November 2024 - Linux-stable-mirror

[PATCH AUTOSEL 6.11 01/20] gpio: free irqs that are still requested when the chip is being removed

by Sasha Levin

From: Bartosz Golaszewski <bartosz.golaszewski(a)linaro.org> [ Upstream commit ec8b6f55b98146c41dcf15e8189eb43291e35e89 ] If we remove a GPIO chip that is also an interrupt controller with users not having freed some interrupts, we'll end up leaking resources as indicated by the following warning: remove_proc_entry: removing non-empty directory 'irq/30', leaking at least 'gpio' As there's no way of notifying interrupt users about the irqchip going away and the interrupt subsystem is not plugged into the driver model and so not all cases can be handled by devlinks, we need to make sure to free all interrupts before the complete the removal of the provider. Reviewed-by: Herve Codina <herve.codina(a)bootlin.com> Tested-by: Herve Codina <herve.codina(a)bootlin.com> Link: https://lore.kernel.org/r/20240919135104.3583-1-brgl@bgdev.pl Signed-off-by: Bartosz Golaszewski <bartosz.golaszewski(a)linaro.org> Signed-off-by: Sasha Levin <sashal(a)kernel.org> --- drivers/gpio/gpiolib.c | 41 +++++++++++++++++++++++++++++++++++++++++ 1 file changed, 41 insertions(+) diff --git a/drivers/gpio/gpiolib.c b/drivers/gpio/gpiolib.c index 337971080dfde..206757d155ef9 100644 --- a/drivers/gpio/gpiolib.c +++ b/drivers/gpio/gpiolib.c @@ -14,6 +14,7 @@ #include <linux/idr.h> #include <linux/interrupt.h> #include <linux/irq.h> +#include <linux/irqdesc.h> #include <linux/kernel.h> #include <linux/list.h> #include <linux/lockdep.h> @@ -710,6 +711,45 @@ bool gpiochip_line_is_valid(const struct gpio_chip *gc, } EXPORT_SYMBOL_GPL(gpiochip_line_is_valid); +static void gpiod_free_irqs(struct gpio_desc *desc) +{ + int irq = gpiod_to_irq(desc); + struct irq_desc *irqd = irq_to_desc(irq); + void *cookie; + + for (;;) { + /* + * Make sure the action doesn't go away while we're + * dereferencing it. Retrieve and store the cookie value. + * If the irq is freed after we release the lock, that's + * alright - the underlying maple tree lookup will return NULL + * and nothing will happen in free_irq(). + */ + scoped_guard(mutex, &irqd->request_mutex) { + if (!irq_desc_has_action(irqd)) + return; + + cookie = irqd->action->dev_id; + } + + free_irq(irq, cookie); + } +} + +/* + * The chip is going away but there may be users who had requested interrupts + * on its GPIO lines who have no idea about its removal and have no way of + * being notified about it. We need to free any interrupts still in use here or + * we'll leak memory and resources (like procfs files). + */ +static void gpiochip_free_remaining_irqs(struct gpio_chip *gc) +{ + struct gpio_desc *desc; + + for_each_gpio_desc_with_flag(gc, desc, FLAG_USED_AS_IRQ) + gpiod_free_irqs(desc); +} + static void gpiodev_release(struct device *dev) { struct gpio_device *gdev = to_gpio_device(dev); @@ -1122,6 +1162,7 @@ void gpiochip_remove(struct gpio_chip *gc) /* FIXME: should the legacy sysfs handling be moved to gpio_device? */ gpiochip_sysfs_unregister(gdev); gpiochip_free_hogs(gc); + gpiochip_free_remaining_irqs(gc); scoped_guard(mutex, &gpio_devices_lock) list_del_rcu(&gdev->list); -- 2.43.0

9 months, 3 weeks

1
19
0 0

[PATCH AUTOSEL 4.19] uprobes: sanitiize xol_free_insn_slot()

by Sasha Levin

From: Oleg Nesterov <oleg(a)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(a)redhat.com> Signed-off-by: Peter Zijlstra (Intel) <peterz(a)infradead.org> Link: https://lore.kernel.org/r/20240929144235.GA9471@redhat.com Signed-off-by: Sasha Levin <sashal(a)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 3ca91daddc9f5..e4156acc004d3 100644 --- a/kernel/events/uprobes.c +++ b/kernel/events/uprobes.c @@ -1332,8 +1332,8 @@ 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; @@ -1342,24 +1342,21 @@ 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

9 months, 3 weeks

1
0
0 0

[PATCH AUTOSEL 5.4 1/2] uprobes: sanitiize xol_free_insn_slot()

by Sasha Levin

From: Oleg Nesterov <oleg(a)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(a)redhat.com> Signed-off-by: Peter Zijlstra (Intel) <peterz(a)infradead.org> Link: https://lore.kernel.org/r/20240929144235.GA9471@redhat.com Signed-off-by: Sasha Levin <sashal(a)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 6285674412f25..d338eeb30145b 100644 --- a/kernel/events/uprobes.c +++ b/kernel/events/uprobes.c @@ -1640,8 +1640,8 @@ 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; @@ -1650,24 +1650,21 @@ 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

9 months, 3 weeks

1
1
0 0

[PATCH] iio: imu: inv_icm42600: fix timestamps after suspend if sensor is on

by Jean-Baptiste Maneyrol via B4 Relay

From: Jean-Baptiste Maneyrol <jean-baptiste.maneyrol(a)tdk.com> Currently suspending while sensors are one will result in timestamping continuing without gap at resume. It can work with monotonic clock but not with other clocks. Fix that by resetting timestamping. Fixes: ec74ae9fd37c ("iio: imu: inv_icm42600: add accurate timestamping") Cc: stable(a)vger.kernel.org Signed-off-by: Jean-Baptiste Maneyrol <jean-baptiste.maneyrol(a)tdk.com> --- drivers/iio/imu/inv_icm42600/inv_icm42600_core.c | 7 ++++++- 1 file changed, 6 insertions(+), 1 deletion(-) diff --git a/drivers/iio/imu/inv_icm42600/inv_icm42600_core.c b/drivers/iio/imu/inv_icm42600/inv_icm42600_core.c index 93b5d7a3339ccff16b21bf6c40ed7b2311317cf4..03139e2e4eddbf37e154de2eb486549bc3bdb284 100644 --- a/drivers/iio/imu/inv_icm42600/inv_icm42600_core.c +++ b/drivers/iio/imu/inv_icm42600/inv_icm42600_core.c @@ -814,6 +814,8 @@ static int inv_icm42600_suspend(struct device *dev) static int inv_icm42600_resume(struct device *dev) { struct inv_icm42600_state *st = dev_get_drvdata(dev); + struct inv_icm42600_sensor_state *gyro_st = iio_priv(st->indio_gyro); + struct inv_icm42600_sensor_state *accel_st = iio_priv(st->indio_accel); int ret; mutex_lock(&st->lock); @@ -834,9 +836,12 @@ static int inv_icm42600_resume(struct device *dev) goto out_unlock; /* restore FIFO data streaming */ - if (st->fifo.on) + if (st->fifo.on) { + inv_sensors_timestamp_reset(&gyro_st->ts); + inv_sensors_timestamp_reset(&accel_st->ts); ret = regmap_write(st->map, INV_ICM42600_REG_FIFO_CONFIG, INV_ICM42600_FIFO_CONFIG_STREAM); + } out_unlock: mutex_unlock(&st->lock); --- base-commit: 9dd2270ca0b38ee16094817f4a53e7ba78e31567 change-id: 20241113-inv_icm42600-fix-timestamps-after-suspend-b7e421eaa40c Best regards, -- Jean-Baptiste Maneyrol <jean-baptiste.maneyrol(a)tdk.com>

9 months, 3 weeks

2
1
0 0

[PATCH AUTOSEL 5.15 1/2] uprobes: sanitiize xol_free_insn_slot()

by Sasha Levin

From: Oleg Nesterov <oleg(a)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(a)redhat.com> Signed-off-by: Peter Zijlstra (Intel) <peterz(a)infradead.org> Link: https://lore.kernel.org/r/20240929144235.GA9471@redhat.com Signed-off-by: Sasha Levin <sashal(a)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 b37a6bde8a915..a0f846b2ac1ea 100644 --- a/kernel/events/uprobes.c +++ b/kernel/events/uprobes.c @@ -1633,8 +1633,8 @@ 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; @@ -1643,24 +1643,21 @@ 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

9 months, 3 weeks

1
1
0 0

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

by Sasha Levin

From: Oleg Nesterov <oleg(a)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(a)redhat.com> Signed-off-by: Peter Zijlstra (Intel) <peterz(a)infradead.org> Link: https://lore.kernel.org/r/20240929144235.GA9471@redhat.com Signed-off-by: Sasha Levin <sashal(a)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

9 months, 3 weeks

1
2
0 0

[PATCH AUTOSEL 5.10] kcsan: Turn report_filterlist_lock into a raw_spinlock

by Sasha Levin

From: Marco Elver <elver(a)google.com> [ Upstream commit 59458fa4ddb47e7891c61b4a928d13d5f5b00aa0 ] Ran Xiaokai reports that with a KCSAN-enabled PREEMPT_RT kernel, we can see splats like: | BUG: sleeping function called from invalid context at kernel/locking/spinlock_rt.c:48 | in_atomic(): 1, irqs_disabled(): 1, non_block: 0, pid: 0, name: swapper/1 | preempt_count: 10002, expected: 0 | RCU nest depth: 0, expected: 0 | no locks held by swapper/1/0. | irq event stamp: 156674 | hardirqs last enabled at (156673): [<ffffffff81130bd9>] do_idle+0x1f9/0x240 | hardirqs last disabled at (156674): [<ffffffff82254f84>] sysvec_apic_timer_interrupt+0x14/0xc0 | softirqs last enabled at (0): [<ffffffff81099f47>] copy_process+0xfc7/0x4b60 | softirqs last disabled at (0): [<0000000000000000>] 0x0 | Preemption disabled at: | [<ffffffff814a3e2a>] paint_ptr+0x2a/0x90 | CPU: 1 UID: 0 PID: 0 Comm: swapper/1 Not tainted 6.11.0+ #3 | Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS rel-1.12.0-0-ga698c8995f-prebuilt.qemu.org 04/01/2014 | Call Trace: | <IRQ> | dump_stack_lvl+0x7e/0xc0 | dump_stack+0x1d/0x30 | __might_resched+0x1a2/0x270 | rt_spin_lock+0x68/0x170 | kcsan_skip_report_debugfs+0x43/0xe0 | print_report+0xb5/0x590 | kcsan_report_known_origin+0x1b1/0x1d0 | kcsan_setup_watchpoint+0x348/0x650 | __tsan_unaligned_write1+0x16d/0x1d0 | hrtimer_interrupt+0x3d6/0x430 | __sysvec_apic_timer_interrupt+0xe8/0x3a0 | sysvec_apic_timer_interrupt+0x97/0xc0 | </IRQ> On a detected data race, KCSAN's reporting logic checks if it should filter the report. That list is protected by the report_filterlist_lock *non-raw* spinlock which may sleep on RT kernels. Since KCSAN may report data races in any context, convert it to a raw_spinlock. This requires being careful about when to allocate memory for the filter list itself which can be done via KCSAN's debugfs interface. Concurrent modification of the filter list via debugfs should be rare: the chosen strategy is to optimistically pre-allocate memory before the critical section and discard if unused. Link: https://lore.kernel.org/all/20240925143154.2322926-1-ranxiaokai627@163.com/ Reported-by: Ran Xiaokai <ran.xiaokai(a)zte.com.cn> Tested-by: Ran Xiaokai <ran.xiaokai(a)zte.com.cn> Signed-off-by: Marco Elver <elver(a)google.com> Signed-off-by: Sasha Levin <sashal(a)kernel.org> --- kernel/kcsan/debugfs.c | 74 ++++++++++++++++++++---------------------- 1 file changed, 36 insertions(+), 38 deletions(-) diff --git a/kernel/kcsan/debugfs.c b/kernel/kcsan/debugfs.c index 62a52be8f6ba9..6a4ecd1a6fa5b 100644 --- a/kernel/kcsan/debugfs.c +++ b/kernel/kcsan/debugfs.c @@ -41,14 +41,8 @@ static struct { int used; /* number of elements used */ bool sorted; /* if elements are sorted */ bool whitelist; /* if list is a blacklist or whitelist */ -} report_filterlist = { - .addrs = NULL, - .size = 8, /* small initial size */ - .used = 0, - .sorted = false, - .whitelist = false, /* default is blacklist */ -}; -static DEFINE_SPINLOCK(report_filterlist_lock); +} report_filterlist; +static DEFINE_RAW_SPINLOCK(report_filterlist_lock); /* * The microbenchmark allows benchmarking KCSAN core runtime only. To run @@ -105,7 +99,7 @@ bool kcsan_skip_report_debugfs(unsigned long func_addr) return false; func_addr -= offset; /* Get function start */ - spin_lock_irqsave(&report_filterlist_lock, flags); + raw_spin_lock_irqsave(&report_filterlist_lock, flags); if (report_filterlist.used == 0) goto out; @@ -122,7 +116,7 @@ bool kcsan_skip_report_debugfs(unsigned long func_addr) ret = !ret; out: - spin_unlock_irqrestore(&report_filterlist_lock, flags); + raw_spin_unlock_irqrestore(&report_filterlist_lock, flags); return ret; } @@ -130,9 +124,9 @@ static void set_report_filterlist_whitelist(bool whitelist) { unsigned long flags; - spin_lock_irqsave(&report_filterlist_lock, flags); + raw_spin_lock_irqsave(&report_filterlist_lock, flags); report_filterlist.whitelist = whitelist; - spin_unlock_irqrestore(&report_filterlist_lock, flags); + raw_spin_unlock_irqrestore(&report_filterlist_lock, flags); } /* Returns 0 on success, error-code otherwise. */ @@ -140,6 +134,9 @@ static ssize_t insert_report_filterlist(const char *func) { unsigned long flags; unsigned long addr = kallsyms_lookup_name(func); + unsigned long *delay_free = NULL; + unsigned long *new_addrs = NULL; + size_t new_size = 0; ssize_t ret = 0; if (!addr) { @@ -147,32 +144,33 @@ static ssize_t insert_report_filterlist(const char *func) return -ENOENT; } - spin_lock_irqsave(&report_filterlist_lock, flags); +retry_alloc: + /* + * Check if we need an allocation, and re-validate under the lock. Since + * the report_filterlist_lock is a raw, cannot allocate under the lock. + */ + if (data_race(report_filterlist.used == report_filterlist.size)) { + new_size = (report_filterlist.size ?: 4) * 2; + delay_free = new_addrs = kmalloc_array(new_size, sizeof(unsigned long), GFP_KERNEL); + if (!new_addrs) + return -ENOMEM; + } - if (report_filterlist.addrs == NULL) { - /* initial allocation */ - report_filterlist.addrs = - kmalloc_array(report_filterlist.size, - sizeof(unsigned long), GFP_ATOMIC); - if (report_filterlist.addrs == NULL) { - ret = -ENOMEM; - goto out; - } - } else if (report_filterlist.used == report_filterlist.size) { - /* resize filterlist */ - size_t new_size = report_filterlist.size * 2; - unsigned long *new_addrs = - krealloc(report_filterlist.addrs, - new_size * sizeof(unsigned long), GFP_ATOMIC); - - if (new_addrs == NULL) { - /* leave filterlist itself untouched */ - ret = -ENOMEM; - goto out; + raw_spin_lock_irqsave(&report_filterlist_lock, flags); + if (report_filterlist.used == report_filterlist.size) { + /* Check we pre-allocated enough, and retry if not. */ + if (report_filterlist.used >= new_size) { + raw_spin_unlock_irqrestore(&report_filterlist_lock, flags); + kfree(new_addrs); /* kfree(NULL) is safe */ + delay_free = new_addrs = NULL; + goto retry_alloc; } + if (report_filterlist.used) + memcpy(new_addrs, report_filterlist.addrs, report_filterlist.used * sizeof(unsigned long)); + delay_free = report_filterlist.addrs; /* free the old list */ + report_filterlist.addrs = new_addrs; /* switch to the new list */ report_filterlist.size = new_size; - report_filterlist.addrs = new_addrs; } /* Note: deduplicating should be done in userspace. */ @@ -180,9 +178,9 @@ static ssize_t insert_report_filterlist(const char *func) kallsyms_lookup_name(func); report_filterlist.sorted = false; -out: - spin_unlock_irqrestore(&report_filterlist_lock, flags); + raw_spin_unlock_irqrestore(&report_filterlist_lock, flags); + kfree(delay_free); return ret; } @@ -199,13 +197,13 @@ static int show_info(struct seq_file *file, void *v) } /* show filter functions, and filter type */ - spin_lock_irqsave(&report_filterlist_lock, flags); + raw_spin_lock_irqsave(&report_filterlist_lock, flags); seq_printf(file, "\n%s functions: %s\n", report_filterlist.whitelist ? "whitelisted" : "blacklisted", report_filterlist.used == 0 ? "none" : ""); for (i = 0; i < report_filterlist.used; ++i) seq_printf(file, " %ps\n", (void *)report_filterlist.addrs[i]); - spin_unlock_irqrestore(&report_filterlist_lock, flags); + raw_spin_unlock_irqrestore(&report_filterlist_lock, flags); return 0; } -- 2.43.0

