[PATCH AUTOSEL 6.1 1/3] uprobes: sanitiize xol_free_insn_slot()

[PATCH AUTOSEL 6.1 1/3] uprobes: sanitiize xol_free_insn_slot()

[Sasha Levin]

From: Oleg Nesterov <oleg@redhat.com>

[ Upstream commit c7b4133c48445dde789ed30b19ccb0448c7593f7 ]

1. Clear utask->xol_vaddr unconditionally, even if this addr is not valid,
   xol_free_insn_slot() should never return with utask->xol_vaddr != NULL.

2. Add a comment to explain why do we need to validate slot_addr.

3. Simplify the validation above. We can simply check offset < PAGE_SIZE,
   unsigned underflows are fine, it should work if slot_addr < area->vaddr.

4. Kill the unnecessary "slot_nr >= UINSNS_PER_PAGE" check, slot_nr must
   be valid if offset < PAGE_SIZE.

The next patches will cleanup this function even more.

Signed-off-by: Oleg Nesterov <oleg@redhat.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lore.kernel.org/r/20240929144235.GA9471@redhat.com
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 kernel/events/uprobes.c | 21 +++++++++------------
 1 file changed, 9 insertions(+), 12 deletions(-)

diff --git a/kernel/events/uprobes.c b/kernel/events/uprobes.c
index 9ee25351cecac..e3c616085137e 100644
--- a/kernel/events/uprobes.c
+++ b/kernel/events/uprobes.c
@@ -1632,8 +1632,8 @@ static unsigned long xol_get_insn_slot(struct uprobe *uprobe)
 static void xol_free_insn_slot(struct task_struct *tsk)
 {
 	struct xol_area *area;
-	unsigned long vma_end;
 	unsigned long slot_addr;
+	unsigned long offset;
 
 	if (!tsk->mm || !tsk->mm->uprobes_state.xol_area || !tsk->utask)
 		return;
@@ -1642,24 +1642,21 @@ static void xol_free_insn_slot(struct task_struct *tsk)
 	if (unlikely(!slot_addr))
 		return;
 
+	tsk->utask->xol_vaddr = 0;
 	area = tsk->mm->uprobes_state.xol_area;
-	vma_end = area->vaddr + PAGE_SIZE;
-	if (area->vaddr <= slot_addr && slot_addr < vma_end) {
-		unsigned long offset;
-		int slot_nr;
-
-		offset = slot_addr - area->vaddr;
-		slot_nr = offset / UPROBE_XOL_SLOT_BYTES;
-		if (slot_nr >= UINSNS_PER_PAGE)
-			return;
+	offset = slot_addr - area->vaddr;
+	/*
+	 * slot_addr must fit into [area->vaddr, area->vaddr + PAGE_SIZE).
+	 * This check can only fail if the "[uprobes]" vma was mremap'ed.
+	 */
+	if (offset < PAGE_SIZE) {
+		int slot_nr = offset / UPROBE_XOL_SLOT_BYTES;
 
 		clear_bit(slot_nr, area->bitmap);
 		atomic_dec(&area->slot_count);
 		smp_mb__after_atomic(); /* pairs with prepare_to_wait() */
 		if (waitqueue_active(&area->wq))
 			wake_up(&area->wq);
-
-		tsk->utask->xol_vaddr = 0;
 	}
 }
 
-- 
2.43.0

[PATCH AUTOSEL 6.1 2/3] perf/x86/amd: Warn only on new bits set

[Sasha Levin]

From: Breno Leitao <leitao@debian.org>

[ Upstream commit de20037e1b3c2f2ca97b8c12b8c7bca8abd509a7 ]

Warning at every leaking bits can cause a flood of message, triggering
various stall-warning mechanisms to fire, including CSD locks, which
makes the machine to be unusable.

Track the bits that are being leaked, and only warn when a new bit is
set.

That said, this patch will help with the following issues:

1) It will tell us which bits are being set, so, it is easy to
   communicate it back to vendor, and to do a root-cause analyzes.

2) It avoid the machine to be unusable, because, worst case
   scenario, the user gets less than 60 WARNs (one per unhandled bit).

Suggested-by: Paul E. McKenney <paulmck@kernel.org>
Signed-off-by: Breno Leitao <leitao@debian.org>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Reviewed-by: Sandipan Das <sandipan.das@amd.com>
Reviewed-by: Paul E. McKenney <paulmck@kernel.org>
Link: https://lkml.kernel.org/r/20241001141020.2620361-1-leitao@debian.org
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 arch/x86/events/amd/core.c | 10 ++++++++--
 1 file changed, 8 insertions(+), 2 deletions(-)

diff --git a/arch/x86/events/amd/core.c b/arch/x86/events/amd/core.c
index 3ac069a4559b0..1282f1a702139 100644
--- a/arch/x86/events/amd/core.c
+++ b/arch/x86/events/amd/core.c
@@ -881,11 +881,12 @@ static int amd_pmu_handle_irq(struct pt_regs *regs)
 static int amd_pmu_v2_handle_irq(struct pt_regs *regs)
 {
 	struct cpu_hw_events *cpuc = this_cpu_ptr(&cpu_hw_events);
+	static atomic64_t status_warned = ATOMIC64_INIT(0);
+	u64 reserved, status, mask, new_bits, prev_bits;
 	struct perf_sample_data data;
 	struct hw_perf_event *hwc;
 	struct perf_event *event;
 	int handled = 0, idx;
-	u64 reserved, status, mask;
 	bool pmu_enabled;
 
 	/*
@@ -952,7 +953,12 @@ static int amd_pmu_v2_handle_irq(struct pt_regs *regs)
 	 * the corresponding PMCs are expected to be inactive according to the
 	 * active_mask
 	 */
-	WARN_ON(status > 0);
+	if (status > 0) {
+		prev_bits = atomic64_fetch_or(status, &status_warned);
+		// A new bit was set for the very first time.
+		new_bits = status & ~prev_bits;
+		WARN(new_bits, "New overflows for inactive PMCs: %llx\n", new_bits);
+	}
 
 	/* Clear overflow and freeze bits */
 	amd_pmu_ack_global_status(~status);
-- 
2.43.0

[PATCH AUTOSEL 6.1 3/3] timekeeping: Always check for negative motion

[Sasha Levin]

From: Thomas Gleixner <tglx@linutronix.de>

[ Upstream commit c163e40af9b2331b2c629fd4ec8b703ed4d4ae39 ]

clocksource_delta() has two variants. One with a check for negative motion,
which is only selected by x86. This is a historic leftover as this function
was previously used in the time getter hot paths.

Since 135225a363ae timekeeping_cycles_to_ns() has unconditional protection
against this as a by-product of the protection against 64bit math overflow.

clocksource_delta() is only used in the clocksource watchdog and in
timekeeping_advance(). The extra conditional there is not hurting anyone.

Remove the config option and unconditionally prevent negative motion of the
readout.

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Acked-by: John Stultz <jstultz@google.com>
Link: https://lore.kernel.org/all/20241031120328.599430157@linutronix.de
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 arch/x86/Kconfig                   | 1 -
 kernel/time/Kconfig                | 5 -----
 kernel/time/timekeeping_internal.h | 7 -------
 3 files changed, 13 deletions(-)

diff --git a/arch/x86/Kconfig b/arch/x86/Kconfig
index 49cea5b81649d..c75e18346c89b 100644
--- a/arch/x86/Kconfig
+++ b/arch/x86/Kconfig
@@ -131,7 +131,6 @@ config X86
 	select ARCH_HAS_PARANOID_L1D_FLUSH
 	select BUILDTIME_TABLE_SORT
 	select CLKEVT_I8253
-	select CLOCKSOURCE_VALIDATE_LAST_CYCLE
 	select CLOCKSOURCE_WATCHDOG
 	# Word-size accesses may read uninitialized data past the trailing \0
 	# in strings and cause false KMSAN reports.
diff --git a/kernel/time/Kconfig b/kernel/time/Kconfig
index a41753be1a2bf..6ce5b66643f0b 100644
--- a/kernel/time/Kconfig
+++ b/kernel/time/Kconfig
@@ -17,11 +17,6 @@ config ARCH_CLOCKSOURCE_DATA
 config ARCH_CLOCKSOURCE_INIT
 	bool
 
-# Clocksources require validation of the clocksource against the last
-# cycle update - x86/TSC misfeature
-config CLOCKSOURCE_VALIDATE_LAST_CYCLE
-	bool
-
 # Timekeeping vsyscall support
 config GENERIC_TIME_VSYSCALL
 	bool
diff --git a/kernel/time/timekeeping_internal.h b/kernel/time/timekeeping_internal.h
index 4ca2787d1642e..1d4854d5c386e 100644
--- a/kernel/time/timekeeping_internal.h
+++ b/kernel/time/timekeeping_internal.h
@@ -15,7 +15,6 @@ extern void tk_debug_account_sleep_time(const struct timespec64 *t);
 #define tk_debug_account_sleep_time(x)
 #endif
 
-#ifdef CONFIG_CLOCKSOURCE_VALIDATE_LAST_CYCLE
 static inline u64 clocksource_delta(u64 now, u64 last, u64 mask)
 {
 	u64 ret = (now - last) & mask;
@@ -26,12 +25,6 @@ static inline u64 clocksource_delta(u64 now, u64 last, u64 mask)
 	 */
 	return ret & ~(mask >> 1) ? 0 : ret;
 }
-#else
-static inline u64 clocksource_delta(u64 now, u64 last, u64 mask)
-{
-	return (now - last) & mask;
-}
-#endif
 
 /* Semi public for serialization of non timekeeper VDSO updates. */
 extern raw_spinlock_t timekeeper_lock;
-- 
2.43.0