Skip to contentSkip to conversation

AMD broadcast TLB invalidation

PATCH Superseded Rik van Riel 7 participants 26 messages Feb 11, 2025 – Feb 13, 2025

Subsystems (8)

v10 of 14 Superseded Latest is v14 →

lore
Share & tools
lore
#0 lore raw

v10 cover AMD broadcast TLB invalidation

Add support for broadcast TLB invalidation using AMD's INVLPGB instruction.

This allows the kernel to invalidate TLB entries on remote CPUs without needing to send IPIs, without having to wait for remote CPUs to handle those interrupts, and with less interruption to what was running on those CPUs.

Because x86 PCID space is limited, and there are some very large systems out there, broadcast TLB invalidation is only used for processes that are active on 3 or more CPUs, with the threshold being gradually increased the more the PCID space gets exhausted.

Combined with the removal of unnecessary lru_add_drain calls (see https://lkml.org/lkml/2024/12/19/1388) this results in a nice performance boost for the will-it-scale tlb_flush2_threads test on an AMD Milan system with 36 cores:

- vanilla kernel:           527k loops/second
- lru_add_drain removal:    731k loops/second
- only INVLPGB:             527k loops/second

- lru_add_drain + INVLPGB: 1157k loops/second

Profiling with only the INVLPGB changes showed while TLB invalidation went down from 40% of the total CPU time to only around 4% of CPU time, the contention simply moved to the LRU lock.

Fixing both at the same time about doubles the
number of iterations per second from this case.

Some numbers closer to real world performance
can be found at Phoronix, thanks to Michael:

https://www.phoronix.com/news/AMD-INVLPGB-Linux-Benefits

My current plan is to implement support for Intel's RAR (Remote Action Request) TLB flushing in a follow-up series, after this thing has been merged into -tip. Making things any larger would just be unwieldy for reviewers.

v10:
- simplify partial pages with min(nr, 1) in the invlpgb loop (Peter)
- document x86 paravirt, AMD invlpgb, and ARM64 flush without IPI (Brendan)
- remove IS_ENABLED(CONFIG_X86_BROADCAST_TLB_FLUSH) (Brendan)
- various cleanups (Brendan)
v9:
- print warning when start or end address was rounded (Peter)
- in the reclaim code, tlbsync at context switch time (Peter)
- fix !CONFIG_CPU_SUP_AMD compile error in arch_tlbbatch_add_pending (Jan)
v8:
- round start & end to handle non-page-aligned callers (Steven & Jan)
- fix up changelog & add tested-by tags (Manali)
v7:
- a few small code cleanups (Nadav)
- fix spurious VM_WARN_ON_ONCE in mm_global_asid
- code simplifications & better barriers (Peter & Dave)
v6:
- fix info->end check in flush_tlb_kernel_range (Michael)
- disable broadcast TLB flushing on 32 bit x86
v5:
- use byte assembly for compatibility with older toolchains (Borislav, Michael)
- ensure a panic on an invalid number of extra pages (Dave, Tom)
- add cant_migrate() assertion to tlbsync (Jann)
- a bunch more cleanups (Nadav)
- key TCE enabling off X86_FEATURE_TCE (Andrew)
- fix a race between reclaim and ASID transition (Jann)
v4:
- Use only bitmaps to track free global ASIDs (Nadav)
- Improved AMD initialization (Borislav & Tom)
- Various naming and documentation improvements (Peter, Nadav, Tom, Dave)
- Fixes for subtle race conditions (Jann)
v3:
- Remove paravirt tlb_remove_table call (thank you Qi Zheng)
- More suggested cleanups and changelog fixes by Peter and Nadav
v2:
- Apply suggestions by Peter and Borislav (thank you!)
- Fix bug in arch_tlbbatch_flush, where we need to do both
the TLBSYNC, and flush the CPUs that are in the cpumask.
- Some updates to comments and changelogs based on questions.

#1 lore raw

PATCH v10 1/12 x86/mm: make MMU_GATHER_RCU_TABLE_FREE unconditional

Currently x86 uses CONFIG_MMU_GATHER_TABLE_FREE when using paravirt, and not when running on bare metal.

There is no real good reason to do things differently for each setup. Make them all the same.

Currently get_user_pages_fast synchronizes against page table freeing in two different ways:
- on bare metal, by blocking IRQs, which block TLB flush IPIs
- on paravirt, with MMU_GATHER_RCU_TABLE_FREE

This is done because some paravirt TLB flush implementations handle the TLB flush in the hypervisor, and will do the flush even when the target CPU has interrupts disabled.

Always handle page table freeing with MMU_GATHER_RCU_TABLE_FREE. Using RCU synchronization between page table freeing and get_user_pages_fast() allows bare metal to also do TLB flushing while interrupts are disabled.

Various places in the mm do still block IRQs or disable preemption as an implicit way to block RCU frees.

That makes it safe to use INVLPGB on AMD CPUs.

---
 arch/x86/Kconfig           |  2 +-
 arch/x86/kernel/paravirt.c |  7 +------
 arch/x86/mm/pgtable.c      | 16 ++++------------
 mm/mmu_gather.c            |  4 ++--
 4 files changed, 8 insertions(+), 21 deletions(-)
arch/x86/Kconfig +1 -1
index 9d7bd0ae48c4..e8743f8c9fd0 100644
@@ -274,7 +274,7 @@ config X86
 	select HAVE_PCI
 	select HAVE_PERF_REGS
 	select HAVE_PERF_USER_STACK_DUMP
-	select MMU_GATHER_RCU_TABLE_FREE	if PARAVIRT
+	select MMU_GATHER_RCU_TABLE_FREE
 	select MMU_GATHER_MERGE_VMAS
 	select HAVE_POSIX_CPU_TIMERS_TASK_WORK
 	select HAVE_REGS_AND_STACK_ACCESS_API
arch/x86/kernel/paravirt.c +1 -6
index fec381533555..2b78a6b466ed 100644
@@ -59,11 +59,6 @@ void __init native_pv_lock_init(void)
 		static_branch_enable(&virt_spin_lock_key);
 }
 
-static void native_tlb_remove_table(struct mmu_gather *tlb, void *table)
-{
-	tlb_remove_page(tlb, table);
-}
-
 struct static_key paravirt_steal_enabled;
 struct static_key paravirt_steal_rq_enabled;
 
@@ -191,7 +186,7 @@ struct paravirt_patch_template pv_ops = {
 	.mmu.flush_tlb_kernel	= native_flush_tlb_global,
 	.mmu.flush_tlb_one_user	= native_flush_tlb_one_user,
 	.mmu.flush_tlb_multi	= native_flush_tlb_multi,
-	.mmu.tlb_remove_table	= native_tlb_remove_table,
+	.mmu.tlb_remove_table	= tlb_remove_table,
 
 	.mmu.exit_mmap		= paravirt_nop,
 	.mmu.notify_page_enc_status_changed	= paravirt_nop,
arch/x86/mm/pgtable.c +4 -12
index 5745a354a241..3dc4af1f7868 100644
@@ -18,14 +18,6 @@ EXPORT_SYMBOL(physical_mask);
 #define PGTABLE_HIGHMEM 0
 #endif
 
-#ifndef CONFIG_PARAVIRT
-static inline
-void paravirt_tlb_remove_table(struct mmu_gather *tlb, void *table)
-{
-	tlb_remove_page(tlb, table);
-}
-#endif
-
 gfp_t __userpte_alloc_gfp = GFP_PGTABLE_USER | PGTABLE_HIGHMEM;
 
 pgtable_t pte_alloc_one(struct mm_struct *mm)
@@ -54,7 +46,7 @@ void ___pte_free_tlb(struct mmu_gather *tlb, struct page *pte)
 {
 	pagetable_pte_dtor(page_ptdesc(pte));
 	paravirt_release_pte(page_to_pfn(pte));
-	paravirt_tlb_remove_table(tlb, pte);
+	tlb_remove_table(tlb, pte);
 }
 
 #if CONFIG_PGTABLE_LEVELS > 2
@@ -70,7 +62,7 @@ void ___pmd_free_tlb(struct mmu_gather *tlb, pmd_t *pmd)
 	tlb->need_flush_all = 1;
 #endif
 	pagetable_pmd_dtor(ptdesc);
-	paravirt_tlb_remove_table(tlb, ptdesc_page(ptdesc));
+	tlb_remove_table(tlb, ptdesc_page(ptdesc));
 }
 
 #if CONFIG_PGTABLE_LEVELS > 3
@@ -80,14 +72,14 @@ void ___pud_free_tlb(struct mmu_gather *tlb, pud_t *pud)
 
 	pagetable_pud_dtor(ptdesc);
 	paravirt_release_pud(__pa(pud) >> PAGE_SHIFT);
-	paravirt_tlb_remove_table(tlb, virt_to_page(pud));
+	tlb_remove_table(tlb, virt_to_page(pud));
 }
 
 #if CONFIG_PGTABLE_LEVELS > 4
 void ___p4d_free_tlb(struct mmu_gather *tlb, p4d_t *p4d)
 {
 	paravirt_release_p4d(__pa(p4d) >> PAGE_SHIFT);
-	paravirt_tlb_remove_table(tlb, virt_to_page(p4d));
+	tlb_remove_table(tlb, virt_to_page(p4d));
 }
 #endif	/* CONFIG_PGTABLE_LEVELS > 4 */
 #endif	/* CONFIG_PGTABLE_LEVELS > 3 */
mm/mmu_gather.c +2 -2
index 99b3e9408aa0..59fd0137af63 100644
@@ -246,8 +246,8 @@ static void __tlb_remove_table_free(struct mmu_table_batch *batch)
  * IRQs delays the completion of the TLB flush we can never observe an already
  * freed page.
  *
- * Architectures that do not have this (PPC) need to delay the freeing by some
- * other means, this is that means.
+ * Architectures that do not use IPIs (PPC, x86 paravirt, AMD INVLPGB, ARM64)
+ * need to delay the freeing by some other means, this is that means.
  *
  * What we do is batch the freed directory pages (tables) and RCU free them.
  * We use the sched RCU variant, as that guarantees that IRQ/preempt disabling
Signature · 1 line
2.47.1
#2 lore raw

PATCH v10 2/12 x86/mm: remove pv_ops.mmu.tlb_remove_table call

In reply to Rik van Riel

Every pv_ops.mmu.tlb_remove_table call ends up calling tlb_remove_table.

Get rid of the indirection by simply calling tlb_remove_table directly, and not going through the paravirt function pointers.

---
 arch/x86/hyperv/mmu.c                 | 1 -
 arch/x86/include/asm/paravirt.h       | 5 -----
 arch/x86/include/asm/paravirt_types.h | 2 --
 arch/x86/kernel/kvm.c                 | 1 -
 arch/x86/kernel/paravirt.c            | 1 -
 arch/x86/xen/mmu_pv.c                 | 1 -
 6 files changed, 11 deletions(-)
arch/x86/hyperv/mmu.c +0 -1
index 1cc113200ff5..cbe6c71e17c1 100644
@@ -240,5 +240,4 @@ void hyperv_setup_mmu_ops(void)
 
 	pr_info("Using hypercall for remote TLB flush\n");
 	pv_ops.mmu.flush_tlb_multi = hyperv_flush_tlb_multi;
-	pv_ops.mmu.tlb_remove_table = tlb_remove_table;
 }
arch/x86/include/asm/paravirt.h +0 -5
index d4eb9e1d61b8..794ba3647c6c 100644
@@ -91,11 +91,6 @@ static inline void __flush_tlb_multi(const struct cpumask *cpumask,
 	PVOP_VCALL2(mmu.flush_tlb_multi, cpumask, info);
 }
 
-static inline void paravirt_tlb_remove_table(struct mmu_gather *tlb, void *table)
-{
-	PVOP_VCALL2(mmu.tlb_remove_table, tlb, table);
-}
-
 static inline void paravirt_arch_exit_mmap(struct mm_struct *mm)
 {
 	PVOP_VCALL1(mmu.exit_mmap, mm);
arch/x86/include/asm/paravirt_types.h +0 -2
index 8d4fbe1be489..13405959e4db 100644
@@ -136,8 +136,6 @@ struct pv_mmu_ops {
 	void (*flush_tlb_multi)(const struct cpumask *cpus,
 				const struct flush_tlb_info *info);
 
-	void (*tlb_remove_table)(struct mmu_gather *tlb, void *table);
-
 	/* Hook for intercepting the destruction of an mm_struct. */
 	void (*exit_mmap)(struct mm_struct *mm);
 	void (*notify_page_enc_status_changed)(unsigned long pfn, int npages, bool enc);
arch/x86/kernel/kvm.c +0 -1
index 7a422a6c5983..3be9b3342c67 100644
@@ -838,7 +838,6 @@ static void __init kvm_guest_init(void)
 #ifdef CONFIG_SMP
 	if (pv_tlb_flush_supported()) {
 		pv_ops.mmu.flush_tlb_multi = kvm_flush_tlb_multi;
-		pv_ops.mmu.tlb_remove_table = tlb_remove_table;
 		pr_info("KVM setup pv remote TLB flush\n");
 	}
 
arch/x86/kernel/paravirt.c +0 -1
index 2b78a6b466ed..c019771e0123 100644
@@ -186,7 +186,6 @@ struct paravirt_patch_template pv_ops = {
 	.mmu.flush_tlb_kernel	= native_flush_tlb_global,
 	.mmu.flush_tlb_one_user	= native_flush_tlb_one_user,
 	.mmu.flush_tlb_multi	= native_flush_tlb_multi,
-	.mmu.tlb_remove_table	= tlb_remove_table,
 
 	.mmu.exit_mmap		= paravirt_nop,
 	.mmu.notify_page_enc_status_changed	= paravirt_nop,
arch/x86/xen/mmu_pv.c +0 -1
index 55a4996d0c04..041e17282af0 100644
@@ -2137,7 +2137,6 @@ static const typeof(pv_ops) xen_mmu_ops __initconst = {
 		.flush_tlb_kernel = xen_flush_tlb,
 		.flush_tlb_one_user = xen_flush_tlb_one_user,
 		.flush_tlb_multi = xen_flush_tlb_multi,
-		.tlb_remove_table = tlb_remove_table,
 
 		.pgd_alloc = xen_pgd_alloc,
 		.pgd_free = xen_pgd_free,
Signature · 1 line
2.47.1
#3 lore raw

PATCH v10 3/12 x86/mm: consolidate full flush threshold decision

In reply to Rik van Riel

Reduce code duplication by consolidating the decision point for whether to do individual invalidations or a full flush inside get_flush_tlb_info.

