Closed Bug 1060268 Opened 11 years ago Closed 5 years ago

transform: skew(90deg) cause FF to crash

Categories

(Core :: Graphics, defect)

Product:
Core
Component:
Graphics
Platform:
x86
All
Type:
defect
Priority:
Not set
Severity:
critical

Tracking

()

Status:
RESOLVED WORKSFORME
Tracking Flags:
Tracking Status
firefox32 --- affected
firefox33 --- affected
firefox34 --- affected
firefox35 --- affected
firefox-esr31 --- affected

People

(Reporter: kleidt, Unassigned)

References

(
URL
)

Details

(Keywords: crash, reproducible)

Crash Data

Attachments

(1 file)

1060268.html
2.77 KB, text/html
Details
User Agent: Mozilla/5.0 (Windows NT 5.1; rv:31.0) Gecko/20100101 Firefox/31.0 (Beta/Release) Build ID: 20140716183446 Steps to reproduce: 1) Visit http://www.css3maker.com/css3-transform.html 2) Play with the "Skew" slider Actual results: Firefox crashes at skew(90deg) and skew(-90deg) Expected results: No Webpage should be able to crash a browser!
WFM using current Nightly on OS X. Can you reproduce with a new profile? Could you add the crash report please?
Flags: needinfo?(kleidt)
Keywords: crash
It's my first bugreport for Mozilla. I hope this is what you want: https://crash-stats.mozilla.com/report/index/25d9a078-5413-46e7-81cc-b07012140829
Flags: needinfo?(kleidt)
(In reply to Michael Kleidt from comment #2) > It's my first bugreport for Mozilla. I hope this is what you want: > https://crash-stats.mozilla.com/report/index/25d9a078-5413-46e7-81cc- > b07012140829 Thanks! There are already a lot of similar reports, but I'm not sure if they have the same cause (See comments in bug 1034254)
Whiteboard: DUPEME
WFM with FF31 on Win 7 with HWA on/off. Reporter, can you test with a clean profile, please. https://support.mozilla.org/en-US/kb/profile-manager-create-and-remove-firefox-profiles
Flags: needinfo?(kleidt)
I Test it with a clean profile: https://crash-stats.mozilla.com/report/index/990ba114-0e23-4d88-acc5-3d9392140829 and I test it on a Win7 Mashine: https://crash-stats.mozilla.com/report/index/facb4ba6-15c7-4281-ae8c-fee712140829 Perhaps a hint to reproduce: 1) Move the "Skew" slider lowly from left to right and reverse - several times. Some times this dose not work, but in this case I always manage to crash, when I "Skew" Radio-Button and use the "Skey Y" slider the same way.
Flags: needinfo?(kleidt)
Is it possible for you to paste here the data from the section "graphics" in the page about:support (type it in the location bar) for your 2 machines.
XP Computer has Grafik Direct2D aktiviert Wurde auf Grund Ihrer Grafiktreiberversion blockiert. DirectWrite aktiviert false (0.0.0.0) Geräte-ID 0x29b2 GPU #2 aktiv false GPU-beschleunigte Fenster 0/1 Basic Karten-Beschreibung Intel(R) Q35 Express Chipset Family Karten-RAM Unknown Karten-Treiber igxprd32 Treiber-Datum 1-13-2010 Treiber-Version 6.14.10.5218 Vendor-ID 0x8086 WebGL-Renderer Google Inc. -- ANGLE (Intel(R) Q35 Express Chipset Family Direct3D9 vs_0_0 ps_2_0) windowLayerManagerRemote false AzureCanvasBackend skia AzureContentBackend cairo AzureFallbackCanvasBackend cairo AzureSkiaAccelerated 0 Win7 VM has Grafik Direct2D aktiviert Wurde auf Grund Ihrer Grafikkarte blockiert, da ungelöste Treiberprobleme bestehen. DirectWrite aktiviert false (6.2.9200.16571) Geräte-ID 0x0000 GPU #2 aktiv false GPU-beschleunigte Fenster 0/1 Basic Wurde auf Grund Ihrer Grafikkarte blockiert, da ungelöste Treiberprobleme bestehen. Karten-Beschreibung RDPDD Chained DD Karten-RAM Unknown Karten-Treiber RDPDD Vendor-ID 0x0000 WebGL-Renderer Wurde auf Grund Ihrer Grafikkarte blockiert, da ungelöste Treiberprobleme bestehen. windowLayerManagerRemote false AzureCanvasBackend skia AzureContentBackend cairo AzureFallbackCanvasBackend cairo AzureSkiaAccelerated 0
On a 64 Bit Win7 mashine, I wasn't able reproduce the problem. (Both XP and Win7 VM are 32 bit)
I also couldn't reproduce it on my Ubuntu System (FF 31.0) Some new findings: When I come near the skew(90deg) value, the rectangle get wider and wider. When it gets wider then the screen, scrollbars are added. When I understand it correctly, the width for skew(90deg) is theretically infinite. But the calculation in Firefox will result in a fine value. But the questions are: what is the maximum width for a webpage? Depends this maximum width on the oerating system? Processor architecture? Graphic card? Graphic driver? ...?
Yes, it's probably this edge case, when the rectangle animation is only a thin line at 90° that disappears from the screen.
And I checked the website of Intel, you have the latest drivers for your old card: https://downloadcenter.intel.com/SearchResult.aspx?lang=eng&ProdId=2880 So it will be difficult to fix something on that side.
But why dose Firefox crashes? There should be a possibility to catch an exception or to prevent the execution like if (width>maxwidht && osIs('xp') [or whatever else is trigering this bug] { width = maxwidth; }
Perhaps an interesting detail: The Win7 VM runs NOT on the XP mashine. Ist a bigger Server (I think Dell) which hosts all customer VM.
Severity: normal → critical
I tried to make the shape as long as possible in firefox-31.0.en-US.linux64, and it crashed with a similar trace: bp-d8eac6ce-8678-40c7-ac93-87a3c2140901 bug 1045588 also has such a trace.
Crash Signature: [@ OOM | small ]
Component: Untriaged → Graphics
OS: Windows XP → All
Product: Firefox → Core
(In reply to [:Aleksej] from comment #14) > as long as possible It could also have ended up very thin.
QA Whiteboard: [bugday-20140901][DUPEME]
Whiteboard: DUPEME
Crash consistently on Mozilla/5.0 (Windows NT 6.1; WOW64; rv:34.0) Gecko/20100101 Firefox/34.0 ID:20140902030202 CSet: c360f3d1c00d Steps: - Load http://www.css3maker.com/css3-transform.html - Select "Skew" radio button - Slide "Skew X" until the rectangle in the "CSS3 Preview Area" becomes a thin line Result: Nightly crashes Expected result: Nightly doesn't crash
URL: http://www.css3maker.com/css3-transfo...
Status: UNCONFIRMED → NEW
Ever confirmed: true
Report ID Date Submitted bp-96f4f0a6-3cdd-4008-8d24-83e312140902 9/2/2014 3:57 PM bp-619985ad-e43a-4f68-83c5-a06b52140902 9/2/2014 3:50 PM bp-88e1cef4-2f6a-45f2-a902-91b502140902 9/2/2014 3:24 PM
status-firefox32: --- → affected
status-firefox33: --- → affected
status-firefox34: --- → affected
status-firefox35: --- → affected
Version: 31 Branch → Trunk
I can reproduce the crash with HWA disabled and STR Comment 16. But I cannot reproduce with HWA enabled. Regression window(m-c) Good http://hg.mozilla.org/mozilla-central/rev/1e33678f89fe Mozilla/5.0 (Windows NT 6.1; WOW64; rv:27.0) Gecko/20100101 Firefox/27.0 ID:20130920190924 Bad: http://hg.mozilla.org/mozilla-central/rev/a2c31dc69ab3 Mozilla/5.0 (Windows NT 6.1; WOW64; rv:27.0) Gecko/20100101 Firefox/27.0 ID:20130920193025 Pushlog: http://hg.mozilla.org/mozilla-central/pushloghtml?fromchange=1e33678f89fe&tochange=a2c31dc69ab3 Regression window(m-i) Good: http://hg.mozilla.org/integration/mozilla-inbound/rev/44952b2191fc Mozilla/5.0 (Windows NT 6.1; WOW64; rv:27.0) Gecko/20100101 Firefox/27.0 ID:20130920023238 Bad: http://hg.mozilla.org/integration/mozilla-inbound/rev/e844f4ed3e1f Mozilla/5.0 (Windows NT 6.1; WOW64; rv:27.0) Gecko/20100101 Firefox/27.0 ID:20130920025128 Pushlog; http://hg.mozilla.org/integration/mozilla-inbound/pushloghtml?fromchange=44952b2191fc&tochange=e844f4ed3e1f In local build: Last Good:ad266fafd429 First Bad:e844f4ed3e1f Triggered by: e844f4ed3e1f Matt Woodrow — Bug 907926 - Enable Azure content for windows. r=nrc
Blocks: 907926
I can also see that on Linux, it OOMs right after.
Matt, you just fixed another OOM|small case in bug 1053934 and the regression range for this points to your patch for enabling Azure - is this case with those steps pointing to something we can easily enough get a fix for as well?
Flags: needinfo?(matt.woodrow)
With HWA off on Win 7, I crashed FF35 too.
Keywords: reproducible
(In reply to Robert Kaiser (:kairo@mozilla.com) from comment #20) > Matt, you just fixed another OOM|small case in bug 1053934 and the > regression range for this points to your patch for enabling Azure - is this > case with those steps pointing to something we can easily enough get a fix > for as well? I strongly suspect that this will be fixed by the same patch.
Flags: needinfo?(matt.woodrow)
It wasn't, still crashes reliably on Mozilla/5.0 (Windows NT 6.1; WOW64; rv:35.0) Gecko/20100101 Firefox/35.0 ID:20140904030202 CSet: 776fa9cf70cd Report ID Date Submitted bp-3089fcec-6bb2-42c9-abd2-a2c802140904 9/4/2014 10:38 AM
QA Whiteboard: [bugday-20140901][DUPEME] → [bugday-20140901]
Flags: needinfo?(matt.woodrow)
WFM Win XP Firefox Nightly 40.0a1 (2015-04-21). Some of the Skew setting become "ugly" but none "crashes" the Browser.
Attached file 1060268.html —
Not causing OOM here, but does freeze the browser. I created a test case. The element must be rotated any degree (other than 0), the element needs to be skewed -90 or 90 degrees, and the element must contain text. The freeze occurs every paint I believe. Resizing the browser and switching tabs triggers it.
Graphics -------- Adapter Description: AMD Radeon R7 200 Series Adapter Drivers: aticfx64 aticfx64 aticfx64 aticfx32 aticfx32 aticfx32 atiumd64 atidxx64 atidxx64 atiumdag atidxx32 atidxx32 atiumdva atiumd6a atitmm64 Adapter RAM: 1024 Asynchronous Pan/Zoom: none Device ID: 0x6613 Direct2D Enabled: true DirectWrite Enabled: true (6.3.9600.17795) Driver Date: 11-20-2014 Driver Version: 14.501.1003.0 GPU #2 Active: false GPU Accelerated Windows: 1/1 Direct3D 11 (OMTC) Subsys ID: 00000000 Supports Hardware H264 Decoding: true Vendor ID: 0x1002 WebGL Renderer: Google Inc. -- ANGLE (AMD Radeon R7 200 Series Direct3D11 vs_5_0 ps_5_0) windowLayerManagerRemote: true AzureCanvasBackend: direct2d 1.1 AzureContentBackend: direct2d 1.1 AzureFallbackCanvasBackend: cairo AzureSkiaAccelerated: 0
Do you get the same freeze with acceleration disabled?
Flags: needinfo?(matt.woodrow)
I'm working on a new machine and I can't reproduce the error any more. When the old system has frozen for ever, the new one just freeze for some seconds. It looks like the error is handled internally: WebGL2-Renderer: WebGL creation failed: * Error during ANGLE OpenGL init. * Error during ANGLE OpenGL init. * Error during ANGLE OpenGL init. * Error during ANGLE OpenGL init. * Error during ANGLE OpenGL init. * Exhausted GL driver caps. Allgemeine Informationen ------------------------ Name: Firefox Version: 51.0.1 Build-ID: 20170125094131 Update-Kanal: release User-Agent: Mozilla/5.0 (Windows NT 6.1; Win64; x64; rv:51.0) Gecko/20100101 Firefox/51.0 Betriebssystem: Windows_NT 6.1 Fenster mit mehreren Prozessen: 1/1 (aktiviert durch Benutzer) Abgesicherter Modus: false Absturzberichte der letzten 3 Tage ---------------------------------- Grafik ------ Allgemeine Merkmale Compositing: Direct3D 11 Asynchrones Wischen und Zoomen: Mausrad-Eingabe aktiviert WebGL-Renderer: Google Inc. -- ANGLE (Intel(R) HD Graphics Direct3D11 vs_4_0 ps_4_0) WebGL2-Renderer: WebGL creation failed: * Error during ANGLE OpenGL init. * Error during ANGLE OpenGL init. * Error during ANGLE OpenGL init. * Error during ANGLE OpenGL init. * Error during ANGLE OpenGL init. * Exhausted GL driver caps. H264-Dekodierung durch Hardware: No; Hardware video decoding disabled or blacklisted Audio-Back-End: wasapi Direct2D: true DirectWrite: true (6.2.9200.21976) GPU 1 Aktiv: Ja Beschreibung: Intel(R) HD Graphics Herstellerkennung: 0x8086 Gerätekennung: 0x0042 Treiber-Version: 8.15.10.2622 Treiber-Datum: 1-10-2012 Treiber: igdumd64 igd10umd64 igdumdx32 igd10umd32 Subsys-ID: 304b103c RAM: Unknown Weitere Informationen ClearType-Parameter: Gamma: 2.2 Pixel Structure: RGB AzureCanvasAccelerated: 0 AzureCanvasBackend: direct2d 1.1 AzureContentBackend: direct2d 1.1 AzureFallbackCanvasBackend: cairo ClearType-Parameter: Gamma: 2.2 Pixel Structure: RGB Entscheidungsprotokoll D3D9_COMPOSITING: disabled by default: Disabled by default

Also no crash here, 80.0.1 with Mac and win10

Status: NEW → RESOLVED
Closed: 5 years ago
Resolution: --- → WORKSFORME
You need to log in before you can comment on or make changes to this bug.

Attachment

General

Creator:
Created:
Updated:
Size: