* [PATCH v4 0/3] target/arm: Add support for FEAT_TLBIOS and FEAT_TLBIRANGE
@ 2021-03-16 15:49 Rebecca Cran
2021-03-16 15:49 ` [PATCH v4 1/3] target/arm: Add support for FEAT_TLBIRANGE Rebecca Cran
` (3 more replies)
0 siblings, 4 replies; 7+ messages in thread
From: Rebecca Cran @ 2021-03-16 15:49 UTC (permalink / raw)
To: Peter Maydell, Richard Henderson, qemu-arm; +Cc: Rebecca Cran, qemu-devel
ARMv8.4 adds the mandatory FEAT_TLBIOS and FEAT_TLBIRANGE.
They provides TLBI maintenance instructions that extend to the Outer
Shareable domain and that apply to a range of input addresses.
Changes from v3 to v4:
o Adapted code from the existing flush_page_bits_* functions to support
flushing a range of addresses.
o Changed the API from num_pages to length.
o Removed the exception generation from tlbi_aa64_range_get_num_pages.
o The addr is no longer passed into tlbi_aa64_range_get_num_pages.
o Other changes based on feedback.
Rebecca Cran (3):
target/arm: Add support for FEAT_TLBIRANGE
target/arm: Add support for FEAT_TLBIOS
target/arm: set ID_AA64ISAR0.TLB to 2 for max AARCH64 CPU type
accel/tcg/cputlb.c | 135 +++++++
include/exec/exec-all.h | 45 +++
target/arm/cpu.h | 10 +
target/arm/cpu64.c | 1 +
target/arm/helper.c | 393 ++++++++++++++++++++
5 files changed, 584 insertions(+)
--
2.26.2
^ permalink raw reply [flat|nested] 7+ messages in thread
* [PATCH v4 1/3] target/arm: Add support for FEAT_TLBIRANGE
2021-03-16 15:49 [PATCH v4 0/3] target/arm: Add support for FEAT_TLBIOS and FEAT_TLBIRANGE Rebecca Cran
@ 2021-03-16 15:49 ` Rebecca Cran
2021-03-16 18:09 ` Richard Henderson
2021-03-16 15:49 ` [PATCH v4 2/3] target/arm: Add support for FEAT_TLBIOS Rebecca Cran
` (2 subsequent siblings)
3 siblings, 1 reply; 7+ messages in thread
From: Rebecca Cran @ 2021-03-16 15:49 UTC (permalink / raw)
To: Peter Maydell, Richard Henderson, qemu-arm; +Cc: Rebecca Cran, qemu-devel
ARMv8.4 adds the mandatory FEAT_TLBIRANGE. It provides TLBI
maintenance instructions that apply to a range of input addresses.
Signed-off-by: Rebecca Cran <rebecca@nuviainc.com>
---
accel/tcg/cputlb.c | 135 +++++++++
include/exec/exec-all.h | 45 +++
target/arm/cpu.h | 5 +
target/arm/helper.c | 318 ++++++++++++++++++++
4 files changed, 503 insertions(+)
diff --git a/accel/tcg/cputlb.c b/accel/tcg/cputlb.c
index 8a7b779270a4..5ef2a562f4d3 100644
--- a/accel/tcg/cputlb.c
+++ b/accel/tcg/cputlb.c
@@ -943,6 +943,141 @@ void tlb_flush_page_bits_by_mmuidx_all_cpus_synced(CPUState *src_cpu,
}
}
+typedef struct {
+ target_ulong addr;
+ target_ulong length;
+ uint16_t idxmap;
+ uint16_t bits;
+} TLBFlushPageRangeBitsByMMUIdxData;
+
+static void
+tlb_flush_page_range_bits_by_mmuidx_async_0(CPUState *cpu,
+ target_ulong addr,
+ target_ulong length,
+ uint16_t idxmap,
+ unsigned bits)
+{
+ CPUArchState *env = cpu->env_ptr;
+ int mmu_idx;
+ target_ulong page;
+
+ assert_cpu_is_self(cpu);
+
+ tlb_debug("page addr:" TARGET_FMT_lx "/%u len: " TARGET_FMT_lx
+ " mmu_map:0x%x\n",
+ addr, bits, length, idxmap);
+
+ qemu_spin_lock(&env_tlb(env)->c.lock);
+ for (page = addr; page < (addr + length); page += TARGET_PAGE_SIZE) {
+ for (mmu_idx = 0; mmu_idx < NB_MMU_MODES; mmu_idx++) {
+ if ((idxmap >> mmu_idx) & 1) {
+ tlb_flush_page_bits_locked(env, mmu_idx, page, bits);
+ tb_flush_jmp_cache(cpu, page);
+ }
+ }
+ }
+ qemu_spin_unlock(&env_tlb(env)->c.lock);
+}
+
+static void
+tlb_flush_page_range_bits_by_mmuidx_async_1(CPUState *cpu,
+ run_on_cpu_data data)
+{
+ TLBFlushPageRangeBitsByMMUIdxData *d = data.host_ptr;
+
+ tlb_flush_page_range_bits_by_mmuidx_async_0(cpu, d->addr, d->length,
+ d->idxmap, d->bits);
+
+ g_free(d);
+}
+
+void tlb_flush_page_range_bits_by_mmuidx(CPUState *cpu,
+ target_ulong addr,
+ target_ulong length,
+ uint16_t idxmap,
+ unsigned bits)
+{
+ TLBFlushPageRangeBitsByMMUIdxData d;
+ TLBFlushPageRangeBitsByMMUIdxData *p;
+
+ /* This should already be page aligned */
+ addr &= TARGET_PAGE_BITS;
+
+ /* If all bits are significant, this devolves to tlb_flush_page. */
+ if (bits >= TARGET_LONG_BITS) {
+ tlb_flush_page_by_mmuidx(cpu, addr, idxmap);
+ return;
+ }
+ /* If no page bits are significant, this devolves to tlb_flush. */
+ if (bits < TARGET_PAGE_BITS) {
+ tlb_flush_by_mmuidx(cpu, idxmap);
+ return;
+ }
+
+ d.addr = addr & TARGET_PAGE_MASK;
+ d.idxmap = idxmap;
+ d.bits = bits;
+ d.length = length;
+
+ if (qemu_cpu_is_self(cpu)) {
+ tlb_flush_page_range_bits_by_mmuidx_async_0(cpu, addr, length,
+ idxmap, bits);
+ } else {
+ p = g_new(TLBFlushPageRangeBitsByMMUIdxData, 1);
+
+ /* Allocate a structure, freed by the worker. */
+ *p = d;
+ async_run_on_cpu(cpu, tlb_flush_page_range_bits_by_mmuidx_async_1,
+ RUN_ON_CPU_HOST_PTR(p));
+ }
+}
+
+void tlb_flush_page_range_bits_by_mmuidx_all_cpus_synced(CPUState *src_cpu,
+ target_ulong addr,
+ target_ulong length,
+ uint16_t idxmap,
+ unsigned bits)
+{
+ TLBFlushPageRangeBitsByMMUIdxData d;
+ TLBFlushPageRangeBitsByMMUIdxData *p;
+ CPUState *dst_cpu;
+
+ /* This should already be page aligned */
+ addr &= TARGET_PAGE_BITS;
+
+ /* If all bits are significant, this devolves to tlb_flush_page. */
+ if (bits >= TARGET_LONG_BITS) {
+ tlb_flush_page_by_mmuidx_all_cpus_synced(src_cpu, addr, idxmap);
+ return;
+ }
+ /* If no page bits are significant, this devolves to tlb_flush. */
+ if (bits < TARGET_PAGE_BITS) {
+ tlb_flush_by_mmuidx_all_cpus_synced(src_cpu, idxmap);
+ return;
+ }
+
+ d.addr = addr;
+ d.idxmap = idxmap;
+ d.bits = bits;
+ d.length = length;
+
+ /* Allocate a separate data block for each destination cpu. */
+ CPU_FOREACH(dst_cpu) {
+ if (dst_cpu != src_cpu) {
+ p = g_new(TLBFlushPageRangeBitsByMMUIdxData, 1);
+ *p = d;
+ async_run_on_cpu(dst_cpu,
+ tlb_flush_page_range_bits_by_mmuidx_async_1,
+ RUN_ON_CPU_HOST_PTR(p));
+ }
+ }
+
+ p = g_new(TLBFlushPageRangeBitsByMMUIdxData, 1);
+ *p = d;
+ async_safe_run_on_cpu(src_cpu, tlb_flush_page_range_bits_by_mmuidx_async_1,
+ RUN_ON_CPU_HOST_PTR(p));
+}
+
/* update the TLBs so that writes to code in the virtual page 'addr'
can be detected */
void tlb_protect_code(ram_addr_t ram_addr)
diff --git a/include/exec/exec-all.h b/include/exec/exec-all.h
index 6b036cae8f65..833e8d654a1c 100644
--- a/include/exec/exec-all.h
+++ b/include/exec/exec-all.h
@@ -212,6 +212,37 @@ void tlb_flush_page_by_mmuidx_all_cpus(CPUState *cpu, target_ulong addr,
*/
void tlb_flush_page_by_mmuidx_all_cpus_synced(CPUState *cpu, target_ulong addr,
uint16_t idxmap);
+/**
+ * tlb_flush_page_range_by_mmuidx:
+ * @cpu: CPU whose TLB should be flushed
+ * @addr: virtual address of start of page range to be flushed
+ * @length: the number of bytes to be flushed
+ * @idxmap: bitmap of MMU indexes to flush
+ *
+ * Flush a range of pages from the TLB of the specified CPU, for the specified
+ * MMU indexes.
+ */
+void tlb_flush_page_range_bits_by_mmuidx(CPUState *cpu, target_ulong addr,
+ target_ulong length, uint16_t idxmap,
+ unsigned bits);
+/**
+ * tlb_flush_page_range_by_mmuidx_all_cpus_synced:
+ * @cpu: Originating CPU of the flush
+ * @addr: virtual address of start of page range to be flushed
+ * @length: the number of bytes to be flushed
+ * @idxmap: bitmap of MMU indexes to flush
+ *
+ * Flush a range of pages from the TLB of all CPUs, for the specified MMU
+ * indexes like tlb_flush_page_by_mmuidx_all_cpus except the source
+ * vCPUs work is scheduled as safe work meaning all flushes will be
+ * complete once the source vCPUs safe work is complete. This will
+ * depend on when the guests translation ends the TB.
+ */
+void tlb_flush_page_range_bits_by_mmuidx_all_cpus_synced(CPUState *cpu,
+ target_ulong addr,
+ target_ulong length,
+ uint16_t idxmap,
+ unsigned bits);
/**
* tlb_flush_by_mmuidx:
* @cpu: CPU whose TLB should be flushed
@@ -313,6 +344,20 @@ static inline void tlb_flush_page_all_cpus_synced(CPUState *src,
target_ulong addr)
{
}
+static inline void tlb_flush_page_range_bits_by_mmuidx(CPUState *cpu,
+ target_ulong addr,
+ target_ulong length,
+ uint16_t idxmap,
+ unsigned bits)
+{
+}
+static inline void tlb_flush_page_range_bits_by_mmuidx_all_cpus_synced(CPUState *src_cpu,
+ target_ulong addr,
+ target_ulong length,
+ uint16_t idxmap,
+ unsigned bits)
+{
+}
static inline void tlb_flush(CPUState *cpu)
{
}
diff --git a/target/arm/cpu.h b/target/arm/cpu.h
index 193a49ec7fac..32b78a4ef587 100644
--- a/target/arm/cpu.h
+++ b/target/arm/cpu.h
@@ -4038,6 +4038,11 @@ static inline bool isar_feature_aa64_pauth_arch(const ARMISARegisters *id)
return FIELD_EX64(id->id_aa64isar1, ID_AA64ISAR1, APA) != 0;
}
+static inline bool isar_feature_aa64_tlbirange(const ARMISARegisters *id)
+{
+ return FIELD_EX64(id->id_aa64isar0, ID_AA64ISAR0, TLB) == 2;
+}
+
static inline bool isar_feature_aa64_sb(const ARMISARegisters *id)
{
return FIELD_EX64(id->id_aa64isar1, ID_AA64ISAR1, SB) != 0;
diff --git a/target/arm/helper.c b/target/arm/helper.c
index d9220be7c5a0..7133454330d3 100644
--- a/target/arm/helper.c
+++ b/target/arm/helper.c
@@ -4759,6 +4759,241 @@ static void tlbi_aa64_vae3is_write(CPUARMState *env, const ARMCPRegInfo *ri,
ARMMMUIdxBit_SE3, bits);
}
+#ifdef TARGET_AARCH64
+static uint64_t tlbi_aa64_range_get_length(CPUARMState *env,
+ uint64_t value)
+{
+ unsigned int page_size;
+ unsigned int page_size_granule;
+ uint64_t num;
+ uint64_t scale;
+ uint64_t exponent;
+ uint64_t length;
+
+ num = extract64(value, 39, 4);
+ scale = extract64(value, 44, 2);
+ page_size_granule = extract64(value, 46, 2);
+
+ switch (page_size_granule) {
+ case 1:
+ page_size = 4096;
+ break;
+ case 2:
+ page_size = 16384;
+ break;
+ case 3:
+ page_size = 65536;
+ break;
+ default:
+ qemu_log_mask(LOG_GUEST_ERROR, "Invalid page size granule %d\n",
+ page_size_granule);
+ return 0;
+ }
+
+ exponent = (5 * scale) + 1;
+ length = ((num + 1) << exponent) * page_size;
+
+ return length;
+}
+
+static void tlbi_aa64_rvae1_write(CPUARMState *env, const ARMCPRegInfo *ri,
+ uint64_t value)
+{
+ /*
+ * Invalidate by VA range, EL1&0.
