summaryrefslogtreecommitdiffstats
path: root/arch/powerpc/kvm
diff options
context:
space:
mode:
Diffstat (limited to 'arch/powerpc/kvm')
-rw-r--r--arch/powerpc/kvm/book3s_64_mmu_radix.c42
-rw-r--r--arch/powerpc/kvm/book3s_hv.c5
-rw-r--r--arch/powerpc/kvm/book3s_hv_rmhandlers.S100
-rw-r--r--arch/powerpc/kvm/book3s_xive_template.c108
4 files changed, 195 insertions, 60 deletions
diff --git a/arch/powerpc/kvm/book3s_64_mmu_radix.c b/arch/powerpc/kvm/book3s_64_mmu_radix.c
index a57eafec4dc2e..481da8f93fa44 100644
--- a/arch/powerpc/kvm/book3s_64_mmu_radix.c
+++ b/arch/powerpc/kvm/book3s_64_mmu_radix.c
@@ -162,7 +162,7 @@ static void kvmppc_radix_tlbie_page(struct kvm *kvm, unsigned long addr,
if (cpu_has_feature(CPU_FTR_P9_TLBIE_BUG))
asm volatile(PPC_TLBIE_5(%0, %1, 0, 0, 1)
: : "r" (addr), "r" (kvm->arch.lpid) : "memory");
- asm volatile("ptesync": : :"memory");
+ asm volatile("eieio ; tlbsync ; ptesync": : :"memory");
}
static void kvmppc_radix_flush_pwc(struct kvm *kvm, unsigned long addr)
@@ -173,7 +173,7 @@ static void kvmppc_radix_flush_pwc(struct kvm *kvm, unsigned long addr)
/* RIC=1 PRS=0 R=1 IS=2 */
asm volatile(PPC_TLBIE_5(%0, %1, 1, 0, 1)
: : "r" (rb), "r" (kvm->arch.lpid) : "memory");
- asm volatile("ptesync": : :"memory");
+ asm volatile("eieio ; tlbsync ; ptesync": : :"memory");
}
unsigned long kvmppc_radix_update_pte(struct kvm *kvm, pte_t *ptep,
@@ -200,6 +200,7 @@ void kvmppc_radix_set_pte_at(struct kvm *kvm, unsigned long addr,
}
static struct kmem_cache *kvm_pte_cache;
+static struct kmem_cache *kvm_pmd_cache;
static pte_t *kvmppc_pte_alloc(void)
{
@@ -217,6 +218,16 @@ static inline int pmd_is_leaf(pmd_t pmd)
return !!(pmd_val(pmd) & _PAGE_PTE);
}
+static pmd_t *kvmppc_pmd_alloc(void)
+{
+ return kmem_cache_alloc(kvm_pmd_cache, GFP_KERNEL);
+}
+
+static void kvmppc_pmd_free(pmd_t *pmdp)
+{
+ kmem_cache_free(kvm_pmd_cache, pmdp);
+}
+
static int kvmppc_create_pte(struct kvm *kvm, pte_t pte, unsigned long gpa,
unsigned int level, unsigned long mmu_seq)
{
@@ -239,7 +250,7 @@ static int kvmppc_create_pte(struct kvm *kvm, pte_t pte, unsigned long gpa,
if (pud && pud_present(*pud) && !pud_huge(*pud))
pmd = pmd_offset(pud, gpa);
else if (level <= 1)
- new_pmd = pmd_alloc_one(kvm->mm, gpa);
+ new_pmd = kvmppc_pmd_alloc();
if (level == 0 && !(pmd && pmd_present(*pmd) && !pmd_is_leaf(*pmd)))
new_ptep = kvmppc_pte_alloc();
@@ -382,7 +393,7 @@ static int kvmppc_create_pte(struct kvm *kvm, pte_t pte, unsigned long gpa,
if (new_pud)
pud_free(kvm->mm, new_pud);
if (new_pmd)
- pmd_free(kvm->mm, new_pmd);
+ kvmppc_pmd_free(new_pmd);
if (new_ptep)
kvmppc_pte_free(new_ptep);
return ret;
@@ -584,7 +595,7 @@ int kvm_unmap_radix(struct kvm *kvm, struct kvm_memory_slot *memslot,
ptep = __find_linux_pte(kvm->arch.pgtable, gpa, NULL, &shift);
if (ptep && pte_present(*ptep)) {
- old = kvmppc_radix_update_pte(kvm, ptep, _PAGE_PRESENT, 0,
+ old = kvmppc_radix_update_pte(kvm, ptep, ~0UL, 0,
gpa, shift);
kvmppc_radix_tlbie_page(kvm, gpa, shift);
if ((old & _PAGE_DIRTY) && memslot->dirty_bitmap) {
@@ -758,7 +769,7 @@ void kvmppc_free_radix(struct kvm *kvm)
kvmppc_pte_free(pte);
pmd_clear(pmd);
}
- pmd_free(kvm->mm, pmd_offset(pud, 0));
+ kvmppc_pmd_free(pmd_offset(pud, 0));
pud_clear(pud);
}
pud_free(kvm->mm, pud_offset(pgd, 0));
@@ -770,20 +781,35 @@ void kvmppc_free_radix(struct kvm *kvm)
static void pte_ctor(void *addr)
{
- memset(addr, 0, PTE_TABLE_SIZE);
+ memset(addr, 0, RADIX_PTE_TABLE_SIZE);
+}
+
+static void pmd_ctor(void *addr)
+{
+ memset(addr, 0, RADIX_PMD_TABLE_SIZE);
}
int kvmppc_radix_init(void)
{
- unsigned long size = sizeof(void *) << PTE_INDEX_SIZE;
+ unsigned long size = sizeof(void *) << RADIX_PTE_INDEX_SIZE;
kvm_pte_cache = kmem_cache_create("kvm-pte", size, size, 0, pte_ctor);
if (!kvm_pte_cache)
return -ENOMEM;
+
+ size = sizeof(void *) << RADIX_PMD_INDEX_SIZE;
+
+ kvm_pmd_cache = kmem_cache_create("kvm-pmd", size, size, 0, pmd_ctor);
+ if (!kvm_pmd_cache) {
+ kmem_cache_destroy(kvm_pte_cache);
+ return -ENOMEM;
+ }
+
return 0;
}
void kvmppc_radix_exit(void)
{
kmem_cache_destroy(kvm_pte_cache);
+ kmem_cache_destroy(kvm_pmd_cache);
}
diff --git a/arch/powerpc/kvm/book3s_hv.c b/arch/powerpc/kvm/book3s_hv.c
index 4d07fca5121c5..cb6d2313b19f4 100644
--- a/arch/powerpc/kvm/book3s_hv.c
+++ b/arch/powerpc/kvm/book3s_hv.c
@@ -2441,6 +2441,7 @@ static void init_vcore_to_run(struct kvmppc_vcore *vc)
vc->in_guest = 0;
vc->napping_threads = 0;
vc->conferring_threads = 0;
+ vc->tb_offset_applied = 0;
}
static bool can_dynamic_split(struct kvmppc_vcore *vc, struct core_info *cip)
@@ -2911,8 +2912,12 @@ static noinline void kvmppc_run_core(struct kvmppc_vcore *vc)
srcu_idx = srcu_read_lock(&vc->kvm->srcu);
+ this_cpu_disable_ftrace();
+
trap = __kvmppc_vcore_entry();
+ this_cpu_enable_ftrace();
+
srcu_read_unlock(&vc->kvm->srcu, srcu_idx);
trace_hardirqs_off();
diff --git a/arch/powerpc/kvm/book3s_hv_rmhandlers.S b/arch/powerpc/kvm/book3s_hv_rmhandlers.S
index bd63fa8a08b5d..b97d261d3b899 100644
--- a/arch/powerpc/kvm/book3s_hv_rmhandlers.S
+++ b/arch/powerpc/kvm/book3s_hv_rmhandlers.S
@@ -342,6 +342,9 @@ kvm_start_guest:
ld r2,PACATOC(r13)
+ li r0,0
+ stb r0,PACA_FTRACE_ENABLED(r13)
+
li r0,KVM_HWTHREAD_IN_KVM
stb r0,HSTATE_HWTHREAD_STATE(r13)
@@ -692,6 +695,7 @@ END_FTR_SECTION_IFSET(CPU_FTR_ARCH_300)
22: ld r8,VCORE_TB_OFFSET(r5)
cmpdi r8,0
beq 37f
+ std r8, VCORE_TB_OFFSET_APPL(r5)
mftb r6 /* current host timebase */
add r8,r8,r6
mtspr SPRN_TBU40,r8 /* update upper 40 bits */
@@ -940,18 +944,6 @@ FTR_SECTION_ELSE
ALT_FTR_SECTION_END_IFCLR(CPU_FTR_ARCH_300)
8:
- /*
- * Set the decrementer to the guest decrementer.
- */
- ld r8,VCPU_DEC_EXPIRES(r4)
- /* r8 is a host timebase value here, convert to guest TB */
- ld r5,HSTATE_KVM_VCORE(r13)
- ld r6,VCORE_TB_OFFSET(r5)
- add r8,r8,r6
- mftb r7
- subf r3,r7,r8
- mtspr SPRN_DEC,r3
-
ld r5, VCPU_SPRG0(r4)
ld r6, VCPU_SPRG1(r4)
ld r7, VCPU_SPRG2(r4)
@@ -1005,6 +997,18 @@ ALT_FTR_SECTION_END_IFCLR(CPU_FTR_ARCH_300)
mtspr SPRN_LPCR,r8
isync
+ /*
+ * Set the decrementer to the guest decrementer.
+ */
+ ld r8,VCPU_DEC_EXPIRES(r4)
+ /* r8 is a host timebase value here, convert to guest TB */
+ ld r5,HSTATE_KVM_VCORE(r13)
+ ld r6,VCORE_TB_OFFSET_APPL(r5)
+ add r8,r8,r6
+ mftb r7
+ subf r3,r7,r8
+ mtspr SPRN_DEC,r3
+
/* Check if HDEC expires soon */
mfspr r3, SPRN_HDEC
EXTEND_HDEC(r3)
@@ -1597,8 +1601,27 @@ END_MMU_FTR_SECTION_IFSET(MMU_FTR_TYPE_RADIX)
guest_bypass:
stw r12, STACK_SLOT_TRAP(r1)
- mr r3, r12
+
+ /* Save DEC */
+ /* Do this before kvmhv_commence_exit so we know TB is guest TB */
+ ld r3, HSTATE_KVM_VCORE(r13)
+ mfspr r5,SPRN_DEC
+ mftb r6
+ /* On P9, if the guest has large decr enabled, don't sign extend */
+BEGIN_FTR_SECTION
+ ld r4, VCORE_LPCR(r3)
+ andis. r4, r4, LPCR_LD@h
+ bne 16f
+END_FTR_SECTION_IFSET(CPU_FTR_ARCH_300)
+ extsw r5,r5
+16: add r5,r5,r6
+ /* r5 is a guest timebase value here, convert to host TB */
+ ld r4,VCORE_TB_OFFSET_APPL(r3)
+ subf r5,r4,r5
+ std r5,VCPU_DEC_EXPIRES(r9)
+
/* Increment exit count, poke other threads to exit */
+ mr r3, r12
bl kvmhv_commence_exit
nop
ld r9, HSTATE_KVM_VCPU(r13)
@@ -1639,23 +1662,6 @@ guest_bypass:
mtspr SPRN_PURR,r3
mtspr SPRN_SPURR,r4
- /* Save DEC */
- ld r3, HSTATE_KVM_VCORE(r13)
- mfspr r5,SPRN_DEC
- mftb r6
- /* On P9, if the guest has large decr enabled, don't sign extend */
-BEGIN_FTR_SECTION
- ld r4, VCORE_LPCR(r3)
- andis. r4, r4, LPCR_LD@h
- bne 16f
-END_FTR_SECTION_IFSET(CPU_FTR_ARCH_300)
- extsw r5,r5
-16: add r5,r5,r6
- /* r5 is a guest timebase value here, convert to host TB */
- ld r4,VCORE_TB_OFFSET(r3)
- subf r5,r4,r5
- std r5,VCPU_DEC_EXPIRES(r9)
-
BEGIN_FTR_SECTION
b 8f
END_FTR_SECTION_IFCLR(CPU_FTR_ARCH_207S)
@@ -1905,6 +1911,14 @@ END_FTR_SECTION_IFSET(CPU_FTR_ARCH_300)
cmpwi cr2, r0, 0
beq cr2, 4f
+ /*
+ * Radix: do eieio; tlbsync; ptesync sequence in case we
+ * interrupted the guest between a tlbie and a ptesync.
+ */
+ eieio
+ tlbsync
+ ptesync
+
/* Radix: Handle the case where the guest used an illegal PID */
LOAD_REG_ADDR(r4, mmu_base_pid)
lwz r3, VCPU_GUEST_PID(r9)
@@ -2017,9 +2031,11 @@ END_FTR_SECTION_IFSET(CPU_FTR_ARCH_207S)
27:
/* Subtract timebase offset from timebase */
- ld r8,VCORE_TB_OFFSET(r5)
+ ld r8, VCORE_TB_OFFSET_APPL(r5)
cmpdi r8,0
beq 17f
+ li r0, 0
+ std r0, VCORE_TB_OFFSET_APPL(r5)
mftb r6 /* current guest timebase */
subf r8,r8,r6
mtspr SPRN_TBU40,r8 /* update upper 40 bits */
@@ -2700,7 +2716,7 @@ END_FTR_SECTION_IFSET(CPU_FTR_ARCH_300)
add r3, r3, r5
ld r4, HSTATE_KVM_VCPU(r13)
ld r5, HSTATE_KVM_VCORE(r13)
- ld r6, VCORE_TB_OFFSET(r5)
+ ld r6, VCORE_TB_OFFSET_APPL(r5)
subf r3, r6, r3 /* convert to host TB value */
std r3, VCPU_DEC_EXPIRES(r4)
@@ -2799,7 +2815,7 @@ END_FTR_SECTION(CPU_FTR_TM | CPU_FTR_P9_TM_HV_ASSIST, 0)
/* Restore guest decrementer */
ld r3, VCPU_DEC_EXPIRES(r4)
ld r5, HSTATE_KVM_VCORE(r13)
- ld r6, VCORE_TB_OFFSET(r5)
+ ld r6, VCORE_TB_OFFSET_APPL(r5)
add r3, r3, r6 /* convert host TB to guest TB value */
mftb r7
subf r3, r7, r3
@@ -3606,12 +3622,9 @@ kvmppc_fix_pmao:
*/
kvmhv_start_timing:
ld r5, HSTATE_KVM_VCORE(r13)
- lbz r6, VCORE_IN_GUEST(r5)
- cmpwi r6, 0
- beq 5f /* if in guest, need to */
- ld r6, VCORE_TB_OFFSET(r5) /* subtract timebase offset */
-5: mftb r5
- subf r5, r6, r5
+ ld r6, VCORE_TB_OFFSET_APPL(r5)
+ mftb r5
+ subf r5, r6, r5 /* subtract current timebase offset */
std r3, VCPU_CUR_ACTIVITY(r4)
std r5, VCPU_ACTIVITY_START(r4)
blr
@@ -3622,15 +3635,12 @@ kvmhv_start_timing:
*/
kvmhv_accumulate_time:
ld r5, HSTATE_KVM_VCORE(r13)
- lbz r8, VCORE_IN_GUEST(r5)
- cmpwi r8, 0
- beq 4f /* if in guest, need to */
- ld r8, VCORE_TB_OFFSET(r5) /* subtract timebase offset */
-4: ld r5, VCPU_CUR_ACTIVITY(r4)
+ ld r8, VCORE_TB_OFFSET_APPL(r5)
+ ld r5, VCPU_CUR_ACTIVITY(r4)
ld r6, VCPU_ACTIVITY_START(r4)
std r3, VCPU_CUR_ACTIVITY(r4)
mftb r7
- subf r7, r8, r7
+ subf r7, r8, r7 /* subtract current timebase offset */
std r7, VCPU_ACTIVITY_START(r4)
cmpdi r5, 0
beqlr
diff --git a/arch/powerpc/kvm/book3s_xive_template.c b/arch/powerpc/kvm/book3s_xive_template.c
index c7a5deadd1cc7..99c3620b40d95 100644
--- a/arch/powerpc/kvm/book3s_xive_template.c
+++ b/arch/powerpc/kvm/book3s_xive_template.c
@@ -11,6 +11,9 @@
#define XGLUE(a,b) a##b
#define GLUE(a,b) XGLUE(a,b)
+/* Dummy interrupt used when taking interrupts out of a queue in H_CPPR */
+#define XICS_DUMMY 1
+
static void GLUE(X_PFX,ack_pending)(struct kvmppc_xive_vcpu *xc)
{
u8 cppr;
@@ -205,6 +208,10 @@ skip_ipi:
goto skip_ipi;
}
+ /* If it's the dummy interrupt, continue searching */
+ if (hirq == XICS_DUMMY)
+ goto skip_ipi;
+
/* If fetching, update queue pointers */
if (scan_type == scan_fetch) {
q->idx = idx;
@@ -385,9 +392,76 @@ static void GLUE(X_PFX,push_pending_to_hw)(struct kvmppc_xive_vcpu *xc)
__x_writeb(prio, __x_tima + TM_SPC_SET_OS_PENDING);
}
+static void GLUE(X_PFX,scan_for_rerouted_irqs)(struct kvmppc_xive *xive,
+ struct kvmppc_xive_vcpu *xc)
+{
+ unsigned int prio;
+
+ /* For each priority that is now masked */
+ for (prio = xc->cppr; prio < KVMPPC_XIVE_Q_COUNT; prio++) {
+ struct xive_q *q = &xc->queues[prio];
+ struct kvmppc_xive_irq_state *state;
+ struct kvmppc_xive_src_block *sb;
+ u32 idx, toggle, entry, irq, hw_num;
+ struct xive_irq_data *xd;
+ __be32 *qpage;
+ u16 src;
+
+ idx = q->idx;
+ toggle = q->toggle;
+ qpage = READ_ONCE(q->qpage);
+ if (!qpage)
+ continue;
+
+ /* For each interrupt in the queue */
+ for (;;) {
+ entry = be32_to_cpup(qpage + idx);
+
+ /* No more ? */
+ if ((entry >> 31) == toggle)
+ break;
+ irq = entry & 0x7fffffff;
+
+ /* Skip dummies and IPIs */
+ if (irq == XICS_DUMMY || irq == XICS_IPI)
+ goto next;
+ sb = kvmppc_xive_find_source(xive, irq, &src);
+ if (!sb)
+ goto next;
+ state = &sb->irq_state[src];
+
+ /* Has it been rerouted ? */
+ if (xc->server_num == state->act_server)
+ goto next;
+
+ /*
+ * Allright, it *has* been re-routed, kill it from
+ * the queue.
+ */
+ qpage[idx] = cpu_to_be32((entry & 0x80000000) | XICS_DUMMY);
+
+ /* Find the HW interrupt */
+ kvmppc_xive_select_irq(state, &hw_num, &xd);
+
+ /* If it's not an LSI, set PQ to 11 the EOI will force a resend */
+ if (!(xd->flags & XIVE_IRQ_FLAG_LSI))
+ GLUE(X_PFX,esb_load)(xd, XIVE_ESB_SET_PQ_11);
+
+ /* EOI the source */
+ GLUE(X_PFX,source_eoi)(hw_num, xd);
+
+ next:
+ idx = (idx + 1) & q->msk;
+ if (idx == 0)
+ toggle ^= 1;
+ }
+ }
+}
+
X_STATIC int GLUE(X_PFX,h_cppr)(struct kvm_vcpu *vcpu, unsigned long cppr)
{
struct kvmppc_xive_vcpu *xc = vcpu->arch.xive_vcpu;
+ struct kvmppc_xive *xive = vcpu->kvm->arch.xive;
u8 old_cppr;
pr_devel("H_CPPR(cppr=%ld)\n", cppr);
@@ -407,14 +481,34 @@ X_STATIC int GLUE(X_PFX,h_cppr)(struct kvm_vcpu *vcpu, unsigned long cppr)
*/
smp_mb();
- /*
- * We are masking less, we need to look for pending things
- * to deliver and set VP pending bits accordingly to trigger
- * a new interrupt otherwise we might miss MFRR changes for
- * which we have optimized out sending an IPI signal.
- */
- if (cppr > old_cppr)
+ if (cppr > old_cppr) {
+ /*
+ * We are masking less, we need to look for pending things
+ * to deliver and set VP pending bits accordingly to trigger
+ * a new interrupt otherwise we might miss MFRR changes for
+ * which we have optimized out sending an IPI signal.
+ */
GLUE(X_PFX,push_pending_to_hw)(xc);
+ } else {
+ /*
+ * We are masking more, we need to check the queue for any
+ * interrupt that has been routed to another CPU, take
+ * it out (replace it with the dummy) and retrigger it.
+ *
+ * This is necessary since those interrupts may otherwise
+ * never be processed, at least not until this CPU restores
+ * its CPPR.
+ *
+ * This is in theory racy vs. HW adding new interrupts to
+ * the queue. In practice this works because the interesting
+ * cases are when the guest has done a set_xive() to move the
+ * interrupt away, which flushes the xive, followed by the
+ * target CPU doing a H_CPPR. So any new interrupt coming into
+ * the queue must still be routed to us and isn't a source
+ * of concern.
+ */
+ GLUE(X_PFX,scan_for_rerouted_irqs)(xive, xc);
+ }
/* Apply new CPPR */
xc->hw_cppr = cppr;
generated by cgit v1.2.3 (git 2.39.1) at 2024-05-01 03:32:16 +0000