May 2022 - openSUSE Kernel - openSUSE Mailing Lists

[kernel] RT
by Felix Miata 24 May '22

24 May '22

# zypper info kernel-rt Retrieving repository... ...[done] Loading repository data... Reading installed packages... Information for package kernel-rt: ---------------------------------- Repository : OSS Name : kernel-rt Version : 5.3.18-8.13.1 Arch : x86_64 Vendor : SUSE LLC <https://www.suse.com/> Installed Size : 279.3 MiB Installed : No Status : not installed Source package : kernel-rt-5.3.18-8.13.1.nosrc Upstream URL : http://www.kernel.org/ Summary : The Linux Kernel Description : The Linux Kernel. Source Timestamp: 2021-06-01 16:17:05 +0000 GIT Revision: ebc4f0cb715d533804b3e9d66899a0470ba951f2 GIT Branch: SLE15-SP2-RT # How does a mere mortal user find out if this package provides a real-time kernel? -- Evolution as taught in public schools is, like religion, based on faith, not based on science. Team OS/2 ** Reg. Linux User #211409 ** a11y rocks! Felix Miata

1 0

VirtualBox fails under kernel 5.18
by Larry Finger 23 May '22

23 May '22

VirtualBox fails to run correctly under kernel 5.18.0. I noticed a problem in the 5.18-rcX cycle and bisected the problem to commit 6e8ec2552c7d ("random: use computational hash for entropy extraction"). That commit has been confirmed as the one causing the problem. We now know the cause. The changes in the random number generator employ the SSSE3 instructions. which in turn affects the FPU. The current belief is that the VirtualBox vcpu is not saving the state of the FPU when a hard IRQ occurs. If any code using SSSE3 runs in the host during that interrupt, the state of the FPU in the guest is incorrect, and the guest gets a kernel exception. Thus far, the only kernel code using that instruction set is the random number generator. It is easy to test for this condition. In the guest start a multi-cpu build of the kernel, or some other large project. Within a few seconds, gcc will get an error. The attached patch will cause rng not to use SSSE3 instructions, thereby bypassing the problem. Generation of random numbers may be a bit slower, but have no loss of robustness. If possible, I would like the patch applied to 5.18.X kernels until Oracle fixes the problem. Finding the fault is beyond my understanding of the implementation of virtual CPUs. Larry Finger - openSUSE VirtualBox Maintainer

1 0

[kernel] TW amdgpu firmware regression?
by Felix Miata 19 May '22

19 May '22

As of kernel-firmware-amdgpu-20220411 (first bad), and sometime after kernel-firmware-amdgpu-20211123 (last available good to revert to), and after TW20220311 (last working), using TW20220517 on: AMD Oland [Radeon HD 8570 / R5 430 OEM R7 240/340 Radeon 520 OEM] vendor: Dell driver: amdgpu v: kernel active: DP-1,DVI-I-1 bus-ID: 01:00.0 chip-ID: 1002:6611 aka Southern Islands GCN 1st gen microarchitecture Display and keyboard lockup around the time KMS should init. Other than reverting the firmware, or the crippling nomodeset option, the only workaround I have found is installing kernel-firmware-radeon, which this PC/GPU combination has never before needed. With kernel-firmware-radeon installed, X nevertheless runs using the amdgpu kernel module and the amdgpu DDX or the modesetting DIX. Before I report a bug, I would like to know if anyone here on the kernel mailing list knows if perchance GCN1 support has been purposely removed somehow from kernel-firmware-amdgpu, thus negating any point for a bug report. dmesg using 20211123 kernel-firmware-amdgpu: https://susepaste.org/6218560 dmesg using 20220411 kernel-firmware-amdgpu: https://susepaste.org/2000260 -- Evolution as taught in public schools is, like religion, based on faith, not based on science. Team OS/2 ** Reg. Linux User #211409 ** a11y rocks! Felix Miata

1 0

Re: Heads-up: Snapshot 20220510 will be larger than usual (GCC12 rebuild)
by Andrei Borzenkov 16 May '22

16 May '22

On 13.05.2022 13:31, Dominique Leuenberger / DimStar wrote: > Dear Tumbleweed users, > > Just a small shout-out: Snapshot 0510 is currently in progress of > syncing out to the mirrors and should become available later this > afternoon. > > This snapshot has been rebuilt entirely using gcc12 Apparently it broke kexec https://bugzilla.opensuse.org/show_bug.cgi?id=1199593

1 0