Skip to contentSkip to conversation

Adding custom bugzilla fields

DISCUSSION Ander Conselvan De Oliveira 5 participants 24 messages Jun 26, 2015 – Aug 28, 2015

lore
Share & tools
lore
#0 lore raw

Hi all,

I've been looking into creating custom fields in Bugzilla to help sort our bugs in a more manageable way. I did some testing in a private installation and came up with this proposal. In a nut shell, we would add the following two fields:

	i915 platform:	list of platforms affected by a bug;
	i915 features:  list of features affected by a bug.

Both would be multiple selection fields. The accepted values would be configured in the Bugzilla installation. The attached screenshots show how this would look like in the bug view and bug list views.

My expectation is that using those fields we could have a clearer view of which areas and/or platforms require more attention. For example, I attached a screenshot of a sample report breaking down the bugs per feature and platform. That report requires Bugzilla 5.0 (which hasn't been released yet) since prior versions didn't support reports with multiple selection fields. However, it is also possible to script a similar report, as the attached python script does. The output looks something like this:

Feature ALL ILK SNB BYT IVB HSW BDW BSW SKL

display - atomic         0    0    1    0    0    0    1    0    0
display - audio          0    0    0    0    0    0    0    0    0
display - DP             0    0    1    0    1    0    1    0    0
display - DP MST         0    0    0    0    0    0    0    0    0
display - DSI            0    0    0    0    0    0    0    0    0
display - eDP            0    0    0    0    0    0    0    0    0
display - fastboot       0    0    0    0    0    0    0    0    0
display - FBC            0    0    0    0    0    0    0    0    0
display - HDMI           0    0    0    0    0    0    0    0    0
display - IPS            0    0    0    0    0    0    0    0    0
display - LVDS           0    0    0    0    0    0    0    0    0
display - PSR            0    0    0    0    0    0    0    0    0
display - Other          0    0    0    0    0    0    0    0    0
GEM - execlists          0    0    0    0    0    0    0    0    0
GEM - PPGTT              0    0    0    0    0    0    0    0    0
GEM - Other              1    0    0    0    0    0    1    0    0
power - RC6              0    0    0    1    0    0    0    0    0
power - RCS              0    0    0    0    0    0    0    0    0
power - Other            0    0    0    0    0    0    0    0    0

So I would like to hear what other people think about this. Specially, about what should be in the features field. The values can change overtime, but would be good to have a good list from the start. The values above are an incomplete list I threw together while looking at different open bugs.

Thanks,
Ander

import bugzilla
import collections

bugzilla_url = "http://192.168.100.244/bugzilla/xmlrpc.cgi"

closed_states = ['RESOLVED', 'VERIFIED', 'CLOSED'] open_states = ["REOPENED", "NEEDINFO", "NEW", "ASSIGNED"]

def get_custom_fields_allowed_values(bz):

	fields = ['cf_i915_platform', 'cf_i915_features']
	r =  bz._proxy.Bug.fields({'names': fields,
				   'include_fields': ['values']})

	list_of_raw_values = [f['values'] for f in r['fields']]
	values = [[v['name'] for v in f] for f in list_of_raw_values]

	assert len(values) == len(fields)
	return dict(zip(fields, values))

def get_drm_intel_bugs(bz):
query = bz.build_query(product="DRI", component="DRM/Intel",
status=open_states)
return bz.query(query)

def split_bugs_per_field(bz, bugs, field):
split = collections.OrderedDict()

allowed_values = get_custom_fields_allowed_values(bz) for v in allowed_values[field]:
split[v] = []

for bug in bugs:
if field in bug.__dict__:
for value in bug.__dict__[field]:
split[value].append(bug)

return split

def intersect_lists(list1, list2):
return [v for v in list1 if v in list2]

def stringfy_list(l, spacing=5):
return ''.join([s.rjust(spacing) for s in map(str, l)])

if __name__ == "__main__":
bz = bugzilla.Bugzilla(url=bugzilla_url)

bugs = get_drm_intel_bugs(bz)

per_platform = split_bugs_per_field(bz, bugs, 'cf_i915_platform') per_feature = split_bugs_per_field(bz, bugs, 'cf_i915_features')

print "Feature".ljust(20), stringfy_list(per_platform)

for feature in per_feature:
common_count = []
for platform in per_platform:
common = intersect_lists(per_feature[feature],
per_platform[platform])
common_count.append(len(common))

print feature.ljust(20), stringfy_list(common_count)

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

#1 lore raw
Quoted 23 lines · Ander Conselvan De Oliveira

Hi all,

I've been looking into creating custom fields in Bugzilla to help sort our bugs in a more manageable way. I did some testing in a private installation and came up with this proposal. In a nut shell, we would add the following two fields:

	i915 platform:	list of platforms affected by a bug;
	i915 features:  list of features affected by a bug.

Both would be multiple selection fields. The accepted values would be configured in the Bugzilla installation. The attached screenshots show how this would look like in the bug view and bug list views.

My expectation is that using those fields we could have a clearer view of which areas and/or platforms require more attention. For example, I attached a screenshot of a sample report breaking down the bugs per feature and platform. That report requires Bugzilla 5.0 (which hasn't been released yet) since prior versions didn't support reports with multiple selection fields. However, it is also possible to script a similar report, as the attached python script does. The output looks something like this:

Quoted 8 lines

Feature ALL ILK SNB BYT IVB HSW BDW BSW SKL

display - atomic         0    0    1    0    0    0    1    0    0
display - audio          0    0    0    0    0    0    0    0    0
display - DP             0    0    1    0    1    0    1    0    0
display - DP MST         0    0    0    0    0    0    0    0    0
display - DSI            0    0    0    0    0    0    0    0    0
display - eDP            0    0    0    0    0    0    0    0    0
display - fastboot       0    0    0    0    0    0    0    0    0

Fastboot is very soonish no more (Maarten has patches to move it all into normal modeset code). I'd drop it.

The other missing bit is all the plane stuff, color manager and similar. Otoh this is new, so not clear yet what kind of bugs will be common. Imo better to wait and then maybe add more categories.

Quoted 9 lines
display - FBC            0    0    0    0    0    0    0    0    0
display - HDMI           0    0    0    0    0    0    0    0    0
display - IPS            0    0    0    0    0    0    0    0    0
display - LVDS           0    0    0    0    0    0    0    0    0
display - PSR            0    0    0    0    0    0    0    0    0
display - Other          0    0    0    0    0    0    0    0    0
GEM - execlists          0    0    0    0    0    0    0    0    0
GEM - PPGTT              0    0    0    0    0    0    0    0    0
GEM - Other              1    0    0    0    0    0    1    0    0

GEM - gpu hang

blows up all the time. And we need a bucket to catch all the userspace hangs which are reported against the kernel.

Quoted 2 lines
power - RC6              0    0    0    1    0    0    0    0    0
power - RCS              0    0    0    0    0    0    0    0    0

RPS is turbo stuff, and it's tightly coupled with rc6. Maybe instead just

power - GT

for all the GT related power saving features?

I'd also add

power - runtime PM

and

power - suspend/resume

here. Tons of stuff blows up here.

Quoted 8 lines

power - Other 0 0 0 0 0 0 0 0 0

So I would like to hear what other people think about this. Specially, about what should be in the features field. The values can change overtime, but would be good to have a good list from the start. The values above are an incomplete list I threw together while looking at different open bugs.

Maybe we need a bit more polish, but probably not worth it to spend too much time on the exact feature list. If we spot serious gaps we can always add more. And remove old ones which have gone out of favour (having that script handy somewhere would be good).
-Daniel

Quoted 75 lines

Thanks,
Ander

import bugzilla
import collections

bugzilla_url = "http://192.168.100.244/bugzilla/xmlrpc.cgi"

closed_states = ['RESOLVED', 'VERIFIED', 'CLOSED'] open_states = ["REOPENED", "NEEDINFO", "NEW", "ASSIGNED"]

def get_custom_fields_allowed_values(bz):

	fields = ['cf_i915_platform', 'cf_i915_features']
	r =  bz._proxy.Bug.fields({'names': fields,
				   'include_fields': ['values']})

	list_of_raw_values = [f['values'] for f in r['fields']]
	values = [[v['name'] for v in f] for f in list_of_raw_values]

	assert len(values) == len(fields)
	return dict(zip(fields, values))

def get_drm_intel_bugs(bz):
query = bz.build_query(product="DRI", component="DRM/Intel",
status=open_states)
return bz.query(query)

def split_bugs_per_field(bz, bugs, field):
split = collections.OrderedDict()

allowed_values = get_custom_fields_allowed_values(bz) for v in allowed_values[field]:
split[v] = []

for bug in bugs:
if field in bug.__dict__:
for value in bug.__dict__[field]:
split[value].append(bug)

return split

def intersect_lists(list1, list2):
return [v for v in list1 if v in list2]

def stringfy_list(l, spacing=5):
return ''.join([s.rjust(spacing) for s in map(str, l)])

if __name__ == "__main__":
bz = bugzilla.Bugzilla(url=bugzilla_url)

bugs = get_drm_intel_bugs(bz)

per_platform = split_bugs_per_field(bz, bugs, 'cf_i915_platform') per_feature = split_bugs_per_field(bz, bugs, 'cf_i915_features')

print "Feature".ljust(20), stringfy_list(per_platform)

for feature in per_feature:
common_count = []
for platform in per_platform:
common = intersect_lists(per_feature[feature],
per_platform[platform])
common_count.append(len(common))

print feature.ljust(20), stringfy_list(common_count)

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

Signature · 7 lines
Daniel Vetter
Software Engineer, Intel Corporation
http://blog.ffwll.ch
_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx
#2 lore raw
Quoted 4 lines · Daniel Vetter

Maybe we need a bit more polish, but probably not worth it to spend too much time on the exact feature list. If we spot serious gaps we can always add more. And remove old ones which have gone out of favour (having that script handy somewhere would be good).

That's the big one where just being strict about using summary keywords pays off - flexibility. From pov, I like the summary keywords as that is shown in the search window - so if custom fields are introduced, I want them visible in the results list.
-Chris

Signature · 5 lines
Chris Wilson, Intel Open Source Technology Centre
_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx
#3 lore raw
Quoted 10 lines · Chris Wilson
Quoted 4 lines · Daniel Vetter

Maybe we need a bit more polish, but probably not worth it to spend too much time on the exact feature list. If we spot serious gaps we can always add more. And remove old ones which have gone out of favour (having that script handy somewhere would be good).

That's the big one where just being strict about using summary keywords pays off - flexibility. From pov, I like the summary keywords as that is shown in the search window - so if custom fields are introduced, I want them visible in the results list.

Yeah visibility in search results was my big concern too. That's why I asked Ander to merge the 3 areas into one, since otherwise the search result page looks super wasteful. It can be shown by just adding more default fields.

Flexibility is another one, but the problem we have with rotating different people through bug duty is that without a list we don't have that consistency - heck even QA engineers fail to be consistent with platform tags.
-Daniel

Signature · 7 lines
Daniel Vetter
Software Engineer, Intel Corporation
http://blog.ffwll.ch
_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx
#4 lore raw
In reply to Chris Wilson
Quoted 10 lines · Chris Wilson
Quoted 4 lines · Daniel Vetter

Maybe we need a bit more polish, but probably not worth it to spend too much time on the exact feature list. If we spot serious gaps we can always add more. And remove old ones which have gone out of favour (having that script handy somewhere would be good).

That's the big one where just being strict about using summary keywords pays off - flexibility. From pov, I like the summary keywords as that is shown in the search window - so if custom fields are introduced, I want them visible in the results list.

You'll be able to add these to the search results list using the "change columns" link at the bottom.

BR,
Jani.

Signature · 5 lines
Jani Nikula, Intel Open Source Technology Center
_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx
#5 lore raw
In reply to Daniel Vetter
Quoted 83 lines · Daniel Vetter
Quoted 23 lines · Ander Conselvan De Oliveira

Hi all,

I've been looking into creating custom fields in Bugzilla to help sort our bugs in a more manageable way. I did some testing in a private installation and came up with this proposal. In a nut shell, we would add the following two fields:

	i915 platform:	list of platforms affected by a bug;
	i915 features:  list of features affected by a bug.

Both would be multiple selection fields. The accepted values would be configured in the Bugzilla installation. The attached screenshots show how this would look like in the bug view and bug list views.

My expectation is that using those fields we could have a clearer view of which areas and/or platforms require more attention. For example, I attached a screenshot of a sample report breaking down the bugs per feature and platform. That report requires Bugzilla 5.0 (which hasn't been released yet) since prior versions didn't support reports with multiple selection fields. However, it is also possible to script a similar report, as the attached python script does. The output looks something like this:

Quoted 8 lines

Feature ALL ILK SNB BYT IVB HSW BDW BSW SKL

display - atomic         0    0    1    0    0    0    1    0    0
display - audio          0    0    0    0    0    0    0    0    0
display - DP             0    0    1    0    1    0    1    0    0
display - DP MST         0    0    0    0    0    0    0    0    0
display - DSI            0    0    0    0    0    0    0    0    0
display - eDP            0    0    0    0    0    0    0    0    0
display - fastboot       0    0    0    0    0    0    0    0    0

Fastboot is very soonish no more (Maarten has patches to move it all into normal modeset code). I'd drop it.

The other missing bit is all the plane stuff, color manager and similar. Otoh this is new, so not clear yet what kind of bugs will be common. Imo better to wait and then maybe add more categories.

Quoted 9 lines
display - FBC            0    0    0    0    0    0    0    0    0
display - HDMI           0    0    0    0    0    0    0    0    0
display - IPS            0    0    0    0    0    0    0    0    0
display - LVDS           0    0    0    0    0    0    0    0    0
display - PSR            0    0    0    0    0    0    0    0    0
display - Other          0    0    0    0    0    0    0    0    0
GEM - execlists          0    0    0    0    0    0    0    0    0
GEM - PPGTT              0    0    0    0    0    0    0    0    0
GEM - Other              1    0    0    0    0    0    1    0    0

GEM - gpu hang

blows up all the time. And we need a bucket to catch all the userspace hangs which are reported against the kernel.

Quoted 2 lines
power - RC6              0    0    0    1    0    0    0    0    0
power - RCS              0    0    0    0    0    0    0    0    0

RPS is turbo stuff, and it's tightly coupled with rc6. Maybe instead just

power - GT

for all the GT related power saving features?

I'd also add

power - runtime PM

and

power - suspend/resume

here. Tons of stuff blows up here.

Quoted 8 lines

power - Other 0 0 0 0 0 0 0 0 0

So I would like to hear what other people think about this. Specially, about what should be in the features field. The values can change overtime, but would be good to have a good list from the start. The values above are an incomplete list I threw together while looking at different open bugs.

Quoted 3 lines

Maybe we need a bit more polish, but probably not worth it to spend too much time on the exact feature list. If we spot serious gaps we can always add more. And remove old ones which have gone out of favour

Here's what I got so far, after updating with your suggestions.

i915 platform:

ALL
SKL
BXT
BDW
BSW
HSW
IVB
BYT
SNB
ILK
I965G
I965GM
G45
GM45
PNV
G33
I945G
I945GM
I915G
I915GM
I865G
I85X
I845G
I830

i915 features:

display - atomic
display - audio
display - DP
display - DP MST
display - DSI
display - eDP
display - FBC
display - HDMI
display - IPS
display - LVDS
display - PSR
display - Other
GEM - execlists
GEM - PPGTT
GEM - GPU hang
GEM - Other
power - GT
power - runtime PM
power - suspend/resume
power - Other

Quoted 1 line

(having that script handy somewhere would be good).

It would be good to have a repository we can dump these to. Or at least a branch in i-g-t.

Ander

Quoted 79 lines

-Daniel

Quoted 75 lines

Thanks,
Ander

import bugzilla
import collections

bugzilla_url = "http://192.168.100.244/bugzilla/xmlrpc.cgi"

closed_states = ['RESOLVED', 'VERIFIED', 'CLOSED'] open_states = ["REOPENED", "NEEDINFO", "NEW", "ASSIGNED"]

def get_custom_fields_allowed_values(bz):

	fields = ['cf_i915_platform', 'cf_i915_features']
	r =  bz._proxy.Bug.fields({'names': fields,
				   'include_fields': ['values']})

	list_of_raw_values = [f['values'] for f in r['fields']]
	values = [[v['name'] for v in f] for f in list_of_raw_values]

	assert len(values) == len(fields)
	return dict(zip(fields, values))

def get_drm_intel_bugs(bz):
query = bz.build_query(product="DRI", component="DRM/Intel",
status=open_states)
return bz.query(query)

def split_bugs_per_field(bz, bugs, field):
split = collections.OrderedDict()

allowed_values = get_custom_fields_allowed_values(bz) for v in allowed_values[field]:
split[v] = []

for bug in bugs:
if field in bug.__dict__:
for value in bug.__dict__[field]:
split[value].append(bug)

return split

def intersect_lists(list1, list2):
return [v for v in list1 if v in list2]

def stringfy_list(l, spacing=5):
return ''.join([s.rjust(spacing) for s in map(str, l)])

if __name__ == "__main__":
bz = bugzilla.Bugzilla(url=bugzilla_url)

bugs = get_drm_intel_bugs(bz)

per_platform = split_bugs_per_field(bz, bugs, 'cf_i915_platform') per_feature = split_bugs_per_field(bz, bugs, 'cf_i915_features')

print "Feature".ljust(20), stringfy_list(per_platform)

for feature in per_feature:
common_count = []
for platform in per_platform:
common = intersect_lists(per_feature[feature],
per_platform[platform])
common_count.append(len(common))

print feature.ljust(20), stringfy_list(common_count)

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

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

#6 lore raw
Quoted 90 lines · Ander Conselvan De Oliveira
Quoted 83 lines · Daniel Vetter
Quoted 23 lines · Ander Conselvan De Oliveira

Hi all,

I've been looking into creating custom fields in Bugzilla to help sort our bugs in a more manageable way. I did some testing in a private installation and came up with this proposal. In a nut shell, we would add the following two fields:

	i915 platform:	list of platforms affected by a bug;
	i915 features:  list of features affected by a bug.

Both would be multiple selection fields. The accepted values would be configured in the Bugzilla installation. The attached screenshots show how this would look like in the bug view and bug list views.

My expectation is that using those fields we could have a clearer view of which areas and/or platforms require more attention. For example, I attached a screenshot of a sample report breaking down the bugs per feature and platform. That report requires Bugzilla 5.0 (which hasn't been released yet) since prior versions didn't support reports with multiple selection fields. However, it is also possible to script a similar report, as the attached python script does. The output looks something like this:

Quoted 8 lines

Feature ALL ILK SNB BYT IVB HSW BDW BSW SKL

display - atomic         0    0    1    0    0    0    1    0    0
display - audio          0    0    0    0    0    0    0    0    0
display - DP             0    0    1    0    1    0    1    0    0
display - DP MST         0    0    0    0    0    0    0    0    0
display - DSI            0    0    0    0    0    0    0    0    0
display - eDP            0    0    0    0    0    0    0    0    0
display - fastboot       0    0    0    0    0    0    0    0    0

Fastboot is very soonish no more (Maarten has patches to move it all into normal modeset code). I'd drop it.

The other missing bit is all the plane stuff, color manager and similar. Otoh this is new, so not clear yet what kind of bugs will be common. Imo better to wait and then maybe add more categories.

Quoted 9 lines
display - FBC            0    0    0    0    0    0    0    0    0
display - HDMI           0    0    0    0    0    0    0    0    0
display - IPS            0    0    0    0    0    0    0    0    0
display - LVDS           0    0    0    0    0    0    0    0    0
display - PSR            0    0    0    0    0    0    0    0    0
display - Other          0    0    0    0    0    0    0    0    0
GEM - execlists          0    0    0    0    0    0    0    0    0
GEM - PPGTT              0    0    0    0    0    0    0    0    0
GEM - Other              1    0    0    0    0    0    1    0    0

GEM - gpu hang

blows up all the time. And we need a bucket to catch all the userspace hangs which are reported against the kernel.

Quoted 2 lines
power - RC6              0    0    0    1    0    0    0    0    0
power - RCS              0    0    0    0    0    0    0    0    0

RPS is turbo stuff, and it's tightly coupled with rc6. Maybe instead just

power - GT

for all the GT related power saving features?

I'd also add

power - runtime PM

and

power - suspend/resume

here. Tons of stuff blows up here.

Quoted 8 lines

power - Other 0 0 0 0 0 0 0 0 0

So I would like to hear what other people think about this. Specially, about what should be in the features field. The values can change overtime, but would be good to have a good list from the start. The values above are an incomplete list I threw together while looking at different open bugs.

Quoted 3 lines

Maybe we need a bit more polish, but probably not worth it to spend too much time on the exact feature list. If we spot serious gaps we can always add more. And remove old ones which have gone out of favour

Here's what I got so far, after updating with your suggestions.

Quoted 7 lines

i915 platform:

ALL
SKL
BXT
BDW
BSW

CHV

Quoted 3 lines

HSW
IVB
BYT

VLV

Quoted 16 lines

SNB
ILK
I965G
I965GM
G45
GM45
PNV
G33
I945G
I945GM
I915G
I915GM
I865G
I85X
I845G
I830

Signature · 6 lines
Ville Syrjälä
Intel OTC
_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx
#7 lore raw
Quoted 93 lines · Ville Syrjälä
Quoted 92 lines · Ander Conselvan De Oliveira
Quoted 83 lines · Daniel Vetter
Quoted 23 lines · Ander Conselvan De Oliveira

Hi all,

I've been looking into creating custom fields in Bugzilla to help sort our bugs in a more manageable way. I did some testing in a private installation and came up with this proposal. In a nut shell, we would add the following two fields:

	i915 platform:	list of platforms affected by a bug;
	i915 features:  list of features affected by a bug.

Both would be multiple selection fields. The accepted values would be configured in the Bugzilla installation. The attached screenshots show how this would look like in the bug view and bug list views.

My expectation is that using those fields we could have a clearer view of which areas and/or platforms require more attention. For example, I attached a screenshot of a sample report breaking down the bugs per feature and platform. That report requires Bugzilla 5.0 (which hasn't been released yet) since prior versions didn't support reports with multiple selection fields. However, it is also possible to script a similar report, as the attached python script does. The output looks something like this:

Quoted 8 lines

Feature ALL ILK SNB BYT IVB HSW BDW BSW SKL

display - atomic         0    0    1    0    0    0    1    0    0
display - audio          0    0    0    0    0    0    0    0    0
display - DP             0    0    1    0    1    0    1    0    0
display - DP MST         0    0    0    0    0    0    0    0    0
display - DSI            0    0    0    0    0    0    0    0    0
display - eDP            0    0    0    0    0    0    0    0    0
display - fastboot       0    0    0    0    0    0    0    0    0

Fastboot is very soonish no more (Maarten has patches to move it all into normal modeset code). I'd drop it.

The other missing bit is all the plane stuff, color manager and similar. Otoh this is new, so not clear yet what kind of bugs will be common. Imo better to wait and then maybe add more categories.

Quoted 9 lines
display - FBC            0    0    0    0    0    0    0    0    0
display - HDMI           0    0    0    0    0    0    0    0    0
display - IPS            0    0    0    0    0    0    0    0    0
display - LVDS           0    0    0    0    0    0    0    0    0
display - PSR            0    0    0    0    0    0    0    0    0
display - Other          0    0    0    0    0    0    0    0    0
GEM - execlists          0    0    0    0    0    0    0    0    0
GEM - PPGTT              0    0    0    0    0    0    0    0    0
GEM - Other              1    0    0    0    0    0    1    0    0

GEM - gpu hang

blows up all the time. And we need a bucket to catch all the userspace hangs which are reported against the kernel.

Quoted 2 lines
power - RC6              0    0    0    1    0    0    0    0    0
power - RCS              0    0    0    0    0    0    0    0    0

RPS is turbo stuff, and it's tightly coupled with rc6. Maybe instead just

power - GT

for all the GT related power saving features?

I'd also add

power - runtime PM

and

power - suspend/resume

here. Tons of stuff blows up here.

Quoted 8 lines

power - Other 0 0 0 0 0 0 0 0 0

So I would like to hear what other people think about this. Specially, about what should be in the features field. The values can change overtime, but would be good to have a good list from the start. The values above are an incomplete list I threw together while looking at different open bugs.

Quoted 3 lines

Maybe we need a bit more polish, but probably not worth it to spend too much time on the exact feature list. If we spot serious gaps we can always add more. And remove old ones which have gone out of favour

Here's what I got so far, after updating with your suggestions.

i915 platform:

Quoted 7 lines
Quoted 5 lines

ALL
SKL
BXT
BDW
BSW

CHV

Or alternatively add CHT, in which case BSW and BYT could make sense.

Quoted 27 lines
Quoted 4 lines

HSW
IVB
BYT

VLV

Quoted 16 lines

SNB
ILK
I965G
I965GM
G45
GM45
PNV
G33
I945G
I945GM
I915G
I915GM
I865G
I85X
I845G
I830

Signature · 2 lines
Ville Syrjälä
Intel OTC
Signature · 6 lines
Ville Syrjälä
Intel OTC
_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx
#8 lore raw
Quoted 12 lines · Ville Syrjälä
Quoted 11 lines · Ville Syrjälä
Quoted 9 lines · Ander Conselvan De Oliveira

Here's what I got so far, after updating with your suggestions.

i915 platform:

ALL
SKL
BXT
BDW
BSW

Quoted 3 lines
Quoted 1 line

CHV

Or alternatively add CHT, in which case BSW and BYT could make sense.

For the record, I had a quick chat with Ville and updated the "BSW" entry to "BSW/CHT".

Ander

Quoted 25 lines
Quoted 25 lines
Quoted 5 lines

HSW
IVB
BYT

VLV

Quoted 16 lines

SNB
ILK
I965G
I965GM
G45
GM45
PNV
G33
I945G
I945GM
I915G
I915GM
I865G
I85X
I845G
I830

Quoted 4 lines
Quoted 3 lines
Signature · 2 lines
Ville Syrjälä
Intel OTC

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

#9 lore raw
In reply to Ander Conselvan De Oliveira
Quoted 4 lines · Ander Conselvan De Oliveira

Hi all,

I've been looking into creating custom fields in Bugzilla to help sort our bugs in a more manageable way.

[...]

Quoted 3 lines

So I would like to hear what other people think about this. Specially, about what should be in the features field. The values can change overtime, but would be good to have a good list from the start.

So here's the list after including Daniel's and Ville's feedback. If there's no more suggestions/objections, I'd like to create those fields already tomorrow. We can edit the list entries afterwards if needed.

i915 platform:

ALL
SKL
BXT
BDW
BSW/CHT
HSW
IVB
BYT
SNB
ILK
I965G
I965GM
G45
GM45
PNV
G33
I945G
I945GM
I915G
I915GM
I865G
I85X
I845G
I830

i915 features:

display - atomic
display - audio
display - DP
display - DP MST
display - DSI
display - eDP
display - FBC
display - HDMI
display - IPS
display - LVDS
display - PSR
display - Other
GEM - execlists
GEM - PPGTT
GEM - GPU hang
GEM - Other
power - GT
power - runtime PM
power - suspend/resume
power - Other

Thanks,
Ander

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

#10 lore raw
Quoted 45 lines · Ander Conselvan De Oliveira
Quoted 4 lines · Ander Conselvan De Oliveira

Hi all,

I've been looking into creating custom fields in Bugzilla to help sort our bugs in a more manageable way.

[...]

Quoted 3 lines

So I would like to hear what other people think about this. Specially, about what should be in the features field. The values can change overtime, but would be good to have a good list from the start.

So here's the list after including Daniel's and Ville's feedback. If there's no more suggestions/objections, I'd like to create those fields already tomorrow. We can edit the list entries afterwards if needed.

i915 platform:

ALL
SKL
BXT
BDW
BSW/CHT
HSW
IVB
BYT
SNB
ILK
I965G
I965GM
G45
GM45
PNV
G33
I945G
I945GM
I915G
I915GM
I865G
I85X
I845G
I830

i915 features:

Quoted 15 lines

display - atomic
display - audio
display - DP
display - DP MST
display - DSI
display - eDP
display - FBC
display - HDMI
display - IPS
display - LVDS
display - PSR
display - Other
GEM - execlists
GEM - PPGTT
GEM - GPU hang

GPU hang's are their own top level category. They can be anything from hardware failure, display faults, driver bugs and last but not least, userspace.

Presumably we can select multiple fields for bugs? -Chris

Signature · 5 lines
Chris Wilson, Intel Open Source Technology Centre
_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx
#11 lore raw
Quoted 62 lines · Chris Wilson
Quoted 45 lines · Ander Conselvan De Oliveira
Quoted 4 lines · Ander Conselvan De Oliveira

Hi all,

I've been looking into creating custom fields in Bugzilla to help sort our bugs in a more manageable way.

[...]

Quoted 3 lines

So I would like to hear what other people think about this. Specially, about what should be in the features field. The values can change overtime, but would be good to have a good list from the start.

So here's the list after including Daniel's and Ville's feedback. If there's no more suggestions/objections, I'd like to create those fields already tomorrow. We can edit the list entries afterwards if needed.

i915 platform:

ALL
SKL
BXT
BDW
BSW/CHT
HSW
IVB
BYT
SNB
ILK
I965G
I965GM
G45
GM45
PNV
G33
I945G
I945GM
I915G
I915GM
I865G
I85X
I845G
I830

i915 features:

Quoted 15 lines

display - atomic
display - audio
display - DP
display - DP MST
display - DSI
display - eDP
display - FBC
display - HDMI
display - IPS
display - LVDS
display - PSR
display - Other
GEM - execlists
GEM - PPGTT
GEM - GPU hang

Quoted 5 lines

GPU hang's are their own top level category. They can be anything from hardware failure, display faults, driver bugs and last but not least, userspace.

Presumably we can select multiple fields for bugs?

Yes, it's a multiple selection field.

Quoted 2 lines

-Chris

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

#12 lore raw
In reply to Ander Conselvan De Oliveira
Quoted 14 lines · Ander Conselvan De Oliveira

i915 features:

display - atomic
display - audio
display - DP
display - DP MST
display - DSI
display - eDP
display - FBC
display - HDMI
display - IPS
display - LVDS
display - PSR
display - Other

*gasp* huge omissions, backlight and hotplug!

Quoted 8 lines

GEM - execlists
GEM - PPGTT
GEM - GPU hang
GEM - Other
power - GT
power - runtime PM
power - suspend/resume
power - Other

Bikeshedding, would it be worth changing s, - ,/, so that it gets slightly more condensed int the search results?

What about perf?

BR,
Jani.

Signature · 5 lines
Jani Nikula, Intel Open Source Technology Center
_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx
#13 lore raw
Quoted 16 lines · Jani Nikula
Quoted 14 lines · Ander Conselvan De Oliveira

i915 features:

display - atomic
display - audio
display - DP
display - DP MST
display - DSI
display - eDP
display - FBC
display - HDMI
display - IPS
display - LVDS
display - PSR
display - Other

Quoted 1 line

*gasp* huge omissions, backlight and hotplug!

Added.

Quoted 11 lines
Quoted 8 lines

GEM - execlists
GEM - PPGTT
GEM - GPU hang
GEM - Other
power - GT
power - runtime PM
power - suspend/resume
power - Other

Bikeshedding, would it be worth changing s, - ,/, so that it gets slightly more condensed int the search results?

Yeah, it does indeed look better.

Quoted 1 line

What about perf?

I haven't thought about that, but I can add a "performance" entry to the list.

Ander

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

#14 lore raw
In reply to Ander Conselvan De Oliveira
Quoted 34 lines · Ander Conselvan De Oliveira
Quoted 4 lines · Ander Conselvan De Oliveira

Hi all,

I've been looking into creating custom fields in Bugzilla to help sort our bugs in a more manageable way.

[...]

Quoted 3 lines

So I would like to hear what other people think about this. Specially, about what should be in the features field. The values can change overtime, but would be good to have a good list from the start.

So here's the list after including Daniel's and Ville's feedback. If there's no more suggestions/objections, I'd like to create those fields already tomorrow. We can edit the list entries afterwards if needed.

i915 platform:

ALL
SKL
BXT
BDW
BSW/CHT
HSW
IVB
BYT
SNB
ILK
I965G
I965GM
G45
GM45
PNV
G33

Quoted 8 lines

I945G
I945GM
I915G
I915GM
I865G
I85X
I845G
I830

I think we can condense the older platforms down a lot, maybe even just GEN2, GEN3, GEN4. Or at least group desktop and mobile together, i.e.

I8XX, I915, I945, G33/PNV, I965, G4X (we put both under that tag usually).

I'd also vote to not split up platforms which are just tailored chips for specific markets, e.g. chv/bsw is pretty much the same thing really from the driver pov.
-Daniel

Quoted 28 lines

i915 features:

display - atomic
display - audio
display - DP
display - DP MST
display - DSI
display - eDP
display - FBC
display - HDMI
display - IPS
display - LVDS
display - PSR
display - Other
GEM - execlists
GEM - PPGTT
GEM - GPU hang
GEM - Other
power - GT
power - runtime PM
power - suspend/resume
power - Other

Thanks,
Ander

Signature · 7 lines
Daniel Vetter
Software Engineer, Intel Corporation
http://blog.ffwll.ch
_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx
#15 lore raw
Quoted 4 lines · Daniel Vetter

I think we can condense the older platforms down a lot, maybe even just GEN2, GEN3, GEN4. Or at least group desktop and mobile together, i.e.

I8XX, I915, I945, G33/PNV, I965, G4X (we put both under that tag usually).

I would actually vote for the opposite. We often find bugs that are GT specific in modern chips, and each of the early gen have different features sets and their own quirks.

So I would say split it by PCI ID, with human readable strings, both codename (inc. target feature set, GT + ULT/ULX/Halo etc) and market name (though given that 50% is just HD Graphics, very meh).

So long as can intelligently group based on the PCI ID, I think it will be of benefit long term.
-Chris

Signature · 5 lines
Chris Wilson, Intel Open Source Technology Centre
_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx
#16 lore raw
Quoted 10 lines · Chris Wilson
Quoted 4 lines · Daniel Vetter

I think we can condense the older platforms down a lot, maybe even just GEN2, GEN3, GEN4. Or at least group desktop and mobile together, i.e.

I8XX, I915, I945, G33/PNV, I965, G4X (we put both under that tag usually).

I would actually vote for the opposite. We often find bugs that are GT specific in modern chips, and each of the early gen have different features sets and their own quirks.

Quoted 6 lines

So I would say split it by PCI ID, with human readable strings, both codename (inc. target feature set, GT + ULT/ULX/Halo etc) and market name (though given that 50% is just HD Graphics, very meh).

So long as can intelligently group based on the PCI ID, I think it will be of benefit long term.

Seems like that'll result in massive lists. And we can always denote special-cases in the summary of the bug. Also maintaining all that information accurately will be a pain and for many bugs it's not needed, hence why I think a bit more coarseness would be good. -Daniel

Signature · 7 lines
Daniel Vetter
Software Engineer, Intel Corporation
http://blog.ffwll.ch
_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx
#17 lore raw
Quoted 15 lines · Daniel Vetter
Quoted 14 lines · Chris Wilson
Quoted 4 lines · Daniel Vetter

I think we can condense the older platforms down a lot, maybe even just GEN2, GEN3, GEN4. Or at least group desktop and mobile together, i.e.

I8XX, I915, I945, G33/PNV, I965, G4X (we put both under that tag usually).

I would actually vote for the opposite. We often find bugs that are GT specific in modern chips, and each of the early gen have different features sets and their own quirks.

So I would say split it by PCI ID, with human readable strings, both codename (inc. target feature set, GT + ULT/ULX/Halo etc) and market name (though given that 50% is just HD Graphics, very meh).

Quoted 7 lines
Quoted 2 lines

So long as can intelligently group based on the PCI ID, I think it will be of benefit long term.

Seems like that'll result in massive lists. And we can always denote special-cases in the summary of the bug. Also maintaining all that information accurately will be a pain and for many bugs it's not needed, hence why I think a bit more coarseness would be good.

On the other hand, the more common task of working out which value should be used is trivial (and it gets harder and harder to remember the mappings over time, so make the computer do it!). I'm in favour of making bug reporting and bug maintenance simple. I'm sure with a bit of perl the bugzilla fields can be maintained directly from i915_pciids.h :-p -Chris

Signature · 5 lines
Chris Wilson, Intel Open Source Technology Centre
_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx
#18 lore raw
Quoted 16 lines · Chris Wilson
Quoted 15 lines · Daniel Vetter
Quoted 14 lines · Chris Wilson
Quoted 4 lines · Daniel Vetter

I think we can condense the older platforms down a lot, maybe even just GEN2, GEN3, GEN4. Or at least group desktop and mobile together, i.e.

I8XX, I915, I945, G33/PNV, I965, G4X (we put both under that tag usually).

I would actually vote for the opposite. We often find bugs that are GT specific in modern chips, and each of the early gen have different features sets and their own quirks.

So I would say split it by PCI ID, with human readable strings, both codename (inc. target feature set, GT + ULT/ULX/Halo etc) and market name (though given that 50% is just HD Graphics, very meh).

Quoted 7 lines
Quoted 7 lines
Quoted 2 lines

So long as can intelligently group based on the PCI ID, I think it will be of benefit long term.

Seems like that'll result in massive lists. And we can always denote special-cases in the summary of the bug. Also maintaining all that information accurately will be a pain and for many bugs it's not needed, hence why I think a bit more coarseness would be good.

Another addendum: Often those patterns are not apparent until we have many reports, at which point we have to go back through all the duplicates to look for patterns. Maintaining the field upfront would make those patterns stand out much earlier.
-Chris

Signature · 5 lines
Chris Wilson, Intel Open Source Technology Centre
_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx
#19 lore raw
In reply to Ander Conselvan De Oliveira
Quoted 8 lines · Ander Conselvan De Oliveira
Quoted 4 lines · Ander Conselvan De Oliveira

Hi all,

I've been looking into creating custom fields in Bugzilla to help sort our bugs in a more manageable way.

[...]

Quoted 7 lines
Quoted 3 lines

So I would like to hear what other people think about this. Specially, about what should be in the features field. The values can change overtime, but would be good to have a good list from the start.

So here's the list after including Daniel's and Ville's feedback. If there's no more suggestions/objections, I'd like to create those fields already tomorrow. We can edit the list entries afterwards if needed.

Thank you, everyone, for the input. The fields are now created.

Cheers,
Ander

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

#20 lore raw
Quoted 13 lines · Ander Conselvan De Oliveira
Quoted 12 lines · Ander Conselvan De Oliveira
Quoted 4 lines · Ander Conselvan De Oliveira

Hi all,

I've been looking into creating custom fields in Bugzilla to help sort our bugs in a more manageable way.

[...]

Quoted 3 lines

So I would like to hear what other people think about this. Specially, about what should be in the features field. The values can change overtime, but would be good to have a good list from the start.

Quoted 5 lines
Quoted 3 lines

So here's the list after including Daniel's and Ville's feedback. If there's no more suggestions/objections, I'd like to create those fields already tomorrow. We can edit the list entries afterwards if needed.

Thank you, everyone, for the input. The fields are now created.

There are some additional feature fields that I've found might be useful:

display/adapter (or display/dongle?)
display/hotplug
display/multi-gpu
GEM/prime (or GEM/dma-buf?)

other/IOMMU (or GEM/IOMMU?)
other/BIOS
other/module-reload (?)

performance?

BR,
Jani.

Signature · 5 lines
Jani Nikula, Intel Open Source Technology Center
_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx
#21 lore raw
Quoted 18 lines · Jani Nikula
Quoted 17 lines · Ander Conselvan De Oliveira
Quoted 15 lines · Ander Conselvan De Oliveira
Quoted 4 lines · Ander Conselvan De Oliveira

Hi all,

I've been looking into creating custom fields in Bugzilla to help sort our bugs in a more manageable way.

[...]

Quoted 3 lines

So I would like to hear what other people think about this. Specially, about what should be in the features field. The values can change overtime, but would be good to have a good list from the start.

So here's the list after including Daniel's and Ville's feedback. If there's no more suggestions/objections, I'd like to create those fields already tomorrow. We can edit the list entries afterwards if needed.

Quoted 6 lines
Quoted 1 line

Thank you, everyone, for the input. The fields are now created.

There are some additional feature fields that I've found might be useful:

display/adapter (or display/dongle?)

display/DP?

Quoted 1 line

display/hotplug

tag it with relevant output port?

Quoted 2 lines

display/multi-gpu
GEM/prime (or GEM/dma-buf?)

Yeah this might be useful, but not sure how much we care.

Quoted 2 lines

other/IOMMU (or GEM/IOMMU?)
other/BIOS

Imo better to label with relevant feature that gets broken. And IOMMU should always have dwmw on cc so he can escalate.

Quoted 1 line

other/module-reload (?)

This is BAT now and should never break ;-)

Quoted 1 line

performance?

Yeah this would be good I guess.

In general I'm vary of adding too many since it'll be hard to maintain. Can you please join the bug scrub org mtg on Thu so we can discuss these?

Thanks, Daniel

Signature · 7 lines
Daniel Vetter
Software Engineer, Intel Corporation
http://blog.ffwll.ch
_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx
#22 lore raw
Quoted 43 lines · Daniel Vetter
Quoted 18 lines · Jani Nikula
Quoted 17 lines · Ander Conselvan De Oliveira
Quoted 15 lines · Ander Conselvan De Oliveira
Quoted 4 lines · Ander Conselvan De Oliveira

Hi all,

I've been looking into creating custom fields in Bugzilla to help sort our bugs in a more manageable way.

[...]

Quoted 3 lines

So I would like to hear what other people think about this. Specially, about what should be in the features field. The values can change overtime, but would be good to have a good list from the start.

So here's the list after including Daniel's and Ville's feedback. If there's no more suggestions/objections, I'd like to create those fields already tomorrow. We can edit the list entries afterwards if needed.

Quoted 6 lines
Quoted 1 line

Thank you, everyone, for the input. The fields are now created.

There are some additional feature fields that I've found might be useful:

display/adapter (or display/dongle?)

display/DP?

Quoted 1 line

display/hotplug

tag it with relevant output port?

Quoted 2 lines

display/multi-gpu
GEM/prime (or GEM/dma-buf?)

Yeah this might be useful, but not sure how much we care.

Quoted 2 lines

other/IOMMU (or GEM/IOMMU?)
other/BIOS

Imo better to label with relevant feature that gets broken. And IOMMU should always have dwmw on cc so he can escalate.

Quoted 1 line

other/module-reload (?)

Quoted 7 lines

This is BAT now and should never break ;-)

Quoted 1 line

performance?

Yeah this would be good I guess.

In general I'm vary of adding too many since it'll be hard to maintain.

All of these spun from a few sessions of actually trying to tag the bugs that were/are still missing the tags.

Quoted 1 line

Can you please join the bug scrub org mtg on Thu so we can discuss these?

Ok.

Jani.

Quoted 6 lines

Thanks, Daniel

Signature · 3 lines
Daniel Vetter
Software Engineer, Intel Corporation
http://blog.ffwll.ch
Signature · 5 lines
Jani Nikula, Intel Open Source Technology Center
_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx
#23 lore raw
In reply to Daniel Vetter
Quoted 15 lines · Daniel Vetter
Quoted 14 lines · Jani Nikula
Quoted 13 lines · Ander Conselvan De Oliveira
Quoted 12 lines · Ander Conselvan De Oliveira
Quoted 4 lines · Ander Conselvan De Oliveira

Hi all,

I've been looking into creating custom fields in Bugzilla to help sort our bugs in a more manageable way.

[...]

Quoted 3 lines

So I would like to hear what other people think about this. Specially, about what should be in the features field. The values can change overtime, but would be good to have a good list from the start.

Quoted 8 lines
Quoted 8 lines
Quoted 5 lines
Quoted 3 lines

So here's the list after including Daniel's and Ville's feedback. If there's no more suggestions/objections, I'd like to create those fields already tomorrow. We can edit the list entries afterwards if needed.

Thank you, everyone, for the input. The fields are now created.

There are some additional feature fields that I've found might be useful:

The discussion yesterday never materialized, so I'll reply here.

Quoted 2 lines
Quoted 1 line

display/adapter (or display/dongle?)

display/DP?

This does not answer the question "I want to see all our adapter bugs".

Quoted 2 lines
Quoted 1 line

display/hotplug

tag it with relevant output port?

This does not answer the question "I want to see all our hotplug bugs".

Quoted 4 lines
Quoted 2 lines

display/multi-gpu
GEM/prime (or GEM/dma-buf?)

Yeah this might be useful, but not sure how much we care.

I might continue like above, because it's useful for finding dupes, but also it's nice to flag bugs we're not sure how much we care. If they end up being flagged, say, DP because gpu switching affects that, it's still a useful distinction to know it's not your average DP bug.

Quoted 4 lines
Quoted 2 lines

other/IOMMU (or GEM/IOMMU?)
other/BIOS

Imo better to label with relevant feature that gets broken.

This does not answer the question "I want to see all our bugs related to VBT/BIOS screwups".

Quoted 1 line

And IOMMU should always have dwmw on cc so he can escalate.

That is totally irrelevant to the discussion about fields.

Quoted 3 lines
Quoted 1 line

other/module-reload (?)

This is BAT now and should never break ;-)

Fine.

Quoted 5 lines
Quoted 1 line

performance?

Yeah this would be good I guess.

In general I'm vary of adding too many since it'll be hard to maintain.

Don't worry, we can remove them too.

Quoted 7 lines

Can you please join the bug scrub org mtg on Thu so we can discuss these?

Thanks, Daniel

Signature · 3 lines
Daniel Vetter
Software Engineer, Intel Corporation
http://blog.ffwll.ch
Signature · 5 lines
Jani Nikula, Intel Open Source Technology Center
_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx