Tobias Koenig wrote: > Well, the backtrace is the same, so I guess the faulty code causes also > the crash mentioned in #428089. > Well, the backtrace in #423485 is about SavageAdjustFrame/SavageSwitchMode, while the only backtrace I found for #428089 is in the upstream bug and talks about ShadowWait/SavageGEReset/SavageWriteMode/SavageCloseScreen. Am I missing something? Cheers Brice