From 6f779e1d359b8d5801f677c1d49dcfa10bf95674 Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?C=C3=A9dric=20Le=20Goater?= Date: Mon, 11 Oct 2021 09:02:03 +0200 Subject: powerpc/xive: Discard disabled interrupts in get_irqchip_state() MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit When an interrupt is passed through, the KVM XIVE device calls the set_vcpu_affinity() handler which raises the P bit to mask the interrupt and to catch any in-flight interrupts while routing the interrupt to the guest. On the guest side, drivers (like some Intels) can request at probe time some MSIs and call synchronize_irq() to check that there are no in flight interrupts. This will call the XIVE get_irqchip_state() handler which will always return true as the interrupt P bit has been set on the host side and lock the CPU in an infinite loop. Fix that by discarding disabled interrupts in get_irqchip_state(). Fixes: da15c03b047d ("powerpc/xive: Implement get_irqchip_state method for XIVE to fix shutdown race") Cc: stable@vger.kernel.org #v5.4+ Signed-off-by: Cédric Le Goater Tested-by: seeteena Signed-off-by: Michael Ellerman Link: https://lore.kernel.org/r/20211011070203.99726-1-clg@kaod.org --- arch/powerpc/sysdev/xive/common.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/arch/powerpc/sysdev/xive/common.c b/arch/powerpc/sysdev/xive/common.c index c732ce5a3e1a..c5d75c02ad8b 100644 --- a/arch/powerpc/sysdev/xive/common.c +++ b/arch/powerpc/sysdev/xive/common.c @@ -945,7 +945,8 @@ static int xive_get_irqchip_state(struct irq_data *data, * interrupt to be inactive in that case. */ *state = (pq != XIVE_ESB_INVALID) && !xd->stale_p && - (xd->saved_p || !!(pq & XIVE_ESB_VAL_P)); + (xd->saved_p || (!!(pq & XIVE_ESB_VAL_P) && + !irqd_irq_disabled(data))); return 0; default: return -EINVAL; -- cgit v1.2.3 From 9b4416c5095c20e110c82ae602c254099b83b72f Mon Sep 17 00:00:00 2001 From: Michael Ellerman Date: Fri, 15 Oct 2021 23:01:48 +1100 Subject: KVM: PPC: Book3S HV: Fix stack handling in idle_kvm_start_guest() In commit 10d91611f426 ("powerpc/64s: Reimplement book3s idle code in C") kvm_start_guest() became idle_kvm_start_guest(). The old code allocated a stack frame on the emergency stack, but didn't use the frame to store anything, and also didn't store anything in its caller's frame. idle_kvm_start_guest() on the other hand is written more like a normal C function, it creates a frame on entry, and also stores CR/LR into its callers frame (per the ABI). The problem is that there is no caller frame on the emergency stack. The emergency stack for a given CPU is allocated with: paca_ptrs[i]->emergency_sp = alloc_stack(limit, i) + THREAD_SIZE; So emergency_sp actually points to the first address above the emergency stack allocation for a given CPU, we must not store above it without first decrementing it to create a frame. This is different to the regular kernel stack, paca->kstack, which is initialised to point at an initial frame that is ready to use. idle_kvm_start_guest() stores the backchain, CR and LR all of which write outside the allocation for the emergency stack. It then creates a stack frame and saves the non-volatile registers. Unfortunately the frame it creates is not large enough to fit the non-volatiles, and so the saving of the non-volatile registers also writes outside the emergency stack allocation. The end result is that we corrupt whatever is at 0-24 bytes, and 112-248 bytes above the emergency stack allocation. In practice this has gone unnoticed because the memory immediately above the emergency stack happens to be used for other stack allocations, either another CPUs mc_emergency_sp or an IRQ stack. See the order of calls to irqstack_early_init() and emergency_stack_init(). The low addresses of another stack are the top of that stack, and so are only used if that stack is under extreme pressue, which essentially never happens in practice - and if it did there's a high likelyhood we'd crash due to that stack overflowing. Still, we shouldn't be corrupting someone else's stack, and it is purely luck that we aren't corrupting something else. To fix it we save CR/LR into the caller's frame using the existing r1 on entry, we then create a SWITCH_FRAME_SIZE frame (which has space for pt_regs) on the emergency stack with the backchain pointing to the existing stack, and then finally we switch to the new frame on the emergency stack. Fixes: 10d91611f426 ("powerpc/64s: Reimplement book3s idle code in C") Cc: stable@vger.kernel.org # v5.2+ Signed-off-by: Michael Ellerman Link: https://lore.kernel.org/r/20211015133929.832061-1-mpe@ellerman.id.au --- arch/powerpc/kvm/book3s_hv_rmhandlers.S | 19 ++++++++++--------- 1 file changed, 10 insertions(+), 9 deletions(-) diff --git a/arch/powerpc/kvm/book3s_hv_rmhandlers.S b/arch/powerpc/kvm/book3s_hv_rmhandlers.S index 90484425a1e6..ec57952b60b0 100644 --- a/arch/powerpc/kvm/book3s_hv_rmhandlers.S +++ b/arch/powerpc/kvm/book3s_hv_rmhandlers.S @@ -255,13 +255,15 @@ kvm_novcpu_exit: * r3 contains the SRR1 wakeup value, SRR1 is trashed. */ _GLOBAL(idle_kvm_start_guest) - ld r4,PACAEMERGSP(r13) mfcr r5 mflr r0 - std r1,0(r4) - std r5,8(r4) - std r0,16(r4) - subi r1,r4,STACK_FRAME_OVERHEAD + std r5, 8(r1) // Save CR in caller's frame + std r0, 16(r1) // Save LR in caller's frame + // Create frame on emergency stack + ld r4, PACAEMERGSP(r13) + stdu r1, -SWITCH_FRAME_SIZE(r4) + // Switch to new frame on emergency stack + mr r1, r4 SAVE_NVGPRS(r1) /* @@ -395,10 +397,9 @@ kvm_no_guest: /* set up r3 for return */ mfspr r3,SPRN_SRR1 REST_NVGPRS(r1) - addi r1, r1, STACK_FRAME_OVERHEAD - ld r0, 16(r1) - ld r5, 8(r1) - ld r1, 0(r1) + ld r1, 0(r1) // Switch back to caller stack + ld r0, 16(r1) // Reload LR + ld r5, 8(r1) // Reload CR mtlr r0 mtcr r5 blr -- cgit v1.2.3 From cdeb5d7d890e14f3b70e8087e745c4a6a7d9f337 Mon Sep 17 00:00:00 2001 From: Michael Ellerman Date: Fri, 15 Oct 2021 23:02:08 +1100 Subject: KVM: PPC: Book3S HV: Make idle_kvm_start_guest() return 0 if it went to guest We call idle_kvm_start_guest() from power7_offline() if the thread has been requested to enter KVM. We pass it the SRR1 value that was returned from power7_idle_insn() which tells us what sort of wakeup we're processing. Depending on the SRR1 value we pass in, the KVM code might enter the guest, or it might return to us to do some host action if the wakeup requires it. If idle_kvm_start_guest() is able to handle the wakeup, and enter the guest it is supposed to indicate that by returning a zero SRR1 value to us. That was the behaviour prior to commit 10d91611f426 ("powerpc/64s: Reimplement book3s idle code in C"), however in that commit the handling of SRR1 was reworked, and the zeroing behaviour was lost. Returning from idle_kvm_start_guest() without zeroing the SRR1 value can confuse the host offline code, causing the guest to crash and other weirdness. Fixes: 10d91611f426 ("powerpc/64s: Reimplement book3s idle code in C") Cc: stable@vger.kernel.org # v5.2+ Signed-off-by: Michael Ellerman Link: https://lore.kernel.org/r/20211015133929.832061-2-mpe@ellerman.id.au --- arch/powerpc/kvm/book3s_hv_rmhandlers.S | 9 +++++++-- 1 file changed, 7 insertions(+), 2 deletions(-) diff --git a/arch/powerpc/kvm/book3s_hv_rmhandlers.S b/arch/powerpc/kvm/book3s_hv_rmhandlers.S index ec57952b60b0..eb776d0c5d8e 100644 --- a/arch/powerpc/kvm/book3s_hv_rmhandlers.S +++ b/arch/powerpc/kvm/book3s_hv_rmhandlers.S @@ -264,6 +264,7 @@ _GLOBAL(idle_kvm_start_guest) stdu r1, -SWITCH_FRAME_SIZE(r4) // Switch to new frame on emergency stack mr r1, r4 + std r3, 32(r1) // Save SRR1 wakeup value SAVE_NVGPRS(r1) /* @@ -315,6 +316,10 @@ kvm_unsplit_wakeup: kvm_secondary_got_guest: + // About to go to guest, clear saved SRR1 + li r0, 0 + std r0, 32(r1) + /* Set HSTATE_DSCR(r13) to something sensible */ ld r6, PACA_DSCR_DEFAULT(r13) std r6, HSTATE_DSCR(r13) @@ -394,8 +399,8 @@ kvm_no_guest: mfspr r4, SPRN_LPCR rlwimi r4, r3, 0, LPCR_PECE0 | LPCR_PECE1 mtspr SPRN_LPCR, r4 - /* set up r3 for return */ - mfspr r3,SPRN_SRR1 + // Return SRR1 wakeup value, or 0 if we went into the guest + ld r3, 32(r1) REST_NVGPRS(r1) ld r1, 0(r1) // Switch back to caller stack ld r0, 16(r1) // Reload LR -- cgit v1.2.3