August 2020 - Linux-stable-mirror

[PATCH 1/4] lpfc: Fix setting irq affinity with an empty cpu mask.

by James Smart

Some systems are reporting the following log message during driver unload or system shutdown: ics_rtas_set_affinity: No online cpus in the mask A prior commit introduced the writing of an empty affinity mask in calls to irq_set_affinity_hint() when disabling interrupts or when there are no remaining online cpus to service an eq interrupt. At least some ppc64 systems are checking whether affinity masks are empty or not. Fix: Do not call irq_set_affinity_hint() with an empty cpu mask. Fixes: dcaa21367938 ("scsi: lