Summary: | [EXA] xtest new failures with XAA->EXA | ||||||||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|---|
Product: | xorg | Reporter: | Pi, Fengming <fengming.pi> | ||||||||||
Component: | Driver/intel | Assignee: | Wang Zhenyu <zhenyu.z.wang> | ||||||||||
Status: | VERIFIED INVALID | QA Contact: | Xorg Project Team <xorg-team> | ||||||||||
Severity: | normal | ||||||||||||
Priority: | medium | CC: | eric, gordon.jin, hong.liu, michael.fu, randrik, zhenyu.z.wang | ||||||||||
Version: | git | ||||||||||||
Hardware: | x86 (IA32) | ||||||||||||
OS: | Linux (All) | ||||||||||||
Whiteboard: | |||||||||||||
i915 platform: | i915 features: | ||||||||||||
Bug Depends on: | |||||||||||||
Bug Blocks: | 13027, 13493 | ||||||||||||
Attachments: |
|
Description
Pi, Fengming
2007-10-25 01:27:39 UTC
following are some UNRES and WARN cases when using EXA but pass in XAA mode: TEST SECTION/CASE PURPOSE XAA NoAccel EXA Xlib8/XSetClipRectangles 3 pass pass UNRES Xlib8/XSetLineAttributes 1 pass pass UNRES Xlib9/XDrawArcs 52 pass pass WARN pls xorg.conf and log as usual... Created attachment 12202 [details]
xorg conf file
Created attachment 12203 [details]
Xtest in XAA mode report log
Created attachment 12204 [details]
Xtest in EXA mode report log
Assuming this doesn't happen with xserver 1.4, can you try and git-bisect the xserver change that introduced this? Created attachment 12231 [details] [review] Test patch Do the failures still occur with this patch? after applying Michel Dänzer's patch,around 90 purposes newly passed,some fail change into UNRES.more detail as folloing: TEST SECTION/CASE PURPOSE old new(applied patch) Xproto/SetModifierMapping 2 FAIL pass Xlib9/XCopyArea 24 FAIL pass Xlib9/XCopyArea 27 FAIL pass Xlib9/XCopyArea 29 FAIL pass Xlib9/XCopyArea 31 FAIL pass Xlib9/XCopyPlane 1 FAIL pass Xlib9/XCopyPlane 27 FAIL pass Xlib9/XCopyPlane 29 FAIL pass Xlib9/XCopyPlane 30 FAIL pass Xlib9/XCopyPlane 33 FAIL pass Xlib9/XCopyPlane 35 FAIL pass Xlib9/XCopyPlane 37 FAIL pass Xlib9/XDrawArc 16 pass FAIL Xlib9/XDrawArc 91 FAIL pass Xlib9/XDrawArcs 55 FAIL pass Xlib9/XDrawArcs 75 FAIL pass Xlib9/XDrawArcs 101 FAIL pass Xlib9/XDrawImageString16 8 FAIL pass Xlib9/XDrawImageString16 17 FAIL pass Xlib9/XDrawImageString16 19 FAIL pass Xlib9/XDrawImageString 8 FAIL pass Xlib9/XDrawImageString 17 FAIL pass Xlib9/XDrawImageString 19 FAIL pass Xlib9/XDrawLine 27 FAIL pass Xlib9/XDrawLine 52 FAIL pass Xlib9/XDrawLine 57 FAIL pass Xlib9/XDrawLine 80 FAIL pass Xlib9/XDrawLines 32 FAIL pass Xlib9/XDrawLines 57 FAIL pass Xlib9/XDrawLines 92 FAIL pass Xlib9/XDrawPoints 24 FAIL pass Xlib9/XDrawRectangle 5 FAIL pass Xlib9/XDrawRectangle 72 FAIL pass Xlib9/XDrawRectangles 27 FAIL pass Xlib9/XDrawRectangles 74 FAIL pass Xlib9/XDrawSegments 6 FAIL pass Xlib9/XDrawSegments 28 FAIL pass Xlib9/XDrawSegments 53 FAIL pass Xlib9/XDrawSegments 58 FAIL pass Xlib9/XDrawSegments 81 FAIL pass Xlib9/XDrawString 27 FAIL pass Xlib9/XDrawString 30 FAIL pass Xlib9/XDrawString 32 FAIL pass Xlib9/XDrawString 34 FAIL pass Xlib9/XDrawString 36 FAIL pass Xlib9/XDrawString 38 FAIL pass Xlib9/XDrawString16 27 FAIL pass Xlib9/XDrawString16 30 FAIL pass Xlib9/XDrawString16 32 FAIL pass Xlib9/XDrawString16 34 FAIL pass Xlib9/XDrawString16 36 FAIL pass Xlib9/XDrawString16 38 FAIL pass Xlib9/XDrawText 1 FAIL pass Xlib9/XDrawText 31 FAIL pass Xlib9/XDrawText 34 FAIL pass Xlib9/XDrawText 36 FAIL pass Xlib9/XDrawText 38 FAIL pass Xlib9/XDrawText 40 FAIL pass Xlib9/XDrawText 42 FAIL pass Xlib9/XDrawText16 31 FAIL pass Xlib9/XDrawText16 34 FAIL pass Xlib9/XDrawText16 36 FAIL pass Xlib9/XDrawText16 38 FAIL pass Xlib9/XDrawText16 40 FAIL pass Xlib9/XDrawText16 42 FAIL pass Xlib9/XFillArc 5 FAIL pass Xlib9/XFillArc 38 FAIL pass Xlib9/XFillArcs 40 FAIL pass Xlib9/XFillPolygon 43 FAIL pass Xlib9/XFillRectangle 37 FAIL pass Xlib9/XFillRectangles 38 FAIL pass Xlib9/XPutImage 3 FAIL pass Xlib11/EnterNotify 3 pass FAIL Xlib13/XSetModifierMapping 5 unsup FAIL Xlib13/XUngrabPointer 3 pass FAIL Xlib14/XmbLookupString 