9 months, 3 weeks

1
0
0 0

[PATCH AUTOSEL 5.15] kcsan: Turn report_filterlist_lock into a raw_spinlock

by Sasha Levin

From: Marco Elver <elver(a)google.com> [ Upstream commit 59458fa4ddb47e7891c61b4a928d13d5f5b00aa0 ] Ran Xiaokai reports that with a KCSAN-enabled PREEMPT_RT kernel, we can see splats like: | BUG: sleeping function called from invalid context at kernel/locking/spinlock_rt.c:48 | in_atomic(): 1, irqs_disabled(): 1, non_block: 0, pid: 0, name: swapper/1 | preempt_count: 10002, expected: 0 | RCU nest depth: 0, expected: 0 | no locks held by swapper/1/0. | irq event stamp: 156674 | hardirqs last enabled at (156673): [<ffffffff81130bd9>] do_idle+0x1f9/0x240 | hardirqs last disabled at (156674): [<ffffffff82254f84>] sysvec_apic_timer_interrupt+0x14/0xc0 | softirqs last enabled at (0): [<ffffffff81099f47>] copy_process+0xfc7/0x4b60 | softirqs last disabled at (0): [<0000000000000000>] 0x0 | Preemption disabled at: | [<ffffffff814a3e2a>] paint_ptr+0x2a/0x90 | CPU: 1 UID: 0 PID: 0 Comm: swapper/1 Not tainted 6.11.0+ #3 | Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS rel-1.12.0-0-ga698c8995f-prebuilt.qemu.org 04/01/2014 | Call Trace: | <IRQ> | dump_stack_lvl+0x7e/0xc0 | dump_stack+0x1d/0x30 | __might_resched+0x1a2/0x270 | rt_spin_lock+0x68/0x170 | kcsan_skip_report_debugfs+0x43/0xe0 | print_report+0xb5/0x590 | kcsan_report_known_origin+0x1b1/0x1d0 | kcsan_setup_watchpoint+0x348/0x650 | __tsan_unaligned_write1+0x16d/0x1d0 | hrtimer_interrupt+0x3d6/0x430 | __sysvec_apic_timer_interrupt+0xe8/0x3a0 | sysvec_apic_timer_interrupt+0x97/0xc0 | </IRQ> On a detected data race, KCSAN's reporting logic checks if it should filter the report. That list is protected by the report_filterlist_lock *non-raw* spinlock which may sleep on RT kernels. Since KCSAN may report data races in any context, convert it to a raw_spinlock. This requires being careful about when to allocate memory for the filter list itself which can be done via KCSAN's debugfs interface. Concurrent modification of the filter list via debugfs should be rare: the chosen strategy is to optimistically pre-allocate memory before the critical section and discard if unused. Link: https://lore.kernel.org/all/20240925143154.2322926-1-ranxiaokai627@163.com/ Reported-by: Ran Xiaokai <ran.xiaokai(a)zte.com.cn> Tested-by: Ran Xiaokai <ran.xiaokai(a)zte.com.cn> Signed-off-by: Marco Elver <elver(a)google.com> Signed-off-by: Sasha Levin <sashal(a)kernel.org> --- kernel/kcsan/debugfs.c | 74 ++++++++++++++++++++---------------------- 1 file changed, 36 insertions(+), 38 deletions(-) diff --git a/kernel/kcsan/debugfs.c b/kernel/kcsan/debugfs.c index 1d1d1b0e42489..f4623910fb1f2 100644 --- a/kernel/kcsan/debugfs.c +++ b/kernel/kcsan/debugfs.c @@ -46,14 +46,8 @@ static struct { int used; /* number of elements used */ bool sorted; /* if elements are sorted */ bool whitelist; /* if list is a blacklist or whitelist */ -} report_filterlist = { - .addrs = NULL, - .size = 8, /* small initial size */ - .used = 0, - .sorted = false, - .whitelist = false, /* default is blacklist */ -}; -static DEFINE_SPINLOCK(report_filterlist_lock); +} report_filterlist; +static DEFINE_RAW_SPINLOCK(report_filterlist_lock); /* * The microbenchmark allows benchmarking KCSAN core runtime only. To run @@ -110,7 +104,7 @@ bool kcsan_skip_report_debugfs(unsigned long func_addr) return false; func_addr -= offset; /* Get function start */ - spin_lock_irqsave(&report_filterlist_lock, flags); + raw_spin_lock_irqsave(&report_filterlist_lock, flags); if (report_filterlist.used == 0) goto out; @@ -127,7 +121,7 @@ bool kcsan_skip_report_debugfs(unsigned long func_addr) ret = !ret; out: - spin_unlock_irqrestore(&report_filterlist_lock, flags); + raw_spin_unlock_irqrestore(&report_filterlist_lock, flags); return ret; } @@ -135,9 +129,9 @@ static void set_report_filterlist_whitelist(bool whitelist) { unsigned long flags; - spin_lock_irqsave(&report_filterlist_lock, flags); + raw_spin_lock_irqsave(&report_filterlist_lock, flags); report_filterlist.whitelist = whitelist; - spin_unlock_irqrestore(&report_filterlist_lock, flags); + raw_spin_unlock_irqrestore(&report_filterlist_lock, flags); } /* Returns 0 on success, error-code otherwise. */ @@ -145,6 +139,9 @@ static ssize_t insert_report_filterlist(const char *func) { unsigned long flags; unsigned long addr = kallsyms_lookup_name(func); + unsigned long *delay_free = NULL; + unsigned long *new_addrs = NULL; + size_t new_size = 0; ssize_t ret = 0; if (!addr) { @@ -152,32 +149,33 @@ static ssize_t insert_report_filterlist(const char *func) return -ENOENT; } - spin_lock_irqsave(&report_filterlist_lock, flags); +retry_alloc: + /* + * Check if we need an allocation, and re-validate under the lock. Since + * the report_filterlist_lock is a raw, cannot allocate under the lock. + */ + if (data_race(report_filterlist.used == report_filterlist.size)) { + new_size = (report_filterlist.size ?: 4) * 2; + delay_free = new_addrs = kmalloc_array(new_size, sizeof(unsigned long), GFP_KERNEL); + if (!new_addrs) + return -ENOMEM; + } - if (report_filterlist.addrs == NULL) { - /* initial allocation */ - report_filterlist.addrs = - kmalloc_array(report_filterlist.size, - sizeof(unsigned long), GFP_ATOMIC); - if (report_filterlist.addrs == NULL) { - ret = -ENOMEM; - goto out; - } - } else if (report_filterlist.used == report_filterlist.size) { - /* resize filterlist */ - size_t new_size = report_filterlist.size * 2; - unsigned long *new_addrs = - krealloc(report_filterlist.addrs, - new_size * sizeof(unsigned long), GFP_ATOMIC); - - if (new_addrs == NULL) { - /* leave filterlist itself untouched */ - ret = -ENOMEM; - goto out; + raw_spin_lock_irqsave(&report_filterlist_lock, flags); + if (report_filterlist.used == report_filterlist.size) { + /* Check we pre-allocated enough, and retry if not. */ + if (report_filterlist.used >= new_size) { + raw_spin_unlock_irqrestore(&report_filterlist_lock, flags); + kfree(new_addrs); /* kfree(NULL) is safe */ + delay_free = new_addrs = NULL; + goto retry_alloc; } + if (report_filterlist.used) + memcpy(new_addrs, report_filterlist.addrs, report_filterlist.used * sizeof(unsigned long)); + delay_free = report_filterlist.addrs; /* free the old list */ + report_filterlist.addrs = new_addrs; /* switch to the new list */ report_filterlist.size = new_size; - report_filterlist.addrs = new_addrs; } /* Note: deduplicating should be done in userspace. */ @@ -185,9 +183,9 @@ static ssize_t insert_report_filterlist(const char *func) kallsyms_lookup_name(func); report_filterlist.sorted = false; -out: - spin_unlock_irqrestore(&report_filterlist_lock, flags); + raw_spin_unlock_irqrestore(&report_filterlist_lock, flags); + kfree(delay_free); return ret; } @@ -204,13 +202,13 @@ static int show_info(struct seq_file *file, void *v) } /* show filter functions, and filter type */ - spin_lock_irqsave(&report_filterlist_lock, flags); + raw_spin_lock_irqsave(&report_filterlist_lock, flags); seq_printf(file, "\n%s functions: %s\n", report_filterlist.whitelist ? "whitelisted" : "blacklisted", report_filterlist.used == 0 ? "none" : ""); for (i = 0; i < report_filterlist.used; ++i) seq_printf(file, " %ps\n", (void *)report_filterlist.addrs[i]); - spin_unlock_irqrestore(&report_filterlist_lock, flags); + raw_spin_unlock_irqrestore(&report_filterlist_lock, flags); return 0; } -- 2.43.0

9 months, 3 weeks

1
0
0 0

[PATCH AUTOSEL 6.1] kcsan: Turn report_filterlist_lock into a raw_spinlock

by Sasha Levin

From: Marco Elver <elver(a)google.com> [ Upstream commit 59458fa4ddb47e7891c61b4a928d13d5f5b00aa0 ] Ran Xiaokai reports that with a KCSAN-enabled PREEMPT_RT kernel, we can see splats like: | BUG: sleeping function called from invalid context at kernel/locking/spinlock_rt.c:48 | in_atomic(): 1, irqs_disabled(): 1, non_block: 0, pid: 0, name: swapper/1 | preempt_count: 10002, expected: 0 | RCU nest depth: 0, expected: 0 | no locks held by swapper/1/0. | irq event stamp: 156674 | hardirqs last enabled at (156673): [<ffffffff81130bd9>] do_idle+0x1f9/0x240 | hardirqs last disabled at (156674): [<ffffffff82254f84>] sysvec_apic_timer_interrupt+0x14/0xc0 | softirqs last enabled at (0): [<ffffffff81099f47>] copy_process+0xfc7/0x4b60 | softirqs last disabled at (0): [<0000000000000000>] 0x0 | Preemption disabled at: | [<ffffffff814a3e2a>] paint_ptr+0x2a/0x90 | CPU: 1 UID: 0 PID: 0 Comm: swapper/1 Not tainted 6.11.0+ #3 | Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS rel-1.12.0-0-ga698c8995f-prebuilt.qemu.org 04/01/2014 | Call Trace: | <IRQ> | dump_stack_lvl+0x7e/0xc0 | dump_stack+0x1d/0x30 | __might_resched+0x1a2/0x270 | rt_spin_lock+0x68/0x170 | kcsan_skip_report_debugfs+0x43/0xe0 | print_report+0xb5/0x590 | kcsan_report_known_origin+0x1b1/0x1d0 | kcsan_setup_watchpoint+0x348/0x650 | __tsan_unaligned_write1+0x16d/0x1d0 | hrtimer_interrupt+0x3d6/0x430 | __sysvec_apic_timer_interrupt+0xe8/0x3a0 | sysvec_apic_timer_interrupt+0x97/0xc0 | </IRQ> On a detected data race, KCSAN's reporting logic checks if it should filter the report. That list is protected by the report_filterlist_lock *non-raw* spinlock which may sleep on RT kernels. Since KCSAN may report data races in any context, convert it to a raw_spinlock. This requires being careful about when to allocate memory for the filter list itself which can be done via KCSAN's debugfs interface. Concurrent modification of the filter list via debugfs should be rare: the chosen strategy is to optimistically pre-allocate memory before the critical section and discard if unused. Link: https://lore.kernel.org/all/20240925143154.2322926-1-ranxiaokai627@163.com/ Reported-by: Ran Xiaokai <ran.xiaokai(a)zte.com.cn> Tested-by: Ran Xiaokai <ran.xiaokai(a)zte.com.cn> Signed-off-by: Marco Elver <elver(a)google.com> Signed-off-by: Sasha Levin <sashal(a)kernel.org> --- kernel/kcsan/debugfs.c | 74 ++++++++++++++++++++---------------------- 1 file changed, 36 insertions(+), 38 deletions(-) diff --git a/kernel/kcsan/debugfs.c b/kernel/kcsan/debugfs.c index 1d1d1b0e42489..f4623910fb1f2 100644 --- a/kernel/kcsan/debugfs.c +++ b/kernel/kcsan/debugfs.c @@ -46,14 +46,8 @@ static struct { int used; /* number of elements used */ bool sorted; /* if elements are sorted */ bool whitelist; /* if list is a blacklist or whitelist */ -} report_filterlist = { - .addrs = NULL, - .size = 8, /* small initial size */ - .used = 0, - .sorted = false, - .whitelist = false, /* default is blacklist */ -}; -static DEFINE_SPINLOCK(report_filterlist_lock); +} report_filterlist; +static DEFINE_RAW_SPINLOCK(report_filterlist_lock); /* * The microbenchmark allows benchmarking KCSAN core runtime only. To run @@ -110,7 +104,7 @@ bool kcsan_skip_report_debugfs(unsigned long func_addr) return false; func_addr -= offset; /* Get function start */ - spin_lock_irqsave(&report_filterlist_lock, flags); + raw_spin_lock_irqsave(&report_filterlist_lock, flags); if (report_filterlist.used == 0) goto out; @@ -127,7 +121,7 @@ bool kcsan_skip_report_debugfs(unsigned long func_addr) ret = !ret; out: - spin_unlock_irqrestore(&report_filterlist_lock, flags); + raw_spin_unlock_irqrestore(&report_filterlist_lock, flags); return ret; } @@ -135,9 +129,9 @@ static void set_report_filterlist_whitelist(bool whitelist) { unsigned long flags; - spin_lock_irqsave(&report_filterlist_lock, flags); + raw_spin_lock_irqsave(&report_filterlist_lock, flags); report_filterlist.whitelist = whitelist; - spin_unlock_irqrestore(&report_filterlist_lock, flags); + raw_spin_unlock_irqrestore(&report_filterlist_lock, flags); } /* Returns 0 on success, error-code otherwise. */ @@ -145,6 +139,9 @@ static ssize_t insert_report_filterlist(const char *func) { unsigned long flags; unsigned long addr = kallsyms_lookup_name(func); + unsigned long *delay_free = NULL; + unsigned long *new_addrs = NULL; + size_t new_size = 0; ssize_t ret = 0; if (!addr) { @@ -152,32 +149,33 @@ static ssize_t insert_report_filterlist(const char *func) return -ENOENT; } - spin_lock_irqsave(&report_filterlist_lock, flags); +retry_alloc: + /* + * Check if we need an allocation, and re-validate under the lock. Since + * the report_filterlist_lock is a raw, cannot allocate under the lock. + */ + if (data_race(report_filterlist.used == report_filterlist.size)) { + new_size = (report_filterlist.size ?: 4) * 2; + delay_free = new_addrs = kmalloc_array(new_size, sizeof(unsigned long), GFP_KERNEL); + if (!new_addrs) + return -ENOMEM; + } - if (report_filterlist.addrs == NULL) { - /* initial allocation */ - report_filterlist.addrs = - kmalloc_array(report_filterlist.size, - sizeof(unsigned long), GFP_ATOMIC); - if (report_filterlist.addrs == NULL) { - ret = -ENOMEM; - goto out; - } - } else if (report_filterlist.used == report_filterlist.size) { - /* resize filterlist */ - size_t new_size = report_filterlist.size * 2; - unsigned long *new_addrs = - krealloc(report_filterlist.addrs, - new_size * sizeof(unsigned long), GFP_ATOMIC); - - if (new_addrs == NULL) { - /* leave filterlist itself untouched */ - ret = -ENOMEM; - goto out; + raw_spin_lock_irqsave(&report_filterlist_lock, flags); + if (report_filterlist.used == report_filterlist.size) { + /* Check we pre-allocated enough, and retry if not. */ + if (report_filterlist.used >= new_size) { + raw_spin_unlock_irqrestore(&report_filterlist_lock, flags); + kfree(new_addrs); /* kfree(NULL) is safe */ + delay_free = new_addrs = NULL; + goto retry_alloc; } + if (report_filterlist.used) + memcpy(new_addrs, report_filterlist.addrs, report_filterlist.used * sizeof(unsigned long)); + delay_free = report_filterlist.addrs; /* free the old list */ + report_filterlist.addrs = new_addrs; /* switch to the new list */ report_filterlist.size = new_size; - report_filterlist.addrs = new_addrs; } /* Note: deduplicating should be done in userspace. */ @@ -185,9 +183,9 @@ static ssize_t insert_report_filterlist(const char *func) kallsyms_lookup_name(func); report_filterlist.sorted = false; -out: - spin_unlock_irqrestore(&report_filterlist_lock, flags); + raw_spin_unlock_irqrestore(&report_filterlist_lock, flags); + kfree(delay_free); return ret; } @@ -204,13 +202,13 @@ static int show_info(struct seq_file *file, void *v) } /* show filter functions, and filter type */ - spin_lock_irqsave(&report_filterlist_lock, flags); + raw_spin_lock_irqsave(&report_filterlist_lock, flags); seq_printf(file, "\n%s functions: %s\n", report_filterlist.whitelist ? "whitelisted" : "blacklisted", report_filterlist.used == 0 ? "none" : ""); for (i = 0; i < report_filterlist.used; ++i) seq_printf(file, " %ps\n", (void *)report_filterlist.addrs[i]); - spin_unlock_irqrestore(&report_filterlist_lock, flags); + raw_spin_unlock_irqrestore(&report_filterlist_lock, flags); return 0; } -- 2.43.0

