Closed kkartaltepe closed 1 year ago
@kkartaltepe The PR is just a very minimal change to the staging patch to fix a conflict caused by a small wine-staging difference
The current child window patch is based from Remi's WIP child window branch (with some Proton 7.0 picks), so that's why it probably isn't very stable
The current child window patch is based from Remi's WIP child window branch (with some Proton 7.0 picks), so that's why it probably isn't very stable
Right im just hoping to find a place to send fixes to, or at worst just get this bug known to whoever is maintaining it.
Right im just hoping to find a place to send fixes to, or at worst just get this bug known to whoever is maintaining it.
@kkartaltepe Remi Bernon is the one who did most of the work on the child window patch, so they're probably the best person to contact (they're in #winehackers channel on Libera Chat as rbernon)
Thanks a bunch ill follow up in #winehackers. And thanks for updating the patchset here as well making it easy to find :D
I'm not sure where this patchset is sourced but I found the recent updated of it here in #1023 so hopefully you can help point me towards where the patchset is maintained.
Its possible for the vkSurface to be freed by
destroy_vk_surface
prior to the client destroying the surface (such as switching between offscreen/onscreen viawine_vk_surface_set_offscreen
.) Then when the client correctly destroys their surfaces during teardown and entersX11DRV_vkDestroySurfaceKHR
they pass in the free'ed and possibly reused memory back into the vulkan driver.This results in a crash in Warcraft 3 1.27 (pre-reforged), where the memory really is reused so the vulkan driver ends up using some random spot in memory as a surface struct and proceeds to implode.