+ * Currently handles all of RVAE1, RVAAE1, RVAALE1 and RVALE1,
+ * since we don't support flush-for-specific-ASID-only or
+ * flush-last-level-only.
+ */
+ ARMMMUIdx mask;
+ int bits;
+ uint64_t pageaddr;
+ uint64_t length;
+
+ CPUState *cs = env_cpu(env);
+ mask = vae1_tlbmask(env);
+ if (regime_has_2_ranges(mask)) {
+ pageaddr = sextract64(value, 0, 37) << TARGET_PAGE_BITS;
+ } else {
+ pageaddr = extract64(value, 0, 37) << TARGET_PAGE_BITS;
+ }
+ length = tlbi_aa64_range_get_length(env, value);
+ bits = tlbbits_for_regime(env, mask, pageaddr);
+
+ if (tlb_force_broadcast(env)) {
+ tlb_flush_page_range_bits_by_mmuidx_all_cpus_synced(cs, pageaddr,
+ length, mask,
+ bits);
+ } else {
+ tlb_flush_page_range_bits_by_mmuidx(cs, pageaddr, length, mask,
+ bits);
+ }
+}
+
+static void tlbi_aa64_rvae1is_write(CPUARMState *env, const ARMCPRegInfo *ri,
+ uint64_t value)
+{
+ /*
+ * Invalidate by VA range, Inner/Outer Shareable EL1&0.
+ * Currently handles all of RVAE1IS, RVAE1OS, RVAAE1IS, RVAAE1OS,
+ * RVAALE1IS, RVAALE1OS, RVALE1IS and RVALE1OS, since we don't support
+ * flush-for-specific-ASID-only, flush-last-level-only or inner/outer
+ * shareable specific flushes.
+ */
+ ARMMMUIdx mask;
+ int bits;
+ uint64_t pageaddr;
+ uint64_t length;
+
+ CPUState *cs = env_cpu(env);
+ mask = vae1_tlbmask(env);
+ if (regime_has_2_ranges(mask)) {
+ pageaddr = sextract64(value, 0, 37) << TARGET_PAGE_BITS;
+ } else {
+ pageaddr = extract64(value, 0, 37) << TARGET_PAGE_BITS;
+ }
+ length = tlbi_aa64_range_get_length(env, value);
+ bits = tlbbits_for_regime(env, mask, pageaddr);
+
+ tlb_flush_page_range_bits_by_mmuidx_all_cpus_synced(cs, pageaddr,
+ length, mask,
+ bits);
+}
+
+static void tlbi_aa64_rvae2_write(CPUARMState *env, const ARMCPRegInfo *ri,
+ uint64_t value)
+{
+ /*
+ * Invalidate by VA range, EL2.
+ * Currently handles all of RVAE2, RVAAE2, RVAALE2 and RVALE2,
+ * since we don't support flush-for-specific-ASID-only or
+ * flush-last-level-only.
+ */
+ ARMMMUIdx mask;
+ bool secure;
+ int bits;
+ uint64_t pageaddr;
+ uint64_t length;
+
+ CPUState *cs = env_cpu(env);
+ secure = arm_is_secure_below_el3(env);
+ if (regime_has_2_ranges(secure ? ARMMMUIdxBit_SE2 : ARMMMUIdxBit_E2)) {
+ pageaddr = sextract64(value, 0, 37) << TARGET_PAGE_BITS;
+ } else {
+ pageaddr = extract64(value, 0, 37) << TARGET_PAGE_BITS;
+ }
+ length = tlbi_aa64_range_get_length(env, value);
+ mask = secure ? ARMMMUIdxBit_SE2 : ARMMMUIdxBit_E2;
+ bits = tlbbits_for_regime(env, secure ? ARMMMUIdxBit_SE2 : ARMMMUIdxBit_E2,
+ pageaddr);
+
+ if (tlb_force_broadcast(env)) {
+ tlb_flush_page_range_bits_by_mmuidx_all_cpus_synced(cs, pageaddr,
+ length, mask,
+ bits);
+ } else {
+ tlb_flush_page_range_bits_by_mmuidx(cs, pageaddr, length, mask,
+ bits);
+ }
+}
+
+static void tlbi_aa64_rvae2is_write(CPUARMState *env, const ARMCPRegInfo *ri,
+ uint64_t value)
+{
+ /*
+ * Invalidate by VA range, Inner/Outer Shareable, EL2.
+ * Currently handles all of RVAE2IS, RVAE2OS, RVAAE2IS, RVAAE2OS,
+ * RVAALE2IS, RVAALE2OS, RVALE2IS and RVALE2OS, since we don't support
+ * flush-for-specific-ASID-only, flush-last-level-only or inner/outer
+ * shareable specific flushes.
+ */
+ ARMMMUIdx mask;
+ bool secure;
+ int bits;
+ uint64_t pageaddr;
+ uint64_t length;
+
+ CPUState *cs = env_cpu(env);
+ secure = arm_is_secure_below_el3(env);
+ if (regime_has_2_ranges(secure ? ARMMMUIdxBit_SE2 : ARMMMUIdxBit_E2)) {
+ pageaddr = sextract64(value, 0, 37) << TARGET_PAGE_BITS;
+ } else {
+ pageaddr = extract64(value, 0, 37) << TARGET_PAGE_BITS;
+ }
+ length = tlbi_aa64_range_get_length(env, value);
+ mask = secure ? ARMMMUIdxBit_SE2 : ARMMMUIdxBit_E2;
+ bits = tlbbits_for_regime(env, secure ? ARMMMUIdxBit_SE2 : ARMMMUIdxBit_E2,
+ pageaddr);
+
+ tlb_flush_page_range_bits_by_mmuidx_all_cpus_synced(cs, pageaddr,
+ length, mask,
+ bits);
+}
+
+static void tlbi_aa64_rvae3_write(CPUARMState *env, const ARMCPRegInfo *ri,
+ uint64_t value)
+{
+ /*
+ * Invalidate by VA range, EL3.