9 months, 3 weeks

1
0
0 0

[PATCH AUTOSEL 6.6 1/3] kcsan: Turn report_filterlist_lock into a raw_spinlock

by Sasha Levin

From: Marco Elver <elver(a)google.com> [ Upstream commit 59458fa4ddb47e7891c61b4a928d13d5f5b00aa0 ] Ran Xiaokai reports that with a KCSAN-enabled PREEMPT_RT kernel, we can see splats like: | BUG: sleeping function called from invalid context at kernel/locking/spinlock_rt.c:48 | in_atomic(): 1, irqs_disabled(): 1, non_block: 0, pid: 0, name: swapper/1 | preempt_count: 10002, expected: 0 | RCU nest depth: 0, expected: 0 | no locks held by swapper/1/0. | irq event stamp: 156674 | hardirqs last enabled at (156673): [<ffffffff81130bd9>] do_idle+0x1f9/0x240 | hardirqs last disabled at (156674): [<ffffffff82254f84>] sysvec_apic_timer_interrupt+0x14/0xc0 | softirqs last enabled at (0): [<ffffffff81099f47>] copy_process+0xfc7/0x4b60 | softirqs last disabled at (0): [<0000000000000000>] 0x0 | Preemption disabled at: | [<ffffffff814a3e2a>] paint_ptr+0x2a/0x90 | CPU: 1 UID: 0 PID: 0 Comm: swapper/1 Not tainted 6.11.0+ #3 | Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS rel-1.12.0-0-ga698c8995f-prebuilt.qemu.org 04/01/2014 | Call Trace: | <IRQ> | dump_stack_lvl+0x7e/0xc0 | dump_stack+0x1d/0x30 | __might_resched+0x1a2/0x270 | rt_spin_lock+0x68/0x170 | kcsan_skip_report_debugfs+0x43/0xe0 | print_report+0xb5/0x590 | kcsan_report_known_origin+0x1b1/0x1d0 | kcsan_setup_watchpoint+0x348/0x650 | __tsan_unaligned_write1+0x16d/0x1d0 | hrtimer_interrupt+0x3d6/0x430 | __sysvec_apic_timer_interrupt+0xe8/0x3a0 | sysvec_apic_timer_interrupt+0x97/0xc0 | </IRQ> On a detected data race, KCSAN's reporting logic checks if it should filter the report. That list is protected by the report_filterlist_lock *non-raw* spinlock which may sleep on RT kernels. Since KCSAN may report data races in any context, convert it to a raw_spinlock. This requires being careful about when to allocate memory for the filter list itself which can be done via KCSAN's debugfs interface. Concurrent modification of the filter list via debugfs should be rare: the chosen strategy is to optimistically pre-allocate memory before the critical section and discard if unused. Link: https://lore.kernel.org/all/20240925143154.2322926-1-ranxiaokai627@163.com/ Reported-by: Ran Xiaokai <ran.xiaokai(a)zte.com.cn> Tested-by: Ran Xiaokai <ran.xiaokai(a)zte.com.cn> Signed-off-by: Marco Elver <elver(a)google.com> Signed-off-by: Sasha Levin <sashal(a)kernel.org> --- kernel/kcsan/debugfs.c | 74 ++++++++++++++++++++---------------------- 1 file changed, 36 insertions(+), 38 deletions(-) diff --git a/kernel/kcsan/debugfs.c b/kernel/kcsan/debugfs.c index 1d1d1b0e42489..f4623910fb1f2 100644 --- a/kernel/kcsan/debugfs.c +++ b/kernel/kcsan/debugfs.c @@ -46,14 +46,8 @@ static struct { int used; /* number of elements used */ bool sorted; /* if elements are sorted */ bool whitelist; /* if list is a blacklist or whitelist */ -} report_filterlist = { - .addrs = NULL, - .size = 8, /* small initial size */ - .used = 0, - .sorted = false, - .whitelist = false, /* default is blacklist */ -}; -static DEFINE_SPINLOCK(report_filterlist_lock); +} report_filterlist; +static DEFINE_RAW_SPINLOCK(report_filterlist_lock); /* * The microbenchmark allows benchmarking KCSAN core runtime only. To run @@ -110,7 +104,7 @@ bool kcsan_skip_report_debugfs(unsigned long func_addr) return false; func_addr -= offset; /* Get function start */ - spin_lock_irqsave(&report_filterlist_lock, flags); + raw_spin_lock_irqsave(&report_filterlist_lock, flags); if (report_filterlist.used == 0) goto out; @@ -127,7 +121,7 @@ bool kcsan_skip_report_debugfs(unsigned long func_addr) ret = !ret; out: - spin_unlock_irqrestore(&report_filterlist_lock, flags); + raw_spin_unlock_irqrestore(&report_filterlist_lock, flags); return ret; } @@ -135,9 +129,9 @@ static void set_report_filterlist_whitelist(bool whitelist) { unsigned long flags; - spin_lock_irqsave(&report_filterlist_lock, flags); + raw_spin_lock_irqsave(&report_filterlist_lock, flags); report_filterlist.whitelist = whitelist; - spin_unlock_irqrestore(&report_filterlist_lock, flags); + raw_spin_unlock_irqrestore(&report_filterlist_lock, flags); } /* Returns 0 on success, error-code otherwise. */ @@ -145,6 +139,9 @@ static ssize_t insert_report_filterlist(const char *func) { unsigned long flags; unsigned long addr = kallsyms_lookup_name(func); + unsigned long *delay_free = NULL; + unsigned long *new_addrs = NULL; + size_t new_size = 0; ssize_t ret = 0; if (!addr) { @@ -152,32 +149,33 @@ static ssize_t insert_report_filterlist(const char *func) return -ENOENT; } - spin_lock_irqsave(&report_filterlist_lock, flags); +retry_alloc: + /* + * Check if we need an allocation, and re-validate under the lock. Since + * the report_filterlist_lock is a raw, cannot allocate under the lock. + */ + if (data_race(report_filterlist.used == report_filterlist.size)) { + new_size = (report_filterlist.size ?: 4) * 2; + delay_free = new_addrs = kmalloc_array(new_size, sizeof(unsigned long), GFP_KERNEL); + if (!new_addrs) + return -ENOMEM; + } - if (report_filterlist.addrs == NULL) { - /* initial allocation */ - report_filterlist.addrs = - kmalloc_array(report_filterlist.size, - sizeof(unsigned long), GFP_ATOMIC); - if (report_filterlist.addrs == NULL) { - ret = -ENOMEM; - goto out; - } - } else if (report_filterlist.used == report_filterlist.size) { - /* resize filterlist */ - size_t new_size = report_filterlist.size * 2; - unsigned long *new_addrs = - krealloc(report_filterlist.addrs, - new_size * sizeof(unsigned long), GFP_ATOMIC); - - if (new_addrs == NULL) { - /* leave filterlist itself untouched */ - ret = -ENOMEM; - goto out; + raw_spin_lock_irqsave(&report_filterlist_lock, flags); + if (report_filterlist.used == report_filterlist.size) { + /* Check we pre-allocated enough, and retry if not. */ + if (report_filterlist.used >= new_size) { + raw_spin_unlock_irqrestore(&report_filterlist_lock, flags); + kfree(new_addrs); /* kfree(NULL) is safe */ + delay_free = new_addrs = NULL; + goto retry_alloc; } + if (report_filterlist.used) + memcpy(new_addrs, report_filterlist.addrs, report_filterlist.used * sizeof(unsigned long)); + delay_free = report_filterlist.addrs; /* free the old list */ + report_filterlist.addrs = new_addrs; /* switch to the new list */ report_filterlist.size = new_size; - report_filterlist.addrs = new_addrs; } /* Note: deduplicating should be done in userspace. */ @@ -185,9 +183,9 @@ static ssize_t insert_report_filterlist(const char *func) kallsyms_lookup_name(func); report_filterlist.sorted = false; -out: - spin_unlock_irqrestore(&report_filterlist_lock, flags); + raw_spin_unlock_irqrestore(&report_filterlist_lock, flags); + kfree(delay_free); return ret; } @@ -204,13 +202,13 @@ static int show_info(struct seq_file *file, void *v) } /* show filter functions, and filter type */ - spin_lock_irqsave(&report_filterlist_lock, flags); + raw_spin_lock_irqsave(&report_filterlist_lock, flags); seq_printf(file, "\n%s functions: %s\n", report_filterlist.whitelist ? "whitelisted" : "blacklisted", report_filterlist.used == 0 ? "none" : ""); for (i = 0; i < report_filterlist.used; ++i) seq_printf(file, " %ps\n", (void *)report_filterlist.addrs[i]); - spin_unlock_irqrestore(&report_filterlist_lock, flags); + raw_spin_unlock_irqrestore(&report_filterlist_lock, flags); return 0; } -- 2.43.0

9 months, 3 weeks

1
2
0 0

2025

2024

2023

2022

2021

2020

2019

2018

2017

Linux-stable-mirror November 2024