- 16 Apr, 2012 7 commits
-
-
Michael S. Tsirkin authored
Intel spec says that TMR needs to be set/cleared when IRR is set, but kvm also clears it on EOI. I did some tests on a real (AMD based) system, and I see same TMR values both before and after EOI, so I think it's a minor bug in kvm. This patch fixes TMR to be set/cleared on IRR set only as per spec. And now that we don't clear TMR, we can save an atomic read of TMR on EOI that's not propagated to ioapic, by checking whether ioapic needs a specific vector first and calculating the mode afterwards. Signed-off-by:
Michael S. Tsirkin <mst@redhat.com> Signed-off-by:
Marcelo Tosatti <mtosatti@redhat.com>
-
Avi Kivity authored
Needed by some framebuffer drivers. See https://bugzilla.kernel.org/show_bug.cgi?id=42779Signed-off-by:
Avi Kivity <avi@redhat.com> Signed-off-by:
-
Avi Kivity authored
General support for the MMX instruction set. Special care is taken to trap pending x87 exceptions so that they are properly reflected to the guest. Signed-off-by:
-
Avi Kivity authored
Used to write to framebuffers (by at least Icaros). Signed-off-by:
-
Stefan Hajnoczi authored
An Ubuntu 9.10 Karmic Koala guest is unable to boot or install due to missing movdqa emulation: kvm_exit: reason EXCEPTION_NMI rip 0x7fef3e025a7b info 7fef3e799000 80000b0e kvm_page_fault: address 7fef3e799000 error_code f kvm_emulate_insn: 0:7fef3e025a7b: 66 0f 7f 07 (prot64) movdqa %xmm0,(%rdi) [avi: mark it explicitly aligned] Signed-off-by:
Stefan Hajnoczi <stefanha@linux.vnet.ibm.com> Signed-off-by:
-
Avi Kivity authored
x86 defines three classes of vector instructions: explicitly aligned (#GP(0) if unaligned, explicitly unaligned, and default (which depends on the encoding: AVX is unaligned, SSE is aligned). Add support for marking an instruction as explicitly aligned or unaligned, and mark MOVDQU as unaligned. Signed-off-by:
-
Josh Triplett authored
Enable x86 feature-based autoloading for the kvm-amd module on CPUs with X86_FEATURE_SVM. Signed-off-by:
Josh Triplett <josh@joshtriplett.org> Signed-off-by:
-
- 12 Apr, 2012 1 commit
-
-
Scott Wood authored
e6500 support (commit 10241842, "powerpc: Add initial e6500 cpu support" and the introduction of CPU_FTR_EMB_HV (commit 73196cd3, "KVM: PPC: e500mc support") collided during merge, leaving e6500's CPU table entry missing CPU_FTR_EMB_HV. Signed-off-by:
Scott Wood <scottwood@freescale.com> Signed-off-by:
-
- 08 Apr, 2012 32 commits
-
-
Takuya Yoshikawa authored
Iteration using rmap_next(), the actual body is pte_list_next(), is inefficient: every time we call it we start from checking whether rmap holds a single spte or points to a descriptor which links more sptes. In the case of shadow paging, this quadratic total iteration cost is a problem. Even for two dimensional paging, with EPT/NPT on, in which we almost always have a single mapping, the extra checks at the end of the iteration should be eliminated. This patch fixes this by introducing rmap_iterator which keeps the iteration context for the next search. Furthermore the implementation of rmap_next() is splitted into two functions, rmap_get_first() and rmap_get_next(), to avoid repeatedly checking whether the rmap being iterated on has only one spte. Although there seemed to be only a slight change for EPT/NPT, the actual improvement was significant: we observed that GET_DIRTY_LOG for 1GB dirty memory became 15% faster than before. This is probably because the new code is easy to make branch predictions. Note: we just remove pte_list_next() because we can think of parent_ptes as a reverse mapping. Signed-off-by:
Takuya Yoshikawa <yoshikawa.takuya@oss.ntt.co.jp> Signed-off-by:
-
Takuya Yoshikawa authored
We have PTE_LIST_EXT + 1 pointers in this structure and these 40/20 bytes do not fit cache lines well. Furthermore, some allocators may use 64/32-byte objects for the pte_list_desc cache. This patch solves this problem by changing PTE_LIST_EXT from 4 to 3. For shadow paging, the new size is still large enough to hold both the kernel and process mappings for usual anonymous pages. For file mappings, there may be a slight change in the cache usage. Note: with EPT/NPT we almost always have a single spte in each reverse mapping and we will not see any change by this. Signed-off-by:
-
Davidlohr Bueso authored
Since most guests will have paging enabled for memory management, add likely() optimization around CR0.PG checks. Signed-off-by:
Davidlohr Bueso <dave@gnu.org> Signed-off-by:
-
Josh Triplett authored
Enable x86 feature-based autoloading for the kvm-intel module on CPUs with X86_FEATURE_VMX. Signed-off-by:
Kay Sievers <kay@vrfy.org> Signed-off-by:
-
Benjamin Herrenschmidt authored
When the kernel calls into RTAS, it switches to 32-bit mode. The magic page was is longer accessible in that case, causing the patched instructions in the RTAS call wrapper to crash. This fixes it by making available a 32-bit mapping of the magic page in that case. This mapping is flushed whenever we switch the kernel back to 64-bit mode. Signed-off-by:
Benjamin Herrenschmidt <benh@kernel.crashing.org> [agraf: add a check if the magic page is mapped] Signed-off-by:
Alexander Graf <agraf@suse.de> Signed-off-by:
-
Alexander Graf authored
When running kvm_vcpu_block and it realizes that the CPU is actually good to run, we get a request bit set for KVM_REQ_UNHALT. Right now, there's nothing we can do with that bit, so let's unset it right after the call again so we don't get confused in our later checks for pending work. Signed-off-by:
-
Alexander Graf authored
When running PR KVM on a p7 system in bare metal, we get HV exits instead of normal supervisor traps. Semantically they are identical though and the HSRR vs SRR difference is already taken care of in the exit code. So all we need to do is handle them in addition to our normal exits. Signed-off-by:
-
Alexander Graf authored
There are 4 conditional trapping instructions: tw, twi, td, tdi. The ones with an i take an immediate comparison, the others compare two registers. All of them arrive in the emulator when the condition to trap was successfully fulfilled. Unfortunately, we were only implementing the i versions so far, so let's also add support for the other two. This fixes kernel booting with recents book3s_32 guest kernels. Reported-by:
Jörg Sommer <joerg@alea.gnuu.de> Signed-off-by:
-
Alexander Graf authored
When emulating updating load/store instructions (lwzu, stwu, ...) we need to write the effective address of the load/store into a register. Currently, we write the physical address in there, which is very wrong. So instead let's save off where the virtual fault was on MMIO and use that information as value to put into the register. While at it, also move the XOP variants of the above instructions to the new scheme of using the already known vaddr instead of calculating it themselves. Reported-by:
-
Paul Mackerras authored
It turns out that on POWER7, writing to the DABR can cause a corrupted value to be written if the PMU is active and updating SDAR in continuous sampling mode. To work around this, we make sure that the PMU is inactive and SDAR updates are disabled (via MMCRA) when we are context-switching DABR. When the guest sets DABR via the H_SET_DABR hypercall, we use a slightly different workaround, which is to read back the DABR and write it again if it got corrupted. While we are at it, make it consistent that the saving and restoring of the guest's non-volatile GPRs and the FPRs are done with the guest setup of the PMU active. Signed-off-by:
Paul Mackerras <paulus@samba.org> Signed-off-by:
-
Paul Mackerras authored
Commits 2f5cdd5487 ("KVM: PPC: Book3S HV: Make secondary threads more robust against stray IPIs") and 1c2066b0f7 ("KVM: PPC: Book3S HV: Make virtual processor area registration more robust") added fields to struct kvm_vcpu_arch inside #ifdef CONFIG_KVM_BOOK3S_64_HV regions, and added lines to arch/powerpc/kernel/asm-offsets.c to generate assembler constants for their offsets. Unfortunately this led to compile errors on Book 3S machines for configs that had KVM enabled but not CONFIG_KVM_BOOK3S_64_HV. This fixes the problem by moving the offending lines inside #ifdef CONFIG_KVM_BOOK3S_64_HV regions. Signed-off-by: -
Bharat Bhushan authored
No instruction which can change Condition Register (CR) should be executed after Guest CR is loaded. So the guest CR is restored after the Exit Timing in lightweight_exit executes cmpw, which can clobber CR. Signed-off-by:
Bharat Bhushan <bharat.bhushan@freescale.com> Signed-off-by:
-
Paul Mackerras authored
This adds code to measure "stolen" time per virtual core in units of timebase ticks, and to report the stolen time to the guest using the dispatch trace log (DTL). The guest can register an area of memory for the DTL for a given vcpu. The DTL is a ring buffer where KVM fills in one entry every time it enters the guest for that vcpu. Stolen time is measured as time when the virtual core is not running, either because the vcore is not runnable (e.g. some of its vcpus are executing elsewhere in the kernel or in userspace), or when the vcpu thread that is running the vcore is preempted. This includes time when all the vcpus are idle (i.e. have executed the H_CEDE hypercall), which is OK because the guest accounts stolen time while idle as idle time. Each vcpu keeps a record of how much stolen time has been reported to the guest for that vcpu so far. When we are about to enter the guest, we create a new DTL entry (if the guest vcpu has a DTL) and report the difference between total stolen time for the vcore and stolen time reported so far for the vcpu as the "enqueue to dispatch" time in the DTL entry. Signed-off-by:
-
Paul Mackerras authored
The PAPR API allows three sorts of per-virtual-processor areas to be registered (VPA, SLB shadow buffer, and dispatch trace log), and furthermore, these can be registered and unregistered for another virtual CPU. Currently we just update the vcpu fields pointing to these areas at the time of registration or unregistration. If this is done on another vcpu, there is the possibility that the target vcpu is using those fields at the time and could end up using a bogus pointer and corrupting memory. This fixes the race by making the target cpu itself do the update, so we can be sure that the update happens at a time when the fields aren't being used. Each area now has a struct kvmppc_vpa which is used to manage these updates. There is also a spinlock which protects access to all of the kvmppc_vpa structs, other than to the pinned_addr fields. (We could have just taken the spinlock when using the vpa, slb_shadow or dtl fields, but that would mean taking the spinlock on every guest entry and exit.) This also changes 'struct dtl' (which was undefined) to 'struct dtl_entry', which is what the rest of the kernel uses. Thanks to Michael Ellerman <michael@ellerman.id.au> for pointing out the need to initialize vcpu->arch.vpa_update_lock. Signed-off-by:
-
Paul Mackerras authored
Currently on POWER7, if we are running the guest on a core and we don't need all the hardware threads, we do nothing to ensure that the unused threads aren't executing in the kernel (other than checking that they are offline). We just assume they're napping and we don't do anything to stop them trying to enter the kernel while the guest is running. This means that a stray IPI can wake up the hardware thread and it will then try to enter the kernel, but since the core is in guest context, it will execute code from the guest in hypervisor mode once it turns the MMU on, which tends to lead to crashes or hangs in the host. This fixes the problem by adding two new one-byte flags in the kvmppc_host_state structure in the PACA which are used to interlock between the primary thread and the unused secondary threads when entering the guest. With these flags, the primary thread can ensure that the unused secondaries are not already in kernel mode (i.e. handling a stray IPI) and then indicate that they should not try to enter the kernel if they do get woken for any reason. Instead they will go into KVM code, find that there is no vcpu to run, acknowledge and clear the IPI and go back to nap mode. Signed-off-by:
-
Alexander Graf authored
On PPC, CR2-CR4 are nonvolatile, thus have to be saved across function calls. We didn't respect that for any architecture until Paul spotted it in his patch for Book3S-HV. This patch saves/restores CR for all KVM capable PPC hosts. Signed-off-by:
-
Matt Evans authored
SPAPR support includes various in-kernel hypercalls, improving performance by cutting out the exit to userspace. H_BULK_REMOVE is implemented in this patch. Signed-off-by:
Matt Evans <matt@ozlabs.org> Signed-off-by:
-
Alexander Graf authored
So far, we've always called prepare_to_enter even when all we did was return to the host. This patch changes that semantic to only call prepare_to_enter when we actually want to get back into the guest. Signed-off-by:
-
Alexander Graf authored
When we get a performance monitor interrupt, we need to make sure that the host receives it. So reinject it like we reinject the other host destined interrupts. Signed-off-by:
-
Alexander Graf authored
When reinjecting an interrupt into the host interrupt handler after we're back in host kernel land, we need to tell the kernel where the interrupt happened. We can't tell it that we were in guest state, because that might lead to random code walking host addresses. So instead, we tell it that we came from the interrupt reinject code. This helps getting reasonable numbers out of perf. Signed-off-by:
-
Alexander Graf authored
When during guest context we get a performance monitor interrupt, we currently bail out and oops. Let's route it to its correct handler instead. Signed-off-by:
-
Alexander Graf authored
The tlbncfg registers should be populated with their respective TLB's values. Fix the obvious typo. Signed-off-by:
-
Alexander Graf authored
There was some unused code in the exit code path that must have been a leftover from earlier iterations. While it did no harm, it's superfluous and thus should be removed. Signed-off-by:
-
Alexander Graf authored
The comment for program interrupts triggered when using bookehv was misleading. Update it to mention why MSR_GS indicates that we have to inject an interrupt into the guest again, not emulate it. Signed-off-by:
-
Alexander Graf authored
When during guest execution we get a machine check interrupt, we don't know how to handle it yet. So let's add the error printing code back again that we dropped accidently earlier and tell user space that something went really wrong. Signed-off-by:
-
Alexander Graf authored
For BookE HV the guest visible MSR is shared->msr and is identical to the MSR that is in use while the guest is running, because we can't trap reads from/to MSR. So shadow_msr is unused there. Indicate that with a comment. Signed-off-by:
-
Alexander Graf authored
We need to make sure that no MAS updates happen automatically while we have the guest MAS registers loaded. So move the disabling code a bit higher up so that it covers the full time we have guest values in MAS registers. The race this patch fixes should never occur, but it makes the code a bit more logical to do it this way around. Signed-off-by:
-
Alexander Graf authored
The SET_VCPU macro is a leftover from times when the vcpu struct wasn't stored in the thread on vcpu_load/put. It's not needed anymore. Remove it. Signed-off-by:
-
Alexander Graf authored
Instead if doing #ifndef CONFIG_64BIT ... #else ... #endif we should rather do #ifdef CONFIG_64BIT ... #else ... #endif which is a lot easier to read. Change the bookehv implementation to stick with this rule. Signed-off-by:
-
Alexander Graf authored
When using exit timing stats, we clobber r9 in the NEED_EMU case, so better move that part down a few lines and fix it that way. Signed-off-by:
-
Alexander Graf authored
The semantics of BOOKE_IRQPRIO_MAX changed to denote the highest available irqprio + 1, so let's reflect that in the code too. Signed-off-by:
-
Alexander Graf authored
Instead of checking whether we should reschedule only when we exited due to an interrupt, let's always check before entering the guest back again. This gets the target more in line with the other archs. Also while at it, generalize the whole thing so that eventually we could have a single kvmppc_prepare_to_enter function for all ppc targets that does signal and reschedule checking for us. Signed-off-by:
-
