kernel: update guest kernel to 4.19.8 on aarch64 #269
Conversation
|
@grahamwhaley @jodh-intel @Weichen81 |
|
As discussed on the issue (and mentioned elsewhere), the default kernel stance is to use the newest longerm kernel (currently 4.14.85). I guess we don't have much choice here and eventually the features needed for aarch64 will end up in a longterm kernel version. But we should get more input on this PR since not only is 4.20 a mainline version, it's an rc version (and not the newest) to boot (pun intended ;) Note that currently this PR will fail CI as you need to increase the number in |
|
@jodh-intel As it's not good to use a RC version kernel. And I remember that @grahamwhaley had said, we are planning to upgrade kernel to 4.19. |
|
/cc @kata-containers/kernel ;) |
Pennyzct
force-pushed
the
build_development_kernel
branch
from
20cb01b to
7161d63
Compare
Pennyzct
changed the title
|
Hi~ all. @kata-containers/kernel |
Pennyzct
force-pushed
the
build_development_kernel
branch
from
7161d63 to
d2e9822
Compare
|
Hi @Pennyzct - 4.19.6 is "stable", not "longterm" fwics? |
|
@jodh-intel yes, you’re right, it’s a stable kernel, not a long term version. I think it’s a typo of @Pennyzct ; ) |
|
It's not ideal to have a deviation between kernel versions for different architectures, but this does seem unavoidable here. Hence, I'm in agreement with @gnawux + @grahamwhaley (see kata-containers/runtime#843). What do others think? /cc @sboeuf, @jcvenegas, @egernst. @WeiZhang555 lgtm |
|
@jodh-intel we had a discussion this morning with a few Intel folks on that, and I think we all agreed that having Kata kernel being stable instead of longterm would be nice. This would be a little bit more of maintenance when bumping versions, but it'd be worth the pain because we want to make sure Kata guest kernel can leverage all the latest features from the Linux kernel. |
|
wfm as long as @gnawux, @WeiZhang555 and @jon raise no objections? We do mention the kernel type briefly in https://github.com/kata-containers/documentation/blob/master/Upgrading.md#guest-kernel but that's not a very prominent document so if/when we're all in agreement I think we should:
|
Pennyzct
force-pushed
the
build_development_kernel
branch
from
d2e9822 to
b089524
Compare
|
@jodh-intel @marcov Hi~I just updated the kernel version to v4.19.8 to follow #278 and deleted all arch-dependent code. once @alicefr landing her PR to remove patch |
For supporting nvdimm, we need to update kernel on aarch64 to the stable version 4.19.8 and backport Suzuki K Poulose's latest Dynamic IPA and 52bit IPA support patch series (https://patchwork.kernel.org/cover/10616271/)which has been included in 4.20-rc3+ to the v4.19.8. Fixes: kata-containers#268 Signed-off-by: Penny Zheng <penny.zheng@arm.com> Signed-off-by: Wei Chen <Wei.Chen@arm.com>
Pennyzct
force-pushed
the
build_development_kernel
branch
from
b089524 to
aee668b
Compare
|
Hi~ all. I have already updated based on the latest master branch. ;) Since amd64 reversing the kernel config back to 4.14.x and ppc64le missing out @nitkon, should we make config arch-dependent to walk around like PR runtime/#973 suggests? |
Pennyzct
changed the title
|
Hi @Pennyzct, I recognize the need to have different kernel versions for different architectures. Forcing a single version just to keep things simpler is going to limit you and others extending kata to other architectures. I just wanted to note that going this road, instead of having a single kernel package built for different architectures, we will likely need one kernel package per architecture, and this will take more effort to maintain. |
|
/test |
|
lgtm |
|
The added patch |
Reverts #269 The added kernel patch doesn't apply and is causing all CI to fail. FYI, the current guest kernel version is 4.14.67. Please rebase and re-submit your PR cleanly @Pennyzct. e.g., http://jenkins.katacontainers.io/job/kata-containers-runtime-centos-7-4-PR/995/console NFO: Applying patch /tmp/jenkins/workspace/kata-containers-runtime-centos-7-4-PR/go/src/github.com/kata-containers/packaging/kernel/patches/0003-backport-Arm64-KVM-Dynamic-IPA-and-52bit-IPA-support.patch patching file arch/arm/include/asm/kvm_arm.h patching file arch/arm/include/asm/kvm_host.h Hunk #1 succeeded at 266 (offset -7 lines). Hunk #2 succeeded at 314 with fuzz 2 (offset -40 lines). patching file arch/arm/include/asm/kvm_mmu.h Hunk #1 FAILED at 35. Hunk #2 succeeded at 47 with fuzz 2 (offset -5 lines). Hunk #3 succeeded at 266 with fuzz 2 (offset -96 lines). 1 out of 3 hunks FAILED -- saving rejects to file arch/arm/include/asm/kvm_mmu.h.rej patching file arch/arm/include/asm/stage2_pgtable.h patching file arch/arm64/include/asm/cpufeature.h Hunk #1 FAILED at 262. Hunk #2 succeeded at 284 (offset -246 lines). 1 out of 2 hunks FAILED -- saving rejects to file arch/arm64/include/asm/cpufeature.h.rej patching file arch/arm64/include/asm/kvm_arm.h Hunk #1 succeeded at 102 (offset -5 lines). Hunk #2 succeeded at 116 (offset -5 lines). Hunk #3 succeeded at 305 (offset -7 lines). patching file arch/arm64/include/asm/kvm_asm.h Hunk #1 succeeded at 27 (offset -3 lines). Hunk #2 FAILED at 73. 1 out of 2 hunks FAILED -- saving rejects to file arch/arm64/include/asm/kvm_asm.h.rej patching file arch/arm64/include/asm/kvm_host.h Hunk #1 succeeded at 48 (offset -5 lines). Hunk #2 FAILED at 61. Hunk #3 succeeded at 399 with fuzz 2 (offset -41 lines). Hunk #4 FAILED at 503. 2 out of 4 hunks FAILED -- saving rejects to file arch/arm64/include/asm/kvm_host.h.rej patching file arch/arm64/include/asm/kvm_hyp.h patching file arch/arm64/include/asm/kvm_mmu.h Hunk #1 succeeded at 154 (offset 13 lines). Hunk #2 succeeded at 229 with fuzz 1 (offset -17 lines). Hunk #3 succeeded at 409 with fuzz 2 (offset -110 lines). patching file arch/arm64/include/asm/stage2_pgtable-nopmd.h patching file arch/arm64/include/asm/stage2_pgtable-nopud.h patching file arch/arm64/include/asm/stage2_pgtable.h patching file arch/arm64/kvm/guest.c Hunk #1 succeeded at 305 (offset -86 lines). patching file arch/arm64/kvm/hyp-init.S Hunk #1 FAILED at 65. 1 out of 1 hunk FAILED -- saving rejects to file arch/arm64/kvm/hyp-init.S.rej patching file arch/arm64/kvm/hyp/Makefile Hunk #1 succeeded at 18 (offset -1 lines). patching file arch/arm64/kvm/hyp/s2-setup.c Reversed (or previously applied) patch detected! Assume -R? [n] Apply anyway? [n] Skipping patch. 1 out of 1 hunk ignored -- saving rejects to file arch/arm64/kvm/hyp/s2-setup.c.rej patching file arch/arm64/kvm/hyp/switch.c Hunk #1 FAILED at 198. Hunk #2 succeeded at 226 (offset -37 lines). 1 out of 2 hunks FAILED -- saving rejects to file arch/arm64/kvm/hyp/switch.c.rej patching file arch/arm64/kvm/hyp/tlb.c Hunk #1 succeeded at 29 (offset -1 lines). Hunk #2 succeeded at 38 (offset -1 lines). patching file arch/arm64/kvm/reset.c Hunk #4 FAILED at 86. Hunk #5 succeeded at 133 (offset -4 lines). 1 out of 5 hunks FAILED -- saving rejects to file arch/arm64/kvm/reset.c.rej patching file include/linux/irqchip/arm-gic-v3.h Hunk #1 succeeded at 374 (offset -14 lines). patching file include/uapi/linux/kvm.h Hunk #1 succeeded at 739 (offset -11 lines). Hunk #2 FAILED at 962. 1 out of 2 hunks FAILED -- saving rejects to file include/uapi/linux/kvm.h.rej patching file virt/kvm/arm/arm.c Hunk #1 succeeded at 115 (offset -5 lines). Hunk #2 succeeded at 207 (offset -6 lines). Hunk #3 succeeded at 236 (offset -6 lines). Hunk #4 succeeded at 463 (offset -35 lines). Hunk #5 FAILED at 546. Hunk #6 succeeded at 1164 with fuzz 2 (offset -145 lines). Hunk #7 succeeded at 1261 (offset -147 lines). 1 out of 7 hunks FAILED -- saving rejects to file virt/kvm/arm/arm.c.rej patching file virt/kvm/arm/mmu.c Hunk #1 succeeded at 43 with fuzz 2 (offset -2 lines). Hunk #2 succeeded at 147 (offset -2 lines). Hunk #3 succeeded at 216 (offset -35 lines). Hunk #4 succeeded at 226 (offset -35 lines). Hunk #5 succeeded at 245 (offset -35 lines). Hunk #6 succeeded at 255 (offset -35 lines). Hunk #7 succeeded at 272 (offset -35 lines). Hunk #8 succeeded at 371 with fuzz 2 (offset 40 lines). Hunk #9 FAILED at 340. Hunk #10 FAILED at 370. Hunk #11 FAILED at 388. Hunk #12 FAILED at 408. Hunk #13 succeeded at 753 (offset -143 lines). Hunk #14 succeeded at 842 (offset -143 lines). Hunk #15 succeeded at 850 (offset -143 lines). Hunk #16 succeeded at 859 (offset -143 lines). Hunk #17 succeeded at 881 (offset -143 lines). Hunk #18 succeeded at 1024 (offset -182 lines). Hunk #19 succeeded at 1114 (offset -188 lines). Hunk #20 succeeded at 1148 (offset -188 lines). Hunk #21 succeeded at 1176 (offset -188 lines). Hunk #22 succeeded at 1190 (offset -188 lines). Hunk #23 succeeded at 1333 (offset -196 lines). Hunk #24 succeeded at 1552 (offset -220 lines). Hunk #25 succeeded at 1852 (offset -219 lines). 4 out of 25 hunks FAILED -- saving rejects to file virt/kvm/arm/mmu.c.rej patching file virt/kvm/arm/vgic/vgic-its.c Hunk #1 succeeded at 233 (offset -8 lines). Hunk #2 succeeded at 696 (offset -56 lines). Hunk #3 succeeded at 721 (offset -56 lines). Hunk #4 succeeded at 736 (offset -56 lines). Hunk #5 succeeded at 746 (offset -56 lines). Hunk #6 succeeded at 1214 (offset -80 lines). Hunk #7 succeeded at 1232 (offset -80 lines). Hunk #8 succeeded at 1279 (offset -80 lines). Hunk #9 succeeded at 2135 (offset -82 lines). Hunk #10 succeeded at 2207 (offset -82 lines). Hunk #11 succeeded at 2256 (offset -82 lines). patching file virt/kvm/arm/vgic/vgic-mmio-v3.c Hunk #1 succeeded at 326 (offset -38 lines). Hunk #2 succeeded at 343 (offset -38 lines). Build step 'Execute shell' marked build as failure Performing Post build task... Match found for :.* : True Logical operation result is TRUE Running script : #!/bin/bash Fixes: #296 Signed-off-by: Peng Tao <bergwolf@gmail.com>
|
This is also something we should in this PR job |
Reverts #269 The added kernel patch doesn't apply and is causing all CI to fail. FYI, the current guest kernel version is 4.14.67. Please rebase and re-submit your PR cleanly @Pennyzct. e.g., http://jenkins.katacontainers.io/job/kata-containers-runtime-centos-7-4-PR/995/console NFO: Applying patch /tmp/jenkins/workspace/kata-containers-runtime-centos-7-4-PR/go/src/github.com/kata-containers/packaging/kernel/patches/0003-backport-Arm64-KVM-Dynamic-IPA-and-52bit-IPA-support.patch patching file arch/arm/include/asm/kvm_arm.h patching file arch/arm/include/asm/kvm_host.h Hunk #1 succeeded at 266 (offset -7 lines). Hunk #2 succeeded at 314 with fuzz 2 (offset -40 lines). patching file arch/arm/include/asm/kvm_mmu.h Hunk #1 FAILED at 35. Hunk #2 succeeded at 47 with fuzz 2 (offset -5 lines). Hunk #3 succeeded at 266 with fuzz 2 (offset -96 lines). 1 out of 3 hunks FAILED -- saving rejects to file arch/arm/include/asm/kvm_mmu.h.rej patching file arch/arm/include/asm/stage2_pgtable.h patching file arch/arm64/include/asm/cpufeature.h Hunk #1 FAILED at 262. Hunk #2 succeeded at 284 (offset -246 lines). 1 out of 2 hunks FAILED -- saving rejects to file arch/arm64/include/asm/cpufeature.h.rej patching file arch/arm64/include/asm/kvm_arm.h Hunk #1 succeeded at 102 (offset -5 lines). Hunk #2 succeeded at 116 (offset -5 lines). Hunk #3 succeeded at 305 (offset -7 lines). patching file arch/arm64/include/asm/kvm_asm.h Hunk #1 succeeded at 27 (offset -3 lines). Hunk #2 FAILED at 73. 1 out of 2 hunks FAILED -- saving rejects to file arch/arm64/include/asm/kvm_asm.h.rej patching file arch/arm64/include/asm/kvm_host.h Hunk #1 succeeded at 48 (offset -5 lines). Hunk #2 FAILED at 61. Hunk #3 succeeded at 399 with fuzz 2 (offset -41 lines). Hunk #4 FAILED at 503. 2 out of 4 hunks FAILED -- saving rejects to file arch/arm64/include/asm/kvm_host.h.rej patching file arch/arm64/include/asm/kvm_hyp.h patching file arch/arm64/include/asm/kvm_mmu.h Hunk #1 succeeded at 154 (offset 13 lines). Hunk #2 succeeded at 229 with fuzz 1 (offset -17 lines). Hunk #3 succeeded at 409 with fuzz 2 (offset -110 lines). patching file arch/arm64/include/asm/stage2_pgtable-nopmd.h patching file arch/arm64/include/asm/stage2_pgtable-nopud.h patching file arch/arm64/include/asm/stage2_pgtable.h patching file arch/arm64/kvm/guest.c Hunk #1 succeeded at 305 (offset -86 lines). patching file arch/arm64/kvm/hyp-init.S Hunk #1 FAILED at 65. 1 out of 1 hunk FAILED -- saving rejects to file arch/arm64/kvm/hyp-init.S.rej patching file arch/arm64/kvm/hyp/Makefile Hunk #1 succeeded at 18 (offset -1 lines). patching file arch/arm64/kvm/hyp/s2-setup.c Reversed (or previously applied) patch detected! Assume -R? [n] Apply anyway? [n] Skipping patch. 1 out of 1 hunk ignored -- saving rejects to file arch/arm64/kvm/hyp/s2-setup.c.rej patching file arch/arm64/kvm/hyp/switch.c Hunk #1 FAILED at 198. Hunk #2 succeeded at 226 (offset -37 lines). 1 out of 2 hunks FAILED -- saving rejects to file arch/arm64/kvm/hyp/switch.c.rej patching file arch/arm64/kvm/hyp/tlb.c Hunk #1 succeeded at 29 (offset -1 lines). Hunk #2 succeeded at 38 (offset -1 lines). patching file arch/arm64/kvm/reset.c Hunk #4 FAILED at 86. Hunk #5 succeeded at 133 (offset -4 lines). 1 out of 5 hunks FAILED -- saving rejects to file arch/arm64/kvm/reset.c.rej patching file include/linux/irqchip/arm-gic-v3.h Hunk #1 succeeded at 374 (offset -14 lines). patching file include/uapi/linux/kvm.h Hunk #1 succeeded at 739 (offset -11 lines). Hunk #2 FAILED at 962. 1 out of 2 hunks FAILED -- saving rejects to file include/uapi/linux/kvm.h.rej patching file virt/kvm/arm/arm.c Hunk #1 succeeded at 115 (offset -5 lines). Hunk #2 succeeded at 207 (offset -6 lines). Hunk #3 succeeded at 236 (offset -6 lines). Hunk #4 succeeded at 463 (offset -35 lines). Hunk #5 FAILED at 546. Hunk #6 succeeded at 1164 with fuzz 2 (offset -145 lines). Hunk #7 succeeded at 1261 (offset -147 lines). 1 out of 7 hunks FAILED -- saving rejects to file virt/kvm/arm/arm.c.rej patching file virt/kvm/arm/mmu.c Hunk #1 succeeded at 43 with fuzz 2 (offset -2 lines). Hunk #2 succeeded at 147 (offset -2 lines). Hunk #3 succeeded at 216 (offset -35 lines). Hunk #4 succeeded at 226 (offset -35 lines). Hunk #5 succeeded at 245 (offset -35 lines). Hunk #6 succeeded at 255 (offset -35 lines). Hunk #7 succeeded at 272 (offset -35 lines). Hunk #8 succeeded at 371 with fuzz 2 (offset 40 lines). Hunk #9 FAILED at 340. Hunk #10 FAILED at 370. Hunk #11 FAILED at 388. Hunk #12 FAILED at 408. Hunk #13 succeeded at 753 (offset -143 lines). Hunk #14 succeeded at 842 (offset -143 lines). Hunk #15 succeeded at 850 (offset -143 lines). Hunk #16 succeeded at 859 (offset -143 lines). Hunk #17 succeeded at 881 (offset -143 lines). Hunk #18 succeeded at 1024 (offset -182 lines). Hunk #19 succeeded at 1114 (offset -188 lines). Hunk #20 succeeded at 1148 (offset -188 lines). Hunk #21 succeeded at 1176 (offset -188 lines). Hunk #22 succeeded at 1190 (offset -188 lines). Hunk #23 succeeded at 1333 (offset -196 lines). Hunk #24 succeeded at 1552 (offset -220 lines). Hunk #25 succeeded at 1852 (offset -219 lines). 4 out of 25 hunks FAILED -- saving rejects to file virt/kvm/arm/mmu.c.rej patching file virt/kvm/arm/vgic/vgic-its.c Hunk #1 succeeded at 233 (offset -8 lines). Hunk #2 succeeded at 696 (offset -56 lines). Hunk #3 succeeded at 721 (offset -56 lines). Hunk #4 succeeded at 736 (offset -56 lines). Hunk #5 succeeded at 746 (offset -56 lines). Hunk #6 succeeded at 1214 (offset -80 lines). Hunk #7 succeeded at 1232 (offset -80 lines). Hunk #8 succeeded at 1279 (offset -80 lines). Hunk #9 succeeded at 2135 (offset -82 lines). Hunk #10 succeeded at 2207 (offset -82 lines). Hunk #11 succeeded at 2256 (offset -82 lines). patching file virt/kvm/arm/vgic/vgic-mmio-v3.c Hunk #1 succeeded at 326 (offset -38 lines). Hunk #2 succeeded at 343 (offset -38 lines). Build step 'Execute shell' marked build as failure Performing Post build task... Match found for :.* : True Logical operation result is TRUE Running script : #!/bin/bash Fixes: #296 Signed-off-by: Peng Tao <bergwolf@gmail.com>
rootfs-builder: add rootfs.sh DOCKER_RUNTIME env var
we should include set-up and install of the dev-kernel, since some siginificant features haven't be supported on stable kernel for now, for instance, the crucial nvdimm-support on aarch64 only
works well on the latest 4.20-rc3+ kernel.
Fixes: #268
Signed-off-by: Penny Zheng penny.zheng@arm.com
Signed-off-by: Wei Chen Wei.Chen@arm.com