---
 arch/x86/mm/tlb.c | 41 +++++++++++++++++++----------------------
 1 file changed, 19 insertions(+), 22 deletions(-)
arch/x86/mm/tlb.c +19 -22
index 6cf881a942bb..58c8a6a94e65 100644
@@ -1009,6 +1009,15 @@ static struct flush_tlb_info *get_flush_tlb_info(struct mm_struct *mm,
 	info->initiating_cpu	= smp_processor_id();
 	info->trim_cpumask	= 0;
 
+	/*
+	 * If the number of flushes is so large that a full flush
+	 * would be faster, do a full flush.
+	 */
+	if ((end - start) >> stride_shift > tlb_single_page_flush_ceiling) {
+		info->start = 0;
+		info->end = TLB_FLUSH_ALL;
+	}
+
 	return info;
 }
 
@@ -1026,17 +1035,8 @@ void flush_tlb_mm_range(struct mm_struct *mm, unsigned long start,
 				bool freed_tables)
 {
 	struct flush_tlb_info *info;
+	int cpu = get_cpu();
 	u64 new_tlb_gen;
-	int cpu;
-
-	cpu = get_cpu();
-
-	/* Should we flush just the requested range? */
-	if ((end == TLB_FLUSH_ALL) ||
-	    ((end - start) >> stride_shift) > tlb_single_page_flush_ceiling) {
-		start = 0;
-		end = TLB_FLUSH_ALL;
-	}
 
 	/* This is also a barrier that synchronizes with switch_mm(). */
 	new_tlb_gen = inc_mm_tlb_gen(mm);
@@ -1089,22 +1089,19 @@ static void do_kernel_range_flush(void *info)
 
 void flush_tlb_kernel_range(unsigned long start, unsigned long end)
 {
-	/* Balance as user space task's flush, a bit conservative */
-	if (end == TLB_FLUSH_ALL ||
-	    (end - start) > tlb_single_page_flush_ceiling << PAGE_SHIFT) {
-		on_each_cpu(do_flush_tlb_all, NULL, 1);
-	} else {
-		struct flush_tlb_info *info;
+	struct flush_tlb_info *info;
 
-		preempt_disable();
-		info = get_flush_tlb_info(NULL, start, end, 0, false,
-					  TLB_GENERATION_INVALID);
+	guard(preempt)();
 
+	info = get_flush_tlb_info(NULL, start, end, PAGE_SHIFT, false,
+				  TLB_GENERATION_INVALID);
+
+	if (info->end == TLB_FLUSH_ALL)
+		on_each_cpu(do_flush_tlb_all, NULL, 1);
+	else
 		on_each_cpu(do_kernel_range_flush, info, 1);
 
-		put_flush_tlb_info();
-		preempt_enable();
-	}
+	put_flush_tlb_info();
 }
 
 /*
Signature · 1 line
2.47.1
#4 lore raw

PATCH v10 4/12 x86/mm: get INVLPGB count max from CPUID

In reply to Rik van Riel

The CPU advertises the maximum number of pages that can be shot down with one INVLPGB instruction in the CPUID data.

Save that information for later use.

---
 arch/x86/Kconfig.cpu               | 5 +++++
 arch/x86/include/asm/cpufeatures.h | 1 +
 arch/x86/include/asm/tlbflush.h    | 7 +++++++
 arch/x86/kernel/cpu/amd.c          | 8 ++++++++
 4 files changed, 21 insertions(+)
arch/x86/Kconfig.cpu +5 -0
index 2a7279d80460..abe013a1b076 100644
@@ -395,6 +395,10 @@ config X86_VMX_FEATURE_NAMES
 	def_bool y
 	depends on IA32_FEAT_CTL
 
+config X86_BROADCAST_TLB_FLUSH
+	def_bool y
+	depends on CPU_SUP_AMD && 64BIT
+
 menuconfig PROCESSOR_SELECT
 	bool "Supported processor vendors" if EXPERT
 	help
@@ -431,6 +435,7 @@ config CPU_SUP_CYRIX_32
 config CPU_SUP_AMD
 	default y
 	bool "Support AMD processors" if PROCESSOR_SELECT
+	select X86_BROADCAST_TLB_FLUSH
 	help
 	  This enables detection, tunings and quirks for AMD processors
 
arch/x86/include/asm/cpufeatures.h +1 -0
index 17b6590748c0..f9b832e971c5 100644
@@ -338,6 +338,7 @@
 #define X86_FEATURE_CLZERO		(13*32+ 0) /* "clzero" CLZERO instruction */
 #define X86_FEATURE_IRPERF		(13*32+ 1) /* "irperf" Instructions Retired Count */
 #define X86_FEATURE_XSAVEERPTR		(13*32+ 2) /* "xsaveerptr" Always save/restore FP error pointers */
+#define X86_FEATURE_INVLPGB		(13*32+ 3) /* INVLPGB and TLBSYNC instruction supported. */
 #define X86_FEATURE_RDPRU		(13*32+ 4) /* "rdpru" Read processor register at user level */
 #define X86_FEATURE_WBNOINVD		(13*32+ 9) /* "wbnoinvd" WBNOINVD instruction */
 #define X86_FEATURE_AMD_IBPB		(13*32+12) /* Indirect Branch Prediction Barrier */
arch/x86/include/asm/tlbflush.h +7 -0
index 02fc2aa06e9e..8fe3b2dda507 100644
@@ -183,6 +183,13 @@ static inline void cr4_init_shadow(void)
 extern unsigned long mmu_cr4_features;
 extern u32 *trampoline_cr4_features;
 
+/* How many pages can we invalidate with one INVLPGB. */
+#ifdef CONFIG_X86_BROADCAST_TLB_FLUSH
+extern u16 invlpgb_count_max;
+#else
+#define invlpgb_count_max 1
+#endif
+
 extern void initialize_tlbstate_and_flush(void);
 
 /*
arch/x86/kernel/cpu/amd.c +8 -0
index 79d2e17f6582..bcf73775b4f8 100644
@@ -29,6 +29,8 @@
 
 #include "cpu.h"
 
+u16 invlpgb_count_max __ro_after_init;
+
 static inline int rdmsrl_amd_safe(unsigned msr, unsigned long long *p)
 {
 	u32 gprs[8] = { 0 };
@@ -1135,6 +1137,12 @@ static void cpu_detect_tlb_amd(struct cpuinfo_x86 *c)
 		tlb_lli_2m[ENTRIES] = eax & mask;
 
 	tlb_lli_4m[ENTRIES] = tlb_lli_2m[ENTRIES] >> 1;
+
+	/* Max number of pages INVLPGB can invalidate in one shot */
+	if (boot_cpu_has(X86_FEATURE_INVLPGB)) {
+		cpuid(0x80000008, &eax, &ebx, &ecx, &edx);
+		invlpgb_count_max = (edx & 0xffff) + 1;
+	}
 }
 
 static const struct cpu_dev amd_cpu_dev = {
Signature · 1 line
2.47.1
#5 lore raw

PATCH v10 5/12 x86/mm: add INVLPGB support code

In reply to Rik van Riel

Add invlpgb.h with the helper functions and definitions needed to use broadcast TLB invalidation on AMD EPYC 3 and newer CPUs.

---
 arch/x86/include/asm/disabled-features.h |   8 +-
 arch/x86/include/asm/invlpgb.h           | 101 +++++++++++++++++++++++
 arch/x86/include/asm/tlbflush.h          |   1 +
 3 files changed, 109 insertions(+), 1 deletion(-)
 create mode 100644 arch/x86/include/asm/invlpgb.h
arch/x86/include/asm/disabled-features.h +7 -1
index c492bdc97b05..625a89259968 100644
@@ -129,6 +129,12 @@
 #define DISABLE_SEV_SNP		(1 << (X86_FEATURE_SEV_SNP & 31))
 #endif
 
+#ifdef CONFIG_X86_BROADCAST_TLB_FLUSH
+#define DISABLE_INVLPGB		0
+#else
+#define DISABLE_INVLPGB		(1 << (X86_FEATURE_INVLPGB & 31))
+#endif
+
 /*
  * Make sure to add features to the correct mask
  */
@@ -146,7 +152,7 @@
 #define DISABLED_MASK11	(DISABLE_RETPOLINE|DISABLE_RETHUNK|DISABLE_UNRET| \
 			 DISABLE_CALL_DEPTH_TRACKING|DISABLE_USER_SHSTK)
 #define DISABLED_MASK12	(DISABLE_FRED|DISABLE_LAM)
-#define DISABLED_MASK13	0
+#define DISABLED_MASK13	(DISABLE_INVLPGB)
 #define DISABLED_MASK14	0
 #define DISABLED_MASK15	0
 #define DISABLED_MASK16	(DISABLE_PKU|DISABLE_OSPKE|DISABLE_LA57|DISABLE_UMIP| \
arch/x86/include/asm/invlpgb.h new file +101 -0

101 lines not on this page · raw message

arch/x86/include/asm/tlbflush.h +1 -0
index 8fe3b2dda507..dba5caa4a9f4 100644
@@ -10,6 +10,7 @@
 #include <asm/cpufeature.h>
 #include <asm/special_insns.h>
 #include <asm/smp.h>
+#include <asm/invlpgb.h>
 #include <asm/invpcid.h>
 #include <asm/pti.h>
 #include <asm/processor-flags.h>
Signature · 1 line
2.47.1
#6 lore raw

PATCH v10 6/12 x86/mm: use INVLPGB for kernel TLB flushes

In reply to Rik van Riel

Use broadcast TLB invalidation for kernel addresses when available.

