Unfortunately we don't know the exact revision when they started to crash, because there is another serious bug which makes all inspector tests crash: https://bugs.webkit.org/show_bug.cgi?id=111441 See https://bugs.webkit.org/show_bug.cgi?id=108632 for previous testing (Comment #0 and Comment #1) I'm going to bisect this regression manually with applying the fix in https://bugs.webkit.org/show_bug.cgi?id=111441 Results on ToT with the fix from bug111441: http://build.webkit.sed.hu/builders/ARMv7%20Linux%20Qt5%20Release%20%28Test%29/builds/7987
The bug was valid on r140000 - http://build.webkit.sed.hu/builders/ARMv7%20Linux%20Qt5%20Release%20%28Test%29/builds/7992 (only inspector tests was ran here)
all inspector tests crashed on r139004 - http://build.webkit.sed.hu/builders/ARMv7%20Linux%20Qt5%20Release%20%28Test%29/builds/7993
all tests crashed on r138500 - http://build.webkit.sed.hu/builders/ARMv7%20Linux%20Qt5%20Release%20%28Test%29/builds/7994
I got it r138609 is the culprit. It caused this regression or revealed a hidden bug. r138608 - http://build.webkit.sed.hu/builders/ARMv7%20Linux%20Qt5%20Release%20%28Test%29/builds/8001 r138609 - http://build.webkit.sed.hu/builders/ARMv7%20Linux%20Qt5%20Release%20%28Test%29/builds/8002 (Of course I applied the fix from https://bugs.webkit.org/show_bug.cgi?id=111441 to unhide this bug)
It might be similar bug on ARM Thumb2 to this ARM Traditional bug - https://bugs.webkit.org/show_bug.cgi?id=105930
The proposed fix in https://bugs.webkit.org/show_bug.cgi?id=103146 fixes these crashes. *** This bug has been marked as a duplicate of bug 103146 ***