Thanks for taking the time to report this bug and helping to make Ubuntu better. We appreciate the difficulties you are facing, but this appears to be a "regular" (non-security) bug. I have unmarked it as a security issue since this bug does not show evidence of allowing attackers to cross privilege boundaries nor directly cause loss of data/privacy. Please feel free to report any other bugs you may find.
** Information type changed from Private Security to Public -- You received this bug notification because you are a member of Kernel Packages, which is subscribed to linux in Ubuntu. https://bugs.launchpad.net/bugs/1977857 Title: [amdgpu] denial of service by WebGL - GPU crashes with system impairment Status in linux package in Ubuntu: New Bug description: Using Ubuntu 20.04.04 with Kernel 5.13.0-46-generic on AMD Ryzen V1605B Platform with integrated Radeon Vega 8 GPU. I am currently working on an Web application using tensorflowjs, which in turn uses WebGL for computations. When "provoking" the GPU by a lot of memory allocations, it can easily be crashed with consequences for the whole system. Symptoms include: -Chromium browser hung in a way that the current tab cant be reloaded. Browser restart fixes this sometimes. Otherwise see below. -Chromium browser exiting, and can't be restarted. Any attempt to run it just hangs without opening a window at all. Using terminal-only applications still works as normal. System reboot fixes this. But.. -After triggering the bugs above, quite often system restart fails (system hang on shutdown). Only power cycling fixes this. -Sometimes whole system freezing, and can't be restarted. SSH connections stall too. All of these can be triggered by a rogue website, and provide risk of denial and data loss. Of course we can blame Chromium first for unsane memory management, but even a rogue userspace application should not render the GPU useless or stalled forever. We can also blame mesa I guess, but as amdgpu is the kernel part is has to handle such attacks in my opinion. I've also changed to updated Mesa from obiaf ppa, which won't fix the whole situation. Maybe also amdgpu firmware is part of the problem. Sadly I didn't collect the dmesg traces of all those crashes, but there where plenty of different messages there. I will add dmesg traces here, if it's happen again (but currently I try to stabilize the system less crashes will happen). To manage notifications about this bug go to: https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1977857/+subscriptions -- Mailing list: https://launchpad.net/~kernel-packages Post to : kernel-packages@lists.launchpad.net Unsubscribe : https://launchpad.net/~kernel-packages More help : https://help.launchpad.net/ListHelp