Remove the need to send IPIs for kernel TLB flushes.

---
 arch/x86/mm/tlb.c | 26 +++++++++++++++++++++++++-
 1 file changed, 25 insertions(+), 1 deletion(-)
arch/x86/mm/tlb.c +25 -1
index 58c8a6a94e65..54a118b22e59 100644
@@ -1077,6 +1077,28 @@ void flush_tlb_all(void)
 	on_each_cpu(do_flush_tlb_all, NULL, 1);
 }
 
+static bool broadcast_kernel_range_flush(struct flush_tlb_info *info)
+{
+	unsigned long addr;
+	unsigned long nr;
+
+	if (!cpu_feature_enabled(X86_FEATURE_INVLPGB))
+		return false;
+
+	if (info->end == TLB_FLUSH_ALL) {
+		invlpgb_flush_all();
+		return true;
+	}
+
+	for (addr = info->start; addr < info->end; addr += nr << PAGE_SHIFT) {
+		nr = min((info->end - addr) >> PAGE_SHIFT, invlpgb_count_max);
+		nr = max(nr, 1); /* Round up if the last page is partial. */
+		invlpgb_flush_addr_nosync(addr, nr);
+	}
+	tlbsync();
+	return true;
+}
+
 static void do_kernel_range_flush(void *info)
 {
 	struct flush_tlb_info *f = info;
@@ -1096,7 +1118,9 @@ void flush_tlb_kernel_range(unsigned long start, unsigned long end)
 	info = get_flush_tlb_info(NULL, start, end, PAGE_SHIFT, false,
 				  TLB_GENERATION_INVALID);
 
-	if (info->end == TLB_FLUSH_ALL)
+	if (broadcast_kernel_range_flush(info))
+		; /* Fall through. */
+	else if (info->end == TLB_FLUSH_ALL)
 		on_each_cpu(do_flush_tlb_all, NULL, 1);
 	else
 		on_each_cpu(do_kernel_range_flush, info, 1);
Signature · 1 line
2.47.1
#7 lore raw

PATCH v10 7/12 x86/mm: use INVLPGB in flush_tlb_all

In reply to Rik van Riel

The flush_tlb_all() function is not used a whole lot, but we might as well use broadcast TLB flushing there, too.

---
 arch/x86/mm/tlb.c | 12 ++++++++++++
 1 file changed, 12 insertions(+)
arch/x86/mm/tlb.c +12 -0
index 54a118b22e59..62b4eaf4e851 100644
@@ -1065,6 +1065,16 @@ void flush_tlb_mm_range(struct mm_struct *mm, unsigned long start,
 }
 
 
+static bool broadcast_flush_tlb_all(void)
+{
+	if (!cpu_feature_enabled(X86_FEATURE_INVLPGB))
+		return false;
+
+	guard(preempt)();
+	invlpgb_flush_all();
+	return true;
+}
+
 static void do_flush_tlb_all(void *info)
 {
 	count_vm_tlb_event(NR_TLB_REMOTE_FLUSH_RECEIVED);
@@ -1073,6 +1083,8 @@ static void do_flush_tlb_all(void *info)
 
 void flush_tlb_all(void)
 {
+	if (broadcast_flush_tlb_all())
+		return;
 	count_vm_tlb_event(NR_TLB_REMOTE_FLUSH);
 	on_each_cpu(do_flush_tlb_all, NULL, 1);
 }
Signature · 1 line
2.47.1
#8 lore raw

PATCH v10 8/12 x86/mm: use broadcast TLB flushing for page reclaim TLB flushing

In reply to Rik van Riel

In the page reclaim code, we only track the CPU(s) where the TLB needs to be flushed, rather than all the individual mappings that may be getting invalidated.

Use broadcast TLB flushing when that is available.

---
 arch/x86/mm/tlb.c | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)
arch/x86/mm/tlb.c +3 -1
index 62b4eaf4e851..55c922083aa2 100644
@@ -1316,7 +1316,9 @@ void arch_tlbbatch_flush(struct arch_tlbflush_unmap_batch *batch)
 	 * a local TLB flush is needed. Optimize this use-case by calling
 	 * flush_tlb_func_local() directly in this case.
 	 */
-	if (cpumask_any_but(&batch->cpumask, cpu) < nr_cpu_ids) {
+	if (cpu_feature_enabled(X86_FEATURE_INVLPGB)) {
+		invlpgb_flush_all_nonglobals();
+	} else if (cpumask_any_but(&batch->cpumask, cpu) < nr_cpu_ids) {
 		flush_tlb_multi(&batch->cpumask, info);
 	} else if (cpumask_test_cpu(cpu, &batch->cpumask)) {
 		lockdep_assert_irqs_enabled();
Signature · 1 line
2.47.1
#9 lore raw

PATCH v10 9/12 x86/mm: enable broadcast TLB invalidation for multi-threaded processes

In reply to Rik van Riel

Use broadcast TLB invalidation, using the INVPLGB instruction, on AMD EPYC 3 and newer CPUs.

In order to not exhaust PCID space, and keep TLB flushes local for single threaded processes, we only hand out broadcast ASIDs to processes active on 4 or more CPUs.

---
 arch/x86/include/asm/mmu.h         |   6 +
 arch/x86/include/asm/mmu_context.h |  14 ++
 arch/x86/include/asm/tlbflush.h    |  73 ++++++
 arch/x86/mm/tlb.c                  | 348 ++++++++++++++++++++++++++++-
 4 files changed, 429 insertions(+), 12 deletions(-)
arch/x86/include/asm/mmu.h +6 -0
index 3b496cdcb74b..d71cd599fec4 100644
@@ -69,6 +69,12 @@ typedef struct {
 	u16 pkey_allocation_map;
 	s16 execute_only_pkey;
 #endif
+
+#ifdef CONFIG_X86_BROADCAST_TLB_FLUSH
+	u16 global_asid;
+	bool asid_transition;
+#endif
+
 } mm_context_t;
 
 #define INIT_MM_CONTEXT(mm)						\
arch/x86/include/asm/mmu_context.h +14 -0
index 795fdd53bd0a..d670699d32c2 100644
@@ -139,6 +139,8 @@ static inline void mm_reset_untag_mask(struct mm_struct *mm)
 #define enter_lazy_tlb enter_lazy_tlb
 extern void enter_lazy_tlb(struct mm_struct *mm, struct task_struct *tsk);
 
+extern void destroy_context_free_global_asid(struct mm_struct *mm);
+
 /*
  * Init a new mm.  Used on mm copies, like at fork()
  * and on mm's that are brand-new, like at execve().
@@ -161,6 +163,14 @@ static inline int init_new_context(struct task_struct *tsk,
 		mm->context.execute_only_pkey = -1;
 	}
 #endif
+
+#ifdef CONFIG_X86_BROADCAST_TLB_FLUSH
+	if (cpu_feature_enabled(X86_FEATURE_INVLPGB)) {
+		mm->context.global_asid = 0;
+		mm->context.asid_transition = false;
+	}
+#endif
+
 	mm_reset_untag_mask(mm);
 	init_new_context_ldt(mm);
 	return 0;
@@ -170,6 +180,10 @@ static inline int init_new_context(struct task_struct *tsk,
 static inline void destroy_context(struct mm_struct *mm)
 {
 	destroy_context_ldt(mm);
+#ifdef CONFIG_X86_BROADCAST_TLB_FLUSH
+	if (cpu_feature_enabled(X86_FEATURE_INVLPGB))
+		destroy_context_free_global_asid(mm);
+#endif
 }
 
 extern void switch_mm(struct mm_struct *prev, struct mm_struct *next,
arch/x86/include/asm/tlbflush.h +73 -0

85 lines not on this page · raw message

arch/x86/mm/tlb.c +336 -12

430 lines not on this page · raw message

Signature · 1 line
2.47.1
maintainer
#10 lore raw

Re: [PATCH v10 09/12] x86/mm: enable broadcast TLB invalidation for multi-threaded processes

arch/x86/mm/tlb.c static void choose_new_asid(struct mm_struct *next, u64 next_tlb_gen, · line 505

Quoted 11 lines · Rik van Riel
arch/x86/mm/tlb.c +11 -0
+static void broadcast_tlb_flush(struct flush_tlb_info *info)
+{
+	bool pmd = info->stride_shift == PMD_SHIFT;
+	unsigned long maxnr = invlpgb_count_max;
+	unsigned long asid = info->mm->context.global_asid;
+	unsigned long addr = info->start;
+	unsigned long nr;
+
+	/* Flushing multiple pages at once is not supported with 1GB pages. */
+	if (info->stride_shift > PMD_SHIFT)
+		maxnr = 1;

How does this work?

Normally, if we get a 1GB range, we'll iterate on the stride and INVLPG each one (just like any other stride).

Should you not instead either force the stride down to PMD level or force a full flush?

Quoted 32 lines
arch/x86/mm/tlb.c +32 -0
+
+	/*
+	 * TLB flushes with INVLPGB are kicked off asynchronously.
+	 * The inc_mm_tlb_gen() guarantees page table updates are done
+	 * before these TLB flushes happen.
+	 */
+	if (info->end == TLB_FLUSH_ALL) {
+		invlpgb_flush_single_pcid_nosync(kern_pcid(asid));
+		/* Do any CPUs supporting INVLPGB need PTI? */
+		if (static_cpu_has(X86_FEATURE_PTI))
+			invlpgb_flush_single_pcid_nosync(user_pcid(asid));
+	} else do {
+		/*
+		 * Calculate how many pages can be flushed at once; if the
+		 * remainder of the range is less than one page, flush one.
+		 */
+		nr = min(maxnr, (info->end - addr) >> info->stride_shift);
+		nr = max(nr, 1);
+
+		invlpgb_flush_user_nr_nosync(kern_pcid(asid), addr, nr, pmd);
+		/* Do any CPUs supporting INVLPGB need PTI? */
+		if (static_cpu_has(X86_FEATURE_PTI))
+			invlpgb_flush_user_nr_nosync(user_pcid(asid), addr, nr, pmd);
+
+		addr += nr << info->stride_shift;
+	} while (addr < info->end);
+
+	finish_asid_transition(info);
+
+	/* Wait for the INVLPGBs kicked off above to finish. */
+	tlbsync();
+}
maintainer
#11 lore raw

Re: [PATCH v10 09/12] x86/mm: enable broadcast TLB invalidation for multi-threaded processes

Quoted 16 lines · Peter Zijlstra

arch/x86/mm/tlb.c static void choose_new_asid(struct mm_struct *next, u64 next_tlb_gen, · line 505

Quoted 11 lines · Rik van Riel
arch/x86/mm/tlb.c +11 -0
+static void broadcast_tlb_flush(struct flush_tlb_info *info)
+{
+	bool pmd = info->stride_shift == PMD_SHIFT;
+	unsigned long maxnr = invlpgb_count_max;
+	unsigned long asid = info->mm->context.global_asid;
+	unsigned long addr = info->start;
+	unsigned long nr;
+
+	/* Flushing multiple pages at once is not supported with 1GB pages. */
+	if (info->stride_shift > PMD_SHIFT)
+		maxnr = 1;

How does this work?

Quoted 5 lines

Normally, if we get a 1GB range, we'll iterate on the stride and INVLPG each one (just like any other stride).

Should you not instead either force the stride down to PMD level or force a full flush?

Oh, n/m, I think I see.

maintainer
#12 lore raw

Re: [PATCH v10 09/12] x86/mm: enable broadcast TLB invalidation for multi-threaded processes

In reply to Rik van Riel

On Tue, Feb 11, 2025 at 04:08:04PM -0500, Rik van Riel wrote:

I poked around at this function a little, and ended up with the below.

As to your question if any INVLPGB capable hardware needs PTI; the answer is no. No AMD machine needs PTI; but it should still work, just in case someone needs to test something.

static void broadcast_tlb_flush(struct flush_tlb_info *info)

{
	unsigned long asid = info->mm->context.global_asid;
	bool pmd = info->stride_shift == PMD_SHIFT;
	unsigned long addr = info->start;

	/*
	 * TLB flushes with INVLPGB are kicked off asynchronously.
	 * The inc_mm_tlb_gen() guarantees page table updates are done
	 * before these TLB flushes happen.
	 */
	if (info->end == TLB_FLUSH_ALL) {
		invlpgb_flush_single_pcid_nosync(kern_pcid(asid));
		if (static_cpu_has(X86_FEATURE_PTI))
			invlpgb_flush_single_pcid_nosync(user_pcid(asid));

	} else do {
		unsigned long nr = 1;

		if (info->stride_shift <= PMD_SHIFT) {
			/*
			 * Calculate how many pages can be flushed at once; if the
			 * remainder of the range is less than one page, flush one.
			 */
			nr = (info->end - addr) >> info->stride_shift);
			nr = clamp_val(nr, 1, invlpgb_count_max);
		}

		invlpgb_flush_user_nr_nosync(kern_pcid(asid), addr, nr, pmd);
		if (static_cpu_has(X86_FEATURE_PTI))
			invlpgb_flush_user_nr_nosync(user_pcid(asid), addr, nr, pmd);

		addr += nr << info->stride_shift;
	} while (addr < info->end);

	finish_asid_transition(info);

	/* Wait for the INVLPGBs kicked off above to finish. */
	tlbsync();
}
Reviewed-by on v10 9/12
#13 lore raw

Re: [PATCH v10 09/12] x86/mm: enable broadcast TLB invalidation for multi-threaded processes

In reply to Rik van Riel

It all looks pretty sane to me.

One small note:

Quoted 6 lines · Rik van Riel
arch/x86/mm/tlb.c +4 -0
+/*
+ * Check whether a process is currently active on more than "threshold" CPUs.
+ * This is a cheap estimation on whether or not it may make sense to assign
+ * a global ASID to this process, and use broadcast TLB invalidation.

arch/x86/mm/tlb.c static void choose_new_asid(struct mm_struct *next, u64 next_tlb_gen, · line 381

Quoted 10 lines
arch/x86/mm/tlb.c +10 -0
+ */
+static bool mm_active_cpus_exceeds(struct mm_struct *mm, int threshold)
+{
+	int count = 0;
+	int cpu;
+
+	/* This quick check should eliminate most single threaded programs. */
+	if (cpumask_weight(mm_cpumask(mm)) <= threshold)
+		return false;
+

Does it make sense to first check mm.mm_users against the threshold? It might be less cache-contended.

Anyhow, for what’s it worth:

Reviewed-by Nadav Amit <[email protected] <mailto:[email protected]>>
#14 lore raw

Re: [PATCH v10 09/12] x86/mm: enable broadcast TLB invalidation for multi-threaded processes

In reply to Rik van Riel

arch/x86/mm/tlb.c static void choose_new_asid(struct mm_struct *next, u64 next_tlb_gen, · line 349

Quoted 3 lines · Rik van Riel
arch/x86/mm/tlb.c +3 -0
+       /* Transition from global->local ASID does not currently happen. */
+       if (!global_asid && is_global_asid(prev_asid))
+               return true;

What about a WARN_ON_ONCE? Then the code contains evidence that the comment is true. IIUC this isn't just a "not implemented feature", rather there is plenty of other code that would need to be updated to make it safe.

arch/x86/mm/tlb.c static void choose_new_asid(struct mm_struct *next, u64 next_tlb_gen, · line 412

Quoted 12 lines
arch/x86/mm/tlb.c +12 -0
+static void use_global_asid(struct mm_struct *mm)
+{
+       u16 asid;
+
+       guard(raw_spinlock_irqsave)(&global_asid_lock);
+
+       /* This process is already using broadcast TLB invalidation. */
+       if (READ_ONCE(mm->context.global_asid))
+               return;
+
+       /* The last global ASID was consumed while waiting for the lock. */
+       if (!READ_ONCE(global_asid_available)) {

I think the READ_ONCE is unnecessary while we have the spinlock, it's fine if this read gets split or whatever?

Ditto for mm->context.global_asid, it's only modified with the lock held.

#15 lore raw

PATCH v10 10/12 x86/mm: do targeted broadcast flushing from tlbbatch code

In reply to Rik van Riel

Instead of doing a system-wide TLB flush from arch_tlbbatch_flush, queue up asynchronous, targeted flushes from arch_tlbbatch_add_pending.

This also allows us to avoid adding the CPUs of processes using broadcast flushing to the batch->cpumask, and will hopefully further reduce TLB flushing from the reclaim and compaction paths.

---
 arch/x86/include/asm/invlpgb.h  | 20 ++++-----
 arch/x86/include/asm/tlbflush.h | 19 ++++----
 arch/x86/mm/tlb.c               | 79 +++++++++++++++++++++++++++++++--
 3 files changed, 96 insertions(+), 22 deletions(-)
arch/x86/include/asm/invlpgb.h +10 -10
index a1d5dedd5217..43c331507cc0 100644
@@ -31,7 +31,7 @@ static inline void __invlpgb(unsigned long asid, unsigned long pcid,
 }
 
 /* Wait for INVLPGB originated by this CPU to complete. */
-static inline void tlbsync(void)
+static inline void __tlbsync(void)
 {
 	cant_migrate();
 	/* TLBSYNC: supported in binutils >= 0.36. */
@@ -61,19 +61,19 @@ static inline void invlpgb_flush_user(unsigned long pcid,
 				      unsigned long addr)
 {
 	__invlpgb(0, pcid, addr, 0, 0, INVLPGB_PCID | INVLPGB_VA);
-	tlbsync();
+	__tlbsync();
 }
 
-static inline void invlpgb_flush_user_nr_nosync(unsigned long pcid,
-						unsigned long addr,
-						u16 nr,
-						bool pmd_stride)
+static inline void __invlpgb_flush_user_nr_nosync(unsigned long pcid,
+						  unsigned long addr,
+						  u16 nr,
+						  bool pmd_stride)
 {
 	__invlpgb(0, pcid, addr, nr - 1, pmd_stride, INVLPGB_PCID | INVLPGB_VA);
 }
 
 /* Flush all mappings for a given PCID, not including globals. */
-static inline void invlpgb_flush_single_pcid_nosync(unsigned long pcid)
+static inline void __invlpgb_flush_single_pcid_nosync(unsigned long pcid)
 {
 	__invlpgb(0, pcid, 0, 0, 0, INVLPGB_PCID);
 }
@@ -82,11 +82,11 @@ static inline void invlpgb_flush_single_pcid_nosync(unsigned long pcid)
 static inline void invlpgb_flush_all(void)
 {
 	__invlpgb(0, 0, 0, 0, 0, INVLPGB_INCLUDE_GLOBAL);
-	tlbsync();
+	__tlbsync();
 }
 
 /* Flush addr, including globals, for all PCIDs. */
-static inline void invlpgb_flush_addr_nosync(unsigned long addr, u16 nr)
+static inline void __invlpgb_flush_addr_nosync(unsigned long addr, u16 nr)
 {
 	__invlpgb(0, 0, addr, nr - 1, 0, INVLPGB_INCLUDE_GLOBAL);
 }
@@ -95,7 +95,7 @@ static inline void invlpgb_flush_addr_nosync(unsigned long addr, u16 nr)
 static inline void invlpgb_flush_all_nonglobals(void)
 {
 	__invlpgb(0, 0, 0, 0, 0, 0);
-	tlbsync();
+	__tlbsync();
 }
 
 #endif /* _ASM_X86_INVLPGB */
arch/x86/include/asm/tlbflush.h +10 -9
index 234277a5ef89..89dddbcd1322 100644
@@ -106,6 +106,9 @@ struct tlb_state {
 	 * need to be invalidated.
 	 */
 	bool invalidate_other;
+#ifdef CONFIG_X86_BROADCAST_TLB_FLUSH
+	bool need_tlbsync;
+#endif
 
 #ifdef CONFIG_ADDRESS_MASKING
 	/*
@@ -310,6 +313,10 @@ static inline void broadcast_tlb_flush(struct flush_tlb_info *info)
 static inline void consider_global_asid(struct mm_struct *mm)
 {
 }
+
+static inline void tlbsync(void)
+{
+}
 #endif
 
 #ifdef CONFIG_PARAVIRT
@@ -359,21 +366,15 @@ static inline u64 inc_mm_tlb_gen(struct mm_struct *mm)
 	return atomic64_inc_return(&mm->context.tlb_gen);
 }
 
-static inline void arch_tlbbatch_add_pending(struct arch_tlbflush_unmap_batch *batch,
-					     struct mm_struct *mm,
-					     unsigned long uaddr)
-{
-	inc_mm_tlb_gen(mm);
-	cpumask_or(&batch->cpumask, &batch->cpumask, mm_cpumask(mm));
-	mmu_notifier_arch_invalidate_secondary_tlbs(mm, 0, -1UL);
-}
-
 static inline void arch_flush_tlb_batched_pending(struct mm_struct *mm)
 {
 	flush_tlb_mm(mm);
 }
 
 extern void arch_tlbbatch_flush(struct arch_tlbflush_unmap_batch *batch);
+extern void arch_tlbbatch_add_pending(struct arch_tlbflush_unmap_batch *batch,
+					     struct mm_struct *mm,
+					     unsigned long uaddr);
 
 static inline bool pte_flags_need_flush(unsigned long oldflags,
 					unsigned long newflags,
arch/x86/mm/tlb.c +76 -3

115 lines not on this page · raw message

Signature · 1 line
2.47.1
#16 lore raw

Re: [PATCH v10 10/12] x86/mm: do targeted broadcast flushing from tlbbatch code

arch/x86/mm/tlb.c void switch_mm_irqs_off(struct mm_struct *unused, struct mm_struct *next, · line 833

Quoted 6 lines · Rik van Riel
arch/x86/mm/tlb.c +2 -0
@@ -798,6 +829,8 @@ void switch_mm_irqs_off(struct mm_struct *unused, struct mm_struct *next,
        if (IS_ENABLED(CONFIG_PROVE_LOCKING))
                WARN_ON_ONCE(!irqs_disabled());
 
+       tlbsync();
+

Reminder to add comments or move this:

https://lore.kernel.org/linux-kernel/[email protected]/

I actually sent another reply which is not yet visible in Lore continuing to advocate for doing this in arch_start_context_switch() - I won't die on that hill but I do think we need either that or the commentary.

#17 lore raw

PATCH v10 11/12 x86/mm: enable AMD translation cache extensions

In reply to Rik van Riel

With AMD TCE (translation cache extensions) only the intermediate mappings that cover the address range zapped by INVLPG / INVLPGB get invalidated, rather than all intermediate mappings getting zapped at every TLB invalidation.

This can help reduce the TLB miss rate, by keeping more intermediate mappings in the cache.

From the AMD manual:

Translation Cache Extension (TCE) Bit. Bit 15, read/write. Setting this bit to 1 changes how the INVLPG, INVLPGB, and INVPCID instructions operate on TLB entries. When this bit is 0, these instructions remove the target PTE from the TLB as well as all upper-level table entries that are cached in the TLB, whether or not they are associated with the target PTE. When this bit is set, these instructions will remove the target PTE and only those upper-level entries that lead to the target PTE in the page table hierarchy, leaving unrelated upper-level entries intact.

---
 arch/x86/include/asm/msr-index.h       | 2 ++
 arch/x86/kernel/cpu/amd.c              | 4 ++++
 tools/arch/x86/include/asm/msr-index.h | 2 ++
 3 files changed, 8 insertions(+)
arch/x86/include/asm/msr-index.h +2 -0
index 3ae84c3b8e6d..dc1c1057f26e 100644
@@ -25,6 +25,7 @@
 #define _EFER_SVME		12 /* Enable virtualization */
 #define _EFER_LMSLE		13 /* Long Mode Segment Limit Enable */
 #define _EFER_FFXSR		14 /* Enable Fast FXSAVE/FXRSTOR */
+#define _EFER_TCE		15 /* Enable Translation Cache Extensions */
 #define _EFER_AUTOIBRS		21 /* Enable Automatic IBRS */
 
 #define EFER_SCE		(1<<_EFER_SCE)
@@ -34,6 +35,7 @@
 #define EFER_SVME		(1<<_EFER_SVME)
 #define EFER_LMSLE		(1<<_EFER_LMSLE)
 #define EFER_FFXSR		(1<<_EFER_FFXSR)
+#define EFER_TCE		(1<<_EFER_TCE)
 #define EFER_AUTOIBRS		(1<<_EFER_AUTOIBRS)
 
 /*
arch/x86/kernel/cpu/amd.c +4 -0
index bcf73775b4f8..21076252a491 100644
@@ -1071,6 +1071,10 @@ static void init_amd(struct cpuinfo_x86 *c)
 
 	/* AMD CPUs don't need fencing after x2APIC/TSC_DEADLINE MSR writes. */
 	clear_cpu_cap(c, X86_FEATURE_APIC_MSRS_FENCE);
+
+	/* Enable Translation Cache Extension */
+	if (cpu_feature_enabled(X86_FEATURE_TCE))
+		msr_set_bit(MSR_EFER, _EFER_TCE);
 }
 
 #ifdef CONFIG_X86_32
tools/arch/x86/include/asm/msr-index.h +2 -0
index 3ae84c3b8e6d..dc1c1057f26e 100644
@@ -25,6 +25,7 @@
 #define _EFER_SVME		12 /* Enable virtualization */
 #define _EFER_LMSLE		13 /* Long Mode Segment Limit Enable */
 #define _EFER_FFXSR		14 /* Enable Fast FXSAVE/FXRSTOR */
+#define _EFER_TCE		15 /* Enable Translation Cache Extensions */
 #define _EFER_AUTOIBRS		21 /* Enable Automatic IBRS */
 
 #define EFER_SCE		(1<<_EFER_SCE)
@@ -34,6 +35,7 @@
 #define EFER_SVME		(1<<_EFER_SVME)
 #define EFER_LMSLE		(1<<_EFER_LMSLE)
 #define EFER_FFXSR		(1<<_EFER_FFXSR)
+#define EFER_TCE		(1<<_EFER_TCE)
 #define EFER_AUTOIBRS		(1<<_EFER_AUTOIBRS)
 
 /*
Signature · 1 line
2.47.1
#18 lore raw

PATCH v10 12/12 x86/mm: only invalidate final translations with INVLPGB

In reply to Rik van Riel

Use the INVLPGB_FINAL_ONLY flag when invalidating mappings with INVPLGB. This way only leaf mappings get removed from the TLB, leaving intermediate translations cached.

On the (rare) occasions where we free page tables we do a full flush, ensuring intermediate translations get flushed from the TLB.

---
 arch/x86/include/asm/invlpgb.h | 10 ++++++++--
 arch/x86/mm/tlb.c              | 13 +++++++------
 2 files changed, 15 insertions(+), 8 deletions(-)
arch/x86/include/asm/invlpgb.h +8 -2
index 43c331507cc0..220aba708b72 100644
@@ -67,9 +67,15 @@ static inline void invlpgb_flush_user(unsigned long pcid,
 static inline void __invlpgb_flush_user_nr_nosync(unsigned long pcid,
 						  unsigned long addr,
 						  u16 nr,
-						  bool pmd_stride)
+						  bool pmd_stride,
+						  bool freed_tables)
 {
-	__invlpgb(0, pcid, addr, nr - 1, pmd_stride, INVLPGB_PCID | INVLPGB_VA);
+	u8 flags = INVLPGB_PCID | INVLPGB_VA;
+
+	if (!freed_tables)
+		flags |= INVLPGB_FINAL_ONLY;
+
+	__invlpgb(0, pcid, addr, nr - 1, pmd_stride, flags);
 }
 
 /* Flush all mappings for a given PCID, not including globals. */
arch/x86/mm/tlb.c +7 -6
index 41ac2d121d76..bef0811dfa27 100644
@@ -502,9 +502,10 @@ static inline void tlbsync(void)
 
 static inline void invlpgb_flush_user_nr_nosync(unsigned long pcid,
 						unsigned long addr,
-						u16 nr, bool pmd_stride)
+						u16 nr, bool pmd_stride,
+						bool freed_tables)
 {
-	__invlpgb_flush_user_nr_nosync(pcid, addr, nr, pmd_stride);
+	__invlpgb_flush_user_nr_nosync(pcid, addr, nr, pmd_stride, freed_tables);
 	if (!this_cpu_read(cpu_tlbstate.need_tlbsync))
 		this_cpu_write(cpu_tlbstate.need_tlbsync, true);
 }
@@ -553,10 +554,10 @@ static void broadcast_tlb_flush(struct flush_tlb_info *info)
 		nr = min(maxnr, (info->end - addr) >> info->stride_shift);
 		nr = max(nr, 1);
 
-		invlpgb_flush_user_nr_nosync(kern_pcid(asid), addr, nr, pmd);
+		invlpgb_flush_user_nr_nosync(kern_pcid(asid), addr, nr, pmd, info->freed_tables);
 		/* Do any CPUs supporting INVLPGB need PTI? */
 		if (static_cpu_has(X86_FEATURE_PTI))
-			invlpgb_flush_user_nr_nosync(user_pcid(asid), addr, nr, pmd);
+			invlpgb_flush_user_nr_nosync(user_pcid(asid), addr, nr, pmd, info->freed_tables);
 
 		addr += nr << info->stride_shift;
 	} while (addr < info->end);
@@ -1704,10 +1705,10 @@ void arch_tlbbatch_add_pending(struct arch_tlbflush_unmap_batch *batch,
 	u16 asid = mm_global_asid(mm);
 
 	if (asid) {
-		invlpgb_flush_user_nr_nosync(kern_pcid(asid), uaddr, 1, false);
+		invlpgb_flush_user_nr_nosync(kern_pcid(asid), uaddr, 1, false, false);
 		/* Do any CPUs supporting INVLPGB need PTI? */
 		if (static_cpu_has(X86_FEATURE_PTI))
-			invlpgb_flush_user_nr_nosync(user_pcid(asid), uaddr, 1, false);
+			invlpgb_flush_user_nr_nosync(user_pcid(asid), uaddr, 1, false, false);
 
 		/*
 		 * Some CPUs might still be using a local ASID for this
Signature · 1 line
2.47.1
maintainer
#19 lore raw
In reply to Rik van Riel
Quoted 1 line · Rik van Riel

Add support for broadcast TLB invalidation using AMD's INVLPGB instruction.

What tree are these patches against? I can't seem to cleanly apply them to anything much :/

#20 lore raw

They apply to 60675d4ca1ef0 ("Merge branch 'linus' into x86/mm, to pick up fixes").

Rik, can I refer you to the BASE TREE INFORMATION section of man git-format-patch. I haven't used that feature lately (b4 takes care of this) but it looks like --base=auto will add the necessary info, or IIRC there's a way to make that behaviour the default.

Quoted 6 lines · Peter Zijlstra
Quoted 1 line · Rik van Riel

Add support for broadcast TLB invalidation using AMD's INVLPGB instruction.

What tree are these patches against? I can't seem to cleanly apply them to anything much :/

maintainer
#21 lore raw
Quoted 2 lines · Brendan Jackman

They apply to 60675d4ca1ef0 ("Merge branch 'linus' into x86/mm, to pick up fixes").

Why some random commit? Shouldn't this be against a sensible branch or something, like perhaps tip/x86/mm ?

#22 lore raw
Quoted 8 lines · Peter Zijlstra
Quoted 2 lines · Brendan Jackman

They apply to 60675d4ca1ef0 ("Merge branch 'linus' into x86/mm, to pick up fixes").

Why some random commit? Shouldn't this be against a sensible branch or
something, like perhaps tip/x86/mm ?

Let me rebase these against current tip/x86/mm
and apply the latest suggested cleanups!

Signature · 1 line
All Rights Reversed.
maintainer
#23 lore raw
Quoted 2 lines · Rik van Riel

Let me rebase these against current tip/x86/mm
and apply the latest suggested cleanups!

Please use tip/master - this is and has always been the branch tip patches should be applied on.

Thx.

Signature · 4 lines
Regards/Gruss,
    Boris.

https://people.kernel.org/tglx/notes-about-netiquette
#24 lore raw
In reply to Brendan Jackman
Quoted 7 lines · Brendan Jackman

They apply to 60675d4ca1ef0 ("Merge branch 'linus' into x86/mm, to pick up fixes").

Rik, can I refer you to the BASE TREE INFORMATION section of man git-format-patch. I haven't used that feature lately (b4 takes care of this) but it looks like --base=auto will add the necessary info, or IIRC there's a way to make that behaviour the default.

IMO, --base=auto is too easy to unintentionally misuse, e.g. it will do the wrong thing if your upstream branch is set to a personal repository. --base itself is fantastic though. I personally do:

  git format-patch --base=HEAD~$nr <bunch of other stuff> -$nr

where $nr is the number of patches in the series. I.e. advertise the base as whatever the series of patches is based on, not what the branch is based on. The only time it doesn't work is if your local branch has a commit that is not in the series, and is not publicly visible. E.g. if the series depends on another in-flight series that you've applied locally. But in that case, you should be explaining what's up in your cover letter no matter what.

Tested-by on the v10 series
#25 lore raw
In reply to Rik van Riel

From: [email protected] <[email protected]> Sent: Tuesday, February 11, 2025 1:08 PM

Quoted 72 lines

Add support for broadcast TLB invalidation using AMD's INVLPGB instruction.

This allows the kernel to invalidate TLB entries on remote CPUs without needing to send IPIs, without having to wait for remote CPUs to handle those interrupts, and with less interruption to what was running on those CPUs.

Because x86 PCID space is limited, and there are some very large systems out there, broadcast TLB invalidation is only used for processes that are active on 3 or more CPUs, with the threshold being gradually increased the more the PCID space gets exhausted.

Combined with the removal of unnecessary lru_add_drain calls (see https://lkml.org/lkml/2024/12/19/1388) this results in a nice performance boost for the will-it-scale tlb_flush2_threads test on an AMD Milan system with 36 cores:

- vanilla kernel:           527k loops/second
- lru_add_drain removal:    731k loops/second
- only INVLPGB:             527k loops/second

- lru_add_drain + INVLPGB: 1157k loops/second

Profiling with only the INVLPGB changes showed while TLB invalidation went down from 40% of the total CPU time to only around 4% of CPU time, the contention simply moved to the LRU lock.

Fixing both at the same time about doubles the
number of iterations per second from this case.

Some numbers closer to real world performance
can be found at Phoronix, thanks to Michael:

https://www.phoronix.com/news/AMD-INVLPGB-Linux-Benefits

My current plan is to implement support for Intel's RAR (Remote Action Request) TLB flushing in a follow-up series, after this thing has been merged into -tip. Making things any larger would just be unwieldy for reviewers.

v10:
- simplify partial pages with min(nr, 1) in the invlpgb loop (Peter)
- document x86 paravirt, AMD invlpgb, and ARM64 flush without IPI (Brendan)
- remove IS_ENABLED(CONFIG_X86_BROADCAST_TLB_FLUSH) (Brendan)
- various cleanups (Brendan)
v9:
- print warning when start or end address was rounded (Peter)
- in the reclaim code, tlbsync at context switch time (Peter)
- fix !CONFIG_CPU_SUP_AMD compile error in arch_tlbbatch_add_pending (Jan)
v8:
- round start & end to handle non-page-aligned callers (Steven & Jan)
- fix up changelog & add tested-by tags (Manali)
v7:
- a few small code cleanups (Nadav)
- fix spurious VM_WARN_ON_ONCE in mm_global_asid
- code simplifications & better barriers (Peter & Dave)
v6:
- fix info->end check in flush_tlb_kernel_range (Michael)
- disable broadcast TLB flushing on 32 bit x86
v5:
- use byte assembly for compatibility with older toolchains (Borislav, Michael)
- ensure a panic on an invalid number of extra pages (Dave, Tom)
- add cant_migrate() assertion to tlbsync (Jann)
- a bunch more cleanups (Nadav)
- key TCE enabling off X86_FEATURE_TCE (Andrew)
- fix a race between reclaim and ASID transition (Jann)
v4:
- Use only bitmaps to track free global ASIDs (Nadav)
- Improved AMD initialization (Borislav & Tom)
- Various naming and documentation improvements (Peter, Nadav, Tom, Dave)
- Fixes for subtle race conditions (Jann)

Quoted 9 lines

v3:
- Remove paravirt tlb_remove_table call (thank you Qi Zheng)
- More suggested cleanups and changelog fixes by Peter and Nadav
v2:
- Apply suggestions by Peter and Borislav (thank you!)
- Fix bug in arch_tlbbatch_flush, where we need to do both the TLBSYNC, and flush the CPUs that are in the cpumask.
- Some updates to comments and changelogs based on questions.

Tested this series in an Azure Confidential VM based on SEV-SNP, which is running on Hyper-V and exposes INVLPGB in the guest VM. I applied the patches to 6.13.0 with one minor fixup, but did not include the patch to remove unnecessary lru_add_drain calls. I also added some custom telemetry to see when INVLPGB is being used vs. Hyper-V's paravirt hypercalls for TLB flushing.

I did not see any problems. The custom telemetry looked about as I expected, showing a mix of INVLPGB and the PV hypercalls. So for the series: