summaryrefslogtreecommitdiffstats
AgeCommit message (Collapse)AuthorFilesLines
2016-07-18KVM: vmx: fix missed cancellation of TSC deadline timerWanpeng Li1-24/+24
INFO: rcu_sched detected stalls on CPUs/tasks: 1-...: (11800 GPs behind) idle=45d/140000000000000/0 softirq=0/0 fqs=21663 (detected by 0, t=65016 jiffies, g=11500, c=11499, q=719) Task dump for CPU 1: qemu-system-x86 R running task 0 3529 3525 0x00080808 ffff8802021791a0 ffff880212895040 0000000000000001 00007f1c2c00db40 ffff8801dd20fcd3 ffffc90002b98000 ffff8801dd20fc88 ffff8801dd20fcf8 0000000000000286 ffff8801dd2ac538 ffff8801dd20fcc0 ffffffffc06949c9 Call Trace: ? kvm_write_guest_cached+0xb9/0x160 [kvm] ? __delay+0xf/0x20 ? wait_lapic_expire+0x14a/0x200 [kvm] ? kvm_arch_vcpu_ioctl_run+0xcbe/0x1b00 [kvm] ? kvm_arch_vcpu_ioctl_run+0xe34/0x1b00 [kvm] ? kvm_vcpu_ioctl+0x2d3/0x7c0 [kvm] ? __fget+0x5/0x210 ? do_vfs_ioctl+0x96/0x6a0 ? __fget_light+0x2a/0x90 ? SyS_ioctl+0x79/0x90 ? do_syscall_64+0x7c/0x1e0 ? entry_SYSCALL64_slow_path+0x25/0x25 This can be reproduced readily by running a full dynticks guest(since hrtimer in guest is heavily used) w/ lapic_timer_advance disabled. If fail to program hardware preemption timer, we will fallback to hrtimer based method, however, a previous programmed preemption timer miss to cancel in this scenario which results in one hardware preemption timer and one hrtimer emulated tsc deadline timer run simultaneously. So sometimes the target guest deadline tsc is earlier than guest tsc, which leads to the computation in vmx_set_hv_timer can underflow and cause delta_tsc to be set a huge value, then host soft lockup as above. This patch fix it by cancelling the previous programmed preemption timer if there is once we failed to program the new preemption timer and fallback to hrtimer based method. Cc: Paolo Bonzini <pbonzini@redhat.com> Cc: Radim Krčmář <rkrcmar@redhat.com> Signed-off-by: Wanpeng Li <wanpeng.li@hotmail.com> Signed-off-by: Paolo Bonzini <pbonzini@redhat.com> Change-Id: I8a2decefab743aecdfab676fb9267324bf42b848 upstream-status: backport Signed-off-by: Yunhong Jiang <yunhong.jiang@linux.intel.com>
2016-07-18KVM: x86: introduce cancel_hv_tscdeadlineWanpeng Li1-8/+10
Introduce cancel_hv_tscdeadline() to encapsulate preemption timer cancel stuff. Cc: Paolo Bonzini <pbonzini@redhat.com> Cc: Radim Krčmář <rkrcmar@redhat.com> Cc: Yunhong Jiang <yunhong.jiang@intel.com> Signed-off-by: Wanpeng Li <wanpeng.li@hotmail.com> Signed-off-by: Paolo Bonzini <pbonzini@redhat.com> Change-Id: Icc038176cbf361a9ecdf37ed3425108db57617f2 upstream-status: backport Signed-off-by: Yunhong Jiang <yunhong.jiang@linux.intel.com>
2016-07-18KVM: vmx: fix underflow in TSC deadline calculationPaolo Bonzini1-3/+3
If the TSC deadline timer is programmed really close to the deadline or even in the past, the computation in vmx_set_hv_timer can underflow and cause delta_tsc to be set to a huge value. This generally results in vmx_set_hv_timer returning -ERANGE, but we can fix it by limiting delta_tsc to be positive or zero. Reported-by: Wanpeng Li <wanpeng.li@hotmail.com> Signed-off-by: Paolo Bonzini <pbonzini@redhat.com> Change-Id: I12eea18c3ec648dbf782d7754b7b574d7d6aa92c upstream-status: backport Signed-off-by: Yunhong Jiang <yunhong.jiang@linux.intel.com>
2016-07-18kvm: vmx: hook preemption timer supportYunhong Jiang3-2/+183
Hook the VMX preemption timer to the "hv timer" functionality added by the previous patch. This includes: checking if the feature is supported, if the feature is broken on the CPU, the hooks to setup/clean the VMX preemption timer, arming the timer on vmentry and handling the vmexit. A module parameter states if the VMX preemption timer should be utilized. Signed-off-by: Yunhong Jiang <yunhong.jiang@intel.com> [Move hv_deadline_tsc to struct vcpu_vmx, use -1 as the "unset" value. Put all VMX bits here. Enable it by default #yolo. - Paolo] Signed-off-by: Paolo Bonzini <pbonzini@redhat.com> Change-Id: Icb8e0b853eedce3d52c394e510fa14d2cdd432e9 upstream-status: backport Signed-off-by: Yunhong Jiang <yunhong.jiang@linux.intel.com>
2016-07-18kvm: vmx: rename vmx_pre/post_block to pi_pre/post_blockYunhong Jiang1-2/+15
Prepare to switch from preemption timer to hrtimer in the vmx_pre/post_block. Current functions are only for posted interrupt, rename them accordingly. upstream-status: backport Change-Id: Ie1dde9be21deeb661de095e07d6c29bcba2e7d73 Signed-off-by: Yunhong Jiang <yunhong.jiang@intel.com> Signed-off-by: Paolo Bonzini <pbonzini@redhat.com> Signed-off-by: Yunhong Jiang <yunhong.jiang@linux.intel.com>
2016-07-18KVM: x86: support using the vmx preemption timer for tsc deadline timerYunhong Jiang5-1/+100
The VMX preemption timer can be used to virtualize the TSC deadline timer. The VMX preemption timer is armed when the vCPU is running, and a VMExit will happen if the virtual TSC deadline timer expires. When the vCPU thread is blocked because of HLT, KVM will switch to use an hrtimer, and then go back to the VMX preemption timer when the vCPU thread is unblocked. This solution avoids the complex OS's hrtimer system, and the host timer interrupt handling cost, replacing them with a little math (for guest->host TSC and host TSC->preemption timer conversion) and a cheaper VMexit. This benefits latency for isolated pCPUs. [A word about performance... Yunhong reported a 30% reduction in average latency from cyclictest. I made a similar test with tscdeadline_latency from kvm-unit-tests, and measured - ~20 clock cycles loss (out of ~3200, so less than 1% but still statistically significant) in the worst case where the test halts just after programming the TSC deadline timer - ~800 clock cycles gain (25% reduction in latency) in the best case where the test busy waits. I removed the VMX bits from Yunhong's patch, to concentrate them in the next patch - Paolo] Signed-off-by: Yunhong Jiang <yunhong.jiang@intel.com> Signed-off-by: Paolo Bonzini <pbonzini@redhat.com> Change-Id: I4aa1ecfa3463d1cbfb317511b45d2074b33d9b6f upstream-status: backport Signed-off-by: Yunhong Jiang <yunhong.jiang@linux.intel.com>
2016-07-18kvm: lapic: separate start_sw_tscdeadline from start_apic_timerYunhong Jiang1-26/+31
The function to start the tsc deadline timer virtualization will be used also by the pre_block hook when we use the preemption timer; change it to a separate function. No logic changes. upstream-status: backport Change-Id: Ie2fc19108c3252f8a299b17aba16c14aa8d31ae8 Signed-off-by: Yunhong Jiang <yunhong.jiang@intel.com> Signed-off-by: Paolo Bonzini <pbonzini@redhat.com> Signed-off-by: Yunhong Jiang <yunhong.jiang@linux.intel.com>
2016-06-20Rebase fuel-plugin for kvmfornfv to fuel 9.0davidjchou11-226/+122
1. Uplifit fuel-plugin for kvmfornfv to fuel 9.0 due to that fuel 9.0 will be the deployment tool of OPNFV Colorado release. 2. Fixed quirk for kernel. 3. Added all tools and libs for building OVS module. 4. Allow KVM developer to build fuel-plugin-kvm after they modify the kerenl code without commiting their change into the repo first. So, they can test their code change by fuel-plugin-kvm till they satisfy with their change, then commit to the repo. 5. The final code built into OPNFV fuel iso will depend on the commit ID for kvmfornfv speicifed in fuel for OPNFV source tree. Change-Id: Iaf9ff49c69df374d0757884cfdac4cccee3eebe4 Signed-off-by: davidjchou <david.j.chou@intel.com>
2016-06-06Move quirks out of build scriptDonald Dugger2-45/+44
Due to expediencyh concerns some quirk code was placed in the build script in order to get the Brahmaputra release out on time. These quirks don't belong in the script itself and, now that there is a mechanisms to applly arbitrary patch files during the build script, this patch moves that quirk code out to the separate patch file area. Upstream status - NA Change-Id: Ib8100eef00009bbaf0e16b28849821ef5878f9f5 Signed-off-by: Donald Dugger <n0ano@n0ano.com>
2016-06-06Merge "Update the files in fuel-plugin to match the upstream"Don Dugger4-8/+9
2016-06-03Update the files in fuel-plugin to match the upstreamdavidjchou4-8/+9
Since the upstream kvmfornfv kernel version and OVS changed, updated the related files in fuel-plugin, so it can build fuel-plugin-kvm with latest kvmfornfv kernel and OVS module. Upstream status: kvmfornfv kernel 4.4.6-rt14nfv Change-Id: I6a7c312f7527acae4d2de64c7b43da4fbec41504 Signed-off-by: davidjchou <david.j.chou@intel.com>
2016-06-03Merge "These changes are the raw update to qemu-2.6."Don Dugger4070-91121/+252483
2016-05-20Remove config file changes from build scriptDonald Dugger1-692/+0
These changes are no longer needed as they are now (correctly) incorporated in the actual config file in the source tree. Upstream status: NA Change-Id: Ide33453789338ba3f7692ca6108de89a4ac3c222 Signed-off-by: Don Dugger <n0ano@n0ano.com>
2016-05-19Merge "Build uio as module to fix initialization of i40e devices in bare ↵Jiang, Yunhong1-1/+1
metal execution of dpdk-16.04."
2016-05-18These changes are the raw update to qemu-2.6.José Pekkarinen4070-91121/+252483
Collission happened in the following patches: migration: do cleanup operation after completion(738df5b9) Bug fix.(1750c932f86) kvmclock: add a new function to update env->tsc.(b52baab2) The code provided by the patches was already in the upstreamed version. Change-Id: I3cc11841a6a76ae20887b2e245710199e1ea7f9a Signed-off-by: José Pekkarinen <jose.pekkarinen@nokia.com>
2016-05-13Add patch to Linux 4.4 to support OVSDonald Dugger2-1/+260
Given that OVS doesn't support Linux 4.4 yet we need to add a patch to the OVS sources so that we can build a 4.4 version of the OVS loadable kernel module that works with the RT Linux 4.4 kernel used in OPNFV. The directory `patches/ovs' contains patches (currently only one) that are applied against the OVS tree. The Fuel build script is modified to apply all of the patches in this directory to the OVS tree. Then a working OVS KLM is created that is then inserted into the RT kernel DEB package so that the end result is an RT kernel that supports OVS. Upstream status: NA Change-Id: I361f92526fb4bcafbeab9ce21570202f4aad1632 Signed-off-by: Don Dugger <n0ano@n0ano.com>
2016-05-13Build uio as module to fix initialization of i40e devicesJosé Pekkarinen1-1/+1
in bare metal execution of dpdk-16.04. Upstream: NA. Change-Id: Ia98461b15348a667c4989dfe1399f0c5bc0f0c12 Signed-off-by: José Pekkarinen <jose.pekkarinen@nokia.com>
2016-05-05Merge "Reverted execution attribute of shell script files"Don Dugger3-0/+0
2016-05-05Merge "Add kernel modules required for OPNFV"Jiang, Yunhong1-130/+732
2016-05-03Reverted execution attribute of shell script filesdavidjchou3-0/+0
Reverted execution attribute of shell script files in fuel-plugin sub-dir. The execution attribute was lost when these files were commited inside fuel-plugin sub-dir, and the executtion attribute of these shell script files is necessary to the build of fuel-plugin-kvm. Upstream status: NA Change-Id: I0ac060a27c4a1570bff17864aa7db5b6b9edf732 Signed-off-by: davidjchou <david.j.chou@intel.com>
2016-05-03Create fuel kvm pluginGuo Ruijing13-0/+1303
Upstream status: NA In Brahmaputra release, KVM plugin functionality was bundled inside the fuel-plugin-qemu. In order for easy maintenance, the KVM plugin functionality would be separated from fuel-plugin-qemu and moved into an independent fuel-plugin-kvm and keep the source code here starting from Colorado release. Change-Id: Id89069234a4529cca40f1887e2d947378f928dd2 Signed-off-by: davidjchou <david.j.chou@intel.com> Signed-off-by: Guo Ruijing <ruijing.guo@intel.com>
2016-04-25Change PTL to raghuveer.reddy@intel.comAric Gardner1-2/+3
Change-Id: Id67d11079e738bc4180c395b86d56cfc81c3770f Signed-off-by: Aric Gardner <agardner@linuxfoundation.org>
2016-04-25Add kernel modules required for OPNFVDon Dugger1-130/+732
The OPNFV environment requires many kernel modules that are not part of the default RT kernel environment. This patch adds those modules back in. Upstream status: NA Change-Id: Id4e63f3d2dd3e19614e9e080adf1cdae9ab26ee1
2016-04-22Update opnfv kernel config for the new kernel 4.4.6-rt14.José Pekkarinen1-67/+237
This config file is based in the previous one, adding the changes needed in the config file for this new kernel version. It has been added in kernel support for CephFS. Upstream: NA. Change-Id: I1de8b4678bdfa81f4fc204f4a02d11f11cb5ae87 Signed-off-by: José Pekkarinen <jose.pekkarinen@nokia.com>
2016-04-13These changes are the raw update to linux-4.4.6-rt14. Kernel sourcesJosé Pekkarinen20902-780094/+2132422
are taken from kernel.org, and rt patch from the rt wiki download page. During the rebasing, the following patch collided: Force tick interrupt and get rid of softirq magic(I70131fb85). Collisions have been removed because its logic was found on the source already. Change-Id: I7f57a4081d9deaa0d9ccfc41a6c8daccdee3b769 Signed-off-by: José Pekkarinen <jose.pekkarinen@nokia.com>
2016-03-01Remove yang.z.zhang@intel.com as committerAric Gardner1-1/+1
Add jose.pekkarinen@nokia.com as committer RT Ticket: 19272, 19265 Change-Id: Ia49e92674f2da2feea53e08231bb4fdd6b54ca88 Signed-off-by: Aric Gardner <agardner@linuxfoundation.org>
2016-02-10Added license informationbrahmaputra.1.0stable/brahmaputraraghu4-1/+17
Upstream Status : Not Applicable For detalied license information please refer to https://wiki.opnfv.org/documentation#licencing_your_documentation Change-Id: Iacba57a9162b25abc474dbac236b4dee5f27544c
2016-01-28Merge "Remove the code to apply yardstick patch"Don Dugger1-3/+0
2016-01-28Fix typos in the documentation.José Pekkarinen3-6/+6
This is a candidate for the Brahmaputra branch. Change-Id: Idbe41ac77f1aae902cd00af4bc9a0e3532f4284a Signed-off-by: José Pekkarinen <jose.pekkarinen@nokia.com>
2016-01-19Add the live migration docuementYunhong Jiang5-1/+109
The document to share the fast live migration environment setup and test method. Change-Id: Icd616f58e2ea39d101fcebc1d760178151d8629f Signed-off-by: Yunhong Jiang <yunhong.jiang@linux.intel.com> Signed-off-by: Liang Li <liang.z.li@intel.com> Signed-off-by: Raghuveer Reddy <raghuveer.reddy@intel.com> Signed-off-by: Yunhong Jiang <yunhong.jiang@linux.intel.com>
2016-01-19Add the documentation for kvm4nfv projectYunhong Jiang6-150/+288
Adding documentation to kvm4nfv project based on https://wiki.opnfv.org/documentation/tools and the contents mostly comes from https://wiki.opnfv.org/nfv-kvm, https://wiki.opnfv.org/nfv-kvm-tuning and https://wiki.opnfv.org/nfv-kvm-test. Change-Id: If321221724ec9b76db065af7cdab97ce981be740 Signed-off-by: Yunhong Jiang <yunhong.jiang@linux.intel.com>
2016-01-15Remove the code to apply yardstick patchYunhong Jiang1-3/+0
The patch to run cyclictest on baremetal (https://gerrit.opnfv.org/gerrit/#/c/3633/) has been merged, thus we don't need apply the patch anymore. Remove it. Change-Id: I7058f9d6c3e873b56be52a0e886fe460506a9911 Signed-off-by: Yunhong Jiang <yunhong.jiang@linux.intel.com>
2016-01-06Merge topic 'yardstick-integ-v2'Don Dugger4-0/+328
* changes: Cyclictest invokation script Add the yardstick invokation script
2016-01-06Merge "Build vfio-pci as a module, as dpdk tools expect it to be."Don Dugger1-2/+2
2016-01-06Merge topic 'yardstick-integ-v2'Don Dugger5-0/+253
* changes: A script to launch real time guest Script to create the guest image Script to create the rt-tests rpm
2016-01-06Merge "Add guest environment setup scripts"Don Dugger2-0/+89
2016-01-06Fix for KVMFORNFV-23(Fuel installed KVM kernel blank screens on boot).José Pekkarinen1-2/+2
The configuration of the kernel enables the framebuffer console without any framebuffer selected other than i915. Adding the VESA compliant framebuffer should fix this issue. Change-Id: Icc384e05774e1de20985aeb19dfef25ae2431bb6 Signed-off-by: José Pekkarinen <jose.pekkarinen@nokia.com>
2016-01-05Cyclictest invokation scriptYunhong Jiang3-0/+180
We run the cyclictest through yardstick, which will help to setup the environment. Environment setup scripts are copied to the yardstick docker images. A yardstick cyclictest yaml file are also used. Change-Id: Iacf1299a38c3c81a08fd5fdbbf64c5a57f30c38b Signed-off-by: Yunhong Jiang <yunhong.jiang@linux.intel.com>
2016-01-05A script to launch real time guestYunhong Jiang1-0/+45
Launch real time guest need special options like lock memory etc. A script is used to achieved this. We invoke the qemu utility directly instead of going through middle layer like libvirt, to get better controll. Change-Id: Ia6ad7313463e2f858516bddd4a4b58e95d8c943e Signed-off-by: David Su <david.w.su@intel.com> Signed-off-by: Yunhong Jiang <yunhong.jiang@linux.intel.com>
2016-01-05Add the yardstick invokation scriptYunhong Jiang1-0/+148
This script does the real yardstick works. It downloads the yardstick code, and run the cyclictest test case. This scripts is copied by the cyclictest.sh to the container image and is executed from the yardstick container. It's based on a script from QiLiang when discussing the integration with yardstick. Change-Id: I5920a21401a3e442d5f4fada05d9e789f2a99add Signed-off-by: Yunhong Jiang <yunhong.jiang@linux.intel.com> Signed-off-by: QiLiang <liangqi1@huawei.com>
2016-01-05Script to create the guest imageYunhong Jiang2-0/+152
This script download the latest CentOS cloud image and modify it to meet the test requirement, including remove cloud related script, setup the ssh keys etc. It derives from yardstick project, which currently only provide support to Ubuntu image. If they support CentOS image, we can possibly switch to them. Change-Id: I9482936615a28da696ff8f51248a62b13e5677f4 Signed-off-by: Yunhong Jiang <yunhong.jiang@linux.intel.com>
2016-01-05Add guest environment setup scriptsYunhong Jiang2-0/+89
To achieve good real time and live migration performance, special setup is needed on guest environment. Two scripts are used to setup the guest environment. The guest-setup0.sh setup the environment that should take effect before the tested kernel is bringup, including install the kernel rpm, the rt-test package and modify the grub entries. The guest-setup1.sh setup the environment that takes effect after the tested kernel is up, like some sysfs entry, interrupt affinity etc. Change-Id: Icaed71e250b314723d6b1814c9ac33c10d99c6a0 Signed-off-by: David Su <david.w.su@intel.com> Signed-off-by: Yunhong Jiang <yunhong.jiang@linux.intel.com>
2016-01-05No change to the grub default settingYunhong Jiang1-1/+0
The kernel rpm installation on rehl/centos is using grubby to setup the default kernel entry, so we should not change the grub default set from saved to 0. Change-Id: I5910f498f5889c052e43d2e1e92b209c05b01455 Signed-off-by: Yunhong Jiang <yunhong.jiang@linux.intel.com>
2016-01-05Script to create the rt-tests rpmYunhong Jiang2-0/+56
There is no rt-tests rpm on centos repo, also we need some special compile flag to enable the -a parameter. So we try to make the rpm ourselves. The version is specified as 0.96. But please notice that the Makefile in the rt-tests tree requires some key, we have to disable that requirement. So this rpm is only for OPNFV testing purpose. Change-Id: Ifdd52649bc14405dbe5ad375dc7fd32087139b18 Signed-off-by: Yunhong Jiang <yunhong.jiang@linux.intel.com>
2016-01-06Merge "Add host environment setup scripts"Jiang, Yunhong3-0/+171
2016-01-05Merge "Add virtual functions support to ixgbe and i40e."Don Dugger1-1/+1
2016-01-05Merge "Add VM preemption latency measuring tool vm-trace"Don Dugger3-0/+706
2016-01-04Add host environment setup scriptsYunhong Jiang3-0/+171
Testing KVM4NFV project requires special host environment to get the best result. Two scrripts are used to setup the environment. The host-setup0.sh setup the environment that should take effect before the kernel is bringup, mostly the grub entries. The host-setup1.sh setup the environment that takes effect after the kernel is up, like some sysfs entry, interrupt affinity etc. The host-config provides the configurations. Change-Id: Ie933ea0089ac82acd39fc48088615215993312f3 Signed-off-by: David Su <david.w.su@intel.com> Signed-off-by: Yunhong Jiang <yunhong.jiang@linux.intel.com>
2015-12-08kvmclock: add a new function to update env->tsc.WW-2015-52Liang Li3-16/+48
(Back port from upstream, commit id: 0fd7e098db30e302d27920487f0afec) The commit 317b0a6d8 fixed an issue which caused by the outdated env->tsc value, but the fix lead to 'cpu_synchronize_all_states()' called twice during live migration. The 'cpu_synchronize_all_states()' takes about 130us for a VM which has 4 vcpus, it's a bit expensive. Synchronize the whole CPU context just for updating env->tsc is too wasting, this patch use a new function to update the env->tsc. Comparing to 'cpu_synchronize_all_states()', it only takes about 20us. Change-Id: I7dc8371aa17b005d073249c3732d9a7424f20a25 Signed-off-by: Liang Li <liang.z.li@intel.com> Message-Id: <1446695464-27116-2-git-send-email-liang.z.li@intel.com> Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
2015-12-04Add VM preemption latency measuring tool vm-traceDavid Su3-0/+706
This tool is to be used to measure VM preemption latencies defined by kvmfornfv requirements. See README.txt for usage of the tool. Change-Id: I78aed07df0af02f2268e36cfeb81923ba2bdb52a Signed-off-by: David Su <david.w.su@intel.com>