1 FAIL pass Xlib14/XwcLookupString 1 FAIL pass Xlib17/XKeycodeToKeysym 1 UNRES pass Xlib17/XKeysymToKeycode 1 FAIL pass Xlib17/XLookupString 1 FAIL pass Xlib17/XLookupString 2 FAIL pass Xlib17/XRefreshKeyboardMapping 1 UNRES pass Xt9/XtGrabKey 1 FAIL UNRES Xt9/XtGrabKey 2 FAIL UNRES Xt9/XtGrabKey 3 FAIL UNRES Xt9/XtUngrabKey 1 FAIL UNRES Xt9/XtUngrabKey 2 FAIL UNRES Xt9/XtGrabKeyboard 1 FAIL pass Xt12/XtSetKeyTranslator 1 UNRES pass Xt12/XtTranslateKeycode 1 UNRES pass Xt12/XtTranslateKeycode 2 UNRES pass Xt12/XtGetActionKeysym 1 UNRES pass Xt12/XtGetActionKeysym 2 UNRES pass Xt12/XtGetActionKeysym 3 UNRES pass Xt12/XtGetActionKeysym 4 UNRES pass Xt12/XtGetActionKeysym 5 UNRES pass Xt12/XtGetActionKeysym 6 UNRES pass Xt12/XtGetActionKeysym 7 UNRES pass Xt12/XtKeysymToKeycodeList 1 UNRES pass Xt12/XtRegisterGrabAction 1 FAIL UNRES (In reply to comment #8) > after applying Michel Dänzer's patch,around 90 purposes newly passed,some fail > change into UNRES.more detail as folloing: > TEST SECTION/CASE PURPOSE old new(applied patch) > > Xproto/SetModifierMapping 2 FAIL pass > Xlib11/EnterNotify 3 pass FAIL > Xlib13/XSetModifierMapping 5 unsup FAIL > Xlib13/XUngrabPointer 3 pass FAIL > Xlib14/XmbLookupString 1 FAIL pass > Xlib14/XwcLookupString 1 FAIL pass > Xlib17/XKeycodeToKeysym 1 UNRES pass > Xlib17/XKeysymToKeycode 1 FAIL pass > Xlib17/XLookupString 1 FAIL pass > Xlib17/XLookupString 2 FAIL pass > Xlib17/XRefreshKeyboardMapping 1 UNRES pass > Xt9/XtGrabKey 1 FAIL UNRES > Xt9/XtGrabKey 2 FAIL UNRES > Xt9/XtGrabKey 3 FAIL UNRES > Xt9/XtUngrabKey 1 FAIL UNRES > Xt9/XtUngrabKey 2 FAIL UNRES > Xt9/XtGrabKeyboard 1 FAIL pass > Xt12/XtSetKeyTranslator 1 UNRES pass > Xt12/XtTranslateKeycode 1 UNRES pass > Xt12/XtTranslateKeycode 2 UNRES pass > Xt12/XtGetActionKeysym 1 UNRES pass > Xt12/XtGetActionKeysym 2 UNRES pass > Xt12/XtGetActionKeysym 3 UNRES pass > Xt12/XtGetActionKeysym 4 UNRES pass > Xt12/XtGetActionKeysym 5 UNRES pass > Xt12/XtGetActionKeysym 6 UNRES pass > Xt12/XtGetActionKeysym 7 UNRES pass > Xt12/XtKeysymToKeycodeList 1 UNRES pass > Xt12/XtRegisterGrabAction 1 FAIL UNRES I assume your test environment was broken, because there's no way EXA could possibly influence these. So it looks like the only remaining relevant failure is:
> Xlib9/XDrawArc 16 pass FAIL
Does this pass as well if you extend the disabled code region to the next if (migrate->pReg) block? If not, what if you disable the whole if (migrate->as_dst) {} else {} block?
For
> Xlib9/XDrawArc 16 pass FAIL
I used the two methods you described "extend the disabled code region to the next if (migrate->pReg) block or disable the whole if (migrate->as_dst) {} else {} block?"
Xlib9/XDrawArc 16 are both FAIL again
*** Bug 13109 has been marked as a duplicate of this bug. *** (In reply to comment #11) > > Xlib9/XDrawArc 16 are both FAIL again Thanks for testing these! I pushed a change which disables the problematic optimization by default. It would be great if you could bisect the change that introduced the above remaining failure. What about current status of this bug? Using Michel's new patch,now "Xlib9/XDrawArc 16" has passed.at the same time,three new failed item was brought in as following: Xlib9/XDrawArcs 55 pass FAIL Xlib9/XDrawArcs 75 pass FAIL Xlib9/XPutImage 3 pass FAIL at the same time,Daniel's comment #9 was verified are right. What about current xtest status? fengming, any update? using the latest git code.xtest can pass around 3770, there are around 80 new passes than before. So, what's the delta between using XAA and EXA now? ( or do we still have the list of passed cases when using XAA? ) I'm closing this deprecated bug as the result has changed much against the original report. Fengming will open a new bug with the latest result. You can see the new Xtest results in bug# 14786:[EXA]xtest new failures now. |
Use of freedesktop.org services, including Bugzilla, is subject to our Code of Conduct. How we collect and use information is described in our Privacy Policy.