+ * Currently handles all of RVAE3, RVAAE3, RVAALE3 and RVALE3,
+ * since we don't support flush-for-specific-ASID-only or
+ * flush-last-level-only.
+ */
+ int bits;
+ uint64_t pageaddr;
+ uint64_t length;
+
+ CPUState *cs = env_cpu(env);
+ if (regime_has_2_ranges(ARMMMUIdxBit_SE3)) {
+ pageaddr = sextract64(value, 0, 37) << TARGET_PAGE_BITS;
+ } else {
+ pageaddr = extract64(value, 0, 37) << TARGET_PAGE_BITS;
+ }
+ length = tlbi_aa64_range_get_length(env, value);
+ bits = tlbbits_for_regime(env, ARMMMUIdx_SE3, pageaddr);
+
+ if (tlb_force_broadcast(env)) {
+ tlb_flush_page_range_bits_by_mmuidx_all_cpus_synced(cs, pageaddr,
+ length,
+ ARMMMUIdxBit_SE3,
+ bits);
+ } else {
+ tlb_flush_page_range_bits_by_mmuidx(cs, pageaddr, length,
+ ARMMMUIdxBit_SE3,
+ bits);
+ }
+}
+
+static void tlbi_aa64_rvae3is_write(CPUARMState *env, const ARMCPRegInfo *ri,
+ uint64_t value)
+{
+ /*
+ * Invalidate by VA range, EL3, Inner/Outer Shareable.
+ * Currently handles all of RVAE3IS, RVAE3OS, RVAAE3IS, RVAAE3OS,
+ * RVAALE3IS, RVAALE3OS, RVALE3IS, and RVALE3OS, since we don't support
+ * flush-for-specific-ASID-only, flush-last-level-only or inner/outer
+ * specific flushes.
+ */
+ int bits;
+ uint64_t pageaddr;
+ uint64_t length;
+
+ CPUState *cs = env_cpu(env);
+ if (regime_has_2_ranges(ARMMMUIdxBit_SE3)) {
+ pageaddr = sextract64(value, 0, 37) << TARGET_PAGE_BITS;
+ } else {
+ pageaddr = extract64(value, 0, 37) << TARGET_PAGE_BITS;
+ }
+ length = tlbi_aa64_range_get_length(env, value);
+ bits = tlbbits_for_regime(env, ARMMMUIdx_SE3, pageaddr);
+
+ tlb_flush_page_range_bits_by_mmuidx_all_cpus_synced(cs, pageaddr,
+ length,
+ ARMMMUIdxBit_SE3,
+ bits);
+}
+#endif
+
static CPAccessResult aa64_zva_access(CPUARMState *env, const ARMCPRegInfo *ri,
bool isread)
{
@@ -6920,6 +7155,86 @@ static const ARMCPRegInfo pauth_reginfo[] = {
REGINFO_SENTINEL
};
+static const ARMCPRegInfo tlbirange_reginfo[] = {
+ { .name = "TLBI_RVAE1IS", .state = ARM_CP_STATE_AA64,
+ .opc0 = 1, .opc1 = 0, .crn = 8, .crm = 2, .opc2 = 1,
+ .access = PL1_W, .type = ARM_CP_NO_RAW,
+ .writefn = tlbi_aa64_rvae1is_write },
+ { .name = "TLBI_RVAAE1IS", .state = ARM_CP_STATE_AA64,
+ .opc0 = 1, .opc1 = 0, .crn = 8, .crm = 2, .opc2 = 3,
+ .access = PL1_W, .type = ARM_CP_NO_RAW,
+ .writefn = tlbi_aa64_rvae1is_write },
+ { .name = "TLBI_RVALE1IS", .state = ARM_CP_STATE_AA64,
+ .opc0 = 1, .opc1 = 0, .crn = 8, .crm = 2, .opc2 = 5,
+ .access = PL1_W, .type = ARM_CP_NO_RAW,
+ .writefn = tlbi_aa64_rvae1is_write },
+ { .name = "TLBI_RVAALE1IS", .state = ARM_CP_STATE_AA64,
+ .opc0 = 1, .opc1 = 0, .crn = 8, .crm = 2, .opc2 = 7,
+ .access = PL1_W, .type = ARM_CP_NO_RAW,
+ .writefn = tlbi_aa64_rvae1is_write },
+ { .name = "TLBI_RVAE1", .state = ARM_CP_STATE_AA64,
+ .opc0 = 1, .opc1 = 0, .crn = 8, .crm = 6, .opc2 = 1,
+ .access = PL1_W, .type = ARM_CP_NO_RAW,
+ .writefn = tlbi_aa64_rvae1_write },
+ { .name = "TLBI_RVAAE1", .state = ARM_CP_STATE_AA64,
+ .opc0 = 1, .opc1 = 0, .crn = 8, .crm = 6, .opc2 = 3,
+ .access = PL1_W, .type = ARM_CP_NO_RAW,
+ .writefn = tlbi_aa64_rvae1_write },
+ { .name = "TLBI_RVALE1", .state = ARM_CP_STATE_AA64,
+ .opc0 = 1, .opc1 = 0, .crn = 8, .crm = 6, .opc2 = 5,
+ .access = PL1_W, .type = ARM_CP_NO_RAW,
+ .writefn = tlbi_aa64_rvae1_write },
+ { .name = "TLBI_RVAALE1", .state = ARM_CP_STATE_AA64,
+ .opc0 = 1, .opc1 = 0, .crn = 8, .crm = 6, .opc2 = 7,
+ .access = PL1_W, .type = ARM_CP_NO_RAW,
+ .writefn = tlbi_aa64_rvae1_write },
+ { .name = "TLBI_RIPAS2E1IS", .state = ARM_CP_STATE_AA64,
+ .opc0 = 1, .opc1 = 4, .crn = 8, .crm = 0, .opc2 = 2,
+ .access = PL2_W, .type = ARM_CP_NOP },
+ { .name = "TLBI_RIPAS2LE1IS", .state = ARM_CP_STATE_AA64,
+ .opc0 = 1, .opc1 = 4, .crn = 8, .crm = 0, .opc2 = 6,
+ .access = PL2_W, .type = ARM_CP_NOP },
+ { .name = "TLBI_RVAE2IS", .state = ARM_CP_STATE_AA64,
+ .opc0 = 1, .opc1 = 4, .crn = 8, .crm = 2, .opc2 = 1,
+ .access = PL2_W, .type = ARM_CP_NO_RAW,
+ .writefn = tlbi_aa64_rvae2is_write },
+ { .name = "TLBI_RVALE2IS", .state = ARM_CP_STATE_AA64,
+ .opc0 = 1, .opc1 = 4, .crn = 8, .crm = 2, .opc2 = 5,
+ .access = PL2_W, .type = ARM_CP_NO_RAW,
+ .writefn = tlbi_aa64_rvae2is_write },
+ { .name = "TLBI_RIPAS2E1", .state = ARM_CP_STATE_AA64,
+ .opc0 = 1, .opc1 = 4, .crn = 8, .crm = 4, .opc2 = 2,
+ .access = PL2_W, .type = ARM_CP_NOP },
+ { .name = "TLBI_RIPAS2LE1", .state = ARM_CP_STATE_AA64,
+ .opc0 = 1, .opc1 = 4, .crn = 8, .crm = 4, .opc2 = 6,
+ .access = PL2_W, .type = ARM_CP_NOP },
+ { .name = "TLBI_RVAE2", .state = ARM_CP_STATE_AA64,
+ .opc0 = 1, .opc1 = 4, .crn = 8, .crm = 6, .opc2 = 1,
+ .access = PL2_W, .type = ARM_CP_NO_RAW,
+ .writefn = tlbi_aa64_rvae2_write },
+ { .name = "TLBI_RVALE2", .state = ARM_CP_STATE_AA64,
+ .opc0 = 1, .opc1 = 4, .crn = 8, .crm = 6, .opc2 = 5,
+ .access = PL2_W, .type = ARM_CP_NO_RAW,
+ .writefn = tlbi_aa64_rvae2_write },
+ { .name = "TLBI_RVAE3IS", .state = ARM_CP_STATE_AA64,
+ .opc0 = 1, .opc1 = 6, .crn = 8, .crm = 2, .opc2 = 1,
+ .access = PL3_W, .type = ARM_CP_NO_RAW,
+ .writefn = tlbi_aa64_rvae3is_write },
+ { .name = "TLBI_RVALE3IS", .state = ARM_CP_STATE_AA64,
+ .opc0 = 1, .opc1 = 6, .crn = 8, .crm = 2, .opc2 = 5,
+ .access = PL3_W, .type = ARM_CP_NO_RAW,
+ .writefn = tlbi_aa64_rvae3is_write },
+ { .name = "TLBI_RVAE3", .state = ARM_CP_STATE_AA64,
+ .opc0 = 1, .opc1 = 6, .crn = 8, .crm = 6, .opc2 = 1,
+ .access = PL3_W, .type = ARM_CP_NO_RAW,
+ .writefn = tlbi_aa64_rvae3_write },
+ { .name = "TLBI_RVALE3", .state = ARM_CP_STATE_AA64,
+ .opc0 = 1, .opc1 = 6, .crn = 8, .crm = 6, .opc2 = 5,
+ .access = PL3_W, .type = ARM_CP_NO_RAW,
+ .writefn = tlbi_aa64_rvae3_write },
+ REGINFO_SENTINEL
+};
+
static uint64_t rndr_readfn(CPUARMState *env, const ARMCPRegInfo *ri)
{
Error *err = NULL;
@@ -8289,6 +8604,9 @@ void register_cp_regs_for_features(ARMCPU *cpu)
if (cpu_isar_feature(aa64_rndr, cpu)) {
define_arm_cp_regs(cpu, rndr_reginfo);
}
+ if (cpu_isar_feature(aa64_tlbirange, cpu)) {
+ define_arm_cp_regs(cpu, tlbirange_reginfo);
+ }
#ifndef CONFIG_USER_ONLY
/* Data Cache clean instructions up to PoP */
if (cpu_isar_feature(aa64_dcpop, cpu)) {
--
2.26.2
^ permalink raw reply related [flat|nested] 7+ messages in thread
* [PATCH v4 2/3] target/arm: Add support for FEAT_TLBIOS
2021-03-16 15:49 [PATCH v4 0/3] target/arm: Add support for FEAT_TLBIOS and FEAT_TLBIRANGE Rebecca Cran
2021-03-16 15:49 ` [PATCH v4 1/3] target/arm: Add support for FEAT_TLBIRANGE Rebecca Cran
@ 2021-03-16 15:49 ` Rebecca Cran
2021-03-16 15:49 ` [PATCH v4 3/3] target/arm: set ID_AA64ISAR0.TLB to 2 for max AARCH64 CPU type Rebecca Cran
2021-03-16 17:03 ` [PATCH v4 0/3] target/arm: Add support for FEAT_TLBIOS and FEAT_TLBIRANGE no-reply
3 siblings, 0 replies; 7+ messages in thread
From: Rebecca Cran @ 2021-03-16 15:49 UTC (permalink / raw)
To: Peter Maydell, Richard Henderson, qemu-arm; +Cc: Rebecca Cran, qemu-devel
ARMv8.4 adds the mandatory FEAT_TLBIOS. It provides TLBI
maintenance instructions that extend to the Outer Shareable domain.
Signed-off-by: Rebecca Cran <rebecca@nuviainc.com>
---
target/arm/cpu.h | 5 ++
target/arm/helper.c | 75 ++++++++++++++++++++
2 files changed, 80 insertions(+)
diff --git a/target/arm/cpu.h b/target/arm/cpu.h
index 32b78a4ef587..272fde83ca4e 100644
--- a/target/arm/cpu.h
+++ b/target/arm/cpu.h
@@ -4043,6 +4043,11 @@ static inline bool isar_feature_aa64_tlbirange(const ARMISARegisters *id)
return FIELD_EX64(id->id_aa64isar0, ID_AA64ISAR0, TLB) == 2;
}
+static inline bool isar_feature_aa64_tlbios(const ARMISARegisters *id)
+{
+ return FIELD_EX64(id->id_aa64isar0, ID_AA64ISAR0, TLB) != 0;
+}
+
static inline bool isar_feature_aa64_sb(const ARMISARegisters *id)
{
return FIELD_EX64(id->id_aa64isar1, ID_AA64ISAR1, SB) != 0;
diff --git a/target/arm/helper.c b/target/arm/helper.c
index 7133454330d3..afcbb5e30c18 100644
--- a/target/arm/helper.c
+++ b/target/arm/helper.c
@@ -7235,6 +7235,78 @@ static const ARMCPRegInfo tlbirange_reginfo[] = {
REGINFO_SENTINEL
};
+static const ARMCPRegInfo tlbios_reginfo[] = {
+ { .name = "TLBI_VMALLE1OS", .state = ARM_CP_STATE_AA64,
+ .opc0 = 1, .opc1 = 0, .crn = 8, .crm = 1, .opc2 = 0,
+ .access = PL1_W, .type = ARM_CP_NO_RAW,
+ .writefn = tlbi_aa64_vmalle1is_write },
+ { .name = "TLBI_ASIDE1OS", .state = ARM_CP_STATE_AA64,
+ .opc0 = 1, .opc1 = 0, .crn = 8, .crm = 1, .opc2 = 2,
+ .access = PL1_W, .type = ARM_CP_NO_RAW,
+ .writefn = tlbi_aa64_vmalle1is_write },
+ { .name = "TLBI_RVAE1OS", .state = ARM_CP_STATE_AA64,
+ .opc0 = 1, .opc1 = 0, .crn = 8, .crm = 5, .opc2 = 1,
+ .access = PL1_W, .type = ARM_CP_NO_RAW,
+ .writefn = tlbi_aa64_rvae1is_write },
+ { .name = "TLBI_RVAAE1OS", .state = ARM_CP_STATE_AA64,
+ .opc0 = 1, .opc1 = 0, .crn = 8, .crm = 5, .opc2 = 3,
+ .access = PL1_W, .type = ARM_CP_NO_RAW,
+ .writefn = tlbi_aa64_rvae1is_write },
+ { .name = "TLBI_RVALE1OS", .state = ARM_CP_STATE_AA64,
+ .opc0 = 1, .opc1 = 0, .crn = 8, .crm = 5, .opc2 = 5,
+ .access = PL1_W, .type = ARM_CP_NO_RAW,
+ .writefn = tlbi_aa64_rvae1is_write },
+ { .name = "TLBI_RVAALE1OS", .state = ARM_CP_STATE_AA64,
+ .opc0 = 1, .opc1 = 0, .crn = 8, .crm = 5, .opc2 = 7,
+ .access = PL1_W, .type = ARM_CP_NO_RAW,
+ .writefn = tlbi_aa64_rvae1is_write },
+ { .name = "TLBI_ALLE2OS", .state = ARM_CP_STATE_AA64,
+ .opc0 = 1, .opc1 = 4, .crn = 8, .crm = 1, .opc2 = 0,
+ .access = PL2_W, .type = ARM_CP_NO_RAW,
+ .writefn = tlbi_aa64_alle2is_write },
+ { .name = "TLBI_ALLE1OS", .state = ARM_CP_STATE_AA64,
+ .opc0 = 1, .opc1 = 4, .crn = 8, .crm = 1, .opc2 = 4,
+ .access = PL2_W, .type = ARM_CP_NO_RAW,
+ .writefn = tlbi_aa64_alle1is_write },
+ { .name = "TLBI_VMALLS12E1OS", .state = ARM_CP_STATE_AA64,
+ .opc0 = 1, .opc1 = 4, .crn = 8, .crm = 1, .opc2 = 6,
+ .access = PL2_W, .type = ARM_CP_NO_RAW,
+ .writefn = tlbi_aa64_alle1is_write },
+ { .name = "TLBI_IPAS2E1OS", .state = ARM_CP_STATE_AA64,
+ .opc0 = 1, .opc1 = 4, .crn = 8, .crm = 4, .opc2 = 0,
+ .access = PL2_W, .type = ARM_CP_NOP },
+ { .name = "TLBI_RIPAS2E1OS", .state = ARM_CP_STATE_AA64,
+ .opc0 = 1, .opc1 = 4, .crn = 8, .crm = 4, .opc2 = 3,
+ .access = PL2_W, .type = ARM_CP_NOP },
+ { .name = "TLBI_IPAS2LE1OS", .state = ARM_CP_STATE_AA64,
+ .opc0 = 1, .opc1 = 4, .crn = 8, .crm = 4, .opc2 = 4,
+ .access = PL2_W, .type = ARM_CP_NOP },
+ { .name = "TLBI_RIPAS2LE1OS", .state = ARM_CP_STATE_AA64,
+ .opc0 = 1, .opc1 = 4, .crn = 8, .crm = 4, .opc2 = 7,
+ .access = PL2_W, .type = ARM_CP_NOP },
+ { .name = "TLBI_RVAE2OS", .state = ARM_CP_STATE_AA64,
+ .opc0 = 1, .opc1 = 4, .crn = 8, .crm = 5, .opc2 = 1,
+ .access = PL2_W, .type = ARM_CP_NO_RAW,
+ .writefn = tlbi_aa64_rvae2is_write },
+ { .name = "TLBI_RVALE2OS", .state = ARM_CP_STATE_AA64,
+ .opc0 = 1, .opc1 = 4, .crn = 8, .crm = 5, .opc2 = 5,
+ .access = PL2_W, .type = ARM_CP_NO_RAW,
+ .writefn = tlbi_aa64_rvae2is_write },
+ { .name = "TLBI_ALLE3OS", .state = ARM_CP_STATE_AA64,
+ .opc0 = 1, .opc1 = 6, .crn = 8, .crm = 1, .opc2 = 0,
+ .access = PL3_W, .type = ARM_CP_NO_RAW,
+ .writefn = tlbi_aa64_alle3is_write },
+ { .name = "TLBI_RVAE3OS", .state = ARM_CP_STATE_AA64,
+ .opc0 = 1, .opc1 = 6, .crn = 8, .crm = 5, .opc2 = 1,
+ .access = PL3_W, .type = ARM_CP_NO_RAW,
+ .writefn = tlbi_aa64_rvae3is_write },
+ { .name = "TLBI_RVALE3OS", .state = ARM_CP_STATE_AA64,
+ .opc0 = 1, .opc1 = 6, .crn = 8, .crm = 5, .opc2 = 5,
+ .access = PL3_W, .type = ARM_CP_NO_RAW,
+ .writefn = tlbi_aa64_rvae3is_write },
+ REGINFO_SENTINEL
+};
+
static uint64_t rndr_readfn(CPUARMState *env, const ARMCPRegInfo *ri)
{
Error *err = NULL;
@@ -8607,6 +8679,9 @@ void register_cp_regs_for_features(ARMCPU *cpu)
if (cpu_isar_feature(aa64_tlbirange, cpu)) {
define_arm_cp_regs(cpu, tlbirange_reginfo);
}
+ if (cpu_isar_feature(aa64_tlbios, cpu)) {
+ define_arm_cp_regs(cpu, tlbios_reginfo);
+ }
#ifndef CONFIG_USER_ONLY
/* Data Cache clean instructions up to PoP */
if (cpu_isar_feature(aa64_dcpop, cpu)) {
--
2.26.2
^ permalink raw reply related [flat|nested] 7+ messages in thread
* [PATCH v4 3/3] target/arm: set ID_AA64ISAR0.TLB to 2 for max AARCH64 CPU type
2021-03-16 15:49 [PATCH v4 0/3] target/arm: Add support for FEAT_TLBIOS and FEAT_TLBIRANGE Rebecca Cran
2021-03-16 15:49 ` [PATCH v4 1/3] target/arm: Add support for FEAT_TLBIRANGE Rebecca Cran
2021-03-16 15:49 ` [PATCH v4 2/3] target/arm: Add support for FEAT_TLBIOS Rebecca Cran
@ 2021-03-16 15:49 ` Rebecca Cran
2021-03-16 17:03 ` [PATCH v4 0/3] target/arm: Add support for FEAT_TLBIOS and FEAT_TLBIRANGE no-reply
3 siblings, 0 replies; 7+ messages in thread
From: Rebecca Cran @ 2021-03-16 15:49 UTC (permalink / raw)
To: Peter Maydell, Richard Henderson, qemu-arm; +Cc: Rebecca Cran, qemu-devel
Indicate support for FEAT_TLBIOS and FEAT_TLBIRANGE by setting
ID_AA64ISAR0.TLB to 2 for the max AARCH64 CPU type.
Signed-off-by: Rebecca Cran <rebecca@nuviainc.com>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
---
target/arm/cpu64.c | 1 +
1 file changed, 1 insertion(+)
diff --git a/target/arm/cpu64.c b/target/arm/cpu64.c
index f0a9e968c9c1..f42803ecaf1d 100644
--- a/target/arm/cpu64.c
+++ b/target/arm/cpu64.c
@@ -651,6 +651,7 @@ static void aarch64_max_initfn(Object *obj)
t = FIELD_DP64(t, ID_AA64ISAR0, DP, 1);
t = FIELD_DP64(t, ID_AA64ISAR0, FHM, 1);
t = FIELD_DP64(t, ID_AA64ISAR0, TS, 2); /* v8.5-CondM */
+ t = FIELD_DP64(t, ID_AA64ISAR0, TLB, 2); /* FEAT_TLBIRANGE */
t = FIELD_DP64(t, ID_AA64ISAR0, RNDR, 1);
cpu->isar.id_aa64isar0 = t;
--
2.26.2
^ permalink raw reply related [flat|nested] 7+ messages in thread
* Re: [PATCH v4 0/3] target/arm: Add support for FEAT_TLBIOS and FEAT_TLBIRANGE
2021-03-16 15:49 [PATCH v4 0/3] target/arm: Add support for FEAT_TLBIOS and FEAT_TLBIRANGE Rebecca Cran
` (2 preceding siblings ...)
2021-03-16 15:49 ` [PATCH v4 3/3] target/arm: set ID_AA64ISAR0.TLB to 2 for max AARCH64 CPU type Rebecca Cran
@ 2021-03-16 17:03 ` no-reply
3 siblings, 0 replies; 7+ messages in thread
From: no-reply @ 2021-03-16 17:03 UTC (permalink / raw)
To: rebecca; +Cc: peter.maydell, qemu-arm, richard.henderson, rebecca, qemu-devel
Patchew URL: https://patchew.org/QEMU/20210316154910.25804-1-rebecca@nuviainc.com/
Hi,
This series seems to have some coding style problems. See output below for
more information:
Type: series
Message-id: 20210316154910.25804-1-rebecca@nuviainc.com
Subject: [PATCH v4 0/3] target/arm: Add support for FEAT_TLBIOS and FEAT_TLBIRANGE
=== TEST SCRIPT BEGIN ===
#!/bin/bash
git rev-parse base > /dev/null || exit 0
git config --local diff.renamelimit 0
git config --local diff.renames True
git config --local diff.algorithm histogram
./scripts/checkpatch.pl --mailback base..
=== TEST SCRIPT END ===
Updating 3c8cf5a9c21ff8782164d1def7f44bd888713384
From https://github.com/patchew-project/qemu
- [tag update] patchew/20210212150256.885-1-zhiwei_liu@c-sky.com -> patchew/20210212150256.885-1-zhiwei_liu@c-sky.com
- [tag update] patchew/20210316120420.19658-1-bmeng.cn@gmail.com -> patchew/20210316120420.19658-1-bmeng.cn@gmail.com
- [tag update] patchew/20210316131353.4533-1-peter.maydell@linaro.org -> patchew/20210316131353.4533-1-peter.maydell@linaro.org
- [tag update] patchew/20210316134456.3243102-1-marcandre.lureau@redhat.com -> patchew/20210316134456.3243102-1-marcandre.lureau@redhat.com
* [new tag] patchew/20210316154910.25804-1-rebecca@nuviainc.com -> patchew/20210316154910.25804-1-rebecca@nuviainc.com
Switched to a new branch 'test'
e9f3e00 target/arm: set ID_AA64ISAR0.TLB to 2 for max AARCH64 CPU type
8bcac4c target/arm: Add support for FEAT_TLBIOS
f171966 target/arm: Add support for FEAT_TLBIRANGE
=== OUTPUT BEGIN ===
1/3 Checking commit f17196657ee3 (target/arm: Add support for FEAT_TLBIRANGE)
WARNING: line over 80 characters
#218: FILE: include/exec/exec-all.h:354:
+static inline void tlb_flush_page_range_bits_by_mmuidx_all_cpus_synced(CPUState *src_cpu,
WARNING: line over 80 characters
#219: FILE: include/exec/exec-all.h:355:
+ target_ulong addr,
ERROR: line over 90 characters
#220: FILE: include/exec/exec-all.h:356:
+ target_ulong length,
WARNING: line over 80 characters
#221: FILE: include/exec/exec-all.h:357:
+ uint16_t idxmap,
WARNING: line over 80 characters
#222: FILE: include/exec/exec-all.h:358:
+ unsigned bits)
total: 1 errors, 4 warnings, 545 lines checked
Patch 1/3 has style problems, please review. If any of these errors
are false positives report them to the maintainer, see
CHECKPATCH in MAINTAINERS.
2/3 Checking commit 8bcac4c04c8e (target/arm: Add support for FEAT_TLBIOS)
3/3 Checking commit e9f3e002d73d (target/arm: set ID_AA64ISAR0.TLB to 2 for max AARCH64 CPU type)
=== OUTPUT END ===
Test command exited with code: 1
The full log is available at
http://patchew.org/logs/20210316154910.25804-1-rebecca@nuviainc.com/testing.checkpatch/?type=message.
---
Email generated automatically by Patchew [https://patchew.org/].
Please send your feedback to patchew-devel@redhat.com
^ permalink raw reply [flat|nested] 7+ messages in thread
* Re: [PATCH v4 1/3] target/arm: Add support for FEAT_TLBIRANGE
2021-03-16 15:49 ` [PATCH v4 1/3] target/arm: Add support for FEAT_TLBIRANGE Rebecca Cran
@ 2021-03-16 18:09 ` Richard Henderson
2021-03-16 21:13 ` Rebecca Cran
0 siblings, 1 reply; 7+ messages in thread
From: Richard Henderson @ 2021-03-16 18:09 UTC (permalink / raw)
To: Rebecca Cran, Peter Maydell, qemu-arm; +Cc: qemu-devel
On 3/16/21 9:49 AM, Rebecca Cran wrote:
> + for (page = addr; page < (addr + length); page += TARGET_PAGE_SIZE) {
This test means that it's impossible to flush the last page of the address
space (addr + length == 0). I think better to do
for (l = 0; l < length; l += TARGET_PAGE_SIZE)
page = addr + l;
...
> + for (mmu_idx = 0; mmu_idx < NB_MMU_MODES; mmu_idx++) {
> + if ((idxmap >> mmu_idx) & 1) {
> + tlb_flush_page_bits_locked(env, mmu_idx, page, bits);
Hmm. I'm not keen on this. You're not able to notice the special cases
within, where we flush the entire tlb -- and therefore you do not need to
continue the outer loop for this mmuidx.
> + tb_flush_jmp_cache(cpu, page);
This does not need to be in the mmuidx loop. But since above means that the
mmuidx loop should be the outer loop, this would go in a separate page loop by
itself.
> +void tlb_flush_page_range_bits_by_mmuidx(CPUState *cpu,
> + target_ulong addr,
> + target_ulong length,
> + uint16_t idxmap,
> + unsigned bits)
> +{
> + TLBFlushPageRangeBitsByMMUIdxData d;
> + TLBFlushPageRangeBitsByMMUIdxData *p;
> +
> + /* This should already be page aligned */
> + addr &= TARGET_PAGE_BITS;
> +
> + /* If all bits are significant, this devolves to tlb_flush_page. */
> + if (bits >= TARGET_LONG_BITS) {
> + tlb_flush_page_by_mmuidx(cpu, addr, idxmap);
> + return;
> + }
This case is incorrect.
The cputlb changes should have remained a separate patch.
> @@ -4759,6 +4759,241 @@ static void tlbi_aa64_vae3is_write(CPUARMState *env, const ARMCPRegInfo *ri,
> ARMMMUIdxBit_SE3, bits);
> }
>
> +#ifdef TARGET_AARCH64
> +static uint64_t tlbi_aa64_range_get_length(CPUARMState *env,
> + uint64_t value)
> +{
> + unsigned int page_size;
> + unsigned int page_size_granule;
> + uint64_t num;
> + uint64_t scale;
> + uint64_t exponent;
> + uint64_t length;
> +
> + num = extract64(value, 39, 4);
> + scale = extract64(value, 44, 2);
> + page_size_granule = extract64(value, 46, 2);
> +
> + switch (page_size_granule) {
> + case 1:
> + page_size = 4096;
Indentation is off?
> + break;
> + case 2:
> + page_size = 16384;
> + break;
> + case 3:
> + page_size = 65536;
You might as well have this as page_shift = {12,14,16}, or perhaps page_shift =
page_size_granule * 2 + 10 instead of the full switch.
> + exponent = (5 * scale) + 1;
> + length = ((num + 1) << exponent) * page_size;
length = (num + 1) << (exponent + page_shift);
> + mask = vae1_tlbmask(env);
> + if (regime_has_2_ranges(mask)) {
You can't pass in mask.
All of the mmuidx will have the same form, so ctz32(mask) would pick out the
mmuidx for the first bit.
> + if (regime_has_2_ranges(secure ? ARMMMUIdxBit_SE2 : ARMMMUIdxBit_E2)) {
again. Only this time we know that E2 & SE2 have one range. Only (S)EL1&0 and
(S)EL2&0 have two ranges.
> + if (regime_has_2_ranges(ARMMMUIdxBit_SE3)) {
Likewise, E3 has only one range.
r~
^ permalink raw reply [flat|nested] 7+ messages in thread
* Re: [PATCH v4 1/3] target/arm: Add support for FEAT_TLBIRANGE
2021-03-16 18:09 ` Richard Henderson
@ 2021-03-16 21:13 ` Rebecca Cran
0 siblings, 0 replies; 7+ messages in thread
From: Rebecca Cran @ 2021-03-16 21:13 UTC (permalink / raw)
To: Richard Henderson, Peter Maydell, qemu-arm; +Cc: qemu-devel
On 3/16/21 12:09 PM, Richard Henderson wrote:
> On 3/16/21 9:49 AM, Rebecca Cran wrote:
>> + for (page = addr; page < (addr + length); page +=
>> TARGET_PAGE_SIZE) {
>
> This test means that it's impossible to flush the last page of the
> address space (addr + length == 0). I think better to do
>
> for (l = 0; l < length; l += TARGET_PAGE_SIZE)
> page = addr + l;
> ...
>
>> + for (mmu_idx = 0; mmu_idx < NB_MMU_MODES; mmu_idx++) {
>> + if ((idxmap >> mmu_idx) & 1) {
>> + tlb_flush_page_bits_locked(env, mmu_idx, page, bits);
>
> Hmm. I'm not keen on this. You're not able to notice the special cases
> within, where we flush the entire tlb -- and therefore you do not need
> to continue the outer loop for this mmuidx.
Would you be happy to have tlb_flush_page_bits_locked updated to return
a bool, indicating if an entire flush has been done?
> The cputlb changes should have remained a separate patch.
Ok, I'll split it out.
--
Rebecca Cran
^ permalink raw reply [flat|nested] 7+ messages in thread
end of thread, other threads:[~2021-03-16 21:14 UTC | newest]
Thread overview: 7+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2021-03-16 15:49 [PATCH v4 0/3] target/arm: Add support for FEAT_TLBIOS and FEAT_TLBIRANGE Rebecca Cran
2021-03-16 15:49 ` [PATCH v4 1/3] target/arm: Add support for FEAT_TLBIRANGE Rebecca Cran
2021-03-16 18:09 ` Richard Henderson
2021-03-16 21:13 ` Rebecca Cran
2021-03-16 15:49 ` [PATCH v4 2/3] target/arm: Add support for FEAT_TLBIOS Rebecca Cran
2021-03-16 15:49 ` [PATCH v4 3/3] target/arm: set ID_AA64ISAR0.TLB to 2 for max AARCH64 CPU type Rebecca Cran
2021-03-16 17:03 ` [PATCH v4 0/3] target/arm: Add support for FEAT_TLBIOS and FEAT_TLBIRANGE no-reply
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.