From: Steven Rostedt (VMware) rostedt@goodmis.org

[ Upstream commit b02414c8f045ab3b9afc816c3735bc98c5c3d262 ]

The recursion protection of the ring buffer depends on preempt_count() to be correct. But it is possible that the ring buffer gets called after an interrupt comes in but before it updates the preempt_count(). This will trigger a false positive in the recursion code.

Use the same trick from the ftrace function callback recursion code which uses a "transition" bit that gets set, to allow for a single recursion for to handle transitions between contexts.

Cc: stable@vger.kernel.org Fixes: 567cd4da54ff4 ("ring-buffer: User context bit recursion checking") Signed-off-by: Steven Rostedt (VMware) rostedt@goodmis.org Signed-off-by: Sasha Levin sashal@kernel.org --- kernel/trace/ring_buffer.c | 54 +++++++++++++++++++++++++++++++------- 1 file changed, 44 insertions(+), 10 deletions(-)

diff --git a/kernel/trace/ring_buffer.c b/kernel/trace/ring_buffer.c index b9b71e7fb6979..8082328eb01a4 100644 --- a/kernel/trace/ring_buffer.c +++ b/kernel/trace/ring_buffer.c @@ -416,14 +416,16 @@ struct rb_event_info {

/* * Used for which event context the event is in. - * NMI = 0 - * IRQ = 1 - * SOFTIRQ = 2 - * NORMAL = 3 + * TRANSITION = 0 + * NMI = 1 + * IRQ = 2 + * SOFTIRQ = 3 + * NORMAL = 4 * * See trace_recursive_lock() comment below for more details. */ enum { + RB_CTX_TRANSITION, RB_CTX_NMI, RB_CTX_IRQ, RB_CTX_SOFTIRQ, @@ -2553,10 +2555,10 @@ rb_wakeups(struct ring_buffer *buffer, struct ring_buffer_per_cpu *cpu_buffer) * a bit of overhead in something as critical as function tracing, * we use a bitmask trick. * - * bit 0 = NMI context - * bit 1 = IRQ context - * bit 2 = SoftIRQ context - * bit 3 = normal context. + * bit 1 = NMI context + * bit 2 = IRQ context + * bit 3 = SoftIRQ context + * bit 4 = normal context. * * This works because this is the order of contexts that can * preempt other contexts. A SoftIRQ never preempts an IRQ @@ -2579,6 +2581,30 @@ rb_wakeups(struct ring_buffer *buffer, struct ring_buffer_per_cpu *cpu_buffer) * The least significant bit can be cleared this way, and it * just so happens that it is the same bit corresponding to * the current context. + * + * Now the TRANSITION bit breaks the above slightly. The TRANSITION bit + * is set when a recursion is detected at the current context, and if + * the TRANSITION bit is already set, it will fail the recursion. + * This is needed because there's a lag between the changing of + * interrupt context and updating the preempt count. In this case, + * a false positive will be found. To handle this, one extra recursion + * is allowed, and this is done by the TRANSITION bit. If the TRANSITION + * bit is already set, then it is considered a recursion and the function + * ends. Otherwise, the TRANSITION bit is set, and that bit is returned. + * + * On the trace_recursive_unlock(), the TRANSITION bit will be the first + * to be cleared. Even if it wasn't the context that set it. That is, + * if an interrupt comes in while NORMAL bit is set and the ring buffer + * is called before preempt_count() is updated, since the check will + * be on the NORMAL bit, the TRANSITION bit will then be set. If an + * NMI then comes in, it will set the NMI bit, but when the NMI code + * does the trace_recursive_unlock() it will clear the TRANSTION bit + * and leave the NMI bit set. But this is fine, because the interrupt + * code that set the TRANSITION bit will then clear the NMI bit when it + * calls trace_recursive_unlock(). If another NMI comes in, it will + * set the TRANSITION bit and continue. + * + * Note: The TRANSITION bit only handles a single transition between context. */

static __always_inline int @@ -2597,8 +2623,16 @@ trace_recursive_lock(struct ring_buffer_per_cpu *cpu_buffer) } else bit = RB_CTX_NORMAL;

- if (unlikely(val & (1 << bit))) - return 1; + if (unlikely(val & (1 << bit))) { + /* + * It is possible that this was called by transitioning + * between interrupt context, and preempt_count() has not + * been updated yet. In this case, use the TRANSITION bit. + */ + bit = RB_CTX_TRANSITION; + if (val & (1 << bit)) + return 1; + }

val |= (1 << bit); cpu_buffer->current_context = val;

From: Zeng Tao prime.zeng@hisilicon.com

[ Upstream commit cb47755725da7b90fecbb2aa82ac3b24a7adb89b ]

UBSAN reports:

Undefined behaviour in ./include/linux/time64.h:127:27 signed integer overflow: 17179869187 * 1000000000 cannot be represented in type 'long long int' Call Trace: timespec64_to_ns include/linux/time64.h:127 [inline] set_cpu_itimer+0x65c/0x880 kernel/time/itimer.c:180 do_setitimer+0x8e/0x740 kernel/time/itimer.c:245 __x64_sys_setitimer+0x14c/0x2c0 kernel/time/itimer.c:336 do_syscall_64+0xa1/0x540 arch/x86/entry/common.c:295

Commit bd40a175769d ("y2038: itimer: change implementation to timespec64") replaced the original conversion which handled time clamping correctly with timespec64_to_ns() which has no overflow protection.

Fix it in timespec64_to_ns() as this is not necessarily limited to the usage in itimers.

[ tglx: Added comment and adjusted the fixes tag ]

Fixes: 361a3bf00582 ("time64: Add time64.h header and define struct timespec64") Signed-off-by: Zeng Tao prime.zeng@hisilicon.com Signed-off-by: Thomas Gleixner tglx@linutronix.de Reviewed-by: Arnd Bergmann arnd@arndb.de Cc: stable@vger.kernel.org Link: https://lore.kernel.org/r/1598952616-6416-1-git-send-email-prime.zeng@hisili... Signed-off-by: Sasha Levin sashal@kernel.org --- include/linux/time64.h | 4 ++++ kernel/time/itimer.c | 4 ---- 2 files changed, 4 insertions(+), 4 deletions(-)

diff --git a/include/linux/time64.h b/include/linux/time64.h index ad33260618f76..99ab4a686c301 100644 --- a/include/linux/time64.h +++ b/include/linux/time64.h @@ -189,6 +189,10 @@ static inline bool timespec64_valid_strict(const struct timespec64 *ts) */ static inline s64 timespec64_to_ns(const struct timespec64 *ts) { + /* Prevent multiplication overflow */ + if ((unsigned long long)ts->tv_sec >= KTIME_SEC_MAX) + return KTIME_MAX; + return ((s64) ts->tv_sec * NSEC_PER_SEC) + ts->tv_nsec; }

diff --git a/kernel/time/itimer.c b/kernel/time/itimer.c index f26acef5d7b48..7f718565507b5 100644 --- a/kernel/time/itimer.c +++ b/kernel/time/itimer.c @@ -153,10 +153,6 @@ static void set_cpu_itimer(struct task_struct *tsk, unsigned int clock_id, u64 oval, nval, ointerval, ninterval; struct cpu_itimer *it = &tsk->signal->it[clock_id];

- /* - * Use the to_ktime conversion because that clamps the maximum - * value to KTIME_MAX and avoid multiplication overflows. - */ nval = ktime_to_ns(timeval_to_ktime(value->it_value)); ninterval = ktime_to_ns(timeval_to_ktime(value->it_interval));

From: Josef Bacik josef@toxicpanda.com

[ Upstream commit ca10845a56856fff4de3804c85e6424d0f6d0cde ]

While running btrfs/061, btrfs/073, btrfs/078, or btrfs/178 we hit the following lockdep splat:

====================================================== WARNING: possible circular locking dependency detected 5.9.0-rc3+ #4 Not tainted ------------------------------------------------------ kswapd0/100 is trying to acquire lock: ffff96ecc22ef4a0 (&delayed_node->mutex){+.+.}-{3:3}, at: __btrfs_release_delayed_node.part.0+0x3f/0x330

but task is already holding lock: ffffffff8dd74700 (fs_reclaim){+.+.}-{0:0}, at: __fs_reclaim_acquire+0x5/0x30

which lock already depends on the new lock.

the existing dependency chain (in reverse order) is:

-> #3 (fs_reclaim){+.+.}-{0:0}: fs_reclaim_acquire+0x65/0x80 slab_pre_alloc_hook.constprop.0+0x20/0x200 kmem_cache_alloc+0x37/0x270 alloc_inode+0x82/0xb0 iget_locked+0x10d/0x2c0 kernfs_get_inode+0x1b/0x130 kernfs_get_tree+0x136/0x240 sysfs_get_tree+0x16/0x40 vfs_get_tree+0x28/0xc0 path_mount+0x434/0xc00 __x64_sys_mount+0xe3/0x120 do_syscall_64+0x33/0x40 entry_SYSCALL_64_after_hwframe+0x44/0xa9

-> #2 (kernfs_mutex){+.+.}-{3:3}: __mutex_lock+0x7e/0x7e0 kernfs_add_one+0x23/0x150 kernfs_create_link+0x63/0xa0 sysfs_do_create_link_sd+0x5e/0xd0 btrfs_sysfs_add_devices_dir+0x81/0x130 btrfs_init_new_device+0x67f/0x1250 btrfs_ioctl+0x1ef/0x2e20 __x64_sys_ioctl+0x83/0xb0 do_syscall_64+0x33/0x40 entry_SYSCALL_64_after_hwframe+0x44/0xa9

-> #1 (&fs_info->chunk_mutex){+.+.}-{3:3}: __mutex_lock+0x7e/0x7e0 btrfs_chunk_alloc+0x125/0x3a0 find_free_extent+0xdf6/0x1210 btrfs_reserve_extent+0xb3/0x1b0 btrfs_alloc_tree_block+0xb0/0x310 alloc_tree_block_no_bg_flush+0x4a/0x60 __btrfs_cow_block+0x11a/0x530 btrfs_cow_block+0x104/0x220 btrfs_search_slot+0x52e/0x9d0 btrfs_insert_empty_items+0x64/0xb0 btrfs_insert_delayed_items+0x90/0x4f0 btrfs_commit_inode_delayed_items+0x93/0x140 btrfs_log_inode+0x5de/0x2020 btrfs_log_inode_parent+0x429/0xc90 btrfs_log_new_name+0x95/0x9b btrfs_rename2+0xbb9/0x1800 vfs_rename+0x64f/0x9f0 do_renameat2+0x320/0x4e0 __x64_sys_rename+0x1f/0x30 do_syscall_64+0x33/0x40 entry_SYSCALL_64_after_hwframe+0x44/0xa9

-> #0 (&delayed_node->mutex){+.+.}-{3:3}: __lock_acquire+0x119c/0x1fc0 lock_acquire+0xa7/0x3d0 __mutex_lock+0x7e/0x7e0 __btrfs_release_delayed_node.part.0+0x3f/0x330 btrfs_evict_inode+0x24c/0x500 evict+0xcf/0x1f0 dispose_list+0x48/0x70 prune_icache_sb+0x44/0x50 super_cache_scan+0x161/0x1e0 do_shrink_slab+0x178/0x3c0 shrink_slab+0x17c/0x290 shrink_node+0x2b2/0x6d0 balance_pgdat+0x30a/0x670 kswapd+0x213/0x4c0 kthread+0x138/0x160 ret_from_fork+0x1f/0x30

other info that might help us debug this:

Chain exists of: &delayed_node->mutex --> kernfs_mutex --> fs_reclaim

Possible unsafe locking scenario:

CPU0 CPU1 ---- ---- lock(fs_reclaim); lock(kernfs_mutex); lock(fs_reclaim); lock(&delayed_node->mutex);

*** DEADLOCK ***

3 locks held by kswapd0/100: #0: ffffffff8dd74700 (fs_reclaim){+.+.}-{0:0}, at: __fs_reclaim_acquire+0x5/0x30 #1: ffffffff8dd65c50 (shrinker_rwsem){++++}-{3:3}, at: shrink_slab+0x115/0x290 #2: ffff96ed2ade30e0 (&type->s_umount_key#36){++++}-{3:3}, at: super_cache_scan+0x38/0x1e0

stack backtrace: CPU: 0 PID: 100 Comm: kswapd0 Not tainted 5.9.0-rc3+ #4 Hardware name: QEMU Standard PC (Q35 + ICH9, 2009), BIOS 1.13.0-2.fc32 04/01/2014 Call Trace: dump_stack+0x8b/0xb8 check_noncircular+0x12d/0x150 __lock_acquire+0x119c/0x1fc0 lock_acquire+0xa7/0x3d0 ? __btrfs_release_delayed_node.part.0+0x3f/0x330 __mutex_lock+0x7e/0x7e0 ? __btrfs_release_delayed_node.part.0+0x3f/0x330 ? __btrfs_release_delayed_node.part.0+0x3f/0x330 ? lock_acquire+0xa7/0x3d0 ? find_held_lock+0x2b/0x80 __btrfs_release_delayed_node.part.0+0x3f/0x330 btrfs_evict_inode+0x24c/0x500 evict+0xcf/0x1f0 dispose_list+0x48/0x70 prune_icache_sb+0x44/0x50 super_cache_scan+0x161/0x1e0 do_shrink_slab+0x178/0x3c0 shrink_slab+0x17c/0x290 shrink_node+0x2b2/0x6d0 balance_pgdat+0x30a/0x670 kswapd+0x213/0x4c0 ? _raw_spin_unlock_irqrestore+0x41/0x50 ? add_wait_queue_exclusive+0x70/0x70 ? balance_pgdat+0x670/0x670 kthread+0x138/0x160 ? kthread_create_worker_on_cpu+0x40/0x40 ret_from_fork+0x1f/0x30

This happens because we are holding the chunk_mutex at the time of adding in a new device. However we only need to hold the device_list_mutex, as we're going to iterate over the fs_devices devices. Move the sysfs init stuff outside of the chunk_mutex to get rid of this lockdep splat.

CC: stable@vger.kernel.org # 4.4.x: f3cd2c58110dad14e: btrfs: sysfs, rename device_link add/remove functions CC: stable@vger.kernel.org # 4.4.x Reported-by: David Sterba dsterba@suse.com Signed-off-by: Josef Bacik josef@toxicpanda.com Reviewed-by: David Sterba dsterba@suse.com Signed-off-by: David Sterba dsterba@suse.com Signed-off-by: Sasha Levin sashal@kernel.org --- fs/btrfs/volumes.c | 7 ++++--- 1 file changed, 4 insertions(+), 3 deletions(-)

diff --git a/fs/btrfs/volumes.c b/fs/btrfs/volumes.c index f3cb042a28d5c..3d92feff249a0 100644 --- a/fs/btrfs/volumes.c +++ b/fs/btrfs/volumes.c @@ -2432,9 +2432,6 @@ int btrfs_init_new_device(struct btrfs_fs_info *fs_info, const char *device_path tmp = btrfs_super_num_devices(fs_info->super_copy); btrfs_set_super_num_devices(fs_info->super_copy, tmp + 1);

- /* add sysfs device entry */ - btrfs_sysfs_add_device_link(fs_info->fs_devices, device); - /* * we've got more storage, clear any full flags on the space * infos @@ -2442,6 +2439,10 @@ int btrfs_init_new_device(struct btrfs_fs_info *fs_info, const char *device_path btrfs_clear_space_info_full(fs_info);

mutex_unlock(&fs_info->chunk_mutex); + + /* Add sysfs device entry */ + btrfs_sysfs_add_device_link(fs_info->fs_devices, device); + mutex_unlock(&fs_info->fs_devices->device_list_mutex);

if (seeding_dev) {

From: Marc Zyngier maz@kernel.org

[ Upstream commit 151a535171be6ff824a0a3875553ea38570f4c05 ]

kernel/irq/ipi.c otherwise fails to compile if nothing else selects it.

Fixes: 379b656446a3 ("genirq: Add GENERIC_IRQ_IPI Kconfig symbol") Reported-by: Pavel Machek pavel@ucw.cz Tested-by: Pavel Machek pavel@ucw.cz Signed-off-by: Marc Zyngier maz@kernel.org Link: https://lore.kernel.org/r/20201015101222.GA32747@amd Signed-off-by: Sasha Levin sashal@kernel.org --- kernel/irq/Kconfig | 1 + 1 file changed, 1 insertion(+)

diff --git a/kernel/irq/Kconfig b/kernel/irq/Kconfig index a117adf7084b8..779d858da2b39 100644 --- a/kernel/irq/Kconfig +++ b/kernel/irq/Kconfig @@ -80,6 +80,7 @@ config IRQ_FASTEOI_HIERARCHY_HANDLERS # Generic IRQ IPI support config GENERIC_IRQ_IPI bool + select IRQ_DOMAIN_HIERARCHY

# Generic MSI interrupt support config GENERIC_MSI_IRQ

From: zhuoliang zhang zhuoliang.zhang@mediatek.com

[ Upstream commit a779d91314ca7208b7feb3ad817b62904397c56d ]

we found that the following race condition exists in xfrm_alloc_userspi flow:

user thread state_hash_work thread ---- ---- xfrm_alloc_userspi() __find_acq_core() /*alloc new xfrm_state:x*/ xfrm_state_alloc() /*schedule state_hash_work thread*/ xfrm_hash_grow_check() xfrm_hash_resize() xfrm_alloc_spi /*hold lock*/ x->id.spi = htonl(spi) spin_lock_bh(&net->xfrm.xfrm_state_lock) /*waiting lock release*/ xfrm_hash_transfer() spin_lock_bh(&net->xfrm.xfrm_state_lock) /*add x into hlist:net->xfrm.state_byspi*/ hlist_add_head_rcu(&x->byspi) spin_unlock_bh(&net->xfrm.xfrm_state_lock)

/*add x into hlist:net->xfrm.state_byspi 2 times*/ hlist_add_head_rcu(&x->byspi)

1. a new state x is alloced in xfrm_state_alloc() and added into the bydst hlist in __find_acq_core() on the LHS; 2. on the RHS, state_hash_work thread travels the old bydst and tranfers every xfrm_state (include x) into the new bydst hlist and new byspi hlist; 3. user thread on the LHS gets the lock and adds x into the new byspi hlist again.

So the same xfrm_state (x) is added into the same list_hash (net->xfrm.state_byspi) 2 times that makes the list_hash become an inifite loop.

To fix the race, x->id.spi = htonl(spi) in the xfrm_alloc_spi() is moved to the back of spin_lock_bh, sothat state_hash_work thread no longer add x which id.spi is zero into the hash_list.

Fixes: f034b5d4efdf ("[XFRM]: Dynamic xfrm_state hash table sizing.") Signed-off-by: zhuoliang zhang zhuoliang.zhang@mediatek.com Acked-by: Herbert Xu herbert@gondor.apana.org.au Signed-off-by: Steffen Klassert steffen.klassert@secunet.com Signed-off-by: Sasha Levin sashal@kernel.org --- net/xfrm/xfrm_state.c | 8 +++++--- 1 file changed, 5 insertions(+), 3 deletions(-)

diff --git a/net/xfrm/xfrm_state.c b/net/xfrm/xfrm_state.c index 05c275a712f11..5164dfe0aa097 100644 --- a/net/xfrm/xfrm_state.c +++ b/net/xfrm/xfrm_state.c @@ -1783,6 +1783,7 @@ int xfrm_alloc_spi(struct xfrm_state *x, u32 low, u32 high) int err = -ENOENT; __be32 minspi = htonl(low); __be32 maxspi = htonl(high); + __be32 newspi = 0; u32 mark = x->mark.v & x->mark.m;

spin_lock_bh(&x->lock); @@ -1801,21 +1802,22 @@ int xfrm_alloc_spi(struct xfrm_state *x, u32 low, u32 high) xfrm_state_put(x0); goto unlock; } - x->id.spi = minspi; + newspi = minspi; } else { u32 spi = 0; for (h = 0; h < high-low+1; h++) { spi = low + prandom_u32()%(high-low+1); x0 = xfrm_state_lookup(net, mark, &x->id.daddr, htonl(spi), x->id.proto, x->props.family); if (x0 == NULL) { - x->id.spi = htonl(spi); + newspi = htonl(spi); break; } xfrm_state_put(x0); } } - if (x->id.spi) { + if (newspi) { spin_lock_bh(&net->xfrm.xfrm_state_lock); + x->id.spi = newspi; h = xfrm_spi_hash(net, &x->id.daddr, x->id.spi, x->id.proto, x->props.family); hlist_add_head_rcu(&x->byspi, net->xfrm.state_byspi + h); spin_unlock_bh(&net->xfrm.xfrm_state_lock);

From: Dan Carpenter dan.carpenter@oracle.com

[ Upstream commit 158e1886b6262c1d1c96a18c85fac5219b8bf804 ]

This is harmless, but the "addr" comes from the user and it could lead to a negative shift or to shift wrapping if it's too high.

Fixes: 0b00a5615dc4 ("ALSA: hdac_ext: add hdac extended controller") Signed-off-by: Dan Carpenter dan.carpenter@oracle.com Link: https://lore.kernel.org/r/20201103101807.GC1127762@mwanda Signed-off-by: Takashi Iwai tiwai@suse.de Signed-off-by: Sasha Levin sashal@kernel.org --- sound/hda/ext/hdac_ext_controller.c | 2 ++ 1 file changed, 2 insertions(+)

diff --git a/sound/hda/ext/hdac_ext_controller.c b/sound/hda/ext/hdac_ext_controller.c index 84f3b81687164..b679d5f37e4d2 100644 --- a/sound/hda/ext/hdac_ext_controller.c +++ b/sound/hda/ext/hdac_ext_controller.c @@ -155,6 +155,8 @@ struct hdac_ext_link *snd_hdac_ext_bus_get_link(struct hdac_ext_bus *ebus, return NULL; if (ebus->idx != bus_idx) return NULL; + if (addr < 0 || addr > 31) + return NULL;

list_for_each_entry(hlink, &ebus->hlink_list, list) { for (i = 0; i < HDA_MAX_CODECS; i++) {

From: Vincent Mailhol mailhol.vincent@wanadoo.fr

[ Upstream commit 2283f79b22684d2812e5c76fc2280aae00390365 ]

If a driver calls can_get_echo_skb() during a hardware IRQ (which is often, but not always, the case), the 'WARN_ON(in_irq)' in net/core/skbuff.c#skb_release_head_state() might be triggered, under network congestion circumstances, together with the potential risk of a NULL pointer dereference.

The root cause of this issue is the call to kfree_skb() instead of dev_kfree_skb_irq() in net/core/dev.c#enqueue_to_backlog().

This patch prevents the skb to be freed within the call to netif_rx() by incrementing its reference count with skb_get(). The skb is finally freed by one of the in-irq-context safe functions: dev_consume_skb_any() or dev_kfree_skb_any(). The "any" version is used because some drivers might call can_get_echo_skb() in a normal context.

The reason for this issue to occur is that initially, in the core network stack, loopback skb were not supposed to be received in hardware IRQ context. The CAN stack is an exeption.

This bug was previously reported back in 2017 in [1] but the proposed patch never got accepted.

While [1] directly modifies net/core/dev.c, we try to propose here a smoother modification local to CAN network stack (the assumption behind is that only CAN devices are affected by this issue).

[1] http://lore.kernel.org/r/57a3ffb6-3309-3ad5-5a34-e93c3fe3614d@cetitec.com

Signed-off-by: Vincent Mailhol mailhol.vincent@wanadoo.fr Link: https://lore.kernel.org/r/20201002154219.4887-2-mailhol.vincent@wanadoo.fr Fixes: 39549eef3587 ("can: CAN Network device driver and Netlink interface") Signed-off-by: Marc Kleine-Budde mkl@pengutronix.de Signed-off-by: Sasha Levin sashal@kernel.org --- drivers/net/can/dev.c | 6 +++++- 1 file changed, 5 insertions(+), 1 deletion(-)

diff --git a/drivers/net/can/dev.c b/drivers/net/can/dev.c index 05ad5ed145a3a..926d663eed37a 100644 --- a/drivers/net/can/dev.c +++ b/drivers/net/can/dev.c @@ -519,7 +519,11 @@ unsigned int can_get_echo_skb(struct net_device *dev, unsigned int idx) if (!skb) return 0;

- netif_rx(skb); + skb_get(skb); + if (netif_rx(skb) == NET_RX_SUCCESS) + dev_consume_skb_any(skb); + else + dev_kfree_skb_any(skb);

return len; }

From: Oleksij Rempel o.rempel@pengutronix.de

[ Upstream commit 286228d382ba6320f04fa2e7c6fc8d4d92e428f4 ]

All user space generated SKBs are owned by a socket (unless injected into the key via AF_PACKET). If a socket is closed, all associated skbs will be cleaned up.

This leads to a problem when a CAN driver calls can_put_echo_skb() on a unshared SKB. If the socket is closed prior to the TX complete handler, can_get_echo_skb() and the subsequent delivering of the echo SKB to all registered callbacks, a SKB with a refcount of 0 is delivered.

To avoid the problem, in can_get_echo_skb() the original SKB is now always cloned, regardless of shared SKB or not. If the process exists it can now safely discard its SKBs, without disturbing the delivery of the echo SKB.

The problem shows up in the j1939 stack, when it clones the incoming skb, which detects the already 0 refcount.

We can easily reproduce this with following example:

testj1939 -B -r can0: & cansend can0 1823ff40#0123

WARNING: CPU: 0 PID: 293 at lib/refcount.c:25 refcount_warn_saturate+0x108/0x174 refcount_t: addition on 0; use-after-free. Modules linked in: coda_vpu imx_vdoa videobuf2_vmalloc dw_hdmi_ahb_audio vcan CPU: 0 PID: 293 Comm: cansend Not tainted 5.5.0-rc6-00376-g9e20dcb7040d #1 Hardware name: Freescale i.MX6 Quad/DualLite (Device Tree) Backtrace: [<c010f570>] (dump_backtrace) from [<c010f90c>] (show_stack+0x20/0x24) [<c010f8ec>] (show_stack) from [<c0c3e1a4>] (dump_stack+0x8c/0xa0) [<c0c3e118>] (dump_stack) from [<c0127fec>] (__warn+0xe0/0x108) [<c0127f0c>] (__warn) from [<c01283c8>] (warn_slowpath_fmt+0xa8/0xcc) [<c0128324>] (warn_slowpath_fmt) from [<c0539c0c>] (refcount_warn_saturate+0x108/0x174) [<c0539b04>] (refcount_warn_saturate) from [<c0ad2cac>] (j1939_can_recv+0x20c/0x210) [<c0ad2aa0>] (j1939_can_recv) from [<c0ac9dc8>] (can_rcv_filter+0xb4/0x268) [<c0ac9d14>] (can_rcv_filter) from [<c0aca2cc>] (can_receive+0xb0/0xe4) [<c0aca21c>] (can_receive) from [<c0aca348>] (can_rcv+0x48/0x98) [<c0aca300>] (can_rcv) from [<c09b1fdc>] (__netif_receive_skb_one_core+0x64/0x88) [<c09b1f78>] (__netif_receive_skb_one_core) from [<c09b2070>] (__netif_receive_skb+0x38/0x94) [<c09b2038>] (__netif_receive_skb) from [<c09b2130>] (netif_receive_skb_internal+0x64/0xf8) [<c09b20cc>] (netif_receive_skb_internal) from [<c09b21f8>] (netif_receive_skb+0x34/0x19c) [<c09b21c4>] (netif_receive_skb) from [<c0791278>] (can_rx_offload_napi_poll+0x58/0xb4)

Fixes: 0ae89beb283a ("can: add destructor for self generated skbs") Signed-off-by: Oleksij Rempel o.rempel@pengutronix.de Link: http://lore.kernel.org/r/20200124132656.22156-1-o.rempel@pengutronix.de Acked-by: Oliver Hartkopp socketcan@hartkopp.net Signed-off-by: Marc Kleine-Budde mkl@pengutronix.de Signed-off-by: Sasha Levin sashal@kernel.org --- include/linux/can/skb.h | 20 ++++++++------------ 1 file changed, 8 insertions(+), 12 deletions(-)

diff --git a/include/linux/can/skb.h b/include/linux/can/skb.h index b3379a97245c1..a34694e675c9a 100644 --- a/include/linux/can/skb.h +++ b/include/linux/can/skb.h @@ -61,21 +61,17 @@ static inline void can_skb_set_owner(struct sk_buff *skb, struct sock *sk) */ static inline struct sk_buff *can_create_echo_skb(struct sk_buff *skb) { - if (skb_shared(skb)) { - struct sk_buff *nskb = skb_clone(skb, GFP_ATOMIC); + struct sk_buff *nskb;

- if (likely(nskb)) { - can_skb_set_owner(nskb, skb->sk); - consume_skb(skb); - return nskb; - } else { - kfree_skb(skb); - return NULL; - } + nskb = skb_clone(skb, GFP_ATOMIC); + if (unlikely(!nskb)) { + kfree_skb(skb); + return NULL; }

- /* we can assume to have an unshared skb with proper owner */ - return skb; + can_skb_set_owner(nskb, skb->sk); + consume_skb(skb); + return nskb; }

#endif /* !_CAN_SKB_H */

From: Stephane Grosjean s.grosjean@peak-system.com

[ Upstream commit ecc7b4187dd388549544195fb13a11b4ea8e6a84 ]

Fabian Inostroza fabianinostrozap@gmail.com has discovered a potential problem in the hardware timestamp reporting from the PCAN-USB USB CAN interface (only), related to the fact that a timestamp of an event may precede the timestamp used for synchronization when both records are part of the same USB packet. However, this case was used to detect the wrapping of the time counter.

This patch details and fixes the two identified cases where this problem can occur.

Reported-by: Fabian Inostroza fabianinostrozap@gmail.com Signed-off-by: Stephane Grosjean s.grosjean@peak-system.com Link: https://lore.kernel.org/r/20201014085631.15128-1-s.grosjean@peak-system.com Fixes: bb4785551f64 ("can: usb: PEAK-System Technik USB adapters driver core") Signed-off-by: Marc Kleine-Budde mkl@pengutronix.de Signed-off-by: Sasha Levin sashal@kernel.org --- drivers/net/can/usb/peak_usb/pcan_usb_core.c | 51 ++++++++++++++++++-- 1 file changed, 46 insertions(+), 5 deletions(-)

diff --git a/drivers/net/can/usb/peak_usb/pcan_usb_core.c b/drivers/net/can/usb/peak_usb/pcan_usb_core.c index 85d92f129af2d..9d78ba7776140 100644 --- a/drivers/net/can/usb/peak_usb/pcan_usb_core.c +++ b/drivers/net/can/usb/peak_usb/pcan_usb_core.c @@ -154,14 +154,55 @@ void peak_usb_get_ts_tv(struct peak_time_ref *time_ref, u32 ts, /* protect from getting timeval before setting now */ if (time_ref->tv_host.tv_sec > 0) { u64 delta_us; + s64 delta_ts = 0; + + /* General case: dev_ts_1 < dev_ts_2 < ts, with: + * + * - dev_ts_1 = previous sync timestamp + * - dev_ts_2 = last sync timestamp + * - ts = event timestamp + * - ts_period = known sync period (theoretical) + * ~ dev_ts2 - dev_ts1 + * *but*: + * + * - time counters wrap (see adapter->ts_used_bits) + * - sometimes, dev_ts_1 < ts < dev_ts2 + * + * "normal" case (sync time counters increase): + * must take into account case when ts wraps (tsw) + * + * < ts_period > < > + * | | | + * ---+--------+----+-------0-+--+--> + * ts_dev_1 | ts_dev_2 | + * ts tsw + */ + if (time_ref->ts_dev_1 < time_ref->ts_dev_2) { + /* case when event time (tsw) wraps */ + if (ts < time_ref->ts_dev_1) + delta_ts = 1 << time_ref->adapter->ts_used_bits; + + /* Otherwise, sync time counter (ts_dev_2) has wrapped: + * handle case when event time (tsn) hasn't. + * + * < ts_period > < > + * | | | + * ---+--------+--0-+---------+--+--> + * ts_dev_1 | ts_dev_2 | + * tsn ts + */ + } else if (time_ref->ts_dev_1 < ts) { + delta_ts = -(1 << time_ref->adapter->ts_used_bits); + }

- delta_us = ts - time_ref->ts_dev_2; - if (ts < time_ref->ts_dev_2) - delta_us &= (1 << time_ref->adapter->ts_used_bits) - 1; + /* add delay between last sync and event timestamps */ + delta_ts += (signed int)(ts - time_ref->ts_dev_2);

- delta_us += time_ref->ts_total; + /* add time from beginning to last sync */ + delta_ts += time_ref->ts_total;

- delta_us *= time_ref->adapter->us_per_ts_scale; + /* convert ticks number into microseconds */ + delta_us = delta_ts * time_ref->adapter->us_per_ts_scale; delta_us >>= time_ref->adapter->us_per_ts_shift;

*tv = time_ref->tv_host_0;

From: Brian Foster bfoster@redhat.com

[ Upstream commit 869ae85dae64b5540e4362d7fe4cd520e10ec05c ]

It is possible to expose non-zeroed post-EOF data in XFS if the new EOF page is dirty, backed by an unwritten block and the truncate happens to race with writeback. iomap_truncate_page() will not zero the post-EOF portion of the page if the underlying block is unwritten. The subsequent call to truncate_setsize() will, but doesn't dirty the page. Therefore, if writeback happens to complete after iomap_truncate_page() (so it still sees the unwritten block) but before truncate_setsize(), the cached page becomes inconsistent with the on-disk block. A mapped read after the associated page is reclaimed or invalidated exposes non-zero post-EOF data.

For example, consider the following sequence when run on a kernel modified to explicitly flush the new EOF page within the race window:

$ xfs_io -fc "falloc 0 4k" -c fsync /mnt/file $ xfs_io -c "pwrite 0 4k" -c "truncate 1k" /mnt/file ... $ xfs_io -c "mmap 0 4k" -c "mread -v 1k 8" /mnt/file 00000400: 00 00 00 00 00 00 00 00 ........ $ umount /mnt/; mount <dev> /mnt/ $ xfs_io -c "mmap 0 4k" -c "mread -v 1k 8" /mnt/file 00000400: cd cd cd cd cd cd cd cd ........

Update xfs_setattr_size() to explicitly flush the new EOF page prior to the page truncate to ensure iomap has the latest state of the underlying block.

Fixes: 68a9f5e7007c ("xfs: implement iomap based buffered write path") Signed-off-by: Brian Foster bfoster@redhat.com Reviewed-by: Darrick J. Wong darrick.wong@oracle.com Signed-off-by: Darrick J. Wong darrick.wong@oracle.com Signed-off-by: Sasha Levin sashal@kernel.org --- fs/xfs/xfs_iops.c | 10 ++++++++++ 1 file changed, 10 insertions(+)

diff --git a/fs/xfs/xfs_iops.c b/fs/xfs/xfs_iops.c index 4e6f2c8574f7e..42c670a313518 100644 --- a/fs/xfs/xfs_iops.c +++ b/fs/xfs/xfs_iops.c @@ -879,6 +879,16 @@ xfs_setattr_size( if (newsize > oldsize) { error = xfs_zero_eof(ip, newsize, oldsize, &did_zeroing); } else { + /* + * iomap won't detect a dirty page over an unwritten block (or a + * cow block over a hole) and subsequently skips zeroing the + * newly post-EOF portion of the page. Flush the new EOF to + * convert the block before the pagecache truncate. + */ + error = filemap_write_and_wait_range(inode->i_mapping, newsize, + newsize); + if (error) + return error; error = iomap_truncate_page(inode, newsize, &did_zeroing, &xfs_iomap_ops); }

From: Masashi Honma masashi.honma@gmail.com

commit 5024f21c159f8c1668f581fff37140741c0b1ba9 upstream.

kernel test robot says: drivers/net/wireless/ath/ath9k/htc_drv_txrx.c:987:20: sparse: warning: incorrect type in assignment (different base types) drivers/net/wireless/ath/ath9k/htc_drv_txrx.c:987:20: sparse: expected restricted __be16 [usertype] rs_datalen drivers/net/wireless/ath/ath9k/htc_drv_txrx.c:987:20: sparse: got unsigned short [usertype] drivers/net/wireless/ath/ath9k/htc_drv_txrx.c:988:13: sparse: warning: restricted __be16 degrades to integer drivers/net/wireless/ath/ath9k/htc_drv_txrx.c:1001:13: sparse: warning: restricted __be16 degrades to integer

Indeed rs_datalen has host byte order, so modify it's own type.

Reported-by: kernel test robot lkp@intel.com Fixes: cd486e627e67 ("ath9k_htc: Discard undersized packets") Signed-off-by: Masashi Honma masashi.honma@gmail.com Signed-off-by: Kalle Valo kvalo@codeaurora.org Link: https://lore.kernel.org/r/20200808233258.4596-1-masashi.honma@gmail.com Signed-off-by: Greg Kroah-Hartman gregkh@linuxfoundation.org

--- drivers/net/wireless/ath/ath9k/htc_drv_txrx.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)

--- a/drivers/net/wireless/ath/ath9k/htc_drv_txrx.c +++ b/drivers/net/wireless/ath/ath9k/htc_drv_txrx.c @@ -973,7 +973,7 @@ static bool ath9k_rx_prepare(struct ath9 struct ath_htc_rx_status *rxstatus; struct ath_rx_status rx_stats; bool decrypt_error = false; - __be16 rs_datalen; + u16 rs_datalen; bool is_phyerr;

if (skb->len < HTC_RX_FRAME_HEADER_SIZE) {

From: Bob Peterson rpeterso@redhat.com

[ Upstream commit d0f17d3883f1e3f085d38572c2ea8edbd5150172 ]

Function gfs2_clear_rgrpd calls kfree(rgd->rd_bits) before calling return_all_reservations, but return_all_reservations still dereferences rgd->rd_bits in __rs_deltree. Fix that by moving the call to kfree below the call to return_all_reservations.

Signed-off-by: Bob Peterson rpeterso@redhat.com Signed-off-by: Andreas Gruenbacher agruenba@redhat.com Signed-off-by: Sasha Levin sashal@kernel.org --- fs/gfs2/rgrp.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/fs/gfs2/rgrp.c b/fs/gfs2/rgrp.c index 7cb0672294dfc..70a344d864447 100644 --- a/fs/gfs2/rgrp.c +++ b/fs/gfs2/rgrp.c @@ -720,9 +720,9 @@ void gfs2_clear_rgrpd(struct gfs2_sbd *sdp) }

gfs2_free_clones(rgd); + return_all_reservations(rgd); kfree(rgd->rd_bits); rgd->rd_bits = NULL; - return_all_reservations(rgd); kmem_cache_free(gfs2_rgrpd_cachep, rgd); } }

From: Bob Peterson rpeterso@redhat.com

[ Upstream commit c5c68724696e7d2f8db58a5fce3673208d35c485 ]

Before this patch, gfs2_fitrim was not properly checking for a "live" file system. If the file system had something to trim and the file system was read-only (or spectator) it would start the trim, but when it starts the transaction, gfs2_trans_begin returns -EROFS (read-only file system) and it errors out. However, if the file system was already trimmed so there's no work to do, it never called gfs2_trans_begin. That code is bypassed so it never returns the error. Instead, it returns a good return code with 0 work. All this makes for inconsistent behavior: The same fstrim command can return -EROFS in one case and 0 in another. This tripped up xfstests generic/537 which reports the error as:

+fstrim with unrecovered metadata just ate your filesystem

This patch adds a check for a "live" (iow, active journal, iow, RW) file system, and if not, returns the error properly.

Signed-off-by: Bob Peterson rpeterso@redhat.com Signed-off-by: Andreas Gruenbacher agruenba@redhat.com Signed-off-by: Sasha Levin sashal@kernel.org --- fs/gfs2/rgrp.c | 3 +++ 1 file changed, 3 insertions(+)

diff --git a/fs/gfs2/rgrp.c b/fs/gfs2/rgrp.c index 70a344d864447..c4eb6a5fcea99 100644 --- a/fs/gfs2/rgrp.c +++ b/fs/gfs2/rgrp.c @@ -1361,6 +1361,9 @@ int gfs2_fitrim(struct file *filp, void __user *argp) if (!capable(CAP_SYS_ADMIN)) return -EPERM;

+ if (!test_bit(SDF_JOURNAL_LIVE, &sdp->sd_flags)) + return -EROFS; + if (!blk_queue_discard(q)) return -EOPNOTSUPP;

From: Evan Quan evan.quan@amd.com

[ Upstream commit 253475c455eb5f8da34faa1af92709e7bb414624 ]

This can address the random SDMA hang after pci config reset seen on Hawaii.

Signed-off-by: Evan Quan evan.quan@amd.com Tested-by: Sandeep Raghuraman sandy.8925@gmail.com Reviewed-by: Alex Deucher alexander.deucher@amd.com Signed-off-by: Alex Deucher alexander.deucher@amd.com Signed-off-by: Sasha Levin sashal@kernel.org --- drivers/gpu/drm/amd/amdgpu/cik_sdma.c | 27 ++++++++++++--------------- 1 file changed, 12 insertions(+), 15 deletions(-)

diff --git a/drivers/gpu/drm/amd/amdgpu/cik_sdma.c b/drivers/gpu/drm/amd/amdgpu/cik_sdma.c index 11beef7c595f2..d35e5d8e8a058 100644 --- a/drivers/gpu/drm/amd/amdgpu/cik_sdma.c +++ b/drivers/gpu/drm/amd/amdgpu/cik_sdma.c @@ -1098,22 +1098,19 @@ static int cik_sdma_soft_reset(void *handle) { u32 srbm_soft_reset = 0; struct amdgpu_device *adev = (struct amdgpu_device *)handle; - u32 tmp = RREG32(mmSRBM_STATUS2); + u32 tmp;

- if (tmp & SRBM_STATUS2__SDMA_BUSY_MASK) { - /* sdma0 */ - tmp = RREG32(mmSDMA0_F32_CNTL + SDMA0_REGISTER_OFFSET); - tmp |= SDMA0_F32_CNTL__HALT_MASK; - WREG32(mmSDMA0_F32_CNTL + SDMA0_REGISTER_OFFSET, tmp); - srbm_soft_reset |= SRBM_SOFT_RESET__SOFT_RESET_SDMA_MASK; - } - if (tmp & SRBM_STATUS2__SDMA1_BUSY_MASK) { - /* sdma1 */ - tmp = RREG32(mmSDMA0_F32_CNTL + SDMA1_REGISTER_OFFSET); - tmp |= SDMA0_F32_CNTL__HALT_MASK; - WREG32(mmSDMA0_F32_CNTL + SDMA1_REGISTER_OFFSET, tmp); - srbm_soft_reset |= SRBM_SOFT_RESET__SOFT_RESET_SDMA1_MASK; - } + /* sdma0 */ + tmp = RREG32(mmSDMA0_F32_CNTL + SDMA0_REGISTER_OFFSET); + tmp |= SDMA0_F32_CNTL__HALT_MASK; + WREG32(mmSDMA0_F32_CNTL + SDMA0_REGISTER_OFFSET, tmp); + srbm_soft_reset |= SRBM_SOFT_RESET__SOFT_RESET_SDMA_MASK; + + /* sdma1 */ + tmp = RREG32(mmSDMA0_F32_CNTL + SDMA1_REGISTER_OFFSET); + tmp |= SDMA0_F32_CNTL__HALT_MASK; + WREG32(mmSDMA0_F32_CNTL + SDMA1_REGISTER_OFFSET, tmp); + srbm_soft_reset |= SRBM_SOFT_RESET__SOFT_RESET_SDMA1_MASK;

if (srbm_soft_reset) { tmp = RREG32(mmSRBM_SOFT_RESET);

From: Johannes Berg johannes.berg@intel.com

[ Upstream commit dcd479e10a0510522a5d88b29b8f79ea3467d501 ]

When (for example) an IBSS station is pre-moved to AUTHORIZED before it's inserted, and then the insertion fails, we don't clean up the fast RX/TX states that might already have been created, since we don't go through all the state transitions again on the way down.

Do that, if it hasn't been done already, when the station is freed. I considered only freeing the fast TX/RX state there, but we might add more state so it's more robust to wind down the state properly.

Note that we warn if the station was ever inserted, it should have been properly cleaned up in that case, and the driver will probably not like things happening out of order.

Reported-by: syzbot+2e293dbd67de2836ba42@syzkaller.appspotmail.com Link: https://lore.kernel.org/r/20201009141710.7223b322a955.I95bd08b9ad0e039c03492... Signed-off-by: Johannes Berg johannes.berg@intel.com Signed-off-by: Sasha Levin sashal@kernel.org --- net/mac80211/sta_info.c | 18 ++++++++++++++++++ 1 file changed, 18 insertions(+)

diff --git a/net/mac80211/sta_info.c b/net/mac80211/sta_info.c index 2a18687019003..b74551323f5fb 100644 --- a/net/mac80211/sta_info.c +++ b/net/mac80211/sta_info.c @@ -244,6 +244,24 @@ struct sta_info *sta_info_get_by_idx(struct ieee80211_sub_if_data *sdata, */ void sta_info_free(struct ieee80211_local *local, struct sta_info *sta) { + /* + * If we had used sta_info_pre_move_state() then we might not + * have gone through the state transitions down again, so do + * it here now (and warn if it's inserted). + * + * This will clear state such as fast TX/RX that may have been + * allocated during state transitions. + */ + while (sta->sta_state > IEEE80211_STA_NONE) { + int ret; + + WARN_ON_ONCE(test_sta_flag(sta, WLAN_STA_INSERTED)); + + ret = sta_info_move_state(sta, sta->sta_state - 1); + if (WARN_ONCE(ret, "sta_info_move_state() returned %d\n", ret)) + break; + } + if (sta->rate_ctrl) rate_control_free_sta(sta);

From: Hannes Reinecke hare@suse.de

[ Upstream commit 5faf50e9e9fdc2117c61ff7e20da49cd6a29e0ca ]

alua_bus_detach() might be running concurrently with alua_rtpg_work(), so we might trip over h->sdev == NULL and call BUG_ON(). The correct way of handling it is to not set h->sdev to NULL in alua_bus_detach(), and call rcu_synchronize() before the final delete to ensure that all concurrent threads have left the critical section. Then we can get rid of the BUG_ON() and replace it with a simple if condition.

Link: https://lore.kernel.org/r/1600167537-12509-1-git-send-email-jitendra.khasdev... Link: https://lore.kernel.org/r/20200924104559.26753-1-hare@suse.de Cc: Brian Bunker brian@purestorage.com Acked-by: Brian Bunker brian@purestorage.com Tested-by: Jitendra Khasdev jitendra.khasdev@oracle.com Reviewed-by: Jitendra Khasdev jitendra.khasdev@oracle.com Signed-off-by: Hannes Reinecke hare@suse.de Signed-off-by: Martin K. Petersen martin.petersen@oracle.com Signed-off-by: Sasha Levin sashal@kernel.org --- drivers/scsi/device_handler/scsi_dh_alua.c | 9 +++++---- 1 file changed, 5 insertions(+), 4 deletions(-)

diff --git a/drivers/scsi/device_handler/scsi_dh_alua.c b/drivers/scsi/device_handler/scsi_dh_alua.c index 135376ee2cbf0..ba68454109bae 100644 --- a/drivers/scsi/device_handler/scsi_dh_alua.c +++ b/drivers/scsi/device_handler/scsi_dh_alua.c @@ -653,8 +653,8 @@ static int alua_rtpg(struct scsi_device *sdev, struct alua_port_group *pg) rcu_read_lock(); list_for_each_entry_rcu(h, &tmp_pg->dh_list, node) { - /* h->sdev should always be valid */ - BUG_ON(!h->sdev); + if (!h->sdev) + continue; h->sdev->access_state = desc[0]; } rcu_read_unlock(); @@ -700,7 +700,8 @@ static int alua_rtpg(struct scsi_device *sdev, struct alua_port_group *pg) pg->expiry = 0; rcu_read_lock(); list_for_each_entry_rcu(h, &pg->dh_list, node) { - BUG_ON(!h->sdev); + if (!h->sdev) + continue; h->sdev->access_state = (pg->state & SCSI_ACCESS_STATE_MASK); if (pg->pref) @@ -1138,7 +1139,6 @@ static void alua_bus_detach(struct scsi_device *sdev) spin_lock(&h->pg_lock); pg = rcu_dereference_protected(h->pg, lockdep_is_held(&h->pg_lock)); rcu_assign_pointer(h->pg, NULL); - h->sdev = NULL; spin_unlock(&h->pg_lock); if (pg) { spin_lock_irq(&pg->lock); @@ -1147,6 +1147,7 @@ static void alua_bus_detach(struct scsi_device *sdev) kref_put(&pg->kref, release_port_group); } sdev->handler_data = NULL; + synchronize_rcu(); kfree(h); }

From: Andy Shevchenko andriy.shevchenko@linux.intel.com

[ Upstream commit f3c75e7a9349d1d33eb53ddc1b31640994969f73 ]

When GPIO library asks pin control to set the bias, it doesn't pass any value of it and argument is considered boolean (and this is true for ACPI GpioIo() / GpioInt() resources, by the way). Thus, individual drivers must behave well, when they got the resistance value of 1 Ohm, i.e. transforming it to sane default.

In case of Intel pin control hardware the 5 kOhm sounds plausible because on one hand it's a minimum of resistors present in all hardware generations and at the same time it's high enough to minimize leakage current (will be only 200 uA with the above choice).

Fixes: e57725eabf87 ("pinctrl: intel: Add support for hardware debouncer") Reported-by: Jamie McClymont jamie@kwiius.com Signed-off-by: Andy Shevchenko andriy.shevchenko@linux.intel.com Acked-by: Mika Westerberg mika.westerberg@linux.intel.com Signed-off-by: Sasha Levin sashal@kernel.org --- drivers/pinctrl/intel/pinctrl-intel.c | 8 ++++++++ 1 file changed, 8 insertions(+)

diff --git a/drivers/pinctrl/intel/pinctrl-intel.c b/drivers/pinctrl/intel/pinctrl-intel.c index 71df0f70b61f0..45b062b0d4188 100644 --- a/drivers/pinctrl/intel/pinctrl-intel.c +++ b/drivers/pinctrl/intel/pinctrl-intel.c @@ -602,6 +602,10 @@ static int intel_config_set_pull(struct intel_pinctrl *pctrl, unsigned pin,

value |= PADCFG1_TERM_UP;

+ /* Set default strength value in case none is given */ + if (arg == 1) + arg = 5000; + switch (arg) { case 20000: value |= PADCFG1_TERM_20K << PADCFG1_TERM_SHIFT; @@ -624,6 +628,10 @@ static int intel_config_set_pull(struct intel_pinctrl *pctrl, unsigned pin, case PIN_CONFIG_BIAS_PULL_DOWN: value &= ~(PADCFG1_TERM_UP | PADCFG1_TERM_MASK);

+ /* Set default strength value in case none is given */ + if (arg == 1) + arg = 5000; + switch (arg) { case 20000: value |= PADCFG1_TERM_20K << PADCFG1_TERM_SHIFT;

From: Billy Tsai billy_tsai@aspeedtech.com

[ Upstream commit 9b92f5c51e9a41352d665f6f956bd95085a56a83 ]

Some gpio pin at aspeed soc is input only and the prefix name of these pin is "GPI" only. This patch fine-tune the condition of GPIO check from "GPIO" to "GPI" and it will fix the usage error of banks D and E in the AST2400/AST2500 and banks T and U in the AST2600.

Fixes: 4d3d0e4272d8 ("pinctrl: Add core support for Aspeed SoCs") Signed-off-by: Billy Tsai billy_tsai@aspeedtech.com Reviewed-by: Andrew Jeffery andrew@aj.id.au Link: https://lore.kernel.org/r/20201030055450.29613-1-billy_tsai@aspeedtech.com Signed-off-by: Linus Walleij linus.walleij@linaro.org Signed-off-by: Sasha Levin sashal@kernel.org --- drivers/pinctrl/aspeed/pinctrl-aspeed.c | 7 ++++--- 1 file changed, 4 insertions(+), 3 deletions(-)

diff --git a/drivers/pinctrl/aspeed/pinctrl-aspeed.c b/drivers/pinctrl/aspeed/pinctrl-aspeed.c index 7f13ce8450a34..5249033ed413e 100644 --- a/drivers/pinctrl/aspeed/pinctrl-aspeed.c +++ b/drivers/pinctrl/aspeed/pinctrl-aspeed.c @@ -458,13 +458,14 @@ int aspeed_pinmux_set_mux(struct pinctrl_dev *pctldev, unsigned int function, static bool aspeed_expr_is_gpio(const struct aspeed_sig_expr *expr) { /* - * The signal type is GPIO if the signal name has "GPIO" as a prefix. + * The signal type is GPIO if the signal name has "GPI" as a prefix. * strncmp (rather than strcmp) is used to implement the prefix * requirement. * - * expr->signal might look like "GPIOT3" in the GPIO case. + * expr->signal might look like "GPIOB1" in the GPIO case. + * expr->signal might look like "GPIT0" in the GPI case. */ - return strncmp(expr->signal, "GPIO", 4) == 0; + return strncmp(expr->signal, "GPI", 3) == 0; }

static bool aspeed_gpio_in_exprs(const struct aspeed_sig_expr **exprs)

From: Darrick J. Wong darrick.wong@oracle.com

[ Upstream commit ea8439899c0b15a176664df62aff928010fad276 ]

Pass the same oldext argument (which contains the existing rmapping's unwritten state) to xfs_rmap_lookup_le_range at the start of xfs_rmap_convert_shared. At this point in the code, flags is zero, which means that we perform lookups using the wrong key.

Fixes: 3f165b334e51 ("xfs: convert unwritten status of reverse mappings for shared files") Signed-off-by: Darrick J. Wong darrick.wong@oracle.com Reviewed-by: Christoph Hellwig hch@lst.de Signed-off-by: Sasha Levin sashal@kernel.org --- fs/xfs/libxfs/xfs_rmap.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/fs/xfs/libxfs/xfs_rmap.c b/fs/xfs/libxfs/xfs_rmap.c index 55c88a732690e..1f16c2da24723 100644 --- a/fs/xfs/libxfs/xfs_rmap.c +++ b/fs/xfs/libxfs/xfs_rmap.c @@ -1319,7 +1319,7 @@ xfs_rmap_convert_shared( * record for our insertion point. This will also give us the record for * start block contiguity tests. */ - error = xfs_rmap_lookup_le_range(cur, bno, owner, offset, flags, + error = xfs_rmap_lookup_le_range(cur, bno, owner, offset, oldext, &PREV, &i); XFS_WANT_CORRUPTED_GOTO(mp, i == 1, done);

From: Christoph Hellwig hch@lst.de

[ Upstream commit 2bd3fa793aaa7e98b74e3653fdcc72fa753913b5 ]

We also need to drop the iolock when invalidate_inode_pages2 fails, not only on all other error or successful cases.

Fixes: 527851124d10 ("xfs: implement pNFS export operations") Signed-off-by: Christoph Hellwig hch@lst.de Reviewed-by: Darrick J. Wong darrick.wong@oracle.com Signed-off-by: Darrick J. Wong darrick.wong@oracle.com Signed-off-by: Sasha Levin sashal@kernel.org --- fs/xfs/xfs_pnfs.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/fs/xfs/xfs_pnfs.c b/fs/xfs/xfs_pnfs.c index aa6c5c193f458..8538916d255ea 100644 --- a/fs/xfs/xfs_pnfs.c +++ b/fs/xfs/xfs_pnfs.c @@ -140,7 +140,7 @@ xfs_fs_map_blocks( goto out_unlock; error = invalidate_inode_pages2(inode->i_mapping); if (WARN_ON_ONCE(error)) - return error; + goto out_unlock;

end_fsb = XFS_B_TO_FSB(mp, (xfs_ufsize_t)offset + length); offset_fsb = XFS_B_TO_FSBT(mp, offset);

From: Wang Hai wanghai38@huawei.com

[ Upstream commit 52755b66ddcef2e897778fac5656df18817b59ab ]

If memory allocation for 'kbuf' succeed, cosa_write() doesn't have a corresponding kfree() in exception handling. Thus add kfree() for this function implementation.

Fixes: 1da177e4c3f4 ("Linux-2.6.12-rc2") Reported-by: Hulk Robot hulkci@huawei.com Signed-off-by: Wang Hai wanghai38@huawei.com Acked-by: Jan "Yenya" Kasprzak kas@fi.muni.cz Link: https://lore.kernel.org/r/20201110144614.43194-1-wanghai38@huawei.com Signed-off-by: Jakub Kicinski kuba@kernel.org Signed-off-by: Sasha Levin sashal@kernel.org --- drivers/net/wan/cosa.c | 1 + 1 file changed, 1 insertion(+)

diff --git a/drivers/net/wan/cosa.c b/drivers/net/wan/cosa.c index 6ea16260ec760..1a1f2d1069fb8 100644 --- a/drivers/net/wan/cosa.c +++ b/drivers/net/wan/cosa.c @@ -902,6 +902,7 @@ static ssize_t cosa_write(struct file *file, chan->tx_status = 1; spin_unlock_irqrestore(&cosa->lock, flags); up(&chan->wsem); + kfree(kbuf); return -ERESTARTSYS; } }

From: Kaixu Xia kaixuxia@tencent.com

commit 174fe5ba2d1ea0d6c5ab2a7d4aa058d6d497ae4d upstream.

The macro MOPT_Q is used to indicates the mount option is related to quota stuff and is defined to be MOPT_NOSUPPORT when CONFIG_QUOTA is disabled. Normally the quota options are handled explicitly, so it didn't matter that the MOPT_STRING flag was missing, even though the usrjquota and grpjquota mount options take a string argument. It's important that's present in the !CONFIG_QUOTA case, since without MOPT_STRING, the mount option matcher will match usrjquota= followed by an integer, and will otherwise skip the table entry, and so "mount option not supported" error message is never reported.

[ Fixed up the commit description to better explain why the fix works. --TYT ]

Fixes: 26092bf52478 ("ext4: use a table-driven handler for mount options") Signed-off-by: Kaixu Xia kaixuxia@tencent.com Link: https://lore.kernel.org/r/1603986396-28917-1-git-send-email-kaixuxia@tencent... Signed-off-by: Theodore Ts'o tytso@mit.edu Cc: stable@kernel.org Signed-off-by: Greg Kroah-Hartman gregkh@linuxfoundation.org

--- fs/ext4/super.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-)

--- a/fs/ext4/super.c +++ b/fs/ext4/super.c @@ -1679,8 +1679,8 @@ static const struct mount_opts { {Opt_noquota, (EXT4_MOUNT_QUOTA | EXT4_MOUNT_USRQUOTA | EXT4_MOUNT_GRPQUOTA | EXT4_MOUNT_PRJQUOTA), MOPT_CLEAR | MOPT_Q}, - {Opt_usrjquota, 0, MOPT_Q}, - {Opt_grpjquota, 0, MOPT_Q}, + {Opt_usrjquota, 0, MOPT_Q | MOPT_STRING}, + {Opt_grpjquota, 0, MOPT_Q | MOPT_STRING}, {Opt_offusrjquota, 0, MOPT_Q}, {Opt_offgrpjquota, 0, MOPT_Q}, {Opt_jqfmt_vfsold, QFMT_VFS_OLD, MOPT_QFMT},

From: Jing Xiangfeng jingxiangfeng@huawei.com

commit 7342ca34d931a357d408aaa25fadd031e46af137 upstream.

ring_request_msix() misses to call ida_simple_remove() in an error path. Add a label 'err_ida_remove' and jump to it.

Fixes: 046bee1f9ab8 ("thunderbolt: Add MSI-X support") Cc: stable@vger.kernel.org Signed-off-by: Jing Xiangfeng jingxiangfeng@huawei.com Reviewed-by: Andy Shevchenko andriy.shevchenko@linux.intel.com Signed-off-by: Mika Westerberg mika.westerberg@linux.intel.com Signed-off-by: Greg Kroah-Hartman gregkh@linuxfoundation.org

--- drivers/thunderbolt/nhi.c | 19 +++++++++++++++---- 1 file changed, 15 insertions(+), 4 deletions(-)

--- a/drivers/thunderbolt/nhi.c +++ b/drivers/thunderbolt/nhi.c @@ -315,12 +315,23 @@ static int ring_request_msix(struct tb_r

ring->vector = ret;

- ring->irq = pci_irq_vector(ring->nhi->pdev, ring->vector); - if (ring->irq < 0) - return ring->irq; + ret = pci_irq_vector(ring->nhi->pdev, ring->vector); + if (ret < 0) + goto err_ida_remove; + + ring->irq = ret;

irqflags = no_suspend ? IRQF_NO_SUSPEND : 0; - return request_irq(ring->irq, ring_msix, irqflags, "thunderbolt", ring); + ret = request_irq(ring->irq, ring_msix, irqflags, "thunderbolt", ring); + if (ret) + goto err_ida_remove; + + return 0; + +err_ida_remove: + ida_simple_remove(&nhi->msix_ida, ring->vector); + + return ret; }

static void ring_release_msix(struct tb_ring *ring)

From: Chris Brandt chris.brandt@renesas.com

commit 6d853c9e4104b4fc8d55dc9cd3b99712aa347174 upstream.

Renesas R-Car and RZ/G SoCs have a firmware download mode over USB. However, on reset a banner string is transmitted out which is not expected to be echoed back and will corrupt the protocol.

Cc: stable stable@vger.kernel.org Acked-by: Oliver Neukum oneukum@suse.com Signed-off-by: Chris Brandt chris.brandt@renesas.com Link: https://lore.kernel.org/r/20201111131209.3977903-1-chris.brandt@renesas.com Signed-off-by: Greg Kroah-Hartman gregkh@linuxfoundation.org

--- drivers/usb/class/cdc-acm.c | 9 +++++++++ 1 file changed, 9 insertions(+)

--- a/drivers/usb/class/cdc-acm.c +++ b/drivers/usb/class/cdc-acm.c @@ -1751,6 +1751,15 @@ static const struct usb_device_id acm_id { USB_DEVICE(0x0870, 0x0001), /* Metricom GS Modem */ .driver_info = NO_UNION_NORMAL, /* has no union descriptor */ }, + { USB_DEVICE(0x045b, 0x023c), /* Renesas USB Download mode */ + .driver_info = DISABLE_ECHO, /* Don't echo banner */ + }, + { USB_DEVICE(0x045b, 0x0248), /* Renesas USB Download mode */ + .driver_info = DISABLE_ECHO, /* Don't echo banner */ + }, + { USB_DEVICE(0x045b, 0x024D), /* Renesas USB Download mode */ + .driver_info = DISABLE_ECHO, /* Don't echo banner */ + }, { USB_DEVICE(0x0e8d, 0x0003), /* FIREFLY, MediaTek Inc; andrey.arapov@gmail.com */ .driver_info = NO_UNION_NORMAL, /* has no union descriptor */ },

From: Dan Carpenter dan.carpenter@oracle.com

commit 1e106aa3509b86738769775969822ffc1ec21bf4 upstream.

The exit_pi_state_list() function calls put_pi_state() with IRQs disabled and is not expecting that IRQs will be enabled inside the function.

Use the _irqsave() variant so that IRQs are restored to the original state instead of being enabled unconditionally.

Fixes: 153fbd1226fb ("futex: Fix more put_pi_state() vs. exit_pi_state_list() races") Signed-off-by: Dan Carpenter dan.carpenter@oracle.com Signed-off-by: Thomas Gleixner tglx@linutronix.de Acked-by: Peter Zijlstra (Intel) peterz@infradead.org Cc: stable@vger.kernel.org Link: https://lore.kernel.org/r/20201106085205.GA1159983@mwanda Signed-off-by: Greg Kroah-Hartman gregkh@linuxfoundation.org

--- kernel/futex.c | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-)

--- a/kernel/futex.c +++ b/kernel/futex.c @@ -862,8 +862,9 @@ static void put_pi_state(struct futex_pi */ if (pi_state->owner) { struct task_struct *owner; + unsigned long flags;

- raw_spin_lock_irq(&pi_state->pi_mutex.wait_lock); + raw_spin_lock_irqsave(&pi_state->pi_mutex.wait_lock, flags); owner = pi_state->owner; if (owner) { raw_spin_lock(&owner->pi_lock); @@ -871,7 +872,7 @@ static void put_pi_state(struct futex_pi raw_spin_unlock(&owner->pi_lock); } rt_mutex_proxy_unlock(&pi_state->pi_mutex, owner); - raw_spin_unlock_irq(&pi_state->pi_mutex.wait_lock); + raw_spin_unlock_irqrestore(&pi_state->pi_mutex.wait_lock, flags); }

if (current->pi_state_cache) {

From: Chen Zhou chenzhou10@huawei.com

commit c350f8bea271782e2733419bd2ab9bf4ec2051ef upstream.

Fix to return a negative error code from the error handling case instead of 0 in function sel_ib_pkey_sid_slow(), as done elsewhere in this function.

Cc: stable@vger.kernel.org Fixes: 409dcf31538a ("selinux: Add a cache for quicker retreival of PKey SIDs") Reported-by: Hulk Robot hulkci@huawei.com Signed-off-by: Chen Zhou chenzhou10@huawei.com Signed-off-by: Paul Moore paul@paul-moore.com Signed-off-by: Greg Kroah-Hartman gregkh@linuxfoundation.org

--- security/selinux/ibpkey.c | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-)

--- a/security/selinux/ibpkey.c +++ b/security/selinux/ibpkey.c @@ -160,8 +160,10 @@ static int sel_ib_pkey_sid_slow(u64 subn * is valid, it just won't be added to the cache. */ new = kzalloc(sizeof(*new), GFP_ATOMIC); - if (!new) + if (!new) { + ret = -ENOMEM; goto out; + }

new->psec.subnet_prefix = subnet_prefix; new->psec.pkey = pkey_num;

From: Thomas Zimmermann tzimmermann@suse.de

commit 06ad8d339524bf94b89859047822c31df6ace239 upstream.

The gma500 driver expects 3 pipelines in several it's IRQ functions. Accessing struct drm_device.vblank[], this fails with devices that only have 2 pipelines. An example KASAN report is shown below.

[ 62.267688] ================================================================== [ 62.268856] BUG: KASAN: slab-out-of-bounds in psb_irq_postinstall+0x250/0x3c0 [gma500_gfx] [ 62.269450] Read of size 1 at addr ffff8880012bc6d0 by task systemd-udevd/285 [ 62.269949] [ 62.270192] CPU: 0 PID: 285 Comm: systemd-udevd Tainted: G E 5.10.0-rc1-1-default+ #572 [ 62.270807] Hardware name: /DN2800MT, BIOS MTCDT10N.86A.0164.2012.1213.1024 12/13/2012 [ 62.271366] Call Trace: [ 62.271705] dump_stack+0xae/0xe5 [ 62.272180] print_address_description.constprop.0+0x17/0xf0 [ 62.272987] ? psb_irq_postinstall+0x250/0x3c0 [gma500_gfx] [ 62.273474] __kasan_report.cold+0x20/0x38 [ 62.273989] ? psb_irq_postinstall+0x250/0x3c0 [gma500_gfx] [ 62.274460] kasan_report+0x3a/0x50 [ 62.274891] psb_irq_postinstall+0x250/0x3c0 [gma500_gfx] [ 62.275380] drm_irq_install+0x131/0x1f0 <...> [ 62.300751] Allocated by task 285: [ 62.301223] kasan_save_stack+0x1b/0x40 [ 62.301731] __kasan_kmalloc.constprop.0+0xbf/0xd0 [ 62.302293] drmm_kmalloc+0x55/0x100 [ 62.302773] drm_vblank_init+0x77/0x210

Resolve the issue by only handling vblank entries up to the number of CRTCs.

I'm adding a Fixes tag for reference, although the bug has been present since the driver's initial commit.

Signed-off-by: Thomas Zimmermann tzimmermann@suse.de Reviewed-by: Daniel Vetter daniel.vetter@ffwll.ch Fixes: 5c49fd3aa0ab ("gma500: Add the core DRM files and headers") Cc: Alan Cox alan@linux.intel.com Cc: Dave Airlie airlied@redhat.com Cc: Patrik Jakobsson patrik.r.jakobsson@gmail.com Cc: dri-devel@lists.freedesktop.org Cc: stable@vger.kernel.org#v3.3+ Link: https://patchwork.freedesktop.org/patch/msgid/20201105190256.3893-1-tzimmerm... Signed-off-by: Greg Kroah-Hartman gregkh@linuxfoundation.org

--- drivers/gpu/drm/gma500/psb_irq.c | 34 ++++++++++++---------------------- 1 file changed, 12 insertions(+), 22 deletions(-)

--- a/drivers/gpu/drm/gma500/psb_irq.c +++ b/drivers/gpu/drm/gma500/psb_irq.c @@ -350,6 +350,7 @@ int psb_irq_postinstall(struct drm_devic { struct drm_psb_private *dev_priv = dev->dev_private; unsigned long irqflags; + unsigned int i;

spin_lock_irqsave(&dev_priv->irqmask_lock, irqflags);

@@ -362,20 +363,12 @@ int psb_irq_postinstall(struct drm_devic PSB_WVDC32(dev_priv->vdc_irq_mask, PSB_INT_ENABLE_R); PSB_WVDC32(0xFFFFFFFF, PSB_HWSTAM);

- if (dev->vblank[0].enabled) - psb_enable_pipestat(dev_priv, 0, PIPE_VBLANK_INTERRUPT_ENABLE); - else - psb_disable_pipestat(dev_priv, 0, PIPE_VBLANK_INTERRUPT_ENABLE); - - if (dev->vblank[1].enabled) - psb_enable_pipestat(dev_priv, 1, PIPE_VBLANK_INTERRUPT_ENABLE); - else - psb_disable_pipestat(dev_priv, 1, PIPE_VBLANK_INTERRUPT_ENABLE); - - if (dev->vblank[2].enabled) - psb_enable_pipestat(dev_priv, 2, PIPE_VBLANK_INTERRUPT_ENABLE); - else - psb_disable_pipestat(dev_priv, 2, PIPE_VBLANK_INTERRUPT_ENABLE); + for (i = 0; i < dev->num_crtcs; ++i) { + if (dev->vblank[i].enabled) + psb_enable_pipestat(dev_priv, i, PIPE_VBLANK_INTERRUPT_ENABLE); + else + psb_disable_pipestat(dev_priv, i, PIPE_VBLANK_INTERRUPT_ENABLE); + }

if (dev_priv->ops->hotplug_enable) dev_priv->ops->hotplug_enable(dev, true); @@ -388,6 +381,7 @@ void psb_irq_uninstall(struct drm_device { struct drm_psb_private *dev_priv = dev->dev_private; unsigned long irqflags; + unsigned int i;

spin_lock_irqsave(&dev_priv->irqmask_lock, irqflags);

@@ -396,14 +390,10 @@ void psb_irq_uninstall(struct drm_device

PSB_WVDC32(0xFFFFFFFF, PSB_HWSTAM);

- if (dev->vblank[0].enabled) - psb_disable_pipestat(dev_priv, 0, PIPE_VBLANK_INTERRUPT_ENABLE); - - if (dev->vblank[1].enabled) - psb_disable_pipestat(dev_priv, 1, PIPE_VBLANK_INTERRUPT_ENABLE); - - if (dev->vblank[2].enabled) - psb_disable_pipestat(dev_priv, 2, PIPE_VBLANK_INTERRUPT_ENABLE); + for (i = 0; i < dev->num_crtcs; ++i) { + if (dev->vblank[i].enabled) + psb_disable_pipestat(dev_priv, i, PIPE_VBLANK_INTERRUPT_ENABLE); + }

dev_priv->vdc_irq_mask &= _PSB_IRQ_SGX_FLAG | _PSB_IRQ_MSVDX_FLAG |

From: Coiby Xu coiby.xu@gmail.com

commit 06abe8291bc31839950f7d0362d9979edc88a666 upstream.

The correct way to disable debounce filter is to clear bit 5 and 6 of the register.

Cc: stable@vger.kerne.org Signed-off-by: Coiby Xu coiby.xu@gmail.com Reviewed-by: Hans de Goede hdegoede@redhat.com Cc: Hans de Goede hdegoede@redhat.com Link: https://lore.kernel.org/linux-gpio/df2c008b-e7b5-4fdd-42ea-4d1c62b52139@redh... Link: https://lore.kernel.org/r/20201105231912.69527-2-coiby.xu@gmail.com Signed-off-by: Linus Walleij linus.walleij@linaro.org Signed-off-by: Greg Kroah-Hartman gregkh@linuxfoundation.org

--- drivers/pinctrl/pinctrl-amd.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-)

--- a/drivers/pinctrl/pinctrl-amd.c +++ b/drivers/pinctrl/pinctrl-amd.c @@ -154,14 +154,14 @@ static int amd_gpio_set_debounce(struct pin_reg |= BIT(DB_TMR_OUT_UNIT_OFF); pin_reg |= BIT(DB_TMR_LARGE_OFF); } else { - pin_reg &= ~DB_CNTRl_MASK; + pin_reg &= ~(DB_CNTRl_MASK << DB_CNTRL_OFF); ret = -EINVAL; } } else { pin_reg &= ~BIT(DB_TMR_OUT_UNIT_OFF); pin_reg &= ~BIT(DB_TMR_LARGE_OFF); pin_reg &= ~DB_TMR_OUT_MASK; - pin_reg &= ~DB_CNTRl_MASK; + pin_reg &= ~(DB_CNTRl_MASK << DB_CNTRL_OFF); } writel(pin_reg, gpio_dev->base + offset * 4); raw_spin_unlock_irqrestore(&gpio_dev->lock, flags);

From: Oliver Herms oliver.peter.herms@gmail.com

[ Upstream commit 8ef9ba4d666614497a057d09b0a6eafc1e34eadf ]

Due to the legacy usage of hard_header_len for SIT tunnels while already using infrastructure from net/ipv4/ip_tunnel.c the calculation of the path MTU in tnl_update_pmtu is incorrect. This leads to unnecessary creation of MTU exceptions for any flow going over a SIT tunnel.

As SIT tunnels do not have a header themsevles other than their transport (L3, L2) headers we're leaving hard_header_len set to zero as tnl_update_pmtu is already taking care of the transport headers sizes.

This will also help avoiding unnecessary IPv6 GC runs and spinlock contention seen when using SIT tunnels and for more than net.ipv6.route.gc_thresh flows.

Fixes: c54419321455 ("GRE: Refactor GRE tunneling code.") Signed-off-by: Oliver Herms oliver.peter.herms@gmail.com Acked-by: Willem de Bruijn willemb@google.com Link: https://lore.kernel.org/r/20201103104133.GA1573211@tws Signed-off-by: Jakub Kicinski kuba@kernel.org Signed-off-by: Greg Kroah-Hartman gregkh@linuxfoundation.org --- net/ipv6/sit.c | 2 -- 1 file changed, 2 deletions(-)

--- a/net/ipv6/sit.c +++ b/net/ipv6/sit.c @@ -1073,7 +1073,6 @@ static void ipip6_tunnel_bind_dev(struct if (tdev && !netif_is_l3_master(tdev)) { int t_hlen = tunnel->hlen + sizeof(struct iphdr);

- dev->hard_header_len = tdev->hard_header_len + sizeof(struct iphdr); dev->mtu = tdev->mtu - t_hlen; if (dev->mtu < IPV6_MIN_MTU) dev->mtu = IPV6_MIN_MTU; @@ -1363,7 +1362,6 @@ static void ipip6_tunnel_setup(struct ne dev->priv_destructor = ipip6_dev_free;

dev->type = ARPHRD_SIT; - dev->hard_header_len = LL_MAX_HEADER + t_hlen; dev->mtu = ETH_DATA_LEN - t_hlen; dev->min_mtu = IPV6_MIN_MTU; dev->max_mtu = IP6_MAX_MTU - t_hlen;

From: Martin Schiller ms@dev.tdt.de

[ Upstream commit 361182308766a265b6c521879b34302617a8c209 ]

This fixes a regression for blocking connects introduced by commit 4becb7ee5b3d ("net/x25: Fix x25_neigh refcnt leak when x25 disconnect").

The x25->neighbour is already set to "NULL" by x25_disconnect() now, while a blocking connect is waiting in x25_wait_for_connection_establishment(). Therefore x25->neighbour must not be accessed here again and x25->state is also already set to X25_STATE_0 by x25_disconnect().

Fixes: 4becb7ee5b3d ("net/x25: Fix x25_neigh refcnt leak when x25 disconnect") Signed-off-by: Martin Schiller ms@dev.tdt.de Reviewed-by: Xie He xie.he.0141@gmail.com Link: https://lore.kernel.org/r/20201109065449.9014-1-ms@dev.tdt.de Signed-off-by: Jakub Kicinski kuba@kernel.org Signed-off-by: Greg Kroah-Hartman gregkh@linuxfoundation.org --- net/x25/af_x25.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)

--- a/net/x25/af_x25.c +++ b/net/x25/af_x25.c @@ -823,7 +823,7 @@ static int x25_connect(struct socket *so sock->state = SS_CONNECTED; rc = 0; out_put_neigh: - if (rc) { + if (rc && x25->neighbour) { read_lock_bh(&x25_list_lock); x25_neigh_put(x25->neighbour); x25->neighbour = NULL;

From: Heiner Kallweit hkallweit1@gmail.com

[ Upstream commit cc6528bc9a0c901c83b8220a2e2617f3354d6dd9 ]

The caller of rtl8169_tso_csum_v2() frees the skb if false is returned. eth_skb_pad() internally frees the skb on error what would result in a double free. Therefore use __skb_put_padto() directly and instruct it to not free the skb on error.

Fixes: b423e9ae49d7 ("r8169: fix offloaded tx checksum for small packets.") Reported-by: Jakub Kicinski kuba@kernel.org Signed-off-by: Heiner Kallweit hkallweit1@gmail.com Link: https://lore.kernel.org/r/f7e68191-acff-9ded-4263-c016428a8762@gmail.com Signed-off-by: Jakub Kicinski kuba@kernel.org Signed-off-by: Greg Kroah-Hartman gregkh@linuxfoundation.org --- drivers/net/ethernet/realtek/r8169.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-)

--- a/drivers/net/ethernet/realtek/r8169.c +++ b/drivers/net/ethernet/realtek/r8169.c @@ -7143,7 +7143,8 @@ static bool rtl8169_tso_csum_v2(struct r opts[1] |= transport_offset << TCPHO_SHIFT; } else { if (unlikely(rtl_test_hw_pad_bug(tp, skb))) - return !eth_skb_pad(skb); + /* eth_skb_pad would free the skb on error */ + return !__skb_put_padto(skb, ETH_ZLEN, false); }

return true;

From: George Spelvin lkml@sdf.org

commit c51f8f88d705e06bd696d7510aff22b33eb8e638 upstream.

Non-cryptographic PRNGs may have great statistical properties, but are usually trivially predictable to someone who knows the algorithm, given a small sample of their output. An LFSR like prandom_u32() is particularly simple, even if the sample is widely scattered bits.

It turns out the network stack uses prandom_u32() for some things like random port numbers which it would prefer are *not* trivially predictable. Predictability led to a practical DNS spoofing attack. Oops.

This patch replaces the LFSR with a homebrew cryptographic PRNG based on the SipHash round function, which is in turn seeded with 128 bits of strong random key. (The authors of SipHash have *not* been consulted about this abuse of their algorithm.) Speed is prioritized over security; attacks are rare, while performance is always wanted.

Replacing all callers of prandom_u32() is the quick fix. Whether to reinstate a weaker PRNG for uses which can tolerate it is an open question.

Commit f227e3ec3b5c ("random32: update the net random state on interrupt and activity") was an earlier attempt at a solution. This patch replaces it.

Reported-by: Amit Klein aksecurity@gmail.com Cc: Willy Tarreau w@1wt.eu Cc: Eric Dumazet edumazet@google.com Cc: "Jason A. Donenfeld" Jason@zx2c4.com Cc: Andy Lutomirski luto@kernel.org Cc: Kees Cook keescook@chromium.org Cc: Thomas Gleixner tglx@linutronix.de Cc: Peter Zijlstra peterz@infradead.org Cc: Linus Torvalds torvalds@linux-foundation.org Cc: tytso@mit.edu Cc: Florian Westphal fw@strlen.de Cc: Marc Plumb lkml.mplumb@gmail.com Fixes: f227e3ec3b5c ("random32: update the net random state on interrupt and activity") Signed-off-by: George Spelvin lkml@sdf.org Link: https://lore.kernel.org/netdev/20200808152628.GA27941@SDF.ORG/ [ willy: partial reversal of f227e3ec3b5c; moved SIPROUND definitions to prandom.h for later use; merged George's prandom_seed() proposal; inlined siprand_u32(); replaced the net_rand_state[] array with 4 members to fix a build issue; cosmetic cleanups to make checkpatch happy; fixed RANDOM32_SELFTEST build ] [wt: backported to 4.14 -- various context adjustments; timer API change] Signed-off-by: Willy Tarreau w@1wt.eu Signed-off-by: Greg Kroah-Hartman gregkh@linuxfoundation.org --- drivers/char/random.c | 1 include/linux/prandom.h | 36 +++ kernel/time/timer.c | 7 lib/random32.c | 462 +++++++++++++++++++++++++++++------------------- 4 files changed, 317 insertions(+), 189 deletions(-)

--- a/drivers/char/random.c +++ b/drivers/char/random.c @@ -1246,7 +1246,6 @@ void add_interrupt_randomness(int irq, i

fast_mix(fast_pool); add_interrupt_bench(cycles); - this_cpu_add(net_rand_state.s1, fast_pool->pool[cycles & 3]);

if (unlikely(crng_init == 0)) { if ((fast_pool->count >= 64) && --- a/include/linux/prandom.h +++ b/include/linux/prandom.h @@ -16,12 +16,44 @@ void prandom_bytes(void *buf, size_t nby void prandom_seed(u32 seed); void prandom_reseed_late(void);

+#if BITS_PER_LONG == 64 +/* + * The core SipHash round function. Each line can be executed in + * parallel given enough CPU resources. + */ +#define PRND_SIPROUND(v0, v1, v2, v3) ( \ + v0 += v1, v1 = rol64(v1, 13), v2 += v3, v3 = rol64(v3, 16), \ + v1 ^= v0, v0 = rol64(v0, 32), v3 ^= v2, \ + v0 += v3, v3 = rol64(v3, 21), v2 += v1, v1 = rol64(v1, 17), \ + v3 ^= v0, v1 ^= v2, v2 = rol64(v2, 32) \ +) + +#define PRND_K0 (0x736f6d6570736575 ^ 0x6c7967656e657261) +#define PRND_K1 (0x646f72616e646f6d ^ 0x7465646279746573) + +#elif BITS_PER_LONG == 32 +/* + * On 32-bit machines, we use HSipHash, a reduced-width version of SipHash. + * This is weaker, but 32-bit machines are not used for high-traffic + * applications, so there is less output for an attacker to analyze. + */ +#define PRND_SIPROUND(v0, v1, v2, v3) ( \ + v0 += v1, v1 = rol32(v1, 5), v2 += v3, v3 = rol32(v3, 8), \ + v1 ^= v0, v0 = rol32(v0, 16), v3 ^= v2, \ + v0 += v3, v3 = rol32(v3, 7), v2 += v1, v1 = rol32(v1, 13), \ + v3 ^= v0, v1 ^= v2, v2 = rol32(v2, 16) \ +) +#define PRND_K0 0x6c796765 +#define PRND_K1 0x74656462 + +#else +#error Unsupported BITS_PER_LONG +#endif + struct rnd_state { __u32 s1, s2, s3, s4; };

-DECLARE_PER_CPU(struct rnd_state, net_rand_state); - u32 prandom_u32_state(struct rnd_state *state); void prandom_bytes_state(struct rnd_state *state, void *buf, size_t nbytes); void prandom_seed_full_state(struct rnd_state __percpu *pcpu_state); --- a/kernel/time/timer.c +++ b/kernel/time/timer.c @@ -1596,13 +1596,6 @@ void update_process_times(int user_tick) scheduler_tick(); if (IS_ENABLED(CONFIG_POSIX_TIMERS)) run_posix_cpu_timers(p); - - /* The current CPU might make use of net randoms without receiving IRQs - * to renew them often enough. Let's update the net_rand_state from a - * non-constant value that's not affine to the number of calls to make - * sure it's updated when there's some activity (we don't care in idle). - */ - this_cpu_add(net_rand_state.s1, rol32(jiffies, 24) + user_tick); }

/** --- a/lib/random32.c +++ b/lib/random32.c @@ -40,16 +40,6 @@ #include <linux/sched.h> #include <asm/unaligned.h>

-#ifdef CONFIG_RANDOM32_SELFTEST -static void __init prandom_state_selftest(void); -#else -static inline void prandom_state_selftest(void) -{ -} -#endif - -DEFINE_PER_CPU(struct rnd_state, net_rand_state) __latent_entropy; - /** * prandom_u32_state - seeded pseudo-random number generator. * @state: pointer to state structure holding seeded state. @@ -70,25 +60,6 @@ u32 prandom_u32_state(struct rnd_state * EXPORT_SYMBOL(prandom_u32_state);

/** - * prandom_u32 - pseudo random number generator - * - * A 32 bit pseudo-random number is generated using a fast - * algorithm suitable for simulation. This algorithm is NOT - * considered safe for cryptographic use. - */ -u32 prandom_u32(void) -{ - struct rnd_state *state = &get_cpu_var(net_rand_state); - u32 res; - - res = prandom_u32_state(state); - put_cpu_var(net_rand_state); - - return res; -} -EXPORT_SYMBOL(prandom_u32); - -/** * prandom_bytes_state - get the requested number of pseudo-random bytes * * @state: pointer to state structure holding seeded state. @@ -119,20 +90,6 @@ void prandom_bytes_state(struct rnd_stat } EXPORT_SYMBOL(prandom_bytes_state);

-/** - * prandom_bytes - get the requested number of pseudo-random bytes - * @buf: where to copy the pseudo-random bytes to - * @bytes: the requested number of bytes - */ -void prandom_bytes(void *buf, size_t bytes) -{ - struct rnd_state *state = &get_cpu_var(net_rand_state); - - prandom_bytes_state(state, buf, bytes); - put_cpu_var(net_rand_state); -} -EXPORT_SYMBOL(prandom_bytes); - static void prandom_warmup(struct rnd_state *state) { /* Calling RNG ten times to satisfy recurrence condition */ @@ -148,96 +105,6 @@ static void prandom_warmup(struct rnd_st prandom_u32_state(state); }

-static u32 __extract_hwseed(void) -{ - unsigned int val = 0; - - (void)(arch_get_random_seed_int(&val) || - arch_get_random_int(&val)); - - return val; -} - -static void prandom_seed_early(struct rnd_state *state, u32 seed, - bool mix_with_hwseed) -{ -#define LCG(x) ((x) * 69069U) /* super-duper LCG */ -#define HWSEED() (mix_with_hwseed ? __extract_hwseed() : 0) - state->s1 = __seed(HWSEED() ^ LCG(seed), 2U); - state->s2 = __seed(HWSEED() ^ LCG(state->s1), 8U); - state->s3 = __seed(HWSEED() ^ LCG(state->s2), 16U); - state->s4 = __seed(HWSEED() ^ LCG(state->s3), 128U); -} - -/** - * prandom_seed - add entropy to pseudo random number generator - * @seed: seed value - * - * Add some additional seeding to the prandom pool. - */ -void prandom_seed(u32 entropy) -{ - int i; - /* - * No locking on the CPUs, but then somewhat random results are, well, - * expected. - */ - for_each_possible_cpu(i) { - struct rnd_state *state = &per_cpu(net_rand_state, i); - - state->s1 = __seed(state->s1 ^ entropy, 2U); - prandom_warmup(state); - } -} -EXPORT_SYMBOL(prandom_seed); - -/* - * Generate some initially weak seeding values to allow - * to start the prandom_u32() engine. - */ -static int __init prandom_init(void) -{ - int i; - - prandom_state_selftest(); - - for_each_possible_cpu(i) { - struct rnd_state *state = &per_cpu(net_rand_state, i); - u32 weak_seed = (i + jiffies) ^ random_get_entropy(); - - prandom_seed_early(state, weak_seed, true); - prandom_warmup(state); - } - - return 0; -} -core_initcall(prandom_init); - -static void __prandom_timer(unsigned long dontcare); - -static DEFINE_TIMER(seed_timer, __prandom_timer, 0, 0); - -static void __prandom_timer(unsigned long dontcare) -{ - u32 entropy; - unsigned long expires; - - get_random_bytes(&entropy, sizeof(entropy)); - prandom_seed(entropy); - - /* reseed every ~60 seconds, in [40 .. 80) interval with slack */ - expires = 40 + prandom_u32_max(40); - seed_timer.expires = jiffies + msecs_to_jiffies(expires * MSEC_PER_SEC); - - add_timer(&seed_timer); -} - -static void __init __prandom_start_seed_timer(void) -{ - seed_timer.expires = jiffies + msecs_to_jiffies(40 * MSEC_PER_SEC); - add_timer(&seed_timer); -} - void prandom_seed_full_state(struct rnd_state __percpu *pcpu_state) { int i; @@ -257,51 +124,6 @@ void prandom_seed_full_state(struct rnd_ } EXPORT_SYMBOL(prandom_seed_full_state);

-/* - * Generate better values after random number generator - * is fully initialized. - */ -static void __prandom_reseed(bool late) -{ - unsigned long flags; - static bool latch = false; - static DEFINE_SPINLOCK(lock); - - /* Asking for random bytes might result in bytes getting - * moved into the nonblocking pool and thus marking it - * as initialized. In this case we would double back into - * this function and attempt to do a late reseed. - * Ignore the pointless attempt to reseed again if we're - * already waiting for bytes when the nonblocking pool - * got initialized. - */ - - /* only allow initial seeding (late == false) once */ - if (!spin_trylock_irqsave(&lock, flags)) - return; - - if (latch && !late) - goto out; - - latch = true; - prandom_seed_full_state(&net_rand_state); -out: - spin_unlock_irqrestore(&lock, flags); -} - -void prandom_reseed_late(void) -{ - __prandom_reseed(true); -} - -static int __init prandom_reseed(void) -{ - __prandom_reseed(false); - __prandom_start_seed_timer(); - return 0; -} -late_initcall(prandom_reseed); - #ifdef CONFIG_RANDOM32_SELFTEST static struct prandom_test1 { u32 seed; @@ -421,7 +243,28 @@ static struct prandom_test2 { { 407983964U, 921U, 728767059U }, };

-static void __init prandom_state_selftest(void) +static u32 __extract_hwseed(void) +{ + unsigned int val = 0; + + (void)(arch_get_random_seed_int(&val) || + arch_get_random_int(&val)); + + return val; +} + +static void prandom_seed_early(struct rnd_state *state, u32 seed, + bool mix_with_hwseed) +{ +#define LCG(x) ((x) * 69069U) /* super-duper LCG */ +#define HWSEED() (mix_with_hwseed ? __extract_hwseed() : 0) + state->s1 = __seed(HWSEED() ^ LCG(seed), 2U); + state->s2 = __seed(HWSEED() ^ LCG(state->s1), 8U); + state->s3 = __seed(HWSEED() ^ LCG(state->s2), 16U); + state->s4 = __seed(HWSEED() ^ LCG(state->s3), 128U); +} + +static int __init prandom_state_selftest(void) { int i, j, errors = 0, runs = 0; bool error = false; @@ -461,5 +304,266 @@ static void __init prandom_state_selftes pr_warn("prandom: %d/%d self tests failed\n", errors, runs); else pr_info("prandom: %d self tests passed\n", runs); + return 0; +} +core_initcall(prandom_state_selftest); +#endif + +/* + * The prandom_u32() implementation is now completely separate from the + * prandom_state() functions, which are retained (for now) for compatibility. + * + * Because of (ab)use in the networking code for choosing random TCP/UDP port + * numbers, which open DoS possibilities if guessable, we want something + * stronger than a standard PRNG. But the performance requirements of + * the network code do not allow robust crypto for this application. + * + * So this is a homebrew Junior Spaceman implementation, based on the + * lowest-latency trustworthy crypto primitive available, SipHash. + * (The authors of SipHash have not been consulted about this abuse of + * their work.) + * + * Standard SipHash-2-4 uses 2n+4 rounds to hash n words of input to + * one word of output. This abbreviated version uses 2 rounds per word + * of output. + */ + +struct siprand_state { + unsigned long v0; + unsigned long v1; + unsigned long v2; + unsigned long v3; +}; + +static DEFINE_PER_CPU(struct siprand_state, net_rand_state) __latent_entropy; + +/* + * This is the core CPRNG function. As "pseudorandom", this is not used + * for truly valuable things, just intended to be a PITA to guess. + * For maximum speed, we do just two SipHash rounds per word. This is + * the same rate as 4 rounds per 64 bits that SipHash normally uses, + * so hopefully it's reasonably secure. + * + * There are two changes from the official SipHash finalization: + * - We omit some constants XORed with v2 in the SipHash spec as irrelevant; + * they are there only to make the output rounds distinct from the input + * rounds, and this application has no input rounds. + * - Rather than returning v0^v1^v2^v3, return v1+v3. + * If you look at the SipHash round, the last operation on v3 is + * "v3 ^= v0", so "v0 ^ v3" just undoes that, a waste of time. + * Likewise "v1 ^= v2". (The rotate of v2 makes a difference, but + * it still cancels out half of the bits in v2 for no benefit.) + * Second, since the last combining operation was xor, continue the + * pattern of alternating xor/add for a tiny bit of extra non-linearity. + */ +static inline u32 siprand_u32(struct siprand_state *s) +{ + unsigned long v0 = s->v0, v1 = s->v1, v2 = s->v2, v3 = s->v3; + + PRND_SIPROUND(v0, v1, v2, v3); + PRND_SIPROUND(v0, v1, v2, v3); + s->v0 = v0; s->v1 = v1; s->v2 = v2; s->v3 = v3; + return v1 + v3; +} + + +/** + * prandom_u32 - pseudo random number generator + * + * A 32 bit pseudo-random number is generated using a fast + * algorithm suitable for simulation. This algorithm is NOT + * considered safe for cryptographic use. + */ +u32 prandom_u32(void) +{ + struct siprand_state *state = get_cpu_ptr(&net_rand_state); + u32 res = siprand_u32(state); + + put_cpu_ptr(&net_rand_state); + return res; +} +EXPORT_SYMBOL(prandom_u32); + +/** + * prandom_bytes - get the requested number of pseudo-random bytes + * @buf: where to copy the pseudo-random bytes to + * @bytes: the requested number of bytes + */ +void prandom_bytes(void *buf, size_t bytes) +{ + struct siprand_state *state = get_cpu_ptr(&net_rand_state); + u8 *ptr = buf; + + while (bytes >= sizeof(u32)) { + put_unaligned(siprand_u32(state), (u32 *)ptr); + ptr += sizeof(u32); + bytes -= sizeof(u32); + } + + if (bytes > 0) { + u32 rem = siprand_u32(state); + + do { + *ptr++ = (u8)rem; + rem >>= BITS_PER_BYTE; + } while (--bytes > 0); + } + put_cpu_ptr(&net_rand_state); } +EXPORT_SYMBOL(prandom_bytes); + +/** + * prandom_seed - add entropy to pseudo random number generator + * @entropy: entropy value + * + * Add some additional seed material to the prandom pool. + * The "entropy" is actually our IP address (the only caller is + * the network code), not for unpredictability, but to ensure that + * different machines are initialized differently. + */ +void prandom_seed(u32 entropy) +{ + int i; + + add_device_randomness(&entropy, sizeof(entropy)); + + for_each_possible_cpu(i) { + struct siprand_state *state = per_cpu_ptr(&net_rand_state, i); + unsigned long v0 = state->v0, v1 = state->v1; + unsigned long v2 = state->v2, v3 = state->v3; + + do { + v3 ^= entropy; + PRND_SIPROUND(v0, v1, v2, v3); + PRND_SIPROUND(v0, v1, v2, v3); + v0 ^= entropy; + } while (unlikely(!v0 || !v1 || !v2 || !v3)); + + WRITE_ONCE(state->v0, v0); + WRITE_ONCE(state->v1, v1); + WRITE_ONCE(state->v2, v2); + WRITE_ONCE(state->v3, v3); + } +} +EXPORT_SYMBOL(prandom_seed); + +/* + * Generate some initially weak seeding values to allow + * the prandom_u32() engine to be started. + */ +static int __init prandom_init_early(void) +{ + int i; + unsigned long v0, v1, v2, v3; + + if (!arch_get_random_long(&v0)) + v0 = jiffies; + if (!arch_get_random_long(&v1)) + v1 = random_get_entropy(); + v2 = v0 ^ PRND_K0; + v3 = v1 ^ PRND_K1; + + for_each_possible_cpu(i) { + struct siprand_state *state; + + v3 ^= i; + PRND_SIPROUND(v0, v1, v2, v3); + PRND_SIPROUND(v0, v1, v2, v3); + v0 ^= i; + + state = per_cpu_ptr(&net_rand_state, i); + state->v0 = v0; state->v1 = v1; + state->v2 = v2; state->v3 = v3; + } + + return 0; +} +core_initcall(prandom_init_early); + + +/* Stronger reseeding when available, and periodically thereafter. */ +static void prandom_reseed(unsigned long dontcare); + +static DEFINE_TIMER(seed_timer, prandom_reseed, 0, 0); + +static void prandom_reseed(unsigned long dontcare) +{ + unsigned long expires; + int i; + + /* + * Reinitialize each CPU's PRNG with 128 bits of key. + * No locking on the CPUs, but then somewhat random results are, + * well, expected. + */ + for_each_possible_cpu(i) { + struct siprand_state *state; + unsigned long v0 = get_random_long(), v2 = v0 ^ PRND_K0; + unsigned long v1 = get_random_long(), v3 = v1 ^ PRND_K1; +#if BITS_PER_LONG == 32 + int j; + + /* + * On 32-bit machines, hash in two extra words to + * approximate 128-bit key length. Not that the hash + * has that much security, but this prevents a trivial + * 64-bit brute force. + */ + for (j = 0; j < 2; j++) { + unsigned long m = get_random_long(); + + v3 ^= m; + PRND_SIPROUND(v0, v1, v2, v3); + PRND_SIPROUND(v0, v1, v2, v3); + v0 ^= m; + } #endif + /* + * Probably impossible in practice, but there is a + * theoretical risk that a race between this reseeding + * and the target CPU writing its state back could + * create the all-zero SipHash fixed point. + * + * To ensure that never happens, ensure the state + * we write contains no zero words. + */ + state = per_cpu_ptr(&net_rand_state, i); + WRITE_ONCE(state->v0, v0 ? v0 : -1ul); + WRITE_ONCE(state->v1, v1 ? v1 : -1ul); + WRITE_ONCE(state->v2, v2 ? v2 : -1ul); + WRITE_ONCE(state->v3, v3 ? v3 : -1ul); + } + + /* reseed every ~60 seconds, in [40 .. 80) interval with slack */ + expires = round_jiffies(jiffies + 40 * HZ + prandom_u32_max(40 * HZ)); + mod_timer(&seed_timer, expires); +} + +/* + * The random ready callback can be called from almost any interrupt. + * To avoid worrying about whether it's safe to delay that interrupt + * long enough to seed all CPUs, just schedule an immediate timer event. + */ +static void prandom_timer_start(struct random_ready_callback *unused) +{ + mod_timer(&seed_timer, jiffies); +} + +/* + * Start periodic full reseeding as soon as strong + * random numbers are available. + */ +static int __init prandom_init_late(void) +{ + static struct random_ready_callback random_ready = { + .func = prandom_timer_start + }; + int ret = add_random_ready_callback(&random_ready); + + if (ret == -EALREADY) { + prandom_timer_start(&random_ready); + ret = 0; + } + return ret; +} +late_initcall(prandom_init_late);

From: Song Liu songliubraving@fb.com

commit 9511bce9fe8e5e6c0f923c09243a713eba560141 upstream

As Miklos reported and suggested:

"This pattern repeats two times in trace_uprobe.c and in kernel/events/core.c as well:

ret = kern_path(filename, LOOKUP_FOLLOW, &path); if (ret) goto fail_address_parse;

inode = igrab(d_inode(path.dentry)); path_put(&path);

And it's wrong. You can only hold a reference to the inode if you have an active ref to the superblock as well (which is normally through path.mnt) or holding s_umount.

This way unmounting the containing filesystem while the tracepoint is active will give you the "VFS: Busy inodes after unmount..." message and a crash when the inode is finally put.

Solution: store path instead of inode."

This patch fixes the issue in kernel/event/core.c.

Reviewed-and-tested-by: Alexander Shishkin alexander.shishkin@linux.intel.com Reported-by: Miklos Szeredi miklos@szeredi.hu Signed-off-by: Song Liu songliubraving@fb.com Signed-off-by: Peter Zijlstra (Intel) peterz@infradead.org Cc: kernel-team@fb.com Cc: Alexander Shishkin alexander.shishkin@linux.intel.com Cc: Arnaldo Carvalho de Melo acme@redhat.com Cc: Jiri Olsa jolsa@redhat.com Cc: Linus Torvalds torvalds@linux-foundation.org Cc: Peter Zijlstra peterz@infradead.org Cc: Stephane Eranian eranian@google.com Cc: Thomas Gleixner tglx@linutronix.de Cc: Vince Weaver vincent.weaver@maine.edu Fixes: 375637bc5249 ("perf/core: Introduce address range filtering") Link: http://lkml.kernel.org/r/20180418062907.3210386-2-songliubraving@fb.com Signed-off-by: Ingo Molnar mingo@kernel.org Signed-off-by: Sudip Mukherjee sudipm.mukherjee@gmail.com Signed-off-by: Greg Kroah-Hartman gregkh@linuxfoundation.org --- arch/x86/events/intel/pt.c | 4 ++-- include/linux/perf_event.h | 2 +- kernel/events/core.c | 21 +++++++++------------ 3 files changed, 12 insertions(+), 15 deletions(-)

--- a/arch/x86/events/intel/pt.c +++ b/arch/x86/events/intel/pt.c @@ -1190,7 +1190,7 @@ static int pt_event_addr_filters_validat if (!filter->range || !filter->size) return -EOPNOTSUPP;

- if (!filter->inode) { + if (!filter->path.dentry) { if (!valid_kernel_ip(filter->offset)) return -EINVAL;

@@ -1217,7 +1217,7 @@ static void pt_event_addr_filters_sync(s return;

list_for_each_entry(filter, &head->list, entry) { - if (filter->inode && !offs[range]) { + if (filter->path.dentry && !offs[range]) { msr_a = msr_b = 0; } else { /* apply the offset */ --- a/include/linux/perf_event.h +++ b/include/linux/perf_event.h @@ -466,7 +466,7 @@ struct pmu { */ struct perf_addr_filter { struct list_head entry; - struct inode *inode; + struct path path; unsigned long offset; unsigned long size; unsigned int range : 1, --- a/kernel/events/core.c +++ b/kernel/events/core.c @@ -6450,7 +6450,7 @@ static void perf_event_addr_filters_exec

raw_spin_lock_irqsave(&ifh->lock, flags); list_for_each_entry(filter, &ifh->list, entry) { - if (filter->inode) { + if (filter->path.dentry) { event->addr_filters_offs[count] = 0; restart++; } @@ -7124,7 +7124,7 @@ static bool perf_addr_filter_match(struc struct file *file, unsigned long offset, unsigned long size) { - if (filter->inode != file_inode(file)) + if (d_inode(filter->path.dentry) != file_inode(file)) return false;

if (filter->offset > offset + size) @@ -8345,8 +8345,7 @@ static void free_filters_list(struct lis struct perf_addr_filter *filter, *iter;

list_for_each_entry_safe(filter, iter, filters, entry) { - if (filter->inode) - iput(filter->inode); + path_put(&filter->path); list_del(&filter->entry); kfree(filter); } @@ -8443,7 +8442,7 @@ static void perf_event_addr_filters_appl * Adjust base offset if the filter is associated to a binary * that needs to be mapped: */ - if (filter->inode) + if (filter->path.dentry) event->addr_filters_offs[count] = perf_addr_filter_apply(filter, mm);

@@ -8516,7 +8515,6 @@ perf_event_parse_addr_filter(struct perf { struct perf_addr_filter *filter = NULL; char *start, *orig, *filename = NULL; - struct path path; substring_t args[MAX_OPT_ARGS]; int state = IF_STATE_ACTION, token; unsigned int kernel = 0; @@ -8620,19 +8618,18 @@ perf_event_parse_addr_filter(struct perf goto fail_free_name;

/* look up the path and grab its inode */ - ret = kern_path(filename, LOOKUP_FOLLOW, &path); + ret = kern_path(filename, LOOKUP_FOLLOW, + &filter->path); if (ret) goto fail_free_name;

- filter->inode = igrab(d_inode(path.dentry)); - path_put(&path); kfree(filename); filename = NULL;

ret = -EINVAL; - if (!filter->inode || - !S_ISREG(filter->inode->i_mode)) - /* free_filters_list() will iput() */ + if (!filter->path.dentry || + !S_ISREG(d_inode(filter->path.dentry) + ->i_mode)) goto fail;

event->addr_filters.nr_file_filters++;

From: "kiyin(尹亮)" kiyin@tencent.com

commit 7bdb157cdebbf95a1cd94ed2e01b338714075d00 upstream

As shown through runtime testing, the "filename" allocation is not always freed in perf_event_parse_addr_filter().

There are three possible ways that this could happen:

- It could be allocated twice on subsequent iterations through the loop, - or leaked on the success path, - or on the failure path.

Clean up the code flow to make it obvious that 'filename' is always freed in the reallocation path and in the two return paths as well.

We rely on the fact that kfree(NULL) is NOP and filename is initialized with NULL.

This fixes the leak. No other side effects expected.

[ Dan Carpenter: cleaned up the code flow & added a changelog. ] [ Ingo Molnar: updated the changelog some more. ]

Fixes: 375637bc5249 ("perf/core: Introduce address range filtering") Signed-off-by: "kiyin(尹亮)" kiyin@tencent.com Signed-off-by: Dan Carpenter dan.carpenter@oracle.com Signed-off-by: Ingo Molnar mingo@kernel.org Cc: "Srivatsa S. Bhat" srivatsa@csail.mit.edu Cc: Anthony Liguori aliguori@amazon.com Signed-off-by: Sudip Mukherjee sudipm.mukherjee@gmail.com Signed-off-by: Greg Kroah-Hartman gregkh@linuxfoundation.org --- kernel/events/core.c | 12 +++++------- 1 file changed, 5 insertions(+), 7 deletions(-)

--- a/kernel/events/core.c +++ b/kernel/events/core.c @@ -8581,6 +8581,7 @@ perf_event_parse_addr_filter(struct perf if (token == IF_SRC_FILE || token == IF_SRC_FILEADDR) { int fpos = filter->range ? 2 : 1;

+ kfree(filename); filename = match_strdup(&args[fpos]); if (!filename) { ret = -ENOMEM; @@ -8619,16 +8620,13 @@ perf_event_parse_addr_filter(struct perf */ ret = -EOPNOTSUPP; if (!event->ctx->task) - goto fail_free_name; + goto fail;

/* look up the path and grab its inode */ ret = kern_path(filename, LOOKUP_FOLLOW, &filter->path); if (ret) - goto fail_free_name; - - kfree(filename); - filename = NULL; + goto fail;

ret = -EINVAL; if (!filter->path.dentry || @@ -8648,13 +8646,13 @@ perf_event_parse_addr_filter(struct perf if (state != IF_STATE_ACTION) goto fail;

+ kfree(filename); kfree(orig);

return 0;

-fail_free_name: - kfree(filename); fail: + kfree(filename); free_filters_list(filters); kfree(orig);

From: Juergen Gross jgross@suse.com

commit 4d3fe31bd993ef504350989786858aefdb877daa upstream.

A follow-up patch will require certain write to happen before an event channel is unmasked.

While the memory barrier is not strictly necessary for all the callers, the main one will need it. In order to avoid an extra memory barrier when using fifo event channels, mandate evtchn_unmask() to provide write ordering.

The 2-level event handling unmask operation is missing an appropriate barrier, so add it. Fifo event channels are fine in this regard due to using sync_cmpxchg().

This is part of XSA-332.

Cc: stable@vger.kernel.org Suggested-by: Julien Grall julien@xen.org Signed-off-by: Juergen Gross jgross@suse.com Reviewed-by: Julien Grall jgrall@amazon.com Reviewed-by: Wei Liu wl@xen.org Signed-off-by: Greg Kroah-Hartman gregkh@linuxfoundation.org --- drivers/xen/events/events_2l.c | 2 ++ 1 file changed, 2 insertions(+)

--- a/drivers/xen/events/events_2l.c +++ b/drivers/xen/events/events_2l.c @@ -91,6 +91,8 @@ static void evtchn_2l_unmask(unsigned po

BUG_ON(!irqs_disabled());

+ smp_wmb(); /* All writes before unmask must be visible. */ + if (unlikely((cpu != cpu_from_evtchn(port)))) do_hypercall = 1; else {

From: Juergen Gross jgross@suse.com

commit 54c9de89895e0a36047fcc4ae754ea5b8655fb9d upstream.

In order to avoid tight event channel related IRQ loops add a new framework of "late EOI" handling: the IRQ the event channel is bound to will be masked until the event has been handled and the related driver is capable to handle another event. The driver is responsible for unmasking the event channel via the new function xen_irq_lateeoi().

This is similar to binding an event channel to a threaded IRQ, but without having to structure the driver accordingly.

In order to support a future special handling in case a rogue guest is sending lots of unsolicited events, add a flag to xen_irq_lateeoi() which can be set by the caller to indicate the event was a spurious one.

This is part of XSA-332.

Cc: stable@vger.kernel.org Reported-by: Julien Grall julien@xen.org Signed-off-by: Juergen Gross jgross@suse.com Reviewed-by: Jan Beulich jbeulich@suse.com Reviewed-by: Stefano Stabellini sstabellini@kernel.org Reviewed-by: Wei Liu wl@xen.org Signed-off-by: Greg Kroah-Hartman gregkh@linuxfoundation.org --- drivers/xen/events/events_base.c | 151 ++++++++++++++++++++++++++++++++++----- include/xen/events.h | 29 ++++++- 2 files changed, 159 insertions(+), 21 deletions(-)

--- a/drivers/xen/events/events_base.c +++ b/drivers/xen/events/events_base.c @@ -111,6 +111,7 @@ static bool (*pirq_needs_eoi)(unsigned i static struct irq_info *legacy_info_ptrs[NR_IRQS_LEGACY];

static struct irq_chip xen_dynamic_chip; +static struct irq_chip xen_lateeoi_chip; static struct irq_chip xen_percpu_chip; static struct irq_chip xen_pirq_chip; static void enable_dynirq(struct irq_data *data); @@ -395,6 +396,33 @@ void notify_remote_via_irq(int irq) } EXPORT_SYMBOL_GPL(notify_remote_via_irq);

+static void xen_irq_lateeoi_locked(struct irq_info *info) +{ + evtchn_port_t evtchn; + + evtchn = info->evtchn; + if (!VALID_EVTCHN(evtchn)) + return; + + unmask_evtchn(evtchn); +} + +void xen_irq_lateeoi(unsigned int irq, unsigned int eoi_flags) +{ + struct irq_info *info; + unsigned long flags; + + read_lock_irqsave(&evtchn_rwlock, flags); + + info = info_for_irq(irq); + + if (info) + xen_irq_lateeoi_locked(info); + + read_unlock_irqrestore(&evtchn_rwlock, flags); +} +EXPORT_SYMBOL_GPL(xen_irq_lateeoi); + static void xen_irq_init(unsigned irq) { struct irq_info *info; @@ -866,7 +894,7 @@ int xen_pirq_from_irq(unsigned irq) } EXPORT_SYMBOL_GPL(xen_pirq_from_irq);

-int bind_evtchn_to_irq(unsigned int evtchn) +static int bind_evtchn_to_irq_chip(evtchn_port_t evtchn, struct irq_chip *chip) { int irq; int ret; @@ -883,7 +911,7 @@ int bind_evtchn_to_irq(unsigned int evtc if (irq < 0) goto out;

- irq_set_chip_and_handler_name(irq, &xen_dynamic_chip, + irq_set_chip_and_handler_name(irq, chip, handle_edge_irq, "event");

ret = xen_irq_info_evtchn_setup(irq, evtchn); @@ -904,8 +932,19 @@ out:

return irq; } + +int bind_evtchn_to_irq(evtchn_port_t evtchn) +{ + return bind_evtchn_to_irq_chip(evtchn, &xen_dynamic_chip); +} EXPORT_SYMBOL_GPL(bind_evtchn_to_irq);

+int bind_evtchn_to_irq_lateeoi(evtchn_port_t evtchn) +{ + return bind_evtchn_to_irq_chip(evtchn, &xen_lateeoi_chip); +} +EXPORT_SYMBOL_GPL(bind_evtchn_to_irq_lateeoi); + static int bind_ipi_to_irq(unsigned int ipi, unsigned int cpu) { struct evtchn_bind_ipi bind_ipi; @@ -947,8 +986,9 @@ static int bind_ipi_to_irq(unsigned int return irq; }

-int bind_interdomain_evtchn_to_irq(unsigned int remote_domain, - unsigned int remote_port) +static int bind_interdomain_evtchn_to_irq_chip(unsigned int remote_domain, + evtchn_port_t remote_port, + struct irq_chip *chip) { struct evtchn_bind_interdomain bind_interdomain; int err; @@ -959,10 +999,26 @@ int bind_interdomain_evtchn_to_irq(unsig err = HYPERVISOR_event_channel_op(EVTCHNOP_bind_interdomain, &bind_interdomain);

- return err ? : bind_evtchn_to_irq(bind_interdomain.local_port); + return err ? : bind_evtchn_to_irq_chip(bind_interdomain.local_port, + chip); +} + +int bind_interdomain_evtchn_to_irq(unsigned int remote_domain, + evtchn_port_t remote_port) +{ + return bind_interdomain_evtchn_to_irq_chip(remote_domain, remote_port, + &xen_dynamic_chip); } EXPORT_SYMBOL_GPL(bind_interdomain_evtchn_to_irq);

+int bind_interdomain_evtchn_to_irq_lateeoi(unsigned int remote_domain, + evtchn_port_t remote_port) +{ + return bind_interdomain_evtchn_to_irq_chip(remote_domain, remote_port, + &xen_lateeoi_chip); +} +EXPORT_SYMBOL_GPL(bind_interdomain_evtchn_to_irq_lateeoi); + static int find_virq(unsigned int virq, unsigned int cpu) { struct evtchn_status status; @@ -1058,14 +1114,15 @@ static void unbind_from_irq(unsigned int mutex_unlock(&irq_mapping_update_lock); }

-int bind_evtchn_to_irqhandler(unsigned int evtchn, - irq_handler_t handler, - unsigned long irqflags, - const char *devname, void *dev_id) +static int bind_evtchn_to_irqhandler_chip(evtchn_port_t evtchn, + irq_handler_t handler, + unsigned long irqflags, + const char *devname, void *dev_id, + struct irq_chip *chip) { int irq, retval;

- irq = bind_evtchn_to_irq(evtchn); + irq = bind_evtchn_to_irq_chip(evtchn, chip); if (irq < 0) return irq; retval = request_irq(irq, handler, irqflags, devname, dev_id); @@ -1076,18 +1133,38 @@ int bind_evtchn_to_irqhandler(unsigned i

return irq; } + +int bind_evtchn_to_irqhandler(evtchn_port_t evtchn, + irq_handler_t handler, + unsigned long irqflags, + const char *devname, void *dev_id) +{ + return bind_evtchn_to_irqhandler_chip(evtchn, handler, irqflags, + devname, dev_id, + &xen_dynamic_chip); +} EXPORT_SYMBOL_GPL(bind_evtchn_to_irqhandler);

-int bind_interdomain_evtchn_to_irqhandler(unsigned int remote_domain, - unsigned int remote_port, - irq_handler_t handler, - unsigned long irqflags, - const char *devname, - void *dev_id) +int bind_evtchn_to_irqhandler_lateeoi(evtchn_port_t evtchn, + irq_handler_t handler, + unsigned long irqflags, + const char *devname, void *dev_id) +{ + return bind_evtchn_to_irqhandler_chip(evtchn, handler, irqflags, + devname, dev_id, + &xen_lateeoi_chip); +} +EXPORT_SYMBOL_GPL(bind_evtchn_to_irqhandler_lateeoi); + +static int bind_interdomain_evtchn_to_irqhandler_chip( + unsigned int remote_domain, evtchn_port_t remote_port, + irq_handler_t handler, unsigned long irqflags, + const char *devname, void *dev_id, struct irq_chip *chip) { int irq, retval;

- irq = bind_interdomain_evtchn_to_irq(remote_domain, remote_port); + irq = bind_interdomain_evtchn_to_irq_chip(remote_domain, remote_port, + chip); if (irq < 0) return irq;

@@ -1099,8 +1176,33 @@ int bind_interdomain_evtchn_to_irqhandle

return irq; } + +int bind_interdomain_evtchn_to_irqhandler(unsigned int remote_domain, + evtchn_port_t remote_port, + irq_handler_t handler, + unsigned long irqflags, + const char *devname, + void *dev_id) +{ + return bind_interdomain_evtchn_to_irqhandler_chip(remote_domain, + remote_port, handler, irqflags, devname, + dev_id, &xen_dynamic_chip); +} EXPORT_SYMBOL_GPL(bind_interdomain_evtchn_to_irqhandler);

+int bind_interdomain_evtchn_to_irqhandler_lateeoi(unsigned int remote_domain, + evtchn_port_t remote_port, + irq_handler_t handler, + unsigned long irqflags, + const char *devname, + void *dev_id) +{ + return bind_interdomain_evtchn_to_irqhandler_chip(remote_domain, + remote_port, handler, irqflags, devname, + dev_id, &xen_lateeoi_chip); +} +EXPORT_SYMBOL_GPL(bind_interdomain_evtchn_to_irqhandler_lateeoi); + int bind_virq_to_irqhandler(unsigned int virq, unsigned int cpu, irq_handler_t handler, unsigned long irqflags, const char *devname, void *dev_id) @@ -1641,6 +1743,21 @@ static struct irq_chip xen_dynamic_chip .irq_mask_ack = mask_ack_dynirq,

.irq_set_affinity = set_affinity_irq, + .irq_retrigger = retrigger_dynirq, +}; + +static struct irq_chip xen_lateeoi_chip __read_mostly = { + /* The chip name needs to contain "xen-dyn" for irqbalance to work. */ + .name = "xen-dyn-lateeoi", + + .irq_disable = disable_dynirq, + .irq_mask = disable_dynirq, + .irq_unmask = enable_dynirq, + + .irq_ack = mask_ack_dynirq, + .irq_mask_ack = mask_ack_dynirq, + + .irq_set_affinity = set_affinity_irq, .irq_retrigger = retrigger_dynirq, };

--- a/include/xen/events.h +++ b/include/xen/events.h @@ -14,11 +14,16 @@

unsigned xen_evtchn_nr_channels(void);

-int bind_evtchn_to_irq(unsigned int evtchn); -int bind_evtchn_to_irqhandler(unsigned int evtchn, +int bind_evtchn_to_irq(evtchn_port_t evtchn); +int bind_evtchn_to_irq_lateeoi(evtchn_port_t evtchn); +int bind_evtchn_to_irqhandler(evtchn_port_t evtchn, irq_handler_t handler, unsigned long irqflags, const char *devname, void *dev_id); +int bind_evtchn_to_irqhandler_lateeoi(evtchn_port_t evtchn, + irq_handler_t handler, + unsigned long irqflags, const char *devname, + void *dev_id); int bind_virq_to_irq(unsigned int virq, unsigned int cpu, bool percpu); int bind_virq_to_irqhandler(unsigned int virq, unsigned int cpu, irq_handler_t handler, @@ -31,13 +36,21 @@ int bind_ipi_to_irqhandler(enum ipi_vect const char *devname, void *dev_id); int bind_interdomain_evtchn_to_irq(unsigned int remote_domain, - unsigned int remote_port); + evtchn_port_t remote_port); +int bind_interdomain_evtchn_to_irq_lateeoi(unsigned int remote_domain, + evtchn_port_t remote_port); int bind_interdomain_evtchn_to_irqhandler(unsigned int remote_domain, - unsigned int remote_port, + evtchn_port_t remote_port, irq_handler_t handler, unsigned long irqflags, const char *devname, void *dev_id); +int bind_interdomain_evtchn_to_irqhandler_lateeoi(unsigned int remote_domain, + evtchn_port_t remote_port, + irq_handler_t handler, + unsigned long irqflags, + const char *devname, + void *dev_id);

/* * Common unbind function for all event sources. Takes IRQ to unbind from. @@ -46,6 +59,14 @@ int bind_interdomain_evtchn_to_irqhandle */ void unbind_from_irqhandler(unsigned int irq, void *dev_id);

+/* + * Send late EOI for an IRQ bound to an event channel via one of the *_lateeoi + * functions above. + */ +void xen_irq_lateeoi(unsigned int irq, unsigned int eoi_flags); +/* Signal an event was spurious, i.e. there was no action resulting from it. */ +#define XEN_EOI_FLAG_SPURIOUS 0x00000001 + #define XEN_IRQ_PRIORITY_MAX EVTCHN_FIFO_PRIORITY_MAX #define XEN_IRQ_PRIORITY_DEFAULT EVTCHN_FIFO_PRIORITY_DEFAULT #define XEN_IRQ_PRIORITY_MIN EVTCHN_FIFO_PRIORITY_MIN

From: Juergen Gross jgross@suse.com

commit 23025393dbeb3b8b3b60ebfa724cdae384992e27 upstream.

In order to reduce the chance for the system becoming unresponsive due to event storms triggered by a misbehaving netfront use the lateeoi irq binding for netback and unmask the event channel only just before going to sleep waiting for new events.

Make sure not to issue an EOI when none is pending by introducing an eoi_pending element to struct xenvif_queue.

When no request has been consumed set the spurious flag when sending the EOI for an interrupt.

This is part of XSA-332.

Cc: stable@vger.kernel.org Reported-by: Julien Grall julien@xen.org Signed-off-by: Juergen Gross jgross@suse.com Reviewed-by: Jan Beulich jbeulich@suse.com Reviewed-by: Wei Liu wl@xen.org Signed-off-by: Greg Kroah-Hartman gregkh@linuxfoundation.org --- drivers/net/xen-netback/common.h | 15 ++++++++ drivers/net/xen-netback/interface.c | 61 ++++++++++++++++++++++++++++++------ drivers/net/xen-netback/netback.c | 11 +++++- drivers/net/xen-netback/rx.c | 13 +++++-- 4 files changed, 86 insertions(+), 14 deletions(-)

--- a/drivers/net/xen-netback/common.h +++ b/drivers/net/xen-netback/common.h @@ -140,6 +140,20 @@ struct xenvif_queue { /* Per-queue data char name[QUEUE_NAME_SIZE]; /* DEVNAME-qN */ struct xenvif *vif; /* Parent VIF */

+ /* + * TX/RX common EOI handling. + * When feature-split-event-channels = 0, interrupt handler sets + * NETBK_COMMON_EOI, otherwise NETBK_RX_EOI and NETBK_TX_EOI are set + * by the RX and TX interrupt handlers. + * RX and TX handler threads will issue an EOI when either + * NETBK_COMMON_EOI or their specific bits (NETBK_RX_EOI or + * NETBK_TX_EOI) are set and they will reset those bits. + */ + atomic_t eoi_pending; +#define NETBK_RX_EOI 0x01 +#define NETBK_TX_EOI 0x02 +#define NETBK_COMMON_EOI 0x04 + /* Use NAPI for guest TX */ struct napi_struct napi; /* When feature-split-event-channels = 0, tx_irq = rx_irq. */ @@ -356,6 +370,7 @@ int xenvif_dealloc_kthread(void *data);

irqreturn_t xenvif_ctrl_irq_fn(int irq, void *data);

+bool xenvif_have_rx_work(struct xenvif_queue *queue, bool test_kthread); void xenvif_rx_action(struct xenvif_queue *queue); void xenvif_rx_queue_tail(struct xenvif_queue *queue, struct sk_buff *skb);

--- a/drivers/net/xen-netback/interface.c +++ b/drivers/net/xen-netback/interface.c @@ -77,12 +77,28 @@ int xenvif_schedulable(struct xenvif *vi !vif->disabled; }

+static bool xenvif_handle_tx_interrupt(struct xenvif_queue *queue) +{ + bool rc; + + rc = RING_HAS_UNCONSUMED_REQUESTS(&queue->tx); + if (rc) + napi_schedule(&queue->napi); + return rc; +} + static irqreturn_t xenvif_tx_interrupt(int irq, void *dev_id) { struct xenvif_queue *queue = dev_id; + int old;

- if (RING_HAS_UNCONSUMED_REQUESTS(&queue->tx)) - napi_schedule(&queue->napi); + old = atomic_fetch_or(NETBK_TX_EOI, &queue->eoi_pending); + WARN(old & NETBK_TX_EOI, "Interrupt while EOI pending\n"); + + if (!xenvif_handle_tx_interrupt(queue)) { + atomic_andnot(NETBK_TX_EOI, &queue->eoi_pending); + xen_irq_lateeoi(irq, XEN_EOI_FLAG_SPURIOUS); + }

return IRQ_HANDLED; } @@ -116,19 +132,46 @@ static int xenvif_poll(struct napi_struc return work_done; }

+static bool xenvif_handle_rx_interrupt(struct xenvif_queue *queue) +{ + bool rc; + + rc = xenvif_have_rx_work(queue, false); + if (rc) + xenvif_kick_thread(queue); + return rc; +} + static irqreturn_t xenvif_rx_interrupt(int irq, void *dev_id) { struct xenvif_queue *queue = dev_id; + int old;

- xenvif_kick_thread(queue); + old = atomic_fetch_or(NETBK_RX_EOI, &queue->eoi_pending); + WARN(old & NETBK_RX_EOI, "Interrupt while EOI pending\n"); + + if (!xenvif_handle_rx_interrupt(queue)) { + atomic_andnot(NETBK_RX_EOI, &queue->eoi_pending); + xen_irq_lateeoi(irq, XEN_EOI_FLAG_SPURIOUS); + }

return IRQ_HANDLED; }

irqreturn_t xenvif_interrupt(int irq, void *dev_id) { - xenvif_tx_interrupt(irq, dev_id); - xenvif_rx_interrupt(irq, dev_id); + struct xenvif_queue *queue = dev_id; + int old; + + old = atomic_fetch_or(NETBK_COMMON_EOI, &queue->eoi_pending); + WARN(old, "Interrupt while EOI pending\n"); + + /* Use bitwise or as we need to call both functions. */ + if ((!xenvif_handle_tx_interrupt(queue) | + !xenvif_handle_rx_interrupt(queue))) { + atomic_andnot(NETBK_COMMON_EOI, &queue->eoi_pending); + xen_irq_lateeoi(irq, XEN_EOI_FLAG_SPURIOUS); + }

return IRQ_HANDLED; } @@ -595,7 +638,7 @@ int xenvif_connect_ctrl(struct xenvif *v shared = (struct xen_netif_ctrl_sring *)addr; BACK_RING_INIT(&vif->ctrl, shared, XEN_PAGE_SIZE);

- err = bind_interdomain_evtchn_to_irq(vif->domid, evtchn); + err = bind_interdomain_evtchn_to_irq_lateeoi(vif->domid, evtchn); if (err < 0) goto err_unmap;

@@ -653,7 +696,7 @@ int xenvif_connect_data(struct xenvif_qu

if (tx_evtchn == rx_evtchn) { /* feature-split-event-channels == 0 */ - err = bind_interdomain_evtchn_to_irqhandler( + err = bind_interdomain_evtchn_to_irqhandler_lateeoi( queue->vif->domid, tx_evtchn, xenvif_interrupt, 0, queue->name, queue); if (err < 0) @@ -664,7 +707,7 @@ int xenvif_connect_data(struct xenvif_qu /* feature-split-event-channels == 1 */ snprintf(queue->tx_irq_name, sizeof(queue->tx_irq_name), "%s-tx", queue->name); - err = bind_interdomain_evtchn_to_irqhandler( + err = bind_interdomain_evtchn_to_irqhandler_lateeoi( queue->vif->domid, tx_evtchn, xenvif_tx_interrupt, 0, queue->tx_irq_name, queue); if (err < 0) @@ -674,7 +717,7 @@ int xenvif_connect_data(struct xenvif_qu

snprintf(queue->rx_irq_name, sizeof(queue->rx_irq_name), "%s-rx", queue->name); - err = bind_interdomain_evtchn_to_irqhandler( + err = bind_interdomain_evtchn_to_irqhandler_lateeoi( queue->vif->domid, rx_evtchn, xenvif_rx_interrupt, 0, queue->rx_irq_name, queue); if (err < 0) --- a/drivers/net/xen-netback/netback.c +++ b/drivers/net/xen-netback/netback.c @@ -162,6 +162,10 @@ void xenvif_napi_schedule_or_enable_even

if (more_to_do) napi_schedule(&queue->napi); + else if (atomic_fetch_andnot(NETBK_TX_EOI | NETBK_COMMON_EOI, + &queue->eoi_pending) & + (NETBK_TX_EOI | NETBK_COMMON_EOI)) + xen_irq_lateeoi(queue->tx_irq, 0); }

static void tx_add_credit(struct xenvif_queue *queue) @@ -1615,9 +1619,14 @@ static bool xenvif_ctrl_work_todo(struct irqreturn_t xenvif_ctrl_irq_fn(int irq, void *data) { struct xenvif *vif = data; + unsigned int eoi_flag = XEN_EOI_FLAG_SPURIOUS;

- while (xenvif_ctrl_work_todo(vif)) + while (xenvif_ctrl_work_todo(vif)) { xenvif_ctrl_action(vif); + eoi_flag = 0; + } + + xen_irq_lateeoi(irq, eoi_flag);

return IRQ_HANDLED; } --- a/drivers/net/xen-netback/rx.c +++ b/drivers/net/xen-netback/rx.c @@ -490,13 +490,13 @@ static bool xenvif_rx_queue_ready(struct return queue->stalled && prod - cons >= 1; }

-static bool xenvif_have_rx_work(struct xenvif_queue *queue) +bool xenvif_have_rx_work(struct xenvif_queue *queue, bool test_kthread) { return xenvif_rx_ring_slots_available(queue) || (queue->vif->stall_timeout && (xenvif_rx_queue_stalled(queue) || xenvif_rx_queue_ready(queue))) || - kthread_should_stop() || + (test_kthread && kthread_should_stop()) || queue->vif->disabled; }

@@ -527,15 +527,20 @@ static void xenvif_wait_for_rx_work(stru { DEFINE_WAIT(wait);

- if (xenvif_have_rx_work(queue)) + if (xenvif_have_rx_work(queue, true)) return;

for (;;) { long ret;

prepare_to_wait(&queue->wq, &wait, TASK_INTERRUPTIBLE); - if (xenvif_have_rx_work(queue)) + if (xenvif_have_rx_work(queue, true)) break; + if (atomic_fetch_andnot(NETBK_RX_EOI | NETBK_COMMON_EOI, + &queue->eoi_pending) & + (NETBK_RX_EOI | NETBK_COMMON_EOI)) + xen_irq_lateeoi(queue->rx_irq, 0); + ret = schedule_timeout(xenvif_rx_queue_timeout(queue)); if (!ret) break;

From: Juergen Gross jgross@suse.com

commit c8d647a326f06a39a8e5f0f1af946eacfa1835f8 upstream.

In order to reduce the chance for the system becoming unresponsive due to event storms triggered by a misbehaving pvcallsfront use the lateeoi irq binding for pvcallsback and unmask the event channel only after handling all write requests, which are the ones coming in via an irq.

This requires modifying the logic a little bit to not require an event for each write request, but to keep the ioworker running until no further data is found on the ring page to be processed.

This is part of XSA-332.

Cc: stable@vger.kernel.org Reported-by: Julien Grall julien@xen.org Signed-off-by: Juergen Gross jgross@suse.com Reviewed-by: Stefano Stabellini sstabellini@kernel.org Reviewed-by: Wei Liu wl@xen.org Signed-off-by: Greg Kroah-Hartman gregkh@linuxfoundation.org --- drivers/xen/pvcalls-back.c | 76 +++++++++++++++++++++++++++------------------ 1 file changed, 46 insertions(+), 30 deletions(-)

--- a/drivers/xen/pvcalls-back.c +++ b/drivers/xen/pvcalls-back.c @@ -75,6 +75,7 @@ struct sock_mapping { atomic_t write; atomic_t io; atomic_t release; + atomic_t eoi; void (*saved_data_ready)(struct sock *sk); struct pvcalls_ioworker ioworker; }; @@ -96,7 +97,7 @@ static int pvcalls_back_release_active(s struct pvcalls_fedata *fedata, struct sock_mapping *map);

-static void pvcalls_conn_back_read(void *opaque) +static bool pvcalls_conn_back_read(void *opaque) { struct sock_mapping *map = (struct sock_mapping *)opaque; struct msghdr msg; @@ -116,17 +117,17 @@ static void pvcalls_conn_back_read(void virt_mb();

if (error) - return; + return false;

size = pvcalls_queued(prod, cons, array_size); if (size >= array_size) - return; + return false; spin_lock_irqsave(&map->sock->sk->sk_receive_queue.lock, flags); if (skb_queue_empty(&map->sock->sk->sk_receive_queue)) { atomic_set(&map->read, 0); spin_unlock_irqrestore(&map->sock->sk->sk_receive_queue.lock, flags); - return; + return true; } spin_unlock_irqrestore(&map->sock->sk->sk_receive_queue.lock, flags); wanted = array_size - size; @@ -154,7 +155,7 @@ static void pvcalls_conn_back_read(void ret = inet_recvmsg(map->sock, &msg, wanted, MSG_DONTWAIT); WARN_ON(ret > wanted); if (ret == -EAGAIN) /* shouldn't happen */ - return; + return true; if (!ret) ret = -ENOTCONN; spin_lock_irqsave(&map->sock->sk->sk_receive_queue.lock, flags); @@ -173,10 +174,10 @@ static void pvcalls_conn_back_read(void virt_wmb(); notify_remote_via_irq(map->irq);

- return; + return true; }

-static void pvcalls_conn_back_write(struct sock_mapping *map) +static bool pvcalls_conn_back_write(struct sock_mapping *map) { struct pvcalls_data_intf *intf = map->ring; struct pvcalls_data *data = &map->data; @@ -193,7 +194,7 @@ static void pvcalls_conn_back_write(stru array_size = XEN_FLEX_RING_SIZE(map->ring_order); size = pvcalls_queued(prod, cons, array_size); if (size == 0) - return; + return false;

memset(&msg, 0, sizeof(msg)); msg.msg_flags |= MSG_DONTWAIT; @@ -215,12 +216,11 @@ static void pvcalls_conn_back_write(stru

atomic_set(&map->write, 0); ret = inet_sendmsg(map->sock, &msg, size); - if (ret == -EAGAIN || (ret >= 0 && ret < size)) { + if (ret == -EAGAIN) { atomic_inc(&map->write); atomic_inc(&map->io); + return true; } - if (ret == -EAGAIN) - return;

/* write the data, then update the indexes */ virt_wmb(); @@ -233,9 +233,13 @@ static void pvcalls_conn_back_write(stru } /* update the indexes, then notify the other end */ virt_wmb(); - if (prod != cons + ret) + if (prod != cons + ret) { atomic_inc(&map->write); + atomic_inc(&map->io); + } notify_remote_via_irq(map->irq); + + return true; }

static void pvcalls_back_ioworker(struct work_struct *work) @@ -244,6 +248,7 @@ static void pvcalls_back_ioworker(struct struct pvcalls_ioworker, register_work); struct sock_mapping *map = container_of(ioworker, struct sock_mapping, ioworker); + unsigned int eoi_flags = XEN_EOI_FLAG_SPURIOUS;

while (atomic_read(&map->io) > 0) { if (atomic_read(&map->release) > 0) { @@ -251,10 +256,18 @@ static void pvcalls_back_ioworker(struct return; }

- if (atomic_read(&map->read) > 0) - pvcalls_conn_back_read(map); - if (atomic_read(&map->write) > 0) - pvcalls_conn_back_write(map); + if (atomic_read(&map->read) > 0 && + pvcalls_conn_back_read(map)) + eoi_flags = 0; + if (atomic_read(&map->write) > 0 && + pvcalls_conn_back_write(map)) + eoi_flags = 0; + + if (atomic_read(&map->eoi) > 0 && !atomic_read(&map->write)) { + atomic_set(&map->eoi, 0); + xen_irq_lateeoi(map->irq, eoi_flags); + eoi_flags = XEN_EOI_FLAG_SPURIOUS; + }

atomic_dec(&map->io); } @@ -351,12 +364,9 @@ static struct sock_mapping *pvcalls_new_ goto out; map->bytes = page;

- ret = bind_interdomain_evtchn_to_irqhandler(fedata->dev->otherend_id, - evtchn, - pvcalls_back_conn_event, - 0, - "pvcalls-backend", - map); + ret = bind_interdomain_evtchn_to_irqhandler_lateeoi( + fedata->dev->otherend_id, evtchn, + pvcalls_back_conn_event, 0, "pvcalls-backend", map); if (ret < 0) goto out; map->irq = ret; @@ -890,15 +900,18 @@ static irqreturn_t pvcalls_back_event(in { struct xenbus_device *dev = dev_id; struct pvcalls_fedata *fedata = NULL; + unsigned int eoi_flags = XEN_EOI_FLAG_SPURIOUS;

- if (dev == NULL) - return IRQ_HANDLED; + if (dev) { + fedata = dev_get_drvdata(&dev->dev); + if (fedata) { + pvcalls_back_work(fedata); + eoi_flags = 0; + } + }

- fedata = dev_get_drvdata(&dev->dev); - if (fedata == NULL) - return IRQ_HANDLED; + xen_irq_lateeoi(irq, eoi_flags);

- pvcalls_back_work(fedata); return IRQ_HANDLED; }

@@ -908,12 +921,15 @@ static irqreturn_t pvcalls_back_conn_eve struct pvcalls_ioworker *iow;

if (map == NULL || map->sock == NULL || map->sock->sk == NULL || - map->sock->sk->sk_user_data != map) + map->sock->sk->sk_user_data != map) { + xen_irq_lateeoi(irq, 0); return IRQ_HANDLED; + }

iow = &map->ioworker;

atomic_inc(&map->write); + atomic_inc(&map->eoi); atomic_inc(&map->io); queue_work(iow->wq, &iow->register_work);

@@ -948,7 +964,7 @@ static int backend_connect(struct xenbus goto error; }

- err = bind_interdomain_evtchn_to_irq(dev->otherend_id, evtchn); + err = bind_interdomain_evtchn_to_irq_lateeoi(dev->otherend_id, evtchn); if (err < 0) goto error; fedata->irq = err;

From: Juergen Gross jgross@suse.com

commit c44b849cee8c3ac587da3b0980e01f77500d158c upstream.

Instead of disabling the irq when an event is received and enabling it again when handled by the user process use the lateeoi model.

This is part of XSA-332.

Cc: stable@vger.kernel.org Reported-by: Julien Grall julien@xen.org Signed-off-by: Juergen Gross jgross@suse.com Tested-by: Stefano Stabellini sstabellini@kernel.org Reviewed-by: Stefano Stabellini sstabellini@kernel.org Reviewed-by: Jan Beulich jbeulich@suse.com Reviewed-by: Wei Liu wl@xen.org Signed-off-by: Greg Kroah-Hartman gregkh@linuxfoundation.org --- drivers/xen/evtchn.c | 7 +++---- 1 file changed, 3 insertions(+), 4 deletions(-)

--- a/drivers/xen/evtchn.c +++ b/drivers/xen/evtchn.c @@ -166,7 +166,6 @@ static irqreturn_t evtchn_interrupt(int "Interrupt for port %d, but apparently not enabled; per-user %p\n", evtchn->port, u);

- disable_irq_nosync(irq); evtchn->enabled = false;

spin_lock(&u->ring_prod_lock); @@ -292,7 +291,7 @@ static ssize_t evtchn_write(struct file evtchn = find_evtchn(u, port); if (evtchn && !evtchn->enabled) { evtchn->enabled = true; - enable_irq(irq_from_evtchn(port)); + xen_irq_lateeoi(irq_from_evtchn(port), 0); } }

@@ -392,8 +391,8 @@ static int evtchn_bind_to_user(struct pe if (rc < 0) goto err;

- rc = bind_evtchn_to_irqhandler(port, evtchn_interrupt, 0, - u->name, evtchn); + rc = bind_evtchn_to_irqhandler_lateeoi(port, evtchn_interrupt, 0, + u->name, evtchn); if (rc < 0) goto err;

From: Juergen Gross jgross@suse.com

commit e99502f76271d6bc4e374fe368c50c67a1fd3070 upstream.

In case rogue guests are sending events at high frequency it might happen that xen_evtchn_do_upcall() won't stop processing events in dom0. As this is done in irq handling a crash might be the result.

In order to avoid that, delay further inter-domain events after some time in xen_evtchn_do_upcall() by forcing eoi processing into a worker on the same cpu, thus inhibiting new events coming in.

The time after which eoi processing is to be delayed is configurable via a new module parameter "event_loop_timeout" which specifies the maximum event loop time in jiffies (default: 2, the value was chosen after some tests showing that a value of 2 was the lowest with an only slight drop of dom0 network throughput while multiple guests performed an event storm).

How long eoi processing will be delayed can be specified via another parameter "event_eoi_delay" (again in jiffies, default 10, again the value was chosen after testing with different delay values).

This is part of XSA-332.

Cc: stable@vger.kernel.org Reported-by: Julien Grall julien@xen.org Signed-off-by: Juergen Gross jgross@suse.com Reviewed-by: Stefano Stabellini sstabellini@kernel.org Reviewed-by: Wei Liu wl@xen.org Signed-off-by: Greg Kroah-Hartman gregkh@linuxfoundation.org --- Documentation/admin-guide/kernel-parameters.txt | 8 + drivers/xen/events/events_2l.c | 7 drivers/xen/events/events_base.c | 189 +++++++++++++++++++++++- drivers/xen/events/events_fifo.c | 30 +-- drivers/xen/events/events_internal.h | 14 + 5 files changed, 216 insertions(+), 32 deletions(-)

--- a/Documentation/admin-guide/kernel-parameters.txt +++ b/Documentation/admin-guide/kernel-parameters.txt @@ -5022,6 +5022,14 @@ Disables the PV optimizations forcing the HVM guest to run as generic HVM guest with no PV drivers.

+ xen.event_eoi_delay= [XEN] + How long to delay EOI handling in case of event + storms (jiffies). Default is 10. + + xen.event_loop_timeout= [XEN] + After which time (jiffies) the event handling loop + should start to delay EOI handling. Default is 2. + xirc2ps_cs= [NET,PCMCIA] Format: <irq>,<irq_mask>,<io>,<full_duplex>,<do_sound>,<lockup_hack>[,<irq2>[,<irq3>[,<irq4>]]] --- a/drivers/xen/events/events_2l.c +++ b/drivers/xen/events/events_2l.c @@ -161,7 +161,7 @@ static inline xen_ulong_t active_evtchns * a bitset of words which contain pending event bits. The second * level is a bitset of pending events themselves. */ -static void evtchn_2l_handle_events(unsigned cpu) +static void evtchn_2l_handle_events(unsigned cpu, struct evtchn_loop_ctrl *ctrl) { int irq; xen_ulong_t pending_words; @@ -242,10 +242,7 @@ static void evtchn_2l_handle_events(unsi

/* Process port. */ port = (word_idx * BITS_PER_EVTCHN_WORD) + bit_idx; - irq = get_evtchn_to_irq(port); - - if (irq != -1) - generic_handle_irq(irq); + handle_irq_for_port(port, ctrl);

bit_idx = (bit_idx + 1) % BITS_PER_EVTCHN_WORD;

--- a/drivers/xen/events/events_base.c +++ b/drivers/xen/events/events_base.c @@ -34,6 +34,8 @@ #include <linux/pci.h> #include <linux/spinlock.h> #include <linux/cpuhotplug.h> +#include <linux/atomic.h> +#include <linux/ktime.h>

#ifdef CONFIG_X86 #include <asm/desc.h> @@ -63,6 +65,15 @@

#include "events_internal.h"

+#undef MODULE_PARAM_PREFIX +#define MODULE_PARAM_PREFIX "xen." + +static uint __read_mostly event_loop_timeout = 2; +module_param(event_loop_timeout, uint, 0644); + +static uint __read_mostly event_eoi_delay = 10; +module_param(event_eoi_delay, uint, 0644); + const struct evtchn_ops *evtchn_ops;

/* @@ -86,6 +97,7 @@ static DEFINE_RWLOCK(evtchn_rwlock); * irq_mapping_update_lock * evtchn_rwlock * IRQ-desc lock + * percpu eoi_list_lock */

static LIST_HEAD(xen_irq_list_head); @@ -118,6 +130,8 @@ static struct irq_chip xen_pirq_chip; static void enable_dynirq(struct irq_data *data); static void disable_dynirq(struct irq_data *data);

+static DEFINE_PER_CPU(unsigned int, irq_epoch); + static void clear_evtchn_to_irq_row(unsigned row) { unsigned col; @@ -397,17 +411,120 @@ void notify_remote_via_irq(int irq) } EXPORT_SYMBOL_GPL(notify_remote_via_irq);

+struct lateeoi_work { + struct delayed_work delayed; + spinlock_t eoi_list_lock; + struct list_head eoi_list; +}; + +static DEFINE_PER_CPU(struct lateeoi_work, lateeoi); + +static void lateeoi_list_del(struct irq_info *info) +{ + struct lateeoi_work *eoi = &per_cpu(lateeoi, info->eoi_cpu); + unsigned long flags; + + spin_lock_irqsave(&eoi->eoi_list_lock, flags); + list_del_init(&info->eoi_list); + spin_unlock_irqrestore(&eoi->eoi_list_lock, flags); +} + +static void lateeoi_list_add(struct irq_info *info) +{ + struct lateeoi_work *eoi = &per_cpu(lateeoi, info->eoi_cpu); + struct irq_info *elem; + u64 now = get_jiffies_64(); + unsigned long delay; + unsigned long flags; + + if (now < info->eoi_time) + delay = info->eoi_time - now; + else + delay = 1; + + spin_lock_irqsave(&eoi->eoi_list_lock, flags); + + if (list_empty(&eoi->eoi_list)) { + list_add(&info->eoi_list, &eoi->eoi_list); + mod_delayed_work_on(info->eoi_cpu, system_wq, + &eoi->delayed, delay); + } else { + list_for_each_entry_reverse(elem, &eoi->eoi_list, eoi_list) { + if (elem->eoi_time <= info->eoi_time) + break; + } + list_add(&info->eoi_list, &elem->eoi_list); + } + + spin_unlock_irqrestore(&eoi->eoi_list_lock, flags); +} + static void xen_irq_lateeoi_locked(struct irq_info *info) { evtchn_port_t evtchn; + unsigned int cpu;

evtchn = info->evtchn; - if (!VALID_EVTCHN(evtchn)) + if (!VALID_EVTCHN(evtchn) || !list_empty(&info->eoi_list)) + return; + + cpu = info->eoi_cpu; + if (info->eoi_time && info->irq_epoch == per_cpu(irq_epoch, cpu)) { + lateeoi_list_add(info); return; + }

+ info->eoi_time = 0; unmask_evtchn(evtchn); }

+static void xen_irq_lateeoi_worker(struct work_struct *work) +{ + struct lateeoi_work *eoi; + struct irq_info *info; + u64 now = get_jiffies_64(); + unsigned long flags; + + eoi = container_of(to_delayed_work(work), struct lateeoi_work, delayed); + + read_lock_irqsave(&evtchn_rwlock, flags); + + while (true) { + spin_lock(&eoi->eoi_list_lock); + + info = list_first_entry_or_null(&eoi->eoi_list, struct irq_info, + eoi_list); + + if (info == NULL || now < info->eoi_time) { + spin_unlock(&eoi->eoi_list_lock); + break; + } + + list_del_init(&info->eoi_list); + + spin_unlock(&eoi->eoi_list_lock); + + info->eoi_time = 0; + + xen_irq_lateeoi_locked(info); + } + + if (info) + mod_delayed_work_on(info->eoi_cpu, system_wq, + &eoi->delayed, info->eoi_time - now); + + read_unlock_irqrestore(&evtchn_rwlock, flags); +} + +static void xen_cpu_init_eoi(unsigned int cpu) +{ + struct lateeoi_work *eoi = &per_cpu(lateeoi, cpu); + + INIT_DELAYED_WORK(&eoi->delayed, xen_irq_lateeoi_worker); + spin_lock_init(&eoi->eoi_list_lock); + INIT_LIST_HEAD(&eoi->eoi_list); +} + void xen_irq_lateeoi(unsigned int irq, unsigned int eoi_flags) { struct irq_info *info; @@ -427,6 +544,7 @@ EXPORT_SYMBOL_GPL(xen_irq_lateeoi); static void xen_irq_init(unsigned irq) { struct irq_info *info; + #ifdef CONFIG_SMP /* By default all event channels notify CPU#0. */ cpumask_copy(irq_get_affinity_mask(irq), cpumask_of(0)); @@ -441,6 +559,7 @@ static void xen_irq_init(unsigned irq)

set_info_for_irq(irq, info);

+ INIT_LIST_HEAD(&info->eoi_list); list_add_tail(&info->list, &xen_irq_list_head); }

@@ -496,6 +615,9 @@ static void xen_free_irq(unsigned irq)

write_lock_irqsave(&evtchn_rwlock, flags);

+ if (!list_empty(&info->eoi_list)) + lateeoi_list_del(info); + list_del(&info->list);

set_info_for_irq(irq, NULL); @@ -1355,6 +1477,54 @@ void xen_send_IPI_one(unsigned int cpu, notify_remote_via_irq(irq); }

+struct evtchn_loop_ctrl { + ktime_t timeout; + unsigned count; + bool defer_eoi; +}; + +void handle_irq_for_port(evtchn_port_t port, struct evtchn_loop_ctrl *ctrl) +{ + int irq; + struct irq_info *info; + + irq = get_evtchn_to_irq(port); + if (irq == -1) + return; + + /* + * Check for timeout every 256 events. + * We are setting the timeout value only after the first 256 + * events in order to not hurt the common case of few loop + * iterations. The 256 is basically an arbitrary value. + * + * In case we are hitting the timeout we need to defer all further + * EOIs in order to ensure to leave the event handling loop rather + * sooner than later. + */ + if (!ctrl->defer_eoi && !(++ctrl->count & 0xff)) { + ktime_t kt = ktime_get(); + + if (!ctrl->timeout) { + kt = ktime_add_ms(kt, + jiffies_to_msecs(event_loop_timeout)); + ctrl->timeout = kt; + } else if (kt > ctrl->timeout) { + ctrl->defer_eoi = true; + } + } + + info = info_for_irq(irq); + + if (ctrl->defer_eoi) { + info->eoi_cpu = smp_processor_id(); + info->irq_epoch = __this_cpu_read(irq_epoch); + info->eoi_time = get_jiffies_64() + event_eoi_delay; + } + + generic_handle_irq(irq); +} + static DEFINE_PER_CPU(unsigned, xed_nesting_count);

static void __xen_evtchn_do_upcall(void) @@ -1362,6 +1532,7 @@ static void __xen_evtchn_do_upcall(void) struct vcpu_info *vcpu_info = __this_cpu_read(xen_vcpu); int cpu = get_cpu(); unsigned count; + struct evtchn_loop_ctrl ctrl = { 0 };

read_lock(&evtchn_rwlock);

@@ -1371,7 +1542,7 @@ static void __xen_evtchn_do_upcall(void) if (__this_cpu_inc_return(xed_nesting_count) - 1) goto out;

- xen_evtchn_handle_events(cpu); + xen_evtchn_handle_events(cpu, &ctrl);

BUG_ON(!irqs_disabled());

@@ -1382,6 +1553,13 @@ static void __xen_evtchn_do_upcall(void) out: read_unlock(&evtchn_rwlock);

+ /* + * Increment irq_epoch only now to defer EOIs only for + * xen_irq_lateeoi() invocations occurring from inside the loop + * above. + */ + __this_cpu_inc(irq_epoch); + put_cpu(); }

@@ -1828,9 +2006,6 @@ void xen_callback_vector(void) void xen_callback_vector(void) {} #endif

-#undef MODULE_PARAM_PREFIX -#define MODULE_PARAM_PREFIX "xen." - static bool fifo_events = true; module_param(fifo_events, bool, 0);

@@ -1838,6 +2013,8 @@ static int xen_evtchn_cpu_prepare(unsign { int ret = 0;

+ xen_cpu_init_eoi(cpu); + if (evtchn_ops->percpu_init) ret = evtchn_ops->percpu_init(cpu);

@@ -1864,6 +2041,8 @@ void __init xen_init_IRQ(void) if (ret < 0) xen_evtchn_2l_init();

+ xen_cpu_init_eoi(smp_processor_id()); + cpuhp_setup_state_nocalls(CPUHP_XEN_EVTCHN_PREPARE, "xen/evtchn:prepare", xen_evtchn_cpu_prepare, xen_evtchn_cpu_dead); --- a/drivers/xen/events/events_fifo.c +++ b/drivers/xen/events/events_fifo.c @@ -275,19 +275,9 @@ static uint32_t clear_linked(volatile ev return w & EVTCHN_FIFO_LINK_MASK; }

-static void handle_irq_for_port(unsigned port) -{ - int irq; - - irq = get_evtchn_to_irq(port); - if (irq != -1) - generic_handle_irq(irq); -} - -static void consume_one_event(unsigned cpu, +static void consume_one_event(unsigned cpu, struct evtchn_loop_ctrl *ctrl, struct evtchn_fifo_control_block *control_block, - unsigned priority, unsigned long *ready, - bool drop) + unsigned priority, unsigned long *ready) { struct evtchn_fifo_queue *q = &per_cpu(cpu_queue, cpu); uint32_t head; @@ -320,16 +310,17 @@ static void consume_one_event(unsigned c clear_bit(priority, ready);

if (evtchn_fifo_is_pending(port) && !evtchn_fifo_is_masked(port)) { - if (unlikely(drop)) + if (unlikely(!ctrl)) pr_warn("Dropping pending event for port %u\n", port); else - handle_irq_for_port(port); + handle_irq_for_port(port, ctrl); }

q->head[priority] = head; }

-static void __evtchn_fifo_handle_events(unsigned cpu, bool drop) +static void __evtchn_fifo_handle_events(unsigned cpu, + struct evtchn_loop_ctrl *ctrl) { struct evtchn_fifo_control_block *control_block; unsigned long ready; @@ -341,14 +332,15 @@ static void __evtchn_fifo_handle_events(

while (ready) { q = find_first_bit(&ready, EVTCHN_FIFO_MAX_QUEUES); - consume_one_event(cpu, control_block, q, &ready, drop); + consume_one_event(cpu, ctrl, control_block, q, &ready); ready |= xchg(&control_block->ready, 0); } }

-static void evtchn_fifo_handle_events(unsigned cpu) +static void evtchn_fifo_handle_events(unsigned cpu, + struct evtchn_loop_ctrl *ctrl) { - __evtchn_fifo_handle_events(cpu, false); + __evtchn_fifo_handle_events(cpu, ctrl); }

static void evtchn_fifo_resume(void) @@ -416,7 +408,7 @@ static int evtchn_fifo_percpu_init(unsig

static int evtchn_fifo_percpu_deinit(unsigned int cpu) { - __evtchn_fifo_handle_events(cpu, true); + __evtchn_fifo_handle_events(cpu, NULL); return 0; }

--- a/drivers/xen/events/events_internal.h +++ b/drivers/xen/events/events_internal.h @@ -32,11 +32,15 @@ enum xen_irq_type { */ struct irq_info { struct list_head list; + struct list_head eoi_list; int refcnt; enum xen_irq_type type; /* type */ unsigned irq; unsigned int evtchn; /* event channel */ unsigned short cpu; /* cpu bound */ + unsigned short eoi_cpu; /* EOI must happen on this cpu */ + unsigned int irq_epoch; /* If eoi_cpu valid: irq_epoch of event */ + u64 eoi_time; /* Time in jiffies when to EOI. */

union { unsigned short virq; @@ -55,6 +59,8 @@ struct irq_info { #define PIRQ_SHAREABLE (1 << 1) #define PIRQ_MSI_GROUP (1 << 2)

+struct evtchn_loop_ctrl; + struct evtchn_ops { unsigned (*max_channels)(void); unsigned (*nr_channels)(void); @@ -69,7 +75,7 @@ struct evtchn_ops { void (*mask)(unsigned port); void (*unmask)(unsigned port);

- void (*handle_events)(unsigned cpu); + void (*handle_events)(unsigned cpu, struct evtchn_loop_ctrl *ctrl); void (*resume)(void);

int (*percpu_init)(unsigned int cpu); @@ -80,6 +86,7 @@ extern const struct evtchn_ops *evtchn_o

extern int **evtchn_to_irq; int get_evtchn_to_irq(unsigned int evtchn); +void handle_irq_for_port(evtchn_port_t port, struct evtchn_loop_ctrl *ctrl);

struct irq_info *info_for_irq(unsigned irq); unsigned cpu_from_irq(unsigned irq); @@ -137,9 +144,10 @@ static inline void unmask_evtchn(unsigne return evtchn_ops->unmask(port); }

-static inline void xen_evtchn_handle_events(unsigned cpu) +static inline void xen_evtchn_handle_events(unsigned cpu, + struct evtchn_loop_ctrl *ctrl) { - return evtchn_ops->handle_events(cpu); + return evtchn_ops->handle_events(cpu, ctrl); }

static inline void xen_evtchn_resume(void)

From: Jiri Olsa jolsa@redhat.com

commit f91072ed1b7283b13ca57fcfbece5a3b92726143 upstream.

There's a possible race in perf_mmap_close() when checking ring buffer's mmap_count refcount value. The problem is that the mmap_count check is not atomic because we call atomic_dec() and atomic_read() separately.

perf_mmap_close: ... atomic_dec(&rb->mmap_count); ... if (atomic_read(&rb->mmap_count)) goto out_put;

<ring buffer detach> free_uid

out_put: ring_buffer_put(rb); /* could be last */

The race can happen when we have two (or more) events sharing same ring buffer and they go through atomic_dec() and then they both see 0 as refcount value later in atomic_read(). Then both will go on and execute code which is meant to be run just once.

The code that detaches ring buffer is probably fine to be executed more than once, but the problem is in calling free_uid(), which will later on demonstrate in related crashes and refcount warnings, like:

refcount_t: addition on 0; use-after-free. ... RIP: 0010:refcount_warn_saturate+0x6d/0xf ... Call Trace: prepare_creds+0x190/0x1e0 copy_creds+0x35/0x172 copy_process+0x471/0x1a80 _do_fork+0x83/0x3a0 __do_sys_wait4+0x83/0x90 __do_sys_clone+0x85/0xa0 do_syscall_64+0x5b/0x1e0 entry_SYSCALL_64_after_hwframe+0x44/0xa9

Using atomic decrease and check instead of separated calls.

Tested-by: Michael Petlan mpetlan@redhat.com Signed-off-by: Jiri Olsa jolsa@kernel.org Signed-off-by: Ingo Molnar mingo@kernel.org Acked-by: Peter Zijlstra a.p.zijlstra@chello.nl Acked-by: Namhyung Kim namhyung@kernel.org Acked-by: Wade Mealing wmealing@redhat.com Fixes: 9bb5d40cd93c ("perf: Fix mmap() accounting hole"); Link: https://lore.kernel.org/r/20200916115311.GE2301783@krava [sudip: used ring_buffer] Signed-off-by: Sudip Mukherjee sudipm.mukherjee@gmail.com Signed-off-by: Greg Kroah-Hartman gregkh@linuxfoundation.org --- kernel/events/core.c | 7 ++++--- 1 file changed, 4 insertions(+), 3 deletions(-)

--- a/kernel/events/core.c +++ b/kernel/events/core.c @@ -5207,11 +5207,11 @@ static void perf_pmu_output_stop(struct static void perf_mmap_close(struct vm_area_struct *vma) { struct perf_event *event = vma->vm_file->private_data; - struct ring_buffer *rb = ring_buffer_get(event); struct user_struct *mmap_user = rb->mmap_user; int mmap_locked = rb->mmap_locked; unsigned long size = perf_data_size(rb); + bool detach_rest = false;

if (event->pmu->event_unmapped) event->pmu->event_unmapped(event, vma->vm_mm); @@ -5242,7 +5242,8 @@ static void perf_mmap_close(struct vm_ar mutex_unlock(&event->mmap_mutex); }

- atomic_dec(&rb->mmap_count); + if (atomic_dec_and_test(&rb->mmap_count)) + detach_rest = true;

if (!atomic_dec_and_mutex_lock(&event->mmap_count, &event->mmap_mutex)) goto out_put; @@ -5251,7 +5252,7 @@ static void perf_mmap_close(struct vm_ar mutex_unlock(&event->mmap_mutex);

/* If there's still other mmap()s of this buffer, we're done. */ - if (atomic_read(&rb->mmap_count)) + if (!detach_rest) goto out_put;

/*

From: Matteo Croce mcroce@microsoft.com

commit df5b0ab3e08a156701b537809914b339b0daa526 upstream.

Limit the CPU number to num_possible_cpus(), because setting it to a value lower than INT_MAX but higher than NR_CPUS produces the following error on reboot and shutdown:

BUG: unable to handle page fault for address: ffffffff90ab1bb0 #PF: supervisor read access in kernel mode #PF: error_code(0x0000) - not-present page PGD 1c09067 P4D 1c09067 PUD 1c0a063 PMD 0 Oops: 0000 [#1] SMP CPU: 1 PID: 1 Comm: systemd-shutdow Not tainted 5.9.0-rc8-kvm #110 Hardware name: QEMU Standard PC (Q35 + ICH9, 2009), BIOS 1.13.0-2.fc32 04/01/2014 RIP: 0010:migrate_to_reboot_cpu+0xe/0x60 Code: ea ea 00 48 89 fa 48 c7 c7 30 57 f1 81 e9 fa ef ff ff 66 2e 0f 1f 84 00 00 00 00 00 53 8b 1d d5 ea ea 00 e8 14 33 fe ff 89 da <48> 0f a3 15 ea fc bd 00 48 89 d0 73 29 89 c2 c1 e8 06 65 48 8b 3c RSP: 0018:ffffc90000013e08 EFLAGS: 00010246 RAX: ffff88801f0a0000 RBX: 0000000077359400 RCX: 0000000000000000 RDX: 0000000077359400 RSI: 0000000000000002 RDI: ffffffff81c199e0 RBP: ffffffff81c1e3c0 R08: ffff88801f41f000 R09: ffffffff81c1e348 R10: 0000000000000000 R11: 0000000000000000 R12: 0000000000000000 R13: 00007f32bedf8830 R14: 00000000fee1dead R15: 0000000000000000 FS: 00007f32bedf8980(0000) GS:ffff88801f480000(0000) knlGS:0000000000000000 CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 CR2: ffffffff90ab1bb0 CR3: 000000001d057000 CR4: 00000000000006a0 DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000 DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400 Call Trace: __do_sys_reboot.cold+0x34/0x5b do_syscall_64+0x2d/0x40

Fixes: 1b3a5d02ee07 ("reboot: move arch/x86 reboot= handling to generic kernel") Signed-off-by: Matteo Croce mcroce@microsoft.com Signed-off-by: Andrew Morton akpm@linux-foundation.org Cc: Arnd Bergmann arnd@arndb.de Cc: Fabian Frederick fabf@skynet.be Cc: Greg Kroah-Hartman gregkh@linuxfoundation.org Cc: Guenter Roeck linux@roeck-us.net Cc: Kees Cook keescook@chromium.org Cc: Mike Rapoport rppt@kernel.org Cc: Pavel Tatashin pasha.tatashin@soleen.com Cc: Petr Mladek pmladek@suse.com Cc: Robin Holt robinmholt@gmail.com Cc: stable@vger.kernel.org Link: https://lkml.kernel.org/r/20201103214025.116799-3-mcroce@linux.microsoft.com Signed-off-by: Linus Torvalds torvalds@linux-foundation.org [sudip: use reboot_mode instead of mode] Signed-off-by: Sudip Mukherjee sudipm.mukherjee@gmail.com Signed-off-by: Greg Kroah-Hartman gregkh@linuxfoundation.org --- kernel/reboot.c | 7 +++++++ 1 file changed, 7 insertions(+)

--- a/kernel/reboot.c +++ b/kernel/reboot.c @@ -519,6 +519,13 @@ static int __init reboot_setup(char *str reboot_cpu = simple_strtoul(str+3, NULL, 0); else reboot_mode = REBOOT_SOFT; + if (reboot_cpu >= num_possible_cpus()) { + pr_err("Ignoring the CPU number in reboot= option. " + "CPU %d exceeds possible cpu number %d\n", + reboot_cpu, num_possible_cpus()); + reboot_cpu = 0; + break; + } break;

case 'g':