[email protected]>
Latest state of patches preceding the virtual engine support.
John Harrison (1):
trace.pl: Improved key/colours
Tvrtko Ursulin (8):
trace.pl: Improve time axis labels
trace.pl: Scale timeline for better precision
trace.pl: Improve readability of graphical timeline representation trace.pl: Improve context colouring for large context id's trace.pl: Context save only applies to last request of a bunch trace.pl: Fix incomplete request handling
trace.pl: Basic preemption support
trace.pl: Fix request split mode
scripts/media-bench.pl | 2 +- scripts/trace.pl | 541 ++++++++++++++++++++++++++++++++--------- 2 files changed, 425 insertions(+), 118 deletions(-)
Signature · 6 lines
2.17.1 _______________________________________________ Intel-gfx mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/intel-gfx
PATCH 1/9 trace.pl: Improve time axis labels
From: Tvrtko Ursulin <[email protected]>
It is possible to customize the axis display so change it to display timestamps in seconds on the major axis (with six decimal spaces) and relative millisecond offsets on the minor axis.
--- scripts/trace.pl | 32 ++++++++++++++++++++++++++++++++ 1 file changed, 32 insertions(+)
scripts/trace.pl +32 -0
@@ -1000,6 +1000,37 @@ $first_ts = ts($first_ts); print <<ENDHTML; ]); + var last_major_ms; + + function majorAxis(date, scale, step) { + var s = date / 1000; + + last_major_ms = date; + + if (scale == 'millisecond') + return s.toFixed(6) + "s"; + else if (scale == 'second') + return s.toFixed(3) + "s"; + else + return s.toFixed(0) + "s"; + } + + function minorAxis(date, scale, step) { + var ms = date; + + ms -= last_major_ms; + + if (ms < 0) + return ''; + + if (scale == 'millisecond') + return "+" + ms.toFixed(3) + "ms"; + else if (scale == 'second') + return ms.toFixed(3) + "s"; + else + return ms.toFixed(0) + "s"; + } + // Configuration for the Timeline var options = { groupOrder: 'content', horizontalScroll: true,
@@ -1007,6 +1038,7 @@ print <<ENDHTML; stackSubgroups: false, zoomKey: 'ctrlKey', orientation: 'top', + format: { majorLabels: majorAxis, minorLabels: minorAxis }, start: '$first_ts', end: '$end_ts'};
Signature · 6 lines
2.17.1 _______________________________________________ Intel-gfx mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/intel-gfx
Re: [igt-dev] [PATCH i-g-t 1/9] trace.pl: Improve time axis labels
Quoted 31 lines · Tvrtko Ursulin
From: Tvrtko Ursulin <[email protected]>
It is possible to customize the axis display so change it to display timestamps in seconds on the major axis (with six decimal spaces) and relative millisecond offsets on the minor axis.
Signed-off-by Tvrtko Ursulin <[email protected]>Suggested-by Chris Wilson <[email protected]>Cc Chris Wilson <[email protected]>Cc John Harrison <[email protected]>--- scripts/trace.pl | 32 ++++++++++++++++++++++++++++++++ 1 file changed, 32 insertions(+)scripts/trace.pl +9
-0@@ -1000,6 +1000,37 @@ $first_ts = ts($first_ts); print <<ENDHTML; ]); + var last_major_ms; + + function majorAxis(date, scale, step) { + var s = date / 1000; + + last_major_ms = date; + + if (scale == 'millisecond') + return s.toFixed(6) + "s";
scripts/trace.pl $first_ts = ts($first_ts); · line 1021
Quoted 10 lines
scripts/trace.pl +10
-0+ else if (scale == 'second') + return s.toFixed(3) + "s"; + else + return s.toFixed(0) + "s"; + } + + function minorAxis(date, scale, step) { + var ms = date; + + ms -= last_major_ms;
This logic is unfortunately a bit dodgy. It looked like major label is getting called before the minor ones, so I thought I could calculate the relative offset for the minor for a more readable display. But after scrolling and zooming around timelines some more I can see cases where that doesn't seem to be the case.
Please have a play and if you think it is bad they only option will be to absolute time for minor labels as well.
Regards,
Tvrtko
Quoted 17 lines
scripts/trace.pl +12
-0+ + if (ms < 0) + return ''; + + if (scale == 'millisecond') + return "+" + ms.toFixed(3) + "ms"; + else if (scale == 'second') + return ms.toFixed(3) + "s"; + else + return ms.toFixed(0) + "s"; + } + // Configuration for the Timeline var options = { groupOrder: 'content', horizontalScroll: true,
scripts/trace.pl print <<ENDHTML; · line 1044
Quoted 7 lines
scripts/trace.pl +1
-0@@ -1007,6 +1038,7 @@ print <<ENDHTML; stackSubgroups: false, zoomKey: 'ctrlKey', orientation: 'top', + format: { majorLabels: majorAxis, minorLabels: minorAxis }, start: '$first_ts', end: '$end_ts'};
_______________________________________________
igt-dev mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/igt-dev
Re: [igt-dev] [PATCH i-g-t 1/9] trace.pl: Improve time axis labels
Quoted 45 lines · Tvrtko Ursulin
Quoted 30 lines · Tvrtko Ursulin
From: Tvrtko Ursulin <[email protected]>
It is possible to customize the axis display so change it to display timestamps in seconds on the major axis (with six decimal spaces) and relative millisecond offsets on the minor axis.
Signed-off-by Tvrtko Ursulin <[email protected]>Suggested-by Chris Wilson <[email protected]>Cc Chris Wilson <[email protected]>Cc John Harrison <[email protected]>---
scripts/trace.pl| 32 ++++++++++++++++++++++++++++++++1 file changed, 32 insertions(+)scripts/trace.pl +8
-0@@ -1000,6 +1000,37 @@ $first_ts = ts($first_ts); print <<ENDHTML; ]); + var last_major_ms; + + function majorAxis(date, scale, step) { + var s = date / 1000; + + last_major_ms = date; + + if (scale == 'millisecond') + return s.toFixed(6) + "s";scripts/trace.pl $first_ts = ts($first_ts); · line 1020
Quoted 10 lines
scripts/trace.pl +10
-0+ else if (scale == 'second') + return s.toFixed(3) + "s"; + else + return s.toFixed(0) + "s"; + } + + function minorAxis(date, scale, step) { + var ms = date; + + ms -= last_major_ms;This logic is unfortunately a bit dodgy. It looked like major label is getting called before the minor ones, so I thought I could calculate the relative offset for the minor for a more readable display. But after scrolling and zooming around timelines some more I can see cases where that doesn't seem to be the case.
Quoted 1 line
Please have a play and if you think it is bad they only option will be to absolute time for minor labels as well.
Something is completely broken for me. I am getting massively out of whack timeline values. For example, a log file entry of 761846.983987 is being displayed as -59196268.216013s! Even ignoring that, I'm not convinced about the relative offset thing. It makes it much harder to correlate the graph with the trace file entries, e.g. if trying to work out exactly what happened when something went peculiar in a test. Would it be possible to make it dynamically switchable between raw absolute values and the funky change-as-you-scroll relative ones?
Also, it doesn't seem especially useful to have the values be relative to the left most edge of the visible portion of the timeline. That means that when you scroll sideways along the timeline, the time-stamp of any given point wanders up and down accordingly. That just seems confusing to me! Maybe it makes sense for some use cases but I would have thought that having the values be relative to the start of the trace would be more use. Or even have clickable anchor points so that you can set the origin to be whatever exciting event you are interested in? Although that might be way too complex to be worth the effort of implementing :).
Quoted 20 lines
Regards,
Tvrtko
Quoted 15 lines
scripts/trace.pl +12
-0+ + if (ms < 0) + return ''; + + if (scale == 'millisecond') + return "+" + ms.toFixed(3) + "ms"; + else if (scale == 'second') + return ms.toFixed(3) + "s"; + else + return ms.toFixed(0) + "s"; + } + // Configuration for the Timeline var options = { groupOrder: 'content', horizontalScroll: true,
Quoted 8 lines
scripts/trace.pl print <<ENDHTML; · line 1044
Quoted 8 lines
scripts/trace.pl +1
-0@@ -1007,6 +1038,7 @@ print <<ENDHTML; stackSubgroups: false, zoomKey: 'ctrlKey', orientation: 'top', + format: { majorLabels: majorAxis, minorLabels: minorAxis }, start: '$first_ts', end: '$end_ts'};
_______________________________________________
igt-dev mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/igt-dev
PATCH v2 1/9 trace.pl: Improve time axis labels
From: Tvrtko Ursulin <[email protected]>
It is possible to customize the axis display so change it to display timestamps in seconds on the major axis (with six decimal spaces) and millisecond offsets on the minor axis.
v2:
* Give up on broken relative timestamps.
--- scripts/trace.pl | 37 +++++++++++++++++++++++++++++++++++++ 1 file changed, 37 insertions(+)
scripts/trace.pl +37 -0
@@ -1000,6 +1000,42 @@ $first_ts = ts($first_ts); print <<ENDHTML; ]); + function majorAxis(date, scale, step) { + var s = date / 1000; + var precision; + + if (scale == 'millisecond') + precision = 6; + else if (scale == 'second') + precision = 3; + else + precision = 0; + + return s.toFixed(precision) + "s"; + } + + function minorAxis(date, scale, step) { + var ms = date; + var precision; + var unit; + + if (scale == 'millisecond') { + ms %= 1000; + precision = 0; + unit = 'ms'; + } else if (scale == 'second') { + ms /= 1000; + precision = 1; + unit = 's'; + } else { + ms /= 1000; + precision = 0; + unit = 's'; + } + + return ms.toFixed(precision) + unit; + } + // Configuration for the Timeline var options = { groupOrder: 'content', horizontalScroll: true,
@@ -1007,6 +1043,7 @@ print <<ENDHTML; stackSubgroups: false, zoomKey: 'ctrlKey', orientation: 'top', + format: { majorLabels: majorAxis, minorLabels: minorAxis }, start: '$first_ts', end: '$end_ts'};
Signature · 6 lines
2.17.1 _______________________________________________ igt-dev mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/igt-dev
Re: [Intel-gfx] [PATCH i-g-t v2 1/9] trace.pl: Improve time axis labels
Quoted 65 lines · Tvrtko Ursulin
From: Tvrtko Ursulin <[email protected]>
It is possible to customize the axis display so change it to display timestamps in seconds on the major axis (with six decimal spaces) and millisecond offsets on the minor axis.
v2:
* Give up on broken relative timestamps.Signed-off-by Tvrtko Ursulin <[email protected]>Suggested-by Chris Wilson <[email protected]>Cc Chris Wilson <[email protected]>Cc John Harrison <[email protected]>--- scripts/trace.pl | 37 +++++++++++++++++++++++++++++++++++++ 1 file changed, 37 insertions(+)scripts/trace.pl +36
-0@@ -1000,6 +1000,42 @@ $first_ts = ts($first_ts); print <<ENDHTML; ]); + function majorAxis(date, scale, step) { + var s = date / 1000; + var precision; + + if (scale == 'millisecond') + precision = 6; + else if (scale == 'second') + precision = 3; + else + precision = 0; + + return s.toFixed(precision) + "s"; + } + + function minorAxis(date, scale, step) { + var ms = date; + var precision; + var unit; + + if (scale == 'millisecond') { + ms %= 1000; + precision = 0; + unit = 'ms'; + } else if (scale == 'second') { + ms /= 1000; + precision = 1; + unit = 's'; + } else { + ms /= 1000; + precision = 0; + unit = 's'; + } + + return ms.toFixed(precision) + unit; + } + // Configuration for the Timeline var options = { groupOrder: 'content', horizontalScroll: true,
scripts/trace.pl print <<ENDHTML; · line 1049
Quoted 7 lines
scripts/trace.pl +1
-0@@ -1007,6 +1043,7 @@ print <<ENDHTML; stackSubgroups: false, zoomKey: 'ctrlKey', orientation: 'top', + format: { majorLabels: majorAxis, minorLabels: minorAxis }, start: '$first_ts', end: '$end_ts'};
I'm still seeing some kind of strange offset. However, it appears to be browser dependent. If I use Chrome then the offset is +28.8 seconds. With Firefox it is -59958115.2 seconds! On the other hand, if I try Edge or IE then I don't get a graph at all. I'm wondering if the issue is with Vis browser compatibility rather than anything in the trace.pl script. Are you seeing anything at all similar?
Hmm, if I comment out the 'format:' line and go back to the unformatted time stamps then IE & Edge still show nothing. However, Firefox shows dates based on a year of 0097 whereas Chrome says 1997.
Either way, I can't spot anything in this patch that could cause a random offset. So...
_______________________________________________
Intel-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/intel-gfx
Re: [Intel-gfx] [igt-dev] [PATCH i-g-t v2 1/9] trace.pl: Improve time axis labels
Quoted 76 lines · John Harrison
Quoted 65 lines · Tvrtko Ursulin
From: Tvrtko Ursulin<[email protected]>
It is possible to customize the axis display so change it to display timestamps in seconds on the major axis (with six decimal spaces) and millisecond offsets on the minor axis.
v2:
* Give up on broken relative timestamps.Signed-off-by Tvrtko Ursulin<[email protected]>Suggested-by Chris Wilson<[email protected]>Cc Chris Wilson<[email protected]>Cc John Harrison<[email protected]>--- scripts/trace.pl | 37 +++++++++++++++++++++++++++++++++++++ 1 file changed, 37 insertions(+)scripts/trace.pl +36
-0@@ -1000,6 +1000,42 @@ $first_ts = ts($first_ts); print <<ENDHTML; ]); + function majorAxis(date, scale, step) { + var s = date / 1000; + var precision; + + if (scale == 'millisecond') + precision = 6; + else if (scale == 'second') + precision = 3; + else + precision = 0; + + return s.toFixed(precision) + "s"; + } + + function minorAxis(date, scale, step) { + var ms = date; + var precision; + var unit; + + if (scale == 'millisecond') { + ms %= 1000; + precision = 0; + unit = 'ms'; + } else if (scale == 'second') { + ms /= 1000; + precision = 1; + unit = 's'; + } else { + ms /= 1000; + precision = 0; + unit = 's'; + } + + return ms.toFixed(precision) + unit; + } + // Configuration for the Timeline var options = { groupOrder: 'content', horizontalScroll: true,scripts/trace.pl print <<ENDHTML; · line 1049
Quoted 7 lines
scripts/trace.pl +1
-0@@ -1007,6 +1043,7 @@ print <<ENDHTML; stackSubgroups: false, zoomKey: 'ctrlKey', orientation: 'top', + format: { majorLabels: majorAxis, minorLabels: minorAxis }, start: '$first_ts', end: '$end_ts'};I'm still seeing some kind of strange offset. However, it appears to be browser dependent. If I use Chrome then the offset is +28.8 seconds. With Firefox it is -59958115.2 seconds! On the other hand, if I try Edge or IE then I don't get a graph at all. I'm wondering if the issue is with Vis browser compatibility rather than anything in the trace.pl script. Are you seeing anything at all similar?
Quoted 3 lines
Hmm, if I comment out the 'format:' line and go back to the unformatted time stamps then IE & Edge still show nothing. However, Firefox shows dates based on a year of 0097 whereas Chrome says 1997.
Either way, I can't spot anything in this patch that could cause a random offset. So...
Yeah, I can see that now that I tried in Firefox. I was using Chromium so far and there timestamps are exactly matching the ones from the tracepoint log. Which is what we want for easy correlation between the log and HTML..
Firefox corrupts that somehow by applying the large negative offset to everyhting. Perhaps around two year worth of negative seconds if my rough calculation can be trusted. Or Vis under Firefox, I wouldn't know really who is to blame.
I have no idea what to do here. :(
Regards,
Tvrtko
_______________________________________________
Intel-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/intel-gfx
Re: [igt-dev] [PATCH i-g-t v2 1/9] trace.pl: Improve time axis labels
Quoted 82 lines · Tvrtko Ursulin
Quoted 74 lines · John Harrison
Quoted 64 lines · Tvrtko Ursulin
From: Tvrtko Ursulin<[email protected]>
It is possible to customize the axis display so change it to display timestamps in seconds on the major axis (with six decimal spaces) and millisecond offsets on the minor axis.
v2:
* Give up on broken relative timestamps.Signed-off-by Tvrtko Ursulin<[email protected]>Suggested-by Chris Wilson<[email protected]>Cc Chris Wilson<[email protected]>Cc John Harrison<[email protected]>---
scripts/trace.pl| 37 +++++++++++++++++++++++++++++++++++++1 file changed, 37 insertions(+)scripts/trace.pl +35
-0@@ -1000,6 +1000,42 @@ $first_ts = ts($first_ts); print <<ENDHTML; ]); + function majorAxis(date, scale, step) { + var s = date / 1000; + var precision; + + if (scale == 'millisecond') + precision = 6; + else if (scale == 'second') + precision = 3; + else + precision = 0; + + return s.toFixed(precision) + "s"; + } + + function minorAxis(date, scale, step) { + var ms = date; + var precision; + var unit; + + if (scale == 'millisecond') { + ms %= 1000; + precision = 0; + unit = 'ms'; + } else if (scale == 'second') { + ms /= 1000; + precision = 1; + unit = 's'; + } else { + ms /= 1000; + precision = 0; + unit = 's'; + } + + return ms.toFixed(precision) + unit; + } + // Configuration for the Timeline var options = { groupOrder: 'content', horizontalScroll: true,scripts/trace.pl print <<ENDHTML; · line 1048
Quoted 6 lines
scripts/trace.pl +1
-0@@ -1007,6 +1043,7 @@ print <<ENDHTML; stackSubgroups: false, zoomKey: 'ctrlKey', orientation: 'top', + format: { majorLabels: majorAxis, minorLabels: minorAxis }, start: '$first_ts', end: '$end_ts'};I'm still seeing some kind of strange offset. However, it appears to be browser dependent. If I use Chrome then the offset is +28.8 seconds. With Firefox it is -59958115.2 seconds! On the other hand, if I try Edge or IE then I don't get a graph at all. I'm wondering if the issue is with Vis browser compatibility rather than anything in the trace.pl script. Are you seeing anything at all similar?
Quoted 3 lines
Hmm, if I comment out the 'format:' line and go back to the unformatted time stamps then IE & Edge still show nothing. However, Firefox shows dates based on a year of 0097 whereas Chrome says 1997.
Either way, I can't spot anything in this patch that could cause a random offset. So...
Yeah, I can see that now that I tried in Firefox. I was using Chromium so far and there timestamps are exactly matching the ones from the tracepoint log. Which is what we want for easy correlation between the log and HTML..
Quoted 7 lines
Firefox corrupts that somehow by applying the large negative offset to everyhting. Perhaps around two year worth of negative seconds if my rough calculation can be trusted. Or Vis under Firefox, I wouldn't know really who is to blame.
I have no idea what to do here. :(
Regards,
Tvrtko
I think ship it for now. It is better than it was. Certainly reporting in date format is vaguely meaningless at best and totally meaningless with the x1000 scale factor.
Note that chromium on Ubuntu 16.04 does the same as Chrome on Windows for me - 28.8 seconds offset. It's not as bad as the 1.9 years of Firefox but it is still out :(. I'm guessing it is a bug in the date -> absolute seconds conversion going on within either Javascript itself or Vis in particular. The timestamps are still encoded as dates in the HTML file (and referenced from 0 not from 1900 or 1970 or whatever). So any difference in calculating leap years between the Perl script and the browser would potentially cause quite a significant delta.
Is it at all possible to put absolute seconds style values in the HTML file instead of dates? That would seem like the obvious answer. I don't know if Vis would cope with that, though?
John.
_______________________________________________
igt-dev mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/igt-dev
Re: [igt-dev] [PATCH i-g-t v2 1/9] trace.pl: Improve time axis labels
Quoted 93 lines · John Harrison
Quoted 82 lines · Tvrtko Ursulin
Quoted 74 lines · John Harrison
Quoted 64 lines · Tvrtko Ursulin
From: Tvrtko Ursulin<[email protected]>
It is possible to customize the axis display so change it to display timestamps in seconds on the major axis (with six decimal spaces) and millisecond offsets on the minor axis.
v2:
* Give up on broken relative timestamps.Signed-off-by Tvrtko Ursulin<[email protected]>Suggested-by Chris Wilson<[email protected]>Cc Chris Wilson<[email protected]>Cc John Harrison<[email protected]>---
scripts/trace.pl| 37 +++++++++++++++++++++++++++++++++++++1 file changed, 37 insertions(+)scripts/trace.pl +35
-0@@ -1000,6 +1000,42 @@ $first_ts = ts($first_ts); print <<ENDHTML; ]); + function majorAxis(date, scale, step) { + var s = date / 1000; + var precision; + + if (scale == 'millisecond') + precision = 6; + else if (scale == 'second') + precision = 3; + else + precision = 0; + + return s.toFixed(precision) + "s"; + } + + function minorAxis(date, scale, step) { + var ms = date; + var precision; + var unit; + + if (scale == 'millisecond') { + ms %= 1000; + precision = 0; + unit = 'ms'; + } else if (scale == 'second') { + ms /= 1000; + precision = 1; + unit = 's'; + } else { + ms /= 1000; + precision = 0; + unit = 's'; + } + + return ms.toFixed(precision) + unit; + } + // Configuration for the Timeline var options = { groupOrder: 'content', horizontalScroll: true,scripts/trace.pl print <<ENDHTML; · line 1048
Quoted 6 lines
scripts/trace.pl +1
-0@@ -1007,6 +1043,7 @@ print <<ENDHTML; stackSubgroups: false, zoomKey: 'ctrlKey', orientation: 'top', + format: { majorLabels: majorAxis, minorLabels: minorAxis }, start: '$first_ts', end: '$end_ts'};I'm still seeing some kind of strange offset. However, it appears to be browser dependent. If I use Chrome then the offset is +28.8 seconds. With Firefox it is -59958115.2 seconds! On the other hand, if I try Edge or IE then I don't get a graph at all. I'm wondering if the issue is with Vis browser compatibility rather than anything in the trace.pl script. Are you seeing anything at all similar?
Quoted 3 lines
Hmm, if I comment out the 'format:' line and go back to the unformatted time stamps then IE & Edge still show nothing. However, Firefox shows dates based on a year of 0097 whereas Chrome says 1997.
Either way, I can't spot anything in this patch that could cause a random offset. So...
Yeah, I can see that now that I tried in Firefox. I was using Chromium so far and there timestamps are exactly matching the ones from the tracepoint log. Which is what we want for easy correlation between the log and HTML..
Quoted 7 lines
Firefox corrupts that somehow by applying the large negative offset to everyhting. Perhaps around two year worth of negative seconds if my rough calculation can be trusted. Or Vis under Firefox, I wouldn't know really who is to blame.
I have no idea what to do here. :(
Regards,
Tvrtko
I think ship it for now. It is better than it was. Certainly reporting in date format is vaguely meaningless at best and totally meaningless with the x1000 scale factor.
Quoted 8 lines
Note that chromium on Ubuntu 16.04 does the same as Chrome on Windows for me - 28.8 seconds offset. It's not as bad as the 1.9 years of Firefox but it is still out :(. I'm guessing it is a bug in the date
-> absolute seconds conversion going on within either Javascript
itself or Vis in particular. The timestamps are still encoded as dates in the HTML file (and referenced from 0 not from 1900 or 1970 or whatever). So any difference in calculating leap years between the Perl script and the browser would potentially cause quite a significant delta.Is it at all possible to put absolute seconds style values in the HTML file instead of dates? That would seem like the obvious answer. I don't know if Vis would cope with that, though?
John.
Hmm. It looks like if I change the 'ts()' function to use 'localtime()' instead of 'gmtime()' and to add on 1900 to the year then it all works fine for me :). So yes, I think it is some incompatibility between the Perl and Javascript implementations of date <-> absolute seconds conversions. Given that the timestamp is no longer being reported as an actual date anymore, the relative value doesn't really matter. So I would go with using whatever scheme produces the least mutation along the way!
I wonder if you see the correct values on Chrome because your logs have smaller timestamps? The ones I am currently testing with are of the order of 856985.688681. With the above tweaks, that comes out as a date of '1997-02-26 11:34:48.681000'. The 'gmtime' version was '1997-02-26 19:34:48.681000' and obviously the non-1900 version was '0097-02-26 19:34:48.681000'. Actually, maybe the Chrome difference is because you are in the UK and don't have a timezone delta? Although I would assume you are on BST not GMT right now?
John.
_______________________________________________
igt-dev mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/igt-dev
Re: [igt-dev] [PATCH i-g-t v2 1/9] trace.pl: Improve time axis labels
Quoted 136 lines · John Harrison
Quoted 119 lines · John Harrison
Quoted 98 lines · Tvrtko Ursulin
Quoted 79 lines · John Harrison
Quoted 64 lines · Tvrtko Ursulin
From: Tvrtko Ursulin<[email protected]>
It is possible to customize the axis display so change it to display timestamps in seconds on the major axis (with six decimal spaces) and millisecond offsets on the minor axis.
v2:
* Give up on broken relative timestamps.Signed-off-by Tvrtko Ursulin<[email protected]>Suggested-by Chris Wilson<[email protected]>Cc Chris Wilson<[email protected]>Cc John Harrison<[email protected]>---
scripts/trace.pl| 37 +++++++++++++++++++++++++++++++++++++1 file changed, 37 insertions(+)scripts/trace.pl +35
-0@@ -1000,6 +1000,42 @@ $first_ts = ts($first_ts); print <<ENDHTML; ]); + function majorAxis(date, scale, step) { + var s = date / 1000; + var precision; + + if (scale == 'millisecond') + precision = 6; + else if (scale == 'second') + precision = 3; + else + precision = 0; + + return s.toFixed(precision) + "s"; + } + + function minorAxis(date, scale, step) { + var ms = date; + var precision; + var unit; + + if (scale == 'millisecond') { + ms %= 1000; + precision = 0; + unit = 'ms'; + } else if (scale == 'second') { + ms /= 1000; + precision = 1; + unit = 's'; + } else { + ms /= 1000; + precision = 0; + unit = 's'; + } + + return ms.toFixed(precision) + unit; + } + // Configuration for the Timeline var options = { groupOrder: 'content', horizontalScroll: true,scripts/trace.pl print <<ENDHTML; · line 1048
Quoted 6 lines
scripts/trace.pl +1
-0@@ -1007,6 +1043,7 @@ print <<ENDHTML; stackSubgroups: false, zoomKey: 'ctrlKey', orientation: 'top', + format: { majorLabels: majorAxis, minorLabels: minorAxis }, start: '$first_ts', end: '$end_ts'};I'm still seeing some kind of strange offset. However, it appears to be browser dependent. If I use Chrome then the offset is +28.8 seconds. With Firefox it is -59958115.2 seconds! On the other hand, if I try Edge or IE then I don't get a graph at all. I'm wondering if the issue is with Vis browser compatibility rather than anything in the trace.pl script. Are you seeing anything at all similar?
Quoted 6 lines
Hmm, if I comment out the 'format:' line and go back to the unformatted time stamps then IE & Edge still show nothing. However, Firefox shows dates based on a year of 0097 whereas Chrome says 1997.
Either way, I can't spot anything in this patch that could cause a random offset. So...
Yeah, I can see that now that I tried in Firefox. I was using Chromium so far and there timestamps are exactly matching the ones from the tracepoint log. Which is what we want for easy correlation between the log and HTML..
Firefox corrupts that somehow by applying the large negative offset to everyhting. Perhaps around two year worth of negative seconds if my rough calculation can be trusted. Or Vis under Firefox, I wouldn't know really who is to blame.
Quoted 5 lines
I have no idea what to do here. :(
Regards,
Tvrtko
I think ship it for now. It is better than it was. Certainly reporting in date format is vaguely meaningless at best and totally meaningless with the x1000 scale factor.
Note that chromium on Ubuntu 16.04 does the same as Chrome on Windows for me - 28.8 seconds offset. It's not as bad as the 1.9 years of Firefox but it is still out :(. I'm guessing it is a bug in the date -> absolute seconds conversion going on within either Javascript itself or Vis in particular. The timestamps are still encoded as dates in the HTML file (and referenced from 0 not from 1900 or 1970 or whatever). So any difference in calculating leap years between the Perl script and the browser would potentially cause quite a significant delta.
Quoted 6 lines
Is it at all possible to put absolute seconds style values in the HTML file instead of dates? That would seem like the obvious answer. I don't know if Vis would cope with that, though?
John.
Hmm. It looks like if I change the '
ts()' function to use 'localtime()' instead of 'gmtime()' and to add on 1900 to the year then it all works fine for me :). So yes, I think it is some incompatibility between the Perl and Javascript implementations of date <-> absolute seconds conversions. Given that the timestamp is no longer being reported as an actual date anymore, the relative value doesn't really matter. So I would go with using whatever scheme produces the least mutation along the way!
Quoted 8 lines
I wonder if you see the correct values on Chrome because your logs have smaller timestamps? The ones I am currently testing with are of the order of 856985.688681. With the above tweaks, that comes out as a date of '1997-02-26 11:34:48.681000'. The 'gmtime' version was '1997-02-26 19:34:48.681000' and obviously the non-1900 version was '0097-02-26 19:34:48.681000'. Actually, maybe the Chrome difference is because you are in the UK and don't have a timezone delta? Although I would assume you are on BST not GMT right now?
Can you try leaving gmtime in ts() and adding this diff:
scripts/trace.pl +4 -0
@@ -1338,6 +1338,10 @@ print <<ENDHTML; zoomKey: 'ctrlKey', orientation: 'top', format: { majorLabels: majorAxis, minorLabels: minorAxis }, + moment: function(date) { + return vis.moment(date).utc(); + }, + start: '$first_ts', end: '$end_ts'};
Could be the gotcha we were missing!
Regards,
Tvrtko
_______________________________________________
igt-dev mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/igt-dev
Re: [igt-dev] [PATCH i-g-t v2 1/9] trace.pl: Improve time axis labels
Quoted 159 lines · Tvrtko Ursulin
Quoted 131 lines · John Harrison
Quoted 115 lines · John Harrison
Quoted 95 lines · Tvrtko Ursulin
Quoted 78 lines · John Harrison
Quoted 64 lines · Tvrtko Ursulin
From: Tvrtko Ursulin<[email protected]>
It is possible to customize the axis display so change it to display timestamps in seconds on the major axis (with six decimal spaces) and millisecond offsets on the minor axis.
v2:
* Give up on broken relative timestamps.Signed-off-by Tvrtko Ursulin<[email protected]>Suggested-by Chris Wilson<[email protected]>Cc Chris Wilson<[email protected]>Cc John Harrison<[email protected]>--- scripts/trace.pl | 37 +++++++++++++++++++++++++++++++++++++ 1 file changed, 37 insertions(+)scripts/trace.pl +35
-0@@ -1000,6 +1000,42 @@ $first_ts = ts($first_ts); print <<ENDHTML; ]); + function majorAxis(date, scale, step) { + var s = date / 1000; + var precision; + + if (scale == 'millisecond') + precision = 6; + else if (scale == 'second') + precision = 3; + else + precision = 0; + + return s.toFixed(precision) + "s"; + } + + function minorAxis(date, scale, step) { + var ms = date; + var precision; + var unit; + + if (scale == 'millisecond') { + ms %= 1000; + precision = 0; + unit = 'ms'; + } else if (scale == 'second') { + ms /= 1000; + precision = 1; + unit = 's'; + } else { + ms /= 1000; + precision = 0; + unit = 's'; + } + + return ms.toFixed(precision) + unit; + } + // Configuration for the Timeline var options = { groupOrder: 'content', horizontalScroll: true,scripts/trace.pl print <<ENDHTML; · line 1048
Quoted 6 lines
scripts/trace.pl +1
-0@@ -1007,6 +1043,7 @@ print <<ENDHTML; stackSubgroups: false, zoomKey: 'ctrlKey', orientation: 'top', + format: { majorLabels: majorAxis, minorLabels: minorAxis }, start: '$first_ts', end: '$end_ts'};I'm still seeing some kind of strange offset. However, it appears to be browser dependent. If I use Chrome then the offset is +28.8 seconds. With Firefox it is -59958115.2 seconds! On the other hand, if I try Edge or IE then I don't get a graph at all. I'm wondering if the issue is with Vis browser compatibility rather than anything in the trace.pl script. Are you seeing anything at all similar?
Quoted 6 lines
Hmm, if I comment out the 'format:' line and go back to the unformatted time stamps then IE & Edge still show nothing. However, Firefox shows dates based on a year of 0097 whereas Chrome says 1997.
Either way, I can't spot anything in this patch that could cause a random offset. So...
Yeah, I can see that now that I tried in Firefox. I was using Chromium so far and there timestamps are exactly matching the ones from the tracepoint log. Which is what we want for easy correlation between the log and HTML..
Firefox corrupts that somehow by applying the large negative offset to everyhting. Perhaps around two year worth of negative seconds if my rough calculation can be trusted. Or Vis under Firefox, I wouldn't know really who is to blame.
Quoted 5 lines
I have no idea what to do here. :(
Regards,
Tvrtko
I think ship it for now. It is better than it was. Certainly reporting in date format is vaguely meaningless at best and totally meaningless with the x1000 scale factor.
Note that chromium on Ubuntu 16.04 does the same as Chrome on Windows for me - 28.8 seconds offset. It's not as bad as the 1.9 years of Firefox but it is still out :(. I'm guessing it is a bug in the date -> absolute seconds conversion going on within either Javascript itself or Vis in particular. The timestamps are still encoded as dates in the HTML file (and referenced from 0 not from 1900 or 1970 or whatever). So any difference in calculating leap years between the Perl script and the browser would potentially cause quite a significant delta.
Quoted 6 lines
Is it at all possible to put absolute seconds style values in the HTML file instead of dates? That would seem like the obvious answer. I don't know if Vis would cope with that, though?
John.
Hmm. It looks like if I change the '
ts()' function to use 'localtime()' instead of 'gmtime()' and to add on 1900 to the year then it all works fine for me :). So yes, I think it is some incompatibility between the Perl and Javascript implementations of date <-> absolute seconds conversions. Given that the timestamp is no longer being reported as an actual date anymore, the relative value doesn't really matter. So I would go with using whatever scheme produces the least mutation along the way!Quoted 8 lines
I wonder if you see the correct values on Chrome because your logs have smaller timestamps? The ones I am currently testing with are of the order of 856985.688681. With the above tweaks, that comes out as a date of '1997-02-26 11:34:48.681000'. The 'gmtime' version was '1997-02-26 19:34:48.681000' and obviously the non-1900 version was '0097-02-26 19:34:48.681000'. Actually, maybe the Chrome difference is because you are in the UK and don't have a timezone delta? Although I would assume you are on BST not GMT right now?
Can you try leaving gmtime in
ts()and adding this diff:scripts/trace.pl +4
-0@@ -1338,6 +1338,10 @@ print <<ENDHTML; zoomKey: 'ctrlKey', orientation: 'top', format: { majorLabels: majorAxis, minorLabels: minorAxis }, + moment: function(date) { + return vis.moment(date).utc(); + }, + start: '$first_ts', end: '$end_ts'};Could be the gotcha we were missing!
Quoted 3 lines
Regards,
Tvrtko
Doesn't seem to make a difference for me :(. The +1900 gets rid of the massive negative offset in Firefox but only the localtime() fixes the 28.8s offset in both Firefox and Chrome.
_______________________________________________
igt-dev mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/igt-dev
PATCH 2/9 trace.pl: Scale timeline for better precision
In reply to Tvrtko UrsulinFrom: Tvrtko Ursulin <[email protected]>
vis library has a limited precision compared to our trace data which prevents zooming into the timeline and seeing the fine detail.
Scale the HTML view by a thousand to work around it.
v2: Rebase for time axis changes.
--- scripts/trace.pl | 46 +++++++++++++++++++++++++++++++++++++++++++--- 1 file changed, 43 insertions(+), 3 deletions(-)
scripts/trace.pl +43 -3
@@ -40,6 +40,7 @@ my $trace = 0; my $avg_delay_stats = 0; my $gpu_timeline = 0; my $colour_contexts = 0; +my $no_timeline_scaling = 0; my @args;
@@ -281,6 +282,18 @@ sub arg_colour_contexts return @_; } +sub arg_no_timeline_scaling +{ + return unless scalar(@_); + + if ($_[0] eq '--no-timeline-scaling') { + shift @_; + $no_timeline_scaling = 1; + } + + return @_; +} + @args = @ARGV; while (@args) { my $left = scalar(@args);
@@ -296,6 +309,7 @@ while (@args) { @args = arg_ignore_ring(@args); @args = arg_skip_box(@args); @args = arg_colour_contexts(@args); + @args = arg_no_timeline_scaling(@args); last if $left == scalar(@args); }
@@ -334,6 +348,8 @@ sub ts my ($us) = @_; my ($y, $mo, $d, $h, $m, $s); + $us *= 1000 unless $no_timeline_scaling; + $s = int($us / 1000000); $us = $us % 1000000;
@@ -1001,11 +1017,28 @@ print <<ENDHTML; ]); var last_major_ms; +ENDHTML + +if ($html) { + if ($no_timeline_scaling) { + say " timeDiv = 1;"; + } else { + say " timeDiv = 1000;"; + } +} +print <<ENDHTML; function majorAxis(date, scale, step) { - var s = date / 1000; + var s = date / 1000 / timeDiv; + + last_major_ms = date / timeDiv; - last_major_ms = date; + if (timeDiv == 1000) { + if (scale == 'second') + scale = 'millisecond'; + else if (scale == 'minute') + scale = 'second' + } if (scale == 'millisecond') return s.toFixed(6) + "s";
@@ -1016,13 +1049,20 @@ print <<ENDHTML; } function minorAxis(date, scale, step) { - var ms = date; + var ms = date / timeDiv; ms -= last_major_ms; if (ms < 0) return ''; + if (timeDiv > 1) { + if (scale == 'second') + scale = 'millisecond'; + else if (scale == 'minute') + scale = 'second'; + } + if (scale == 'millisecond') return "+" + ms.toFixed(3) + "ms"; else if (scale == 'second')
Signature · 6 lines
2.17.1 _______________________________________________ igt-dev mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/igt-dev
Re: [igt-dev] [PATCH i-g-t 2/9] trace.pl: Scale timeline for better precision
Quoted 108 lines · Tvrtko Ursulin
From: Tvrtko Ursulin <[email protected]>
vis library has a limited precision compared to our trace data which prevents zooming into the timeline and seeing the fine detail.
Scale the HTML view by a thousand to work around it.
v2: Rebase for time axis changes.
Signed-off-by Tvrtko Ursulin <[email protected]>Suggested-by John Harrison <[email protected]>Cc John Harrison <[email protected]>Reviewed-by John Harrison <[email protected]> # v1--- scripts/trace.pl | 46 +++++++++++++++++++++++++++++++++++++++++++--- 1 file changed, 43 insertions(+), 3 deletions(-)scripts/trace.pl +36
-3@@ -40,6 +40,7 @@ my $trace = 0; my $avg_delay_stats = 0; my $gpu_timeline = 0; my $colour_contexts = 0; +my $no_timeline_scaling = 0; my @args;@@ -281,6 +282,18 @@ sub arg_colour_contexts return @_; } +sub arg_no_timeline_scaling +{ + return unless scalar(@_); + + if ($_[0] eq '--no-timeline-scaling') { + shift @_; + $no_timeline_scaling = 1; + } + + return @_; +} + @args = @ARGV; while (@args) { my $left = scalar(@args);@@ -296,6 +309,7 @@ while (@args) { @args = arg_ignore_ring(@args); @args = arg_skip_box(@args); @args = arg_colour_contexts(@args); + @args = arg_no_timeline_scaling(@args); last if $left == scalar(@args); }@@ -334,6 +348,8 @@ sub ts my ($us) = @_; my ($y, $mo, $d, $h, $m, $s); + $us *= 1000 unless $no_timeline_scaling; + $s = int($us / 1000000); $us = $us % 1000000;@@ -1001,11 +1017,28 @@ print <<ENDHTML; ]); var last_major_ms; +ENDHTML + +if ($html) { + if ($no_timeline_scaling) { + say " timeDiv = 1;"; + } else { + say " timeDiv = 1000;"; + } +} +print <<ENDHTML; function majorAxis(date, scale, step) { - var s = date / 1000; + var s = date / 1000 / timeDiv; + + last_major_ms = date / timeDiv; - last_major_ms = date; + if (timeDiv == 1000) { + if (scale == 'second') + scale = 'millisecond'; + else if (scale == 'minute') + scale = 'second' + } if (scale == 'millisecond') return s.toFixed(6) + "s";@@ -1016,13 +1049,20 @@ print <<ENDHTML; } function minorAxis(date, scale, step) { - var ms = date; + var ms = date / timeDiv; ms -= last_major_ms; if (ms < 0) return '';
scripts/trace.pl print <<ENDHTML; · line 1068
Quoted 10 lines
scripts/trace.pl +7
-0+ if (timeDiv > 1) { + if (scale == 'second') + scale = 'millisecond'; + else if (scale == 'minute') + scale = 'second'; + } + if (scale == 'millisecond') return "+" + ms.toFixed(3) + "ms"; else if (scale == 'second')
_______________________________________________
igt-dev mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/igt-dev
PATCH v3 2/9 trace.pl: Scale timeline for better precision
In reply to Tvrtko UrsulinFrom: Tvrtko Ursulin <[email protected]>
vis library has a limited precision compared to our trace data which prevents zooming into the timeline and seeing the fine detail.
Scale the HTML view by a thousand to work around it.
v2: Rebase for time axis changes.
v3: Another tweak for time axis changes.
--- scripts/trace.pl | 54 ++++++++++++++++++++++++++++++++++++++++++++---- 1 file changed, 50 insertions(+), 4 deletions(-)
scripts/trace.pl +50 -4
@@ -40,6 +40,7 @@ my $trace = 0; my $avg_delay_stats = 0; my $gpu_timeline = 0; my $colour_contexts = 0; +my $no_timeline_scaling = 0; my @args;
@@ -281,6 +282,18 @@ sub arg_colour_contexts return @_; } +sub arg_no_timeline_scaling +{ + return unless scalar(@_); + + if ($_[0] eq '--no-timeline-scaling') { + shift @_; + $no_timeline_scaling = 1; + } + + return @_; +} + @args = @ARGV; while (@args) { my $left = scalar(@args);
@@ -296,6 +309,7 @@ while (@args) { @args = arg_ignore_ring(@args); @args = arg_skip_box(@args); @args = arg_colour_contexts(@args); + @args = arg_no_timeline_scaling(@args); last if $left == scalar(@args); }
@@ -334,6 +348,8 @@ sub ts my ($us) = @_; my ($y, $mo, $d, $h, $m, $s); + $us *= 1000 unless $no_timeline_scaling; + $s = int($us / 1000000); $us = $us % 1000000;
@@ -999,11 +1015,28 @@ $first_ts = ts($first_ts); print <<ENDHTML; ]); +ENDHTML +if ($html) { + if ($no_timeline_scaling) { + say " timeDiv = 1;"; + } else { + say " timeDiv = 1000;"; + } +} + +print <<ENDHTML; function majorAxis(date, scale, step) { - var s = date / 1000; + var s = date / 1000 / timeDiv; var precision; + if (timeDiv == 1000) { + if (scale == 'second') + scale = 'millisecond'; + else if (scale == 'minute') + scale = 'second' + } + if (scale == 'millisecond') precision = 6; else if (scale == 'second')
@@ -1015,14 +1048,27 @@ print <<ENDHTML; } function minorAxis(date, scale, step) { - var ms = date; + var ms = date / timeDiv; + var orig_scale = scale; var precision; var unit; + if (timeDiv == 1000) { + if (scale == 'second') + scale = 'millisecond'; + else if (scale == 'minute') + scale = 'second' + } + if (scale == 'millisecond') { - ms %= 1000; precision = 0; - unit = 'ms'; + if (orig_scale == 'millisecond' && step <= 200) { + ms = date % 1000; + unit = 'us'; + } else { + ms %= 1000; + unit = 'ms'; + } } else if (scale == 'second') { ms /= 1000; precision = 1;
Signature · 6 lines
2.17.1 _______________________________________________ igt-dev mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/igt-dev
Re: [igt-dev] [PATCH i-g-t v3 2/9] trace.pl: Scale timeline for better precision
Quoted 117 lines · Tvrtko Ursulin
From: Tvrtko Ursulin <[email protected]>
vis library has a limited precision compared to our trace data which prevents zooming into the timeline and seeing the fine detail.
Scale the HTML view by a thousand to work around it.
v2: Rebase for time axis changes.
v3: Another tweak for time axis changes.Signed-off-by Tvrtko Ursulin <[email protected]>Suggested-by John Harrison <[email protected]>Cc John Harrison <[email protected]>Reviewed-by John Harrison <[email protected]> # v2--- scripts/trace.pl | 54 ++++++++++++++++++++++++++++++++++++++++++++---- 1 file changed, 50 insertions(+), 4 deletions(-)scripts/trace.pl +43
-4@@ -40,6 +40,7 @@ my $trace = 0; my $avg_delay_stats = 0; my $gpu_timeline = 0; my $colour_contexts = 0; +my $no_timeline_scaling = 0; my @args;@@ -281,6 +282,18 @@ sub arg_colour_contexts return @_; } +sub arg_no_timeline_scaling +{ + return unless scalar(@_); + + if ($_[0] eq '--no-timeline-scaling') { + shift @_; + $no_timeline_scaling = 1; + } + + return @_; +} + @args = @ARGV; while (@args) { my $left = scalar(@args);@@ -296,6 +309,7 @@ while (@args) { @args = arg_ignore_ring(@args); @args = arg_skip_box(@args); @args = arg_colour_contexts(@args); + @args = arg_no_timeline_scaling(@args); last if $left == scalar(@args); }@@ -334,6 +348,8 @@ sub ts my ($us) = @_; my ($y, $mo, $d, $h, $m, $s); + $us *= 1000 unless $no_timeline_scaling; + $s = int($us / 1000000); $us = $us % 1000000;@@ -999,11 +1015,28 @@ $first_ts = ts($first_ts); print <<ENDHTML; ]); +ENDHTML +if ($html) { + if ($no_timeline_scaling) { + say " timeDiv = 1;"; + } else { + say " timeDiv = 1000;"; + } +} + +print <<ENDHTML; function majorAxis(date, scale, step) { - var s = date / 1000; + var s = date / 1000 / timeDiv; var precision; + if (timeDiv == 1000) { + if (scale == 'second') + scale = 'millisecond'; + else if (scale == 'minute') + scale = 'second' + } + if (scale == 'millisecond') precision = 6; else if (scale == 'second')@@ -1015,14 +1048,27 @@ print <<ENDHTML; } function minorAxis(date, scale, step) { - var ms = date; + var ms = date / timeDiv; + var orig_scale = scale; var precision; var unit; + if (timeDiv == 1000) { + if (scale == 'second') + scale = 'millisecond'; + else if (scale == 'minute') + scale = 'second' + } + if (scale == 'millisecond') { - ms %= 1000; precision = 0; - unit = 'ms';
scripts/trace.pl print <<ENDHTML; · line 1074
Quoted 10 lines
scripts/trace.pl +7
-0+ if (orig_scale == 'millisecond' && step <= 200) { + ms = date % 1000; + unit = 'us'; + } else { + ms %= 1000; + unit = 'ms'; + } } else if (scale == 'second') { ms /= 1000; precision = 1;
_______________________________________________
igt-dev mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/igt-dev
PATCH 3/9 trace.pl: Improve readability of graphical timeline representation
In reply to Tvrtko UrsulinFrom: Tvrtko Ursulin <[email protected]>
We add stripes for different stages of request execution so it is easier to follow one context in the multi-colour mode.
Vertical stripe pattern indicates pipeline "blockages" - requests waiting for dependencies before they are runnable.
Diagonal stripes indicate runnable requests waiting for GPU time.
Horizontal strips are requests executing on the GPU.
Also use this new multi-coloured mode from media-bench.pl.
v2:
John Harrison:
* Mention media-bench.pl in the commit.
* Fix HTML for single colour mode.
v3:
* Rebase.
* Apply stripes to legacy colouring as well.
v4:
John Harrison:
* Use per context colours for ctxsave and incomplete boxes.
* Clearer timeline legend.
--- scripts/media-bench.pl | 2 +- scripts/trace.pl | 112 ++++++++++++++++++++++++++++++++--------- 2 files changed, 90 insertions(+), 24 deletions(-)
scripts/media-bench.pl +1 -1
@@ -207,7 +207,7 @@ sub trace_workload show_cmd($cmd); system($cmd); - $cmd = "perf script | $tracepl --html -x ctxsave -s "; + $cmd = "perf script | $tracepl --html -x ctxsave -s -c "; $cmd .= join ' ', map("-i $_", @skip_engine); $cmd .= " > ${file}.html"; show_cmd($cmd);
scripts/trace.pl +89 -23
171 lines not on this page · raw message
Signature · 6 lines
2.17.1 _______________________________________________ Intel-gfx mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/intel-gfx
Re: [Intel-gfx] [PATCH i-g-t 3/9] trace.pl: Improve readability of graphical timeline representation
Quoted 23 lines · Tvrtko Ursulin
From: Tvrtko Ursulin <[email protected]>
We add stripes for different stages of request execution so it is easier to follow one context in the multi-colour mode.
Vertical stripe pattern indicates pipeline "blockages" - requests waiting for dependencies before they are runnable.
Diagonal stripes indicate runnable requests waiting for GPU time.
Horizontal strips are requests executing on the GPU.
Also use this new multi-coloured mode from media-bench.pl.
v2:
John Harrison:
* Mention media-bench.pl in the commit.
* Fix HTML for single colour mode.v3:
* Rebase.
* Apply stripes to legacy colouring as well.
Quoted 7 lines
v4:
John Harrison:
* Use per context colours for ctxsave and incomplete boxes.
* Clearer timeline legend.Signed-off-by Tvrtko Ursulin <[email protected]>Cc John Harrison <[email protected]>
_______________________________________________
Intel-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/intel-gfx
PATCH 4/9 trace.pl: Improve context colouring for large context id's
In reply to Tvrtko UrsulinFrom: Tvrtko Ursulin <[email protected]>
John reports that on a long runnning systems the huge disparity between kernel context and user context id's causes all interesting colours to be clustered too close together.
Fix this by assigning colours to seen contexts instead of basing purely on context id's.
--- scripts/trace.pl | 25 ++++++++++++++++++++----- 1 file changed, 20 insertions(+), 5 deletions(-)
scripts/trace.pl +20 -5
@@ -336,7 +336,7 @@ sub sanitize_ctx { my ($ctx, $ring) = @_; - if (exists $ctxdb{$ctx}) { + if (exists $ctxdb{$ctx} and $ctxdb{$ctx} > 1) { return $ctx . '.' . $ctxdb{$ctx}; } else { return $ctx;
@@ -425,6 +425,8 @@ while (<>) { $ctxdb{$orig_ctx}++; $ctx = sanitize_ctx($orig_ctx, $ring); $key = db_key($ring, $ctx, $seqno); + } else { + $ctxdb{$orig_ctx} = 1; } $queue{$key} = $time;
@@ -592,7 +594,7 @@ foreach my $key (@sorted_keys) { my $last_ts = 0; my $first_ts; -my ($min_ctx, $max_ctx); +my $min_ctx; foreach my $key (@sorted_keys) { my $ring = $db{$key}->{'ring'};
@@ -604,8 +606,6 @@ foreach my $key (@sorted_keys) { $last_ts = $end if $end > $last_ts; $min_ctx = $db{$key}->{'ctx'} if not defined $min_ctx or $db{$key}->{'ctx'} < $min_ctx; - $max_ctx = $db{$key}->{'ctx'} if not defined $max_ctx or - $db{$key}->{'ctx'} > $max_ctx; $db{$key}->{'context-complete-delay'} = $end - $notify; $db{$key}->{'execute-delay'} = $start - $db{$key}->{'submit'};
@@ -740,6 +740,21 @@ say sprintf('GPU: %.2f%% idle, %.2f%% busy', my $timeline_text = $colour_contexts ? 'Per context coloured shading like:' : 'Box shading like:'; +my %ctx_colours; + +sub generate_ctx_colours +{ + my $num_ctx = keys %ctxdb; + my $i = 0; + + foreach my $ctx (sort keys %ctxdb) { + $ctx_colours{$ctx} = int(360 / $num_ctx * $i++); + } +} + + +generate_ctx_colours() if $html and $colour_contexts; + my $queued_style = box_style($min_ctx, 'queue'); my $ready_style = box_style($min_ctx, 'ready'); my $execute_style = box_style($min_ctx, 'execute');
@@ -948,7 +963,7 @@ sub ctx_colour $l = 25; } - $val = int(360 / ($max_ctx - $min_ctx + 1)) * ($ctx - $min_ctx); + $val = $ctx_colours{$ctx}; } $l = int($l * $lfac);
Signature · 6 lines
2.17.1 _______________________________________________ igt-dev mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/igt-dev
Re: [Intel-gfx] [PATCH i-g-t 4/9] trace.pl: Improve context colouring for large context id's
Quoted 6 lines · Tvrtko Ursulin
From: Tvrtko Ursulin <[email protected]>
John reports that on a long runnning systems the huge disparity between kernel context and user context id's causes all interesting colours to be clustered too close together.
Quoted 8 lines
Fix this by assigning colours to seen contexts instead of basing purely on context id's.
Signed-off-by Tvrtko Ursulin <[email protected]>Suggested-by John Harrison <[email protected]>Cc John Harrison <[email protected]>---
_______________________________________________
Intel-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/intel-gfx
PATCH 5/9 trace.pl: Improved key/colours
In reply to Tvrtko UrsulinFrom: John Harrison <[email protected]>
Improve the timeline legend to show actual context colours.
v2: (Tvrtko Ursulin)
* Commit msg.
* Tweak layout for more compactness and more readability.
--- scripts/trace.pl | 69 ++++++++++++++++++++++++++++++------------------ 1 file changed, 43 insertions(+), 26 deletions(-)
scripts/trace.pl +43 -26
@@ -738,9 +738,35 @@ say sprintf('GPU: %.2f%% idle, %.2f%% busy', $flat_busy{'gpu-idle'}, $flat_busy{'gpu-busy'}) unless $html; my $timeline_text = $colour_contexts ? - 'Per context coloured shading like:' : 'Box shading like:'; + 'per context coloured shading like' : 'box shading like'; my %ctx_colours; +my $ctx_table; + +sub generate_ctx_table +{ + my @states = ('queue', 'ready', 'execute', 'ctxsave', 'incomplete'); + my @ctxts; + + if( $colour_contexts ) { + @ctxts = sort keys %ctxdb; + } else { + @ctxts = ($min_ctx); + } + + $ctx_table .= '<table>'; + + foreach my $ctx (@ctxts) { + $ctx_table .= "<tr>\n"; + $ctx_table .= " <td>Context $ctx</td>\n" if $colour_contexts; + foreach my $state (@states) { + $ctx_table .= " <td align='center' valign='middle'><div style='" . box_style($ctx, $state) . " padding-top: 6px; padding-bottom: 6px; padding-left: 6x; padding-right: 6px;'>" . uc($state) . "</div></td>\n"; + } + $ctx_table .= "</tr>\n"; + } + + $ctx_table .= '</table>'; +} sub generate_ctx_colours {
@@ -754,12 +780,7 @@ sub generate_ctx_colours generate_ctx_colours() if $html and $colour_contexts; - -my $queued_style = box_style($min_ctx, 'queue'); -my $ready_style = box_style($min_ctx, 'ready'); -my $execute_style = box_style($min_ctx, 'execute'); -my $ctxsave_style = box_style($min_ctx, 'ctxsave'); -my $incomplete_style = box_style($min_ctx, 'incomplete'); +generate_ctx_table() if $html; print <<ENDHTML if $html; <!DOCTYPE HTML>
@@ -778,35 +799,27 @@ print <<ENDHTML if $html; </head> <body> <p> -<b>Timeline request view:</b> +<b>Timeline request view is $timeline_text:</b> <table> -<tr><td colspan='4'>$timeline_text</td></tr> <tr> -<td align='center'><div style='$queued_style'>QUEUED</div></td> -<td align='center'><div style='$ready_style'>READY</div></td> -<td align='center'><div style='$execute_style'>EXECUTE</div></td> -<td align='center'><div style='$ctxsave_style'>CTXSAVE</div></td> -</tr><tr> -<td></td> -<td></td> -<td align='center'><div style='$incomplete_style'>(INCOMPLETE)</div></td> -<td></td> -</tr/></table> -</p> -<p> -<small> -QUEUED = requests executing on the GPU<br> +<td> +$ctx_table +</td> +<td> +QUEUE = requests executing on the GPU<br> READY = runnable requests waiting for a slot on GPU<br> EXECUTE = requests waiting on fences and dependencies before they are runnable<br> CTXSAVE = GPU saving the context image<br> -</small> -</p> +INCOMPLETE = request of unknown completion time <p> Boxes are in format 'ctx-id/seqno'. </p> <p> Use Ctrl+scroll-action to zoom-in/out and scroll-action or dragging to move around the timeline. </p> +</td> +</tr> +</table> <p> <b>GPU idle: $flat_busy{'gpu-idle'}%</b> <br>
@@ -975,20 +988,24 @@ sub box_style { my ($ctx, $stage) = @_; my $deg; + my $text_col = 'white'; if ($stage eq 'queue') { $deg = 90; + $text_col = 'black' if $colour_contexts; } elsif ($stage eq 'ready') { $deg = 45; } elsif ($stage eq 'execute') { $deg = 0; + $text_col = 'black' if $colour_contexts; } elsif ($stage eq 'ctxsave') { $deg = 105; + $text_col = 'black' if $colour_contexts; } elsif ($stage eq 'incomplete') { $deg = 0; } - return 'color: black; background: repeating-linear-gradient(' . + return "color: $text_col; background: repeating-linear-gradient(" . $deg . 'deg, ' . ctx_colour($ctx, $stage, 1.0) . ', ' . ctx_colour($ctx, $stage, 1.0) . ' 10px, ' .
Signature · 6 lines
2.17.1 _______________________________________________ igt-dev mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/igt-dev
Re: [igt-dev] [PATCH i-g-t 5/9] trace.pl: Improved key/colours
Quoted 9 lines · Tvrtko Ursulin
From: John Harrison <[email protected]>
Improve the timeline legend to show actual context colours.
v2: (Tvrtko Ursulin)
* Commit msg.
* Tweak layout for more compactness and more readability.Signed-off-by Tvrtko Ursulin <[email protected]>
I've picked up this patch of yours and made some tweaks. Hopefully you will find them agreeable, even if as a compromise to what we consider more readable. :)
But more importantly please add your s-o-b for it.
R-b we can both add if happy.
Regards,
Tvrtko
Quoted 124 lines
Cc John Harrison <[email protected]>Cc Tvrtko Ursulin <[email protected]>--- scripts/trace.pl | 69 ++++++++++++++++++++++++++++++------------------ 1 file changed, 43 insertions(+), 26 deletions(-)scripts/trace.pl +41
-25@@ -738,9 +738,35 @@ say sprintf('GPU: %.2f%% idle, %.2f%% busy', $flat_busy{'gpu-idle'}, $flat_busy{'gpu-busy'}) unless $html; my $timeline_text = $colour_contexts ? - 'Per context coloured shading like:' : 'Box shading like:'; + 'per context coloured shading like' : 'box shading like'; my %ctx_colours; +my $ctx_table; + +sub generate_ctx_table +{ + my @states = ('queue', 'ready', 'execute', 'ctxsave', 'incomplete'); + my @ctxts; + + if( $colour_contexts ) { + @ctxts = sort keys %ctxdb; + } else { + @ctxts = ($min_ctx); + } + + $ctx_table .= '<table>'; + + foreach my $ctx (@ctxts) { + $ctx_table .= "<tr>\n"; + $ctx_table .= " <td>Context $ctx</td>\n" if $colour_contexts; + foreach my $state (@states) { + $ctx_table .= " <td align='center' valign='middle'><div style='" . box_style($ctx, $state) . " padding-top: 6px; padding-bottom: 6px; padding-left: 6x; padding-right: 6px;'>" . uc($state) . "</div></td>\n"; + } + $ctx_table .= "</tr>\n"; + } + + $ctx_table .= '</table>'; +} sub generate_ctx_colours {@@ -754,12 +780,7 @@ sub generate_ctx_colours generate_ctx_colours() if $html and $colour_contexts; - -my $queued_style = box_style($min_ctx, 'queue'); -my $ready_style = box_style($min_ctx, 'ready'); -my $execute_style = box_style($min_ctx, 'execute'); -my $ctxsave_style = box_style($min_ctx, 'ctxsave'); -my $incomplete_style = box_style($min_ctx, 'incomplete'); +generate_ctx_table() if $html; print <<ENDHTML if $html; <!DOCTYPE HTML>@@ -778,35 +799,27 @@ print <<ENDHTML if $html; </head> <body> <p> -<b>Timeline request view:</b> +<b>Timeline request view is $timeline_text:</b> <table> -<tr><td colspan='4'>$timeline_text</td></tr> <tr> -<td align='center'><div style='$queued_style'>QUEUED</div></td> -<td align='center'><div style='$ready_style'>READY</div></td> -<td align='center'><div style='$execute_style'>EXECUTE</div></td> -<td align='center'><div style='$ctxsave_style'>CTXSAVE</div></td> -</tr><tr> -<td></td> -<td></td> -<td align='center'><div style='$incomplete_style'>(INCOMPLETE)</div></td> -<td></td> -</tr/></table> -</p> -<p> -<small> -QUEUED = requests executing on the GPU<br> +<td> +$ctx_table +</td> +<td> +QUEUE = requests executing on the GPU<br> READY = runnable requests waiting for a slot on GPU<br> EXECUTE = requests waiting on fences and dependencies before they are runnable<br> CTXSAVE = GPU saving the context image<br> -</small> -</p> +INCOMPLETE = request of unknown completion time <p> Boxes are in format 'ctx-id/seqno'. </p> <p> Use Ctrl+scroll-action to zoom-in/out and scroll-action or dragging to move around the timeline. </p> +</td> +</tr> +</table> <p> <b>GPU idle: $flat_busy{'gpu-idle'}%</b> <br>@@ -975,20 +988,24 @@ sub box_style { my ($ctx, $stage) = @_; my $deg; + my $text_col = 'white'; if ($stage eq 'queue') { $deg = 90; + $text_col = 'black' if $colour_contexts; } elsif ($stage eq 'ready') { $deg = 45; } elsif ($stage eq 'execute') { $deg = 0; + $text_col = 'black' if $colour_contexts; } elsif ($stage eq 'ctxsave') { $deg = 105;
scripts/trace.pl sub box_style · line 1011
Quoted 10 lines
scripts/trace.pl +2
-1+ $text_col = 'black' if $colour_contexts; } elsif ($stage eq 'incomplete') { $deg = 0; } - return 'color: black; background: repeating-linear-gradient(' . + return "color: $text_col; background: repeating-linear-gradient(" . $deg . 'deg, ' . ctx_colour($ctx, $stage, 1.0) . ', ' . ctx_colour($ctx, $stage, 1.0) . ' 10px, ' .
_______________________________________________
igt-dev mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/igt-dev
Re: [igt-dev] [PATCH i-g-t 5/9] trace.pl: Improved key/colours
In reply to Tvrtko UrsulinQuoted 39 lines · Tvrtko Ursulin
From: John Harrison <[email protected]>
Improve the timeline legend to show actual context colours.
v2: (Tvrtko Ursulin)
* Commit msg.
* Tweak layout for more compactness and more readability.Signed-off-by Tvrtko Ursulin <[email protected]>Cc John Harrison <[email protected]>Cc Tvrtko Ursulin <[email protected]>--- scripts/trace.pl | 69 ++++++++++++++++++++++++++++++------------------ 1 file changed, 43 insertions(+), 26 deletions(-)scripts/trace.pl +13
-1@@ -738,9 +738,35 @@ say sprintf('GPU: %.2f%% idle, %.2f%% busy', $flat_busy{'gpu-idle'}, $flat_busy{'gpu-busy'}) unless $html; my $timeline_text = $colour_contexts ? - 'Per context coloured shading like:' : 'Box shading like:'; + 'per context coloured shading like' : 'box shading like'; my %ctx_colours; +my $ctx_table; + +sub generate_ctx_table +{ + my @states = ('queue', 'ready', 'execute', 'ctxsave', 'incomplete'); + my @ctxts; + + if( $colour_contexts ) { + @ctxts = sort keys %ctxdb; + } else { + @ctxts = ($min_ctx); + }
scripts/trace.pl say sprintf('GPU: %.2f%% idle, %.2f%% busy', · line 765
Quoted 10 lines
scripts/trace.pl +10
-0+ + $ctx_table .= '<table>'; + + foreach my $ctx (@ctxts) { + $ctx_table .= "<tr>\n"; + $ctx_table .= " <td>Context $ctx</td>\n" if $colour_contexts; + foreach my $state (@states) { + $ctx_table .= " <td align='center' valign='middle'><div style='" . box_style($ctx, $state) . " padding-top: 6px; padding-bottom: 6px; padding-left: 6x; padding-right: 6px;'>" . uc($state) . "</div></td>\n"; + } + $ctx_table .= "</tr>\n";
As mentioned in the email, I would definitely add something like:
my $i = 0;
...
if( $i++ > 5) {
$ctx_table .= "<tr><td>etc...</td></tr>";
last;
}
I've had to analyse traces with hundreds of contexts in them before now. Which would produce an HTML file swamped by a huge legend (albeit with a beautifully smoothly shaded rainbow!).
With that...
Quoted 83 lines
scripts/trace.pl +18
-24+ } + + $ctx_table .= '</table>'; +} sub generate_ctx_colours {@@ -754,12 +780,7 @@ sub generate_ctx_colours generate_ctx_colours() if $html and $colour_contexts; - -my $queued_style = box_style($min_ctx, 'queue'); -my $ready_style = box_style($min_ctx, 'ready'); -my $execute_style = box_style($min_ctx, 'execute'); -my $ctxsave_style = box_style($min_ctx, 'ctxsave'); -my $incomplete_style = box_style($min_ctx, 'incomplete'); +generate_ctx_table() if $html; print <<ENDHTML if $html; <!DOCTYPE HTML>@@ -778,35 +799,27 @@ print <<ENDHTML if $html; </head> <body> <p> -<b>Timeline request view:</b> +<b>Timeline request view is $timeline_text:</b> <table> -<tr><td colspan='4'>$timeline_text</td></tr> <tr> -<td align='center'><div style='$queued_style'>QUEUED</div></td> -<td align='center'><div style='$ready_style'>READY</div></td> -<td align='center'><div style='$execute_style'>EXECUTE</div></td> -<td align='center'><div style='$ctxsave_style'>CTXSAVE</div></td> -</tr><tr> -<td></td> -<td></td> -<td align='center'><div style='$incomplete_style'>(INCOMPLETE)</div></td> -<td></td> -</tr/></table> -</p> -<p> -<small> -QUEUED = requests executing on the GPU<br> +<td> +$ctx_table +</td> +<td> +QUEUE = requests executing on the GPU<br> READY = runnable requests waiting for a slot on GPU<br> EXECUTE = requests waiting on fences and dependencies before they are runnable<br> CTXSAVE = GPU saving the context image<br> -</small> -</p> +INCOMPLETE = request of unknown completion time <p> Boxes are in format 'ctx-id/seqno'. </p> <p> Use Ctrl+scroll-action to zoom-in/out and scroll-action or dragging to move around the timeline. </p> +</td> +</tr> +</table> <p> <b>GPU idle: $flat_busy{'gpu-idle'}%</b> <br>@@ -975,20 +988,24 @@ sub box_style { my ($ctx, $stage) = @_; my $deg; + my $text_col = 'white'; if ($stage eq 'queue') { $deg = 90; + $text_col = 'black' if $colour_contexts; } elsif ($stage eq 'ready') { $deg = 45; } elsif ($stage eq 'execute') { $deg = 0; + $text_col = 'black' if $colour_contexts; } elsif ($stage eq 'ctxsave') { $deg = 105;
scripts/trace.pl sub box_style · line 1011
Quoted 10 lines
scripts/trace.pl +2
-1+ $text_col = 'black' if $colour_contexts; } elsif ($stage eq 'incomplete') { $deg = 0; } - return 'color: black; background: repeating-linear-gradient(' . + return "color: $text_col; background: repeating-linear-gradient(" . $deg . 'deg, ' . ctx_colour($ctx, $stage, 1.0) . ', ' . ctx_colour($ctx, $stage, 1.0) . ' 10px, ' .
_______________________________________________
igt-dev mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/igt-dev
PATCH v3 5/9 trace.pl: Improved key/colours
From: John Harrison <[email protected]>
Improve the timeline legend to show actual context colours.
v2: (Tvrtko Ursulin)
* Commit msg.
* Tweak layout for more compactness and more readability.
v3:
* Limit number of shown contexts in the legend. (John Harrison)
--- scripts/trace.pl | 80 ++++++++++++++++++++++++++++++++---------------- 1 file changed, 54 insertions(+), 26 deletions(-)
scripts/trace.pl +54 -26
@@ -738,9 +738,46 @@ say sprintf('GPU: %.2f%% idle, %.2f%% busy', $flat_busy{'gpu-idle'}, $flat_busy{'gpu-busy'}) unless $html; my $timeline_text = $colour_contexts ? - 'Per context coloured shading like:' : 'Box shading like:'; + 'per context coloured shading like' : 'box shading like'; my %ctx_colours; +my $ctx_table; + +sub generate_ctx_table +{ + my @states = ('queue', 'ready', 'execute', 'ctxsave', 'incomplete'); + my $max_show = 6; + my (@ctxts, @disp_ctxts); + my $step; + + if( $colour_contexts ) { + @ctxts = sort keys %ctxdb; + } else { + @ctxts = ($min_ctx); + } + + # Limit number of shown context examples + $step = int(scalar(@ctxts) / $max_show); + if ($step) { + foreach my $i (0..$#ctxts) { + push @disp_ctxts, $ctxts[$i] unless $i % $step; + last if scalar(@disp_ctxts) == $max_show; + } + } + + $ctx_table .= '<table>'; + + foreach my $ctx (@disp_ctxts) { + $ctx_table .= "<tr>\n"; + $ctx_table .= " <td>Context $ctx</td>\n" if $colour_contexts; + foreach my $state (@states) { + $ctx_table .= " <td align='center' valign='middle'><div style='" . box_style($ctx, $state) . " padding-top: 6px; padding-bottom: 6px; padding-left: 6x; padding-right: 6px;'>" . uc($state) . "</div></td>\n"; + } + $ctx_table .= "</tr>\n"; + } + + $ctx_table .= '</table>'; +} sub generate_ctx_colours {
@@ -754,12 +791,7 @@ sub generate_ctx_colours generate_ctx_colours() if $html and $colour_contexts; - -my $queued_style = box_style($min_ctx, 'queue'); -my $ready_style = box_style($min_ctx, 'ready'); -my $execute_style = box_style($min_ctx, 'execute'); -my $ctxsave_style = box_style($min_ctx, 'ctxsave'); -my $incomplete_style = box_style($min_ctx, 'incomplete'); +generate_ctx_table() if $html; print <<ENDHTML if $html; <!DOCTYPE HTML>
@@ -778,35 +810,27 @@ print <<ENDHTML if $html; </head> <body> <p> -<b>Timeline request view:</b> +<b>Timeline request view is $timeline_text:</b> <table> -<tr><td colspan='4'>$timeline_text</td></tr> <tr> -<td align='center'><div style='$queued_style'>QUEUED</div></td> -<td align='center'><div style='$ready_style'>READY</div></td> -<td align='center'><div style='$execute_style'>EXECUTE</div></td> -<td align='center'><div style='$ctxsave_style'>CTXSAVE</div></td> -</tr><tr> -<td></td> -<td></td> -<td align='center'><div style='$incomplete_style'>(INCOMPLETE)</div></td> -<td></td> -</tr/></table> -</p> -<p> -<small> -QUEUED = requests executing on the GPU<br> +<td> +$ctx_table +</td> +<td> +QUEUE = requests executing on the GPU<br> READY = runnable requests waiting for a slot on GPU<br> EXECUTE = requests waiting on fences and dependencies before they are runnable<br> CTXSAVE = GPU saving the context image<br> -</small> -</p> +INCOMPLETE = request of unknown completion time <p> Boxes are in format 'ctx-id/seqno'. </p> <p> Use Ctrl+scroll-action to zoom-in/out and scroll-action or dragging to move around the timeline. </p> +</td> +</tr> +</table> <p> <b>GPU idle: $flat_busy{'gpu-idle'}%</b> <br>
@@ -975,20 +999,24 @@ sub box_style { my ($ctx, $stage) = @_; my $deg; + my $text_col = 'white'; if ($stage eq 'queue') { $deg = 90; + $text_col = 'black' if $colour_contexts; } elsif ($stage eq 'ready') { $deg = 45; } elsif ($stage eq 'execute') { $deg = 0; + $text_col = 'black' if $colour_contexts; } elsif ($stage eq 'ctxsave') { $deg = 105; + $text_col = 'black' if $colour_contexts; } elsif ($stage eq 'incomplete') { $deg = 0; } - return 'color: black; background: repeating-linear-gradient(' . + return "color: $text_col; background: repeating-linear-gradient(" . $deg . 'deg, ' . ctx_colour($ctx, $stage, 1.0) . ', ' . ctx_colour($ctx, $stage, 1.0) . ' 10px, ' .
Signature · 6 lines
2.17.1 _______________________________________________ igt-dev mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/igt-dev
Re: [igt-dev] [PATCH i-g-t v3 5/9] trace.pl: Improved key/colours
Quoted 45 lines · Tvrtko Ursulin
From: John Harrison <[email protected]>
Improve the timeline legend to show actual context colours.
v2: (Tvrtko Ursulin)
* Commit msg.
* Tweak layout for more compactness and more readability.v3:
* Limit number of shown contexts in the legend. (John Harrison)Signed-off-by John Harrison <[email protected]>Signed-off-by Tvrtko Ursulin <[email protected]>Cc John Harrison <[email protected]>Cc Tvrtko Ursulin <[email protected]>Reviewed-by John Harrison <[email protected]>--- scripts/trace.pl | 80 ++++++++++++++++++++++++++++++++---------------- 1 file changed, 54 insertions(+), 26 deletions(-)scripts/trace.pl +14
-1@@ -738,9 +738,46 @@ say sprintf('GPU: %.2f%% idle, %.2f%% busy', $flat_busy{'gpu-idle'}, $flat_busy{'gpu-busy'}) unless $html; my $timeline_text = $colour_contexts ? - 'Per context coloured shading like:' : 'Box shading like:'; + 'per context coloured shading like' : 'box shading like'; my %ctx_colours; +my $ctx_table; + +sub generate_ctx_table +{ + my @states = ('queue', 'ready', 'execute', 'ctxsave', 'incomplete'); + my $max_show = 6; + my (@ctxts, @disp_ctxts); + my $step; + + if( $colour_contexts ) { + @ctxts = sort keys %ctxdb; + } else { + @ctxts = ($min_ctx);
scripts/trace.pl say sprintf('GPU: %.2f%% idle, %.2f%% busy', · line 766
Quoted 10 lines
scripts/trace.pl +10
-0+ } + + # Limit number of shown context examples + $step = int(scalar(@ctxts) / $max_show); + if ($step) { + foreach my $i (0..$#ctxts) { + push @disp_ctxts, $ctxts[$i] unless $i % $step; + last if scalar(@disp_ctxts) == $max_show; + } + }
else { @disp_ctxts = @ctxts;} ???
Currently, if step is zero then there is no legend. Otherwise...
Quoted 93 lines
scripts/trace.pl +28
-24+ + $ctx_table .= '<table>'; + + foreach my $ctx (@disp_ctxts) { + $ctx_table .= "<tr>\n"; + $ctx_table .= " <td>Context $ctx</td>\n" if $colour_contexts; + foreach my $state (@states) { + $ctx_table .= " <td align='center' valign='middle'><div style='" . box_style($ctx, $state) . " padding-top: 6px; padding-bottom: 6px; padding-left: 6x; padding-right: 6px;'>" . uc($state) . "</div></td>\n"; + } + $ctx_table .= "</tr>\n"; + } + + $ctx_table .= '</table>'; +} sub generate_ctx_colours {@@ -754,12 +791,7 @@ sub generate_ctx_colours generate_ctx_colours() if $html and $colour_contexts; - -my $queued_style = box_style($min_ctx, 'queue'); -my $ready_style = box_style($min_ctx, 'ready'); -my $execute_style = box_style($min_ctx, 'execute'); -my $ctxsave_style = box_style($min_ctx, 'ctxsave'); -my $incomplete_style = box_style($min_ctx, 'incomplete'); +generate_ctx_table() if $html; print <<ENDHTML if $html; <!DOCTYPE HTML>@@ -778,35 +810,27 @@ print <<ENDHTML if $html; </head> <body> <p> -<b>Timeline request view:</b> +<b>Timeline request view is $timeline_text:</b> <table> -<tr><td colspan='4'>$timeline_text</td></tr> <tr> -<td align='center'><div style='$queued_style'>QUEUED</div></td> -<td align='center'><div style='$ready_style'>READY</div></td> -<td align='center'><div style='$execute_style'>EXECUTE</div></td> -<td align='center'><div style='$ctxsave_style'>CTXSAVE</div></td> -</tr><tr> -<td></td> -<td></td> -<td align='center'><div style='$incomplete_style'>(INCOMPLETE)</div></td> -<td></td> -</tr/></table> -</p> -<p> -<small> -QUEUED = requests executing on the GPU<br> +<td> +$ctx_table +</td> +<td> +QUEUE = requests executing on the GPU<br> READY = runnable requests waiting for a slot on GPU<br> EXECUTE = requests waiting on fences and dependencies before they are runnable<br> CTXSAVE = GPU saving the context image<br> -</small> -</p> +INCOMPLETE = request of unknown completion time <p> Boxes are in format 'ctx-id/seqno'. </p> <p> Use Ctrl+scroll-action to zoom-in/out and scroll-action or dragging to move around the timeline. </p> +</td> +</tr> +</table> <p> <b>GPU idle: $flat_busy{'gpu-idle'}%</b> <br>@@ -975,20 +999,24 @@ sub box_style { my ($ctx, $stage) = @_; my $deg; + my $text_col = 'white'; if ($stage eq 'queue') { $deg = 90; + $text_col = 'black' if $colour_contexts; } elsif ($stage eq 'ready') { $deg = 45; } elsif ($stage eq 'execute') { $deg = 0; + $text_col = 'black' if $colour_contexts; } elsif ($stage eq 'ctxsave') { $deg = 105;
scripts/trace.pl sub box_style · line 1022
Quoted 10 lines
scripts/trace.pl +2
-1+ $text_col = 'black' if $colour_contexts; } elsif ($stage eq 'incomplete') { $deg = 0; } - return 'color: black; background: repeating-linear-gradient(' . + return "color: $text_col; background: repeating-linear-gradient(" . $deg . 'deg, ' . ctx_colour($ctx, $stage, 1.0) . ', ' . ctx_colour($ctx, $stage, 1.0) . ' 10px, ' .
_______________________________________________
igt-dev mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/igt-dev
PATCH v4 5/9 trace.pl: Improved key/colours
From: John Harrison <[email protected]>
Improve the timeline legend to show actual context colours.
v2: (Tvrtko Ursulin)
* Commit msg.
* Tweak layout for more compactness and more readability.
v3:
* Limit number of shown contexts in the legend. (John Harrison)
v4:
* Unbreak legend display with small context counts. (John Harrison)
--- scripts/trace.pl | 82 +++++++++++++++++++++++++++++++++--------------- 1 file changed, 56 insertions(+), 26 deletions(-)
scripts/trace.pl +56 -26
@@ -738,9 +738,48 @@ say sprintf('GPU: %.2f%% idle, %.2f%% busy', $flat_busy{'gpu-idle'}, $flat_busy{'gpu-busy'}) unless $html; my $timeline_text = $colour_contexts ? - 'Per context coloured shading like:' : 'Box shading like:'; + 'per context coloured shading like' : 'box shading like'; my %ctx_colours; +my $ctx_table; + +sub generate_ctx_table +{ + my @states = ('queue', 'ready', 'execute', 'ctxsave', 'incomplete'); + my $max_show = 6; + my (@ctxts, @disp_ctxts); + my $step; + + if( $colour_contexts ) { + @ctxts = sort keys %ctxdb; + } else { + @ctxts = ($min_ctx); + } + + # Limit number of shown context examples + $step = int(scalar(@ctxts) / $max_show); + if ($step) { + foreach my $i (0..$#ctxts) { + push @disp_ctxts, $ctxts[$i] unless $i % $step; + last if scalar(@disp_ctxts) == $max_show; + } + } else { + @disp_ctxts = @ctxts; + } + + $ctx_table .= '<table>'; + + foreach my $ctx (@disp_ctxts) { + $ctx_table .= "<tr>\n"; + $ctx_table .= " <td>Context $ctx</td>\n" if $colour_contexts; + foreach my $state (@states) { + $ctx_table .= " <td align='center' valign='middle'><div style='" . box_style($ctx, $state) . " padding-top: 6px; padding-bottom: 6px; padding-left: 6x; padding-right: 6px;'>" . uc($state) . "</div></td>\n"; + } + $ctx_table .= "</tr>\n"; + } + + $ctx_table .= '</table>'; +} sub generate_ctx_colours {
@@ -754,12 +793,7 @@ sub generate_ctx_colours generate_ctx_colours() if $html and $colour_contexts; - -my $queued_style = box_style($min_ctx, 'queue'); -my $ready_style = box_style($min_ctx, 'ready'); -my $execute_style = box_style($min_ctx, 'execute'); -my $ctxsave_style = box_style($min_ctx, 'ctxsave'); -my $incomplete_style = box_style($min_ctx, 'incomplete'); +generate_ctx_table() if $html; print <<ENDHTML if $html; <!DOCTYPE HTML>
@@ -778,35 +812,27 @@ print <<ENDHTML if $html; </head> <body> <p> -<b>Timeline request view:</b> +<b>Timeline request view is $timeline_text:</b> <table> -<tr><td colspan='4'>$timeline_text</td></tr> <tr> -<td align='center'><div style='$queued_style'>QUEUED</div></td> -<td align='center'><div style='$ready_style'>READY</div></td> -<td align='center'><div style='$execute_style'>EXECUTE</div></td> -<td align='center'><div style='$ctxsave_style'>CTXSAVE</div></td> -</tr><tr> -<td></td> -<td></td> -<td align='center'><div style='$incomplete_style'>(INCOMPLETE)</div></td> -<td></td> -</tr/></table> -</p> -<p> -<small> -QUEUED = requests executing on the GPU<br> +<td> +$ctx_table +</td> +<td> +QUEUE = requests executing on the GPU<br> READY = runnable requests waiting for a slot on GPU<br> EXECUTE = requests waiting on fences and dependencies before they are runnable<br> CTXSAVE = GPU saving the context image<br> -</small> -</p> +INCOMPLETE = request of unknown completion time <p> Boxes are in format 'ctx-id/seqno'. </p> <p> Use Ctrl+scroll-action to zoom-in/out and scroll-action or dragging to move around the timeline. </p> +</td> +</tr> +</table> <p> <b>GPU idle: $flat_busy{'gpu-idle'}%</b> <br>
@@ -975,20 +1001,24 @@ sub box_style { my ($ctx, $stage) = @_; my $deg; + my $text_col = 'white'; if ($stage eq 'queue') { $deg = 90; + $text_col = 'black' if $colour_contexts; } elsif ($stage eq 'ready') { $deg = 45; } elsif ($stage eq 'execute') { $deg = 0; + $text_col = 'black' if $colour_contexts; } elsif ($stage eq 'ctxsave') { $deg = 105; + $text_col = 'black' if $colour_contexts; } elsif ($stage eq 'incomplete') { $deg = 0; } - return 'color: black; background: repeating-linear-gradient(' . + return "color: $text_col; background: repeating-linear-gradient(" . $deg . 'deg, ' . ctx_colour($ctx, $stage, 1.0) . ', ' . ctx_colour($ctx, $stage, 1.0) . ' 10px, ' .
Signature · 6 lines
2.17.1 _______________________________________________ Intel-gfx mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/intel-gfx
Re: [igt-dev] [PATCH i-g-t v4 5/9] trace.pl: Improved key/colours
Quoted 153 lines · Tvrtko Ursulin
From: John Harrison <[email protected]>
Improve the timeline legend to show actual context colours.
v2: (Tvrtko Ursulin)
* Commit msg.
* Tweak layout for more compactness and more readability.v3:
* Limit number of shown contexts in the legend. (John Harrison)v4:
* Unbreak legend display with small context counts. (John Harrison)Signed-off-by John Harrison <[email protected]>Signed-off-by Tvrtko Ursulin <[email protected]>Cc John Harrison <[email protected]>Cc Tvrtko Ursulin <[email protected]>Reviewed-by John Harrison <[email protected]>--- scripts/trace.pl | 82 +++++++++++++++++++++++++++++++++--------------- 1 file changed, 56 insertions(+), 26 deletions(-)scripts/trace.pl +54
-25@@ -738,9 +738,48 @@ say sprintf('GPU: %.2f%% idle, %.2f%% busy', $flat_busy{'gpu-idle'}, $flat_busy{'gpu-busy'}) unless $html; my $timeline_text = $colour_contexts ? - 'Per context coloured shading like:' : 'Box shading like:'; + 'per context coloured shading like' : 'box shading like'; my %ctx_colours; +my $ctx_table; + +sub generate_ctx_table +{ + my @states = ('queue', 'ready', 'execute', 'ctxsave', 'incomplete'); + my $max_show = 6; + my (@ctxts, @disp_ctxts); + my $step; + + if( $colour_contexts ) { + @ctxts = sort keys %ctxdb; + } else { + @ctxts = ($min_ctx); + } + + # Limit number of shown context examples + $step = int(scalar(@ctxts) / $max_show); + if ($step) { + foreach my $i (0..$#ctxts) { + push @disp_ctxts, $ctxts[$i] unless $i % $step; + last if scalar(@disp_ctxts) == $max_show; + } + } else { + @disp_ctxts = @ctxts; + } + + $ctx_table .= '<table>'; + + foreach my $ctx (@disp_ctxts) { + $ctx_table .= "<tr>\n"; + $ctx_table .= " <td>Context $ctx</td>\n" if $colour_contexts; + foreach my $state (@states) { + $ctx_table .= " <td align='center' valign='middle'><div style='" . box_style($ctx, $state) . " padding-top: 6px; padding-bottom: 6px; padding-left: 6x; padding-right: 6px;'>" . uc($state) . "</div></td>\n"; + } + $ctx_table .= "</tr>\n"; + } + + $ctx_table .= '</table>'; +} sub generate_ctx_colours {@@ -754,12 +793,7 @@ sub generate_ctx_colours generate_ctx_colours() if $html and $colour_contexts; - -my $queued_style = box_style($min_ctx, 'queue'); -my $ready_style = box_style($min_ctx, 'ready'); -my $execute_style = box_style($min_ctx, 'execute'); -my $ctxsave_style = box_style($min_ctx, 'ctxsave'); -my $incomplete_style = box_style($min_ctx, 'incomplete'); +generate_ctx_table() if $html; print <<ENDHTML if $html; <!DOCTYPE HTML>@@ -778,35 +812,27 @@ print <<ENDHTML if $html; </head> <body> <p> -<b>Timeline request view:</b> +<b>Timeline request view is $timeline_text:</b> <table> -<tr><td colspan='4'>$timeline_text</td></tr> <tr> -<td align='center'><div style='$queued_style'>QUEUED</div></td> -<td align='center'><div style='$ready_style'>READY</div></td> -<td align='center'><div style='$execute_style'>EXECUTE</div></td> -<td align='center'><div style='$ctxsave_style'>CTXSAVE</div></td> -</tr><tr> -<td></td> -<td></td> -<td align='center'><div style='$incomplete_style'>(INCOMPLETE)</div></td> -<td></td> -</tr/></table> -</p> -<p> -<small> -QUEUED = requests executing on the GPU<br> +<td> +$ctx_table +</td> +<td> +QUEUE = requests executing on the GPU<br> READY = runnable requests waiting for a slot on GPU<br> EXECUTE = requests waiting on fences and dependencies before they are runnable<br> CTXSAVE = GPU saving the context image<br> -</small> -</p> +INCOMPLETE = request of unknown completion time <p> Boxes are in format 'ctx-id/seqno'. </p> <p> Use Ctrl+scroll-action to zoom-in/out and scroll-action or dragging to move around the timeline. </p> +</td> +</tr> +</table> <p> <b>GPU idle: $flat_busy{'gpu-idle'}%</b> <br>@@ -975,20 +1001,24 @@ sub box_style { my ($ctx, $stage) = @_; my $deg; + my $text_col = 'white'; if ($stage eq 'queue') { $deg = 90; + $text_col = 'black' if $colour_contexts; } elsif ($stage eq 'ready') { $deg = 45; } elsif ($stage eq 'execute') { $deg = 0; + $text_col = 'black' if $colour_contexts; } elsif ($stage eq 'ctxsave') { $deg = 105;
scripts/trace.pl sub box_style · line 1024
Quoted 10 lines
scripts/trace.pl +2
-1+ $text_col = 'black' if $colour_contexts; } elsif ($stage eq 'incomplete') { $deg = 0; } - return 'color: black; background: repeating-linear-gradient(' . + return "color: $text_col; background: repeating-linear-gradient(" . $deg . 'deg, ' . ctx_colour($ctx, $stage, 1.0) . ', ' . ctx_colour($ctx, $stage, 1.0) . ' 10px, ' .
_______________________________________________
igt-dev mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/igt-dev
PATCH 6/9 trace.pl: Context save only applies to last request of a bunch
In reply to Tvrtko UrsulinFrom: Tvrtko Ursulin <[email protected]>
Skip accounting the context save time for anything but the last request of the coalesced bunch, and also skip drawing those boxes on the timeline.
--- scripts/trace.pl | 10 +++++++--- 1 file changed, 7 insertions(+), 3 deletions(-)
scripts/trace.pl +7 -3
@@ -607,7 +607,11 @@ foreach my $key (@sorted_keys) { $min_ctx = $db{$key}->{'ctx'} if not defined $min_ctx or $db{$key}->{'ctx'} < $min_ctx; - $db{$key}->{'context-complete-delay'} = $end - $notify; + unless (exists $db{$key}->{'no-end'}) { + $db{$key}->{'context-complete-delay'} = $end - $notify; + } else { + $db{$key}->{'context-complete-delay'} = 0; + } $db{$key}->{'execute-delay'} = $start - $db{$key}->{'submit'}; $db{$key}->{'submit-delay'} = $db{$key}->{'submit'} - $db{$key}->{'queue'}; $db{$key}->{'duration'} = $notify - $start;
@@ -624,7 +628,7 @@ foreach my $key (@sorted_keys) { $submit_avg{$ring} += $db{$key}->{'submit-delay'}; $execute_avg{$ring} += $db{$key}->{'execute-delay'}; - $ctxsave_avg{$ring} += $end - $notify; + $ctxsave_avg{$ring} += $db{$key}->{'context-complete-delay'}; } foreach my $ring (sort keys %batch_avg) {
@@ -1062,7 +1066,7 @@ foreach my $key (sort sortQueue keys %db) { } # user interrupt to context complete - unless (exists $skip_box{'ctxsave'}) { + unless (exists $skip_box{'ctxsave'} or exists $db{$key}->{'no-end'}) { $skey = -2 * $max_seqno * $ctx - 2 * $seqno; $style = box_style($ctx, 'ctxsave'); my $ctxsave = $db{$key}->{'end'} - $db{$key}->{'notify'};
Signature · 6 lines
2.17.1 _______________________________________________ Intel-gfx mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/intel-gfx
Re: [igt-dev] [PATCH i-g-t 6/9] trace.pl: Context save only applies to last request of a bunch
Quoted 8 lines · Tvrtko Ursulin
From: Tvrtko Ursulin <[email protected]>
Skip accounting the context save time for anything but the last request of the coalesced bunch, and also skip drawing those boxes on the timeline.
Signed-off-by Tvrtko Ursulin <[email protected]>---
_______________________________________________
igt-dev mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/igt-dev
PATCH 7/9 trace.pl: Fix incomplete request handling
In reply to Tvrtko UrsulinFrom: Tvrtko Ursulin <[email protected]>
Incomplete requests (no notify, no context complete) have to be corrected by looking at the engine timeline, and not the sorted-by-start-time view as was previously used.
Per-engine timelines are generated on demand and cached for later use.
v2: Find end of current context on the engine timeline instead of just
using the next request for adjusting the incomplete start time.
v3: Improve scaling with large datasets by only walking each engine
timeline once and some caching. (John Harrison)
v4:
* Fix logic fail from v3.
* Refactor the code a bit to separate the stages better.
* Do not account batches with unknown duration in avg stats.
* Handle two user interrupts with the same seqno.
* Handle user interrupt arriving after request_out.
v5:
* Update for class:instance.
--- scripts/trace.pl | 163 +++++++++++++++++++++++++++++++---------------- 1 file changed, 108 insertions(+), 55 deletions(-)
scripts/trace.pl +108 -55
@@ -462,16 +462,11 @@ while (<>) { die if exists $db{$key}->{'end'}; $db{$key}->{'end'} = $time; - if (exists $notify{$gkey}) { - $db{$key}->{'notify'} = $notify{$gkey}; - } else { - # No notify so far. Maybe it will arrive later which - # will be handled in the sanitation loop below. - $db{$key}->{'notify'} = $db{$key}->{'end'}; - $db{$key}->{'no-notify'} = 1; - } + $db{$key}->{'notify'} = $notify{$gkey} if exists $notify{$gkey}; } elsif ($tp_name eq 'i915:intel_engine_notify:') { - $notify{global_key($ring, $seqno)} = $time; + my $gkey = global_key($ring, $seqno); + + $notify{$gkey} = $time unless exists $notify{$gkey}; } elsif ($tp_name eq 'i915:intel_gpu_freq_change:') { push @freqs, [$prev_freq_ts, $time, $prev_freq] if $prev_freq; $prev_freq_ts = $time;
@@ -484,66 +479,116 @@ while (<>) { my $max_seqno = 0; foreach my $key (keys %db) { my $gkey = global_key($db{$key}->{'ring'}, $db{$key}->{'global'}); - my $notify = $notify{$gkey}; die unless exists $db{$key}->{'start'}; $max_seqno = $db{$key}->{'seqno'} if $db{$key}->{'seqno'} > $max_seqno; - unless (exists $db{$key}->{'end'}) { - # Context complete not received. - $db{$key}->{'no-end'} = 1; + # Notify arrived after context complete? + $db{$key}->{'notify'} = $notify{$gkey} if not exists $db{$key}->{'notify'} + and exists $notify{$gkey}; - if (defined($notify)) { - # No context complete due req merging - use notify. - $db{$key}->{'notify'} = $notify; - $db{$key}->{'end'} = $notify; - } else { - # No notify and no context complete - give up for now. - $db{$key}->{'incomplete'} = 1; - } - } else { - # Notify arrived after context complete. - if (exists $db{$key}->{'no-notify'} and defined($notify)) { - delete $db{$key}->{'no-notify'}; - $db{$key}->{'notify'} = $notify; - } + # No notify but we have end? + $db{$key}->{'notify'} = $db{$key}->{'end'} if exists $db{$key}->{'end'} and + not exists $db{$key}->{'notify'}; + + # If user interrupt arrived out of order push it back to be no later + # than request out. + if (exists $db{$key}->{'end'} and exists $db{$key}->{'notify'} and + $db{$key}->{'notify'} > $db{$key}->{'end'}) { + $db{$key}->{'notify'} = $db{$key}->{'end'}; } } -# Fix up incompletes my $key_count = scalar(keys %db); -foreach my $key (keys %db) { - next unless exists $db{$key}->{'incomplete'}; - # End the incomplete batch at the time next one starts - my $ring = $db{$key}->{'ring'}; - my $ctx = $db{$key}->{'ctx'}; - my $seqno = $db{$key}->{'seqno'}; - my $next_key; - my $i = 1; - my $end; - - do { - $next_key = db_key($ring, $ctx, $seqno + $i); - $i++; - } until ((exists $db{$next_key} and not exists $db{$next_key}->{'incomplete'}) - or $i > $key_count); # ugly stop hack +my %engine_timelines; - if (exists $db{$next_key}) { - $end = $db{$next_key}->{'end'}; - } else { - # No info at all, fake it: - $end = $db{$key}->{'start'} + 999; - } +sub sortEngine { + my $as = $db{$a}->{'global'}; + my $bs = $db{$b}->{'global'}; + my $val; + + $val = $as <=> $bs; - $db{$key}->{'notify'} = $end; - $db{$key}->{'end'} = $end; + die if $val == 0; + + return $val; } -# GPU time accounting -my (%running, %runnable, %queued, %batch_avg, %batch_total_avg, %batch_count); -my (%submit_avg, %execute_avg, %ctxsave_avg); +sub get_engine_timeline { + my ($ring) = @_; + my @timeline; + + return $engine_timelines{$ring} if exists $engine_timelines{$ring}; + + @timeline = grep { $db{$_}->{'ring'} eq $ring } keys %db; + # FIXME seqno restart + @timeline = sort sortEngine @timeline; + + $engine_timelines{$ring} = \@timeline; + + return \@timeline; +} + +# Fix up coalesced requests by ending them either when the following same +# context request with known end ends, or when a different context starts. +foreach my $gid (sort keys %rings) { + my $ring = $ringmap{$rings{$gid}}; + my $timeline = get_engine_timeline($ring); + my $last_complete = -1; + my $last_ctx = -1; + my $complete; + + foreach my $pos (0..$#{$timeline}) { + my $key = @{$timeline}[$pos]; + my ($ctx, $end); + + next if exists $db{$key}->{'end'}; + + $db{$key}->{'no-end'} = 1; + $ctx = $db{$key}->{'ctx'}; + + if ($pos > $last_complete or $ctx != $last_ctx) { + my $next = $pos; + + undef $complete; + + while ($next < $#{$timeline}) { + my $next_key = ${$timeline}[++$next]; + if ($ctx == $db{$next_key}->{'ctx'} and + exists $db{$next_key}->{'end'}) { + $last_ctx = $db{$next_key}->{'ctx'}; + $last_complete = $next; + $complete = $next_key; + last; + } + } + } + + if (defined $complete) { + if ($ctx == $db{$complete}->{'ctx'}) { + $end = $db{$complete}->{'end'}; + } else { + $end = $db{$complete}->{'start'}; + } + } else { + # No next submission. Use notify if available or give up. + if (exists $db{$key}->{'notify'}) { + $end = $db{$key}->{'notify'}; + } else { + $end = $db{$key}->{'start'}; + $db{$key}->{'incomplete'} = 1; + } + } + + unless (exists $db{$key}->{'notify'}) { + $db{$key}->{'notify'} = $end; + $db{$key}->{'no-notify'} = 1; + } + $db{$key}->{'end'} = $end; + } +} sub sortStart { my $as = $db{$a}->{'start'};
@@ -592,6 +637,10 @@ foreach my $key (@sorted_keys) { @sorted_keys = sort sortStart keys %db if $re_sort; +# GPU time accounting +my (%running, %runnable, %queued, %batch_avg, %batch_total_avg, %batch_count); +my (%submit_avg, %execute_avg, %ctxsave_avg); + my $last_ts = 0; my $first_ts; my $min_ctx;
@@ -614,7 +663,11 @@ foreach my $key (@sorted_keys) { } $db{$key}->{'execute-delay'} = $start - $db{$key}->{'submit'}; $db{$key}->{'submit-delay'} = $db{$key}->{'submit'} - $db{$key}->{'queue'}; - $db{$key}->{'duration'} = $notify - $start; + unless (exists $db{$key}->{'no-notify'}) { + $db{$key}->{'duration'} = $notify - $start; + } else { + $db{$key}->{'duration'} = 0; + } $running{$ring} += $end - $start if $correct_durations or not exists $db{$key}->{'no-end'};
Signature · 6 lines
2.17.1 _______________________________________________ Intel-gfx mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/intel-gfx
Re: [igt-dev] [PATCH i-g-t 7/9] trace.pl: Fix incomplete request handling
Quoted 21 lines · Tvrtko Ursulin
From: Tvrtko Ursulin <[email protected]>
Incomplete requests (no notify, no context complete) have to be corrected by looking at the engine timeline, and not the sorted-by-start-time view as was previously used.
Per-engine timelines are generated on demand and cached for later use.
v2: Find end of current context on the engine timeline instead of just
using the next request for adjusting the incomplete start time.v3: Improve scaling with large datasets by only walking each engine
timeline once and some caching. (John Harrison)v4:
* Fix logic fail from v3.
* Refactor the code a bit to separate the stages better.
* Do not account batches with unknown duration in avg stats.
* Handle two user interrupts with the same seqno.
* Handle user interrupt arriving afterrequest_out.
Quoted 7 lines
v5:
* Update for class:instance.Signed-off-by Tvrtko Ursulin <[email protected]>Cc John Harrison <[email protected]>---
_______________________________________________
igt-dev mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/igt-dev
PATCH 8/9 trace.pl: Basic preemption support
In reply to Tvrtko UrsulinFrom: Tvrtko Ursulin <[email protected]>
Just forget about earlier request_in events.
--- scripts/trace.pl | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-)
scripts/trace.pl +3 -1
@@ -438,7 +438,9 @@ while (<>) { } elsif ($tp_name eq 'i915:i915_request_in:') { my %req; - die if exists $db{$key}; + # preemption + delete $db{$key} if exists $db{$key}; + die unless exists $queue{$key}; die unless exists $submit{$key};
Signature · 6 lines
2.17.1 _______________________________________________ igt-dev mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/igt-dev
Re: [igt-dev] [PATCH i-g-t 8/9] trace.pl: Basic preemption support
Quoted 14 lines · Tvrtko Ursulin
From: Tvrtko Ursulin <[email protected]>
Just forget about earlier
request_inevents.Signed-off-by Tvrtko Ursulin <[email protected]>--- scripts/trace.pl | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-)scripts/trace.pl +0
-0
scripts/trace.pl while (<>) { · line 446
Quoted 10 lines
scripts/trace.pl +3
-1@@ -438,7 +438,9 @@ while (<>) { } elsif ($tp_name eq 'i915:i915_request_in:') { my %req; - die if exists $db{$key}; + # preemption + delete $db{$key} if exists $db{$key}; + die unless exists $queue{$key}; die unless exists $submit{$key};
_______________________________________________
igt-dev mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/igt-dev
PATCH 9/9 trace.pl: Fix request split mode
In reply to Tvrtko UrsulinFrom: Tvrtko Ursulin <[email protected]>
Request split mode had several bugs, both in the original version and also after the recent refactorings.
One big one was that it wasn't considering different submit ports as a reason to split execution, and also that it was too time based instead of looking at relevant timelines.
In this refactoring we address the former by using the engine timelines introduced in the previous patch. Secondary port submissions are moved to follow the preceding submission as a first step in the correction process.
In the second step, we add context timelines and use then in a similar fashion to separate start and end time of coalesced requests. For each coalesced request we know its boundaries by looking at the engine timeline (via global seqnos), and we know the previous request it should only start after, by looking at the context timeline.
v2:
* Remove some dead code.
* Fix !port0 shifting logic.
v3:
* Refactor for less list walking as with incomplete handling.
v4:
* Database of context timelines should not contain duplicates! (Converted from array into a hash.)
v5:
* Avoid over-accounting runnable time for a coalesced group by recording the time first request entered the GPU and ending the execute delay at that point for the whole group.
v6:
* Update for engine class:instance.
--- scripts/trace.pl | 138 ++++++++++++++++++++++++++++++++++++----------- 1 file changed, 108 insertions(+), 30 deletions(-)
scripts/trace.pl +108 -30
@@ -27,7 +27,7 @@ use warnings; use 5.010; my $gid = 0; -my (%db, %queue, %submit, %notify, %rings, %ctxdb, %ringmap, %reqwait); +my (%db, %queue, %submit, %notify, %rings, %ctxdb, %ringmap, %reqwait, %ctxtimelines); my @freqs; my $max_items = 3000;
@@ -448,6 +448,7 @@ while (<>) { $req{'ring'} = $ring; $req{'seqno'} = $seqno; $req{'ctx'} = $ctx; + $ctxtimelines{$ctx . '/' . $ring} = 1; $req{'name'} = $ctx . '/' . $seqno; $req{'global'} = $tp{'global'}; $req{'port'} = $tp{'port'};
@@ -603,41 +604,113 @@ sub sortStart { return $val; } -my @sorted_keys = sort sortStart keys %db; -my $re_sort = 0; +my $re_sort = 1; +my @sorted_keys; -die "Database changed size?!" unless scalar(@sorted_keys) == $key_count; +sub maybe_sort_keys +{ + if ($re_sort) { + @sorted_keys = sort sortStart keys %db; + $re_sort = 0; + die "Database changed size?!" unless scalar(@sorted_keys) == + $key_count; + } +} -foreach my $key (@sorted_keys) { - my $ring = $db{$key}->{'ring'}; - my $end = $db{$key}->{'end'}; +maybe_sort_keys(); + +my %ctx_timelines; + +sub sortContext { + my $as = $db{$a}->{'seqno'}; + my $bs = $db{$b}->{'seqno'}; + my $val; + + $val = $as <=> $bs; + + die if $val == 0; + + return $val; +} + +sub get_ctx_timeline { + my ($ctx, $ring, $key) = @_; + my @timeline; + + return $ctx_timelines{$key} if exists $ctx_timelines{$key}; + + @timeline = grep { $db{$_}->{'ring'} eq $ring and + $db{$_}->{'ctx'} == $ctx } @sorted_keys; + # FIXME seqno restart + @timeline = sort sortContext @timeline; + + $ctx_timelines{$key} = \@timeline; + + return \@timeline; +} + +# Split out merged batches if requested. +if ($correct_durations) { + # Shift !port0 requests start time to after the previous context on the + # same timeline has finished. + foreach my $gid (sort keys %rings) { + my $ring = $ringmap{$rings{$gid}}; + my $timeline = get_engine_timeline($ring); + my $complete; + + foreach my $pos (0..$#{$timeline}) { + my $key = @{$timeline}[$pos]; + my $prev = $complete; + my $pkey; + + $complete = $key unless exists $db{$key}->{'no-end'}; + $pkey = $complete; + + next if $db{$key}->{'port'} == 0; + + $pkey = $prev if $complete eq $key; + + die unless defined $pkey; + + $db{$key}->{'start'} = $db{$pkey}->{'end'}; + $db{$key}->{'start'} = $db{$pkey}->{'notify'} if $db{$key}->{'start'} > $db{$key}->{'end'}; + + die if $db{$key}->{'start'} > $db{$key}->{'end'}; - # correct duration of merged batches - if ($correct_durations and exists $db{$key}->{'no-end'}) { - my $ctx = $db{$key}->{'ctx'}; - my $seqno = $db{$key}->{'seqno'}; - my $start = $db{$key}->{'start'}; - my $next_key; - my $i = 1; - - do { - $next_key = db_key($ring, $ctx, $seqno + $i); - $i++; - } until (exists $db{$next_key} or $i > $key_count); # ugly stop hack - - # 20us tolerance - if (exists $db{$next_key} and $db{$next_key}->{'start'} < $start + 20) { - my $notify = $db{$key}->{'notify'}; $re_sort = 1; - $db{$next_key}->{'start'} = $notify; - $db{$next_key}->{'start'} = $db{$next_key}->{'end'} if $db{$next_key}->{'start'} > $db{$next_key}->{'end'}; - die if $db{$next_key}->{'start'} > $db{$next_key}->{'end'}; } - die if $start > $end; + } + + maybe_sort_keys(); + + # Batch with no-end (no request_out) means it was submitted as part of + # coalesced context. This means it's start time should be set to the end + # time of a following request on this context timeline. + foreach my $tkey (sort keys %ctxtimelines) { + my ($ctx, $ring) = split '/', $tkey; + my $timeline = get_ctx_timeline($ctx, $ring, $tkey); + my $last_complete = -1; + my $complete; + + foreach my $pos (0..$#{$timeline}) { + my $key = @{$timeline}[$pos]; + my $next_key; + + next unless exists $db{$key}->{'no-end'}; + last if $pos == $#{$timeline}; + + # Shift following request to start after the current one + $next_key = ${$timeline}[$pos + 1]; + if (exists $db{$key}->{'notify'}) { + $db{$next_key}->{'engine-start'} = $db{$next_key}->{'start'}; + $db{$next_key}->{'start'} = $db{$key}->{'notify'}; + $re_sort = 1; + } + } } } -@sorted_keys = sort sortStart keys %db if $re_sort; +maybe_sort_keys(); # GPU time accounting my (%running, %runnable, %queued, %batch_avg, %batch_total_avg, %batch_count);
@@ -651,6 +724,7 @@ foreach my $key (@sorted_keys) { my $ring = $db{$key}->{'ring'}; my $end = $db{$key}->{'end'}; my $start = $db{$key}->{'start'}; + my $engine_start = $db{$key}->{'engine_start'}; my $notify = $db{$key}->{'notify'}; $first_ts = $db{$key}->{'queue'} if not defined $first_ts or $db{$key}->{'queue'} < $first_ts;
@@ -663,7 +737,9 @@ foreach my $key (@sorted_keys) { } else { $db{$key}->{'context-complete-delay'} = 0; } - $db{$key}->{'execute-delay'} = $start - $db{$key}->{'submit'}; + + $engine_start = $db{$key}->{'start'} unless defined $engine_start; + $db{$key}->{'execute-delay'} = $engine_start - $db{$key}->{'submit'}; $db{$key}->{'submit-delay'} = $db{$key}->{'submit'} - $db{$key}->{'queue'}; unless (exists $db{$key}->{'no-notify'}) { $db{$key}->{'duration'} = $notify - $start;
@@ -1076,6 +1152,7 @@ my $i = 0; foreach my $key (sort sortQueue keys %db) { my ($name, $ctx, $seqno) = ($db{$key}->{'name'}, $db{$key}->{'ctx'}, $db{$key}->{'seqno'}); my ($queue, $start, $notify, $end) = ($db{$key}->{'queue'}, $db{$key}->{'start'}, $db{$key}->{'notify'}, $db{$key}->{'end'}); + my $engine_start = $db{$key}->{'engine-start'}; my $submit = $queue + $db{$key}->{'submit-delay'}; my ($content, $style); my $group = $engine_start_id + $rings{$db{$key}->{'ring'}};
@@ -1095,11 +1172,12 @@ foreach my $key (sort sortQueue keys %db) { } # execute to start + $engine_start = $db{$key}->{'start'} unless defined $engine_start; unless (exists $skip_box{'ready'}) { $skey = 2 * $max_seqno * $ctx + 2 * $seqno + 1; $style = box_style($ctx, 'ready'); $content = "<small>$name<br>$db{$key}->{'execute-delay'}us</small>"; - $startend = 'start: \'' . ts($submit) . '\', end: \'' . ts($start) . '\''; + $startend = 'start: \'' . ts($submit) . '\', end: \'' . ts($engine_start) . '\''; print "\t{id: $i, key: $skey, $type group: $group, subgroup: $subgroup, subgroupOrder: $subgroup, content: '$content', $startend, style: \'$style\'},\n"; $i++; }
Signature · 6 lines
2.17.1 _______________________________________________ igt-dev mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/igt-dev
Patchwork
CI[igt-dev] ✓ Fi.CI.BAT: success for trace.pl fixes and improvements (rev3)
In reply to Tvrtko Ursulin== Series Details ==
== Series Details ==
Series: trace.pl fixes and improvements (rev3)
URL : https://patchwork.freedesktop.org/series/46177/
State : success
== Summary ==
= CI Bug Log - changes from CI_DRM_4475 -> IGTPW_1564 =
== Summary - SUCCESS ==
No regressions found.
External URL: https://patchwork.freedesktop.org/api/1.0/series/46177/revisions/3/mbox/
== Known issues ==
Here are the changes found in IGTPW_1564 that come from known issues:
=== IGT changes ===
==== Issues hit ====
igt@kms_pipe_crc_basic@suspend-read-crc-pipe-b:
fi-snb-2520m: PASS -> INCOMPLETE (fdo#103713)
==== Warnings ====
igt@gem_exec_suspend@basic-s4-devices:
{fi-kbl-8809g}: DMESG-WARN (fdo#107139) -> INCOMPLETE (fdo#107139) {name}: This element is suppressed. This means it is ignored when computing
the status of the difference (SUCCESS, WARNING, or FAILURE).
fdo#103713 https://bugs.freedesktop.org/show_bug.cgi?id=103713
fdo#107139 https://bugs.freedesktop.org/show_bug.cgi?id=107139
== Participating hosts (45 -> 40) ==
Additional (1): fi-elk-e7500
Missing (6): fi-ilk-m540 fi-bxt-dsi fi-hsw-4200u fi-byt-squawks fi-ctg-p8600 fi-kbl-7560u
== Build changes ==
* IGT: IGT_4551 -> IGTPW_1564
CI_DRM_4475: 1b6f049d73237a170919604538e747b0282b0109 @ git://anongit.freedesktop.org/gfx-ci/linux
IGTPW_1564: https://intel-gfx-ci.01.org/tree/drm-tip/IGTPW_1564/
IGT_4551: 93cf6931b33e2c0f5b89c89b65817fe245ecc391 @ git://anongit.freedesktop.org/xorg/app/intel-gpu-tools
== Logs ==
For more details see: https://intel-gfx-ci.01.org/tree/drm-tip/IGTPW_1564/issues.html
_______________________________________________
igt-dev mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/igt-dev
Patchwork
CI[igt-dev] ✓ Fi.CI.IGT: success for trace.pl fixes and improvements (rev3)
In reply to Tvrtko Ursulin== Series Details ==
== Series Details ==
Series: trace.pl fixes and improvements (rev3)
URL : https://patchwork.freedesktop.org/series/46177/
State : success
== Summary ==
= CI Bug Log - changes from IGT_4551_full -> IGTPW_1564_full =
== Summary - WARNING ==
Minor unknown changes coming with IGTPW_1564_full need to be verified manually.
If you think the reported changes have nothing to do with the changes introduced in IGTPW_1564_full, please notify your bug team to allow them to document this new failure mode, which will reduce false positives in CI.
External URL: https://patchwork.freedesktop.org/api/1.0/series/46177/revisions/3/mbox/
== Possible new issues ==
Here are the unknown changes that may have been introduced in IGTPW_1564_full:
=== IGT changes ===
==== Warnings ====
igt@gem_exec_schedule@deep-bsd1:
shard-kbl: SKIP -> PASS +2== Known issues ==
Here are the changes found in IGTPW_1564_full that come from known issues:
=== IGT changes ===
==== Issues hit ====
igt@gem_userptr_blits@unsync-unmap-after-close:
shard-snb: PASS -> INCOMPLETE (fdo#105411)
igt@kms_flip@2x-dpms-vs-vblank-race:
shard-hsw: PASS -> FAIL (fdo#103060)
igt@kms_flip_tiling@flip-x-tiled:
shard-glk: PASS -> FAIL (fdo#107161)
igt@kms_rotation_crc@primary-rotation-180:
shard-snb: NOTRUN -> FAIL (fdo#103925)
==== Possible fixes ====
igt@gem_ctx_isolation@bcs0-s3:
shard-glk: FAIL (fdo#103375) -> PASS
igt@gem_partial_pwrite_pread@reads-uncached:
shard-snb: INCOMPLETE (fdo#105411) -> PASS
igt@kms_cursor_legacy@all-pipes-torture-move:
shard-snb: DMESG-WARN -> PASS
igt@kms_cursor_legacy@pipe-c-torture-move:
shard-apl: DMESG-WARN (fdo#107122) -> PASS
igt@kms_flip@2x-plain-flip-ts-check:
shard-glk: FAIL (fdo#100368) -> PASS
igt@kms_flip@flip-vs-expired-vblank-interruptible:
shard-glk: FAIL (fdo#102887) -> PASS
igt@kms_flip_tiling@flip-to-x-tiled:
shard-glk: FAIL (fdo#107161) -> PASS
igt@kms_plane_multiple@atomic-pipe-a-tiling-x:
shard-snb: FAIL (fdo#103166) -> PASS
igt@kms_setmode@basic:
shard-apl: FAIL (fdo#99912) -> PASS
igt@kms_vblank@pipe-a-ts-continuation-dpms-suspend:
shard-kbl: INCOMPLETE (fdo#103665) -> PASS
igt@kms_vblank@pipe-b-ts-continuation-suspend:
shard-apl: FAIL (fdo#103375) -> PASS fdo#100368 https://bugs.freedesktop.org/show_bug.cgi?id=100368
fdo#102887 https://bugs.freedesktop.org/show_bug.cgi?id=102887
fdo#103060 https://bugs.freedesktop.org/show_bug.cgi?id=103060
fdo#103166 https://bugs.freedesktop.org/show_bug.cgi?id=103166
fdo#103375 https://bugs.freedesktop.org/show_bug.cgi?id=103375
fdo#103665 https://bugs.freedesktop.org/show_bug.cgi?id=103665
fdo#103925 https://bugs.freedesktop.org/show_bug.cgi?id=103925
fdo#105411 https://bugs.freedesktop.org/show_bug.cgi?id=105411
fdo#107122 https://bugs.freedesktop.org/show_bug.cgi?id=107122
fdo#107161 https://bugs.freedesktop.org/show_bug.cgi?id=107161
fdo#99912 https://bugs.freedesktop.org/show_bug.cgi?id=99912
== Participating hosts (5 -> 5) ==
No changes in participating hosts
== Build changes ==
* IGT: IGT_4551 -> IGTPW_1564
* Linux: CI_DRM_4474 -> CI_DRM_4475
CI_DRM_4474: c298e0700edde3d016ae9b16d0ce2a098bee1022 @ git://anongit.freedesktop.org/gfx-ci/linux
CI_DRM_4475: 1b6f049d73237a170919604538e747b0282b0109 @ git://anongit.freedesktop.org/gfx-ci/linux
IGTPW_1564: https://intel-gfx-ci.01.org/tree/drm-tip/IGTPW_1564/
IGT_4551: 93cf6931b33e2c0f5b89c89b65817fe245ecc391 @ git://anongit.freedesktop.org/xorg/app/intel-gpu-tools
== Logs ==
For more details see: https://intel-gfx-ci.01.org/tree/drm-tip/IGTPW_1564/shards.html
_______________________________________________
igt-dev mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/igt-dev
Patchwork
CI[igt-dev] ✓ Fi.CI.BAT: success for trace.pl fixes and improvements (rev6)
In reply to Tvrtko Ursulin== Series Details ==
== Series Details ==
Series: trace.pl fixes and improvements (rev6)
URL : https://patchwork.freedesktop.org/series/46177/
State : success
== Summary ==
= CI Bug Log - changes from CI_DRM_4479 -> IGTPW_1577 =
== Summary - SUCCESS ==
No regressions found.
External URL: https://patchwork.freedesktop.org/api/1.0/series/46177/revisions/6/mbox/
== Known issues ==
Here are the changes found in IGTPW_1577 that come from known issues:
=== IGT changes ===
==== Issues hit ====
igt@kms_flip@basic-flip-vs-modeset:
fi-skl-6700hq: PASS -> DMESG-WARN (fdo#105998)
==== Possible fixes ====
igt@drv_module_reload@basic-reload:
fi-ilk-650: DMESG-WARN (fdo#106387) -> PASS +2
igt@kms_chamelium@hdmi-hpd-fast:
fi-kbl-7500u: FAIL (fdo#103841, fdo#102672) -> SKIP
==== Warnings ====
igt@gem_exec_suspend@basic-s4-devices:
{fi-kbl-8809g}: INCOMPLETE (fdo#107139) -> DMESG-WARN (fdo#107139) {name}: This element is suppressed. This means it is ignored when computing
the status of the difference (SUCCESS, WARNING, or FAILURE).
fdo#102672 https://bugs.freedesktop.org/show_bug.cgi?id=102672
fdo#103841 https://bugs.freedesktop.org/show_bug.cgi?id=103841
fdo#105998 https://bugs.freedesktop.org/show_bug.cgi?id=105998
fdo#106387 https://bugs.freedesktop.org/show_bug.cgi?id=106387
fdo#107139 https://bugs.freedesktop.org/show_bug.cgi?id=107139
== Participating hosts (46 -> 41) ==
Additional (1): fi-byt-j1900
Missing (6): fi-ilk-m540 fi-hsw-4200u fi-byt-squawks fi-bsw-cyan fi-ctg-p8600 fi-ivb-3520m
== Build changes ==
* IGT: IGT_4553 -> IGTPW_1577
CI_DRM_4479: c9c54a1c37adefd414d51ed919aab3f94794ee35 @ git://anongit.freedesktop.org/gfx-ci/linux
IGTPW_1577: https://intel-gfx-ci.01.org/tree/drm-tip/IGTPW_1577/
IGT_4553: 9d25d62bb6ff62694e5226b02d79075eab304402 @ git://anongit.freedesktop.org/xorg/app/intel-gpu-tools
== Logs ==
For more details see: https://intel-gfx-ci.01.org/tree/drm-tip/IGTPW_1577/issues.html
_______________________________________________
igt-dev mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/igt-dev
Patchwork
CI[igt-dev] ✗ Fi.CI.IGT: failure for trace.pl fixes and improvements (rev6)
In reply to Tvrtko Ursulin== Series Details ==
== Series Details ==
Series: trace.pl fixes and improvements (rev6)
URL : https://patchwork.freedesktop.org/series/46177/
State : failure
== Summary ==
= CI Bug Log - changes from IGT_4553_full -> IGTPW_1577_full =
== Summary - FAILURE ==
Serious unknown changes coming with IGTPW_1577_full absolutely need to be verified manually.
If you think the reported changes have nothing to do with the changes introduced in IGTPW_1577_full, please notify your bug team to allow them to document this new failure mode, which will reduce false positives in CI.
External URL: https://patchwork.freedesktop.org/api/1.0/series/46177/revisions/6/mbox/
== Possible new issues ==
Here are the unknown changes that may have been introduced in IGTPW_1577_full:
=== IGT changes ===
==== Possible regressions ====
igt@kms_draw_crc@fill-fb:
shard-glk: PASS -> FAIL
==== Warnings ====
igt@gem_exec_schedule@deep-blt:
shard-kbl: SKIP -> PASS +1
igt@pm_rc6_residency@rc6-accuracy:
shard-kbl: PASS -> SKIP +2
shard-snb: SKIP -> PASS== Known issues ==
Here are the changes found in IGTPW_1577_full that come from known issues:
=== IGT changes ===
==== Issues hit ====
igt@drv_suspend@forcewake:
shard-kbl: PASS -> INCOMPLETE (fdo#103665)
igt@kms_atomic_transition@1x-modeset-transitions-nonblocking-fencing:
shard-glk: PASS -> FAIL (fdo#105703)
igt@kms_busy@basic-flip-e:
shard-snb: SKIP -> INCOMPLETE (fdo#105411)
igt@kms_cursor_legacy@2x-nonblocking-modeset-vs-cursor-atomic:
shard-glk: PASS -> FAIL (fdo#105454, fdo#106509)
igt@kms_flip@2x-plain-flip-fb-recreate:
shard-hsw: PASS -> FAIL (fdo#100368)
igt@kms_flip_tiling@flip-y-tiled:
shard-glk: NOTRUN -> FAIL (fdo#107161)
igt@kms_frontbuffer_tracking@fbc-2p-primscrn-pri-shrfb-draw-mmap-cpu:
shard-hsw: PASS -> FAIL (fdo#103167, fdo#105682)
igt@kms_setmode@basic:
shard-apl: PASS -> FAIL (fdo#99912)
shard-glk: PASS -> FAIL (fdo#99912)
==== Possible fixes ====
igt@kms_flip@plain-flip-ts-check-interruptible:
shard-glk: FAIL (fdo#100368) -> PASS
igt@testdisplay:
shard-glk: INCOMPLETE (fdo#103359, k.org#198133) -> PASS fdo#100368 https://bugs.freedesktop.org/show_bug.cgi?id=100368
fdo#103167 https://bugs.freedesktop.org/show_bug.cgi?id=103167
fdo#103359 https://bugs.freedesktop.org/show_bug.cgi?id=103359
fdo#103665 https://bugs.freedesktop.org/show_bug.cgi?id=103665
fdo#105411 https://bugs.freedesktop.org/show_bug.cgi?id=105411
fdo#105454 https://bugs.freedesktop.org/show_bug.cgi?id=105454
fdo#105682 https://bugs.freedesktop.org/show_bug.cgi?id=105682
fdo#105703 https://bugs.freedesktop.org/show_bug.cgi?id=105703
fdo#106509 https://bugs.freedesktop.org/show_bug.cgi?id=106509
fdo#107161 https://bugs.freedesktop.org/show_bug.cgi?id=107161
fdo#99912 https://bugs.freedesktop.org/show_bug.cgi?id=99912
k.org#198133 https://bugzilla.kernel.org/show_bug.cgi?id=198133
== Participating hosts (5 -> 5) ==
No changes in participating hosts
== Build changes ==
* IGT: IGT_4553 -> IGTPW_1577
* Linux: CI_DRM_4476 -> CI_DRM_4479
CI_DRM_4476: b818fac0878147c6df45338cb515b9b7bd878b7f @ git://anongit.freedesktop.org/gfx-ci/linux
CI_DRM_4479: c9c54a1c37adefd414d51ed919aab3f94794ee35 @ git://anongit.freedesktop.org/gfx-ci/linux
IGTPW_1577: https://intel-gfx-ci.01.org/tree/drm-tip/IGTPW_1577/
IGT_4553: 9d25d62bb6ff62694e5226b02d79075eab304402 @ git://anongit.freedesktop.org/xorg/app/intel-gpu-tools
== Logs ==
For more details see: https://intel-gfx-ci.01.org/tree/drm-tip/IGTPW_1577/shards.html
_______________________________________________
igt-dev mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/igt-dev
Patchwork
CI[igt-dev] ✓ Fi.CI.BAT: success for trace.pl fixes and improvements (rev7)
In reply to Tvrtko Ursulin== Series Details ==
== Series Details ==
Series: trace.pl fixes and improvements (rev7)
URL : https://patchwork.freedesktop.org/series/46177/
State : success
== Summary ==
= CI Bug Log - changes from CI_DRM_4488 -> IGTPW_1592 =
== Summary - SUCCESS ==
No regressions found.
External URL: https://patchwork.freedesktop.org/api/1.0/series/46177/revisions/7/mbox/
== Known issues ==
Here are the changes found in IGTPW_1592 that come from known issues:
=== IGT changes ===
==== Possible fixes ====
igt@debugfs_test@read_all_entries:
fi-snb-2520m: INCOMPLETE (fdo#103713) -> PASS
igt@kms_chamelium@hdmi-hpd-fast:
fi-kbl-7500u: FAIL (fdo#103841, fdo#102672) -> SKIP fdo#102672 https://bugs.freedesktop.org/show_bug.cgi?id=102672
fdo#103713 https://bugs.freedesktop.org/show_bug.cgi?id=103713
fdo#103841 https://bugs.freedesktop.org/show_bug.cgi?id=103841
== Participating hosts (46 -> 40) ==
Missing (6): fi-ilk-m540 fi-hsw-4200u fi-byt-j1900 fi-byt-squawks fi-bsw-cyan fi-skl-6700hq
== Build changes ==
* IGT: IGT_4557 -> IGTPW_1592
CI_DRM_4488: a3a527faf129e7901fdff5309a7e6df8abd72d13 @ git://anongit.freedesktop.org/gfx-ci/linux
IGTPW_1592: https://intel-gfx-ci.01.org/tree/drm-tip/IGTPW_1592/
IGT_4557: 140a67c13aad2595ee6c72e41d14d35a793158b5 @ git://anongit.freedesktop.org/xorg/app/intel-gpu-tools
== Logs ==
For more details see: https://intel-gfx-ci.01.org/tree/drm-tip/IGTPW_1592/issues.html
_______________________________________________
igt-dev mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/igt-dev
Patchwork
CI[igt-dev] ✗ Fi.CI.IGT: failure for trace.pl fixes and improvements (rev7)
In reply to Tvrtko Ursulin== Series Details ==
== Series Details ==
Series: trace.pl fixes and improvements (rev7)
URL : https://patchwork.freedesktop.org/series/46177/
State : failure
== Summary ==
= CI Bug Log - changes from IGT_4557_full -> IGTPW_1592_full =
== Summary - FAILURE ==
Serious unknown changes coming with IGTPW_1592_full absolutely need to be verified manually.
If you think the reported changes have nothing to do with the changes introduced in IGTPW_1592_full, please notify your bug team to allow them to document this new failure mode, which will reduce false positives in CI.
External URL: https://patchwork.freedesktop.org/api/1.0/series/46177/revisions/7/mbox/
== Possible new issues ==
Here are the unknown changes that may have been introduced in IGTPW_1592_full:
=== IGT changes ===
==== Possible regressions ====
igt@gem_reg_read@timestamp-monotonic:
shard-snb: PASS -> DMESG-WARN
igt@kms_vblank@pipe-a-wait-busy-hang:
shard-snb: PASS -> FAIL
==== Warnings ====
igt@gem_busy@busy-render:
shard-snb: PASS -> SKIP +11
igt@gem_exec_schedule@deep-bsd2:
shard-kbl: PASS -> SKIP +3
igt@gem_mocs_settings@mocs-rc6-vebox:
shard-kbl: SKIP -> PASS== Known issues ==
Here are the changes found in IGTPW_1592_full that come from known issues:
=== IGT changes ===
==== Issues hit ====
igt@drv_suspend@shrink:
shard-snb: PASS -> FAIL (fdo#106886)
igt@kms_cursor_legacy@cursora-vs-flipa-toggle:
shard-glk: PASS -> DMESG-WARN (fdo#105763)
igt@kms_fbcon_fbt@fbc:
shard-glk: PASS -> FAIL (fdo#103833, fdo#105681)
igt@kms_frontbuffer_tracking@fbc-1p-offscren-pri-indfb-draw-pwrite:
shard-snb: NOTRUN -> DMESG-FAIL (fdo#103167)
igt@kms_frontbuffer_tracking@fbc-1p-primscrn-pri-indfb-draw-mmap-cpu:
shard-snb: PASS -> INCOMPLETE (fdo#105411)
igt@kms_rotation_crc@sprite-rotation-180:
shard-snb: PASS -> FAIL (fdo#103925)
igt@kms_universal_plane@cursor-fb-leak-pipe-c:
shard-apl: PASS -> FAIL (fdo#107241)
==== Possible fixes ====
{igt@gem_userptr_blits@readonly-mmap-unsync}:
shard-snb: INCOMPLETE (fdo#105411) -> SKIP
igt@kms_flip@2x-plain-flip-ts-check-interruptible:
shard-glk: FAIL (fdo#100368) -> PASS
igt@kms_plane@plane-panning-bottom-right-suspend-pipe-c-planes:
shard-glk: FAIL (fdo#103375) -> PASS {name}: This element is suppressed. This means it is ignored when computing
the status of the difference (SUCCESS, WARNING, or FAILURE).
fdo#100368 https://bugs.freedesktop.org/show_bug.cgi?id=100368
fdo#103167 https://bugs.freedesktop.org/show_bug.cgi?id=103167
fdo#103375 https://bugs.freedesktop.org/show_bug.cgi?id=103375
fdo#103833 https://bugs.freedesktop.org/show_bug.cgi?id=103833
fdo#103925 https://bugs.freedesktop.org/show_bug.cgi?id=103925
fdo#105411 https://bugs.freedesktop.org/show_bug.cgi?id=105411
fdo#105681 https://bugs.freedesktop.org/show_bug.cgi?id=105681
fdo#105763 https://bugs.freedesktop.org/show_bug.cgi?id=105763
fdo#106886 https://bugs.freedesktop.org/show_bug.cgi?id=106886
fdo#107241 https://bugs.freedesktop.org/show_bug.cgi?id=107241
== Participating hosts (5 -> 5) ==
No changes in participating hosts
== Build changes ==
* IGT: IGT_4557 -> IGTPW_1592
* Linux: CI_DRM_4487 -> CI_DRM_4488
CI_DRM_4487: 627ed020cac6a73f0a014537dac7191efbb57084 @ git://anongit.freedesktop.org/gfx-ci/linux
CI_DRM_4488: a3a527faf129e7901fdff5309a7e6df8abd72d13 @ git://anongit.freedesktop.org/gfx-ci/linux
IGTPW_1592: https://intel-gfx-ci.01.org/tree/drm-tip/IGTPW_1592/
IGT_4557: 140a67c13aad2595ee6c72e41d14d35a793158b5 @ git://anongit.freedesktop.org/xorg/app/intel-gpu-tools
== Logs ==
For more details see: https://intel-gfx-ci.01.org/tree/drm-tip/IGTPW_1592/shards.html
_______________________________________________
igt-dev mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/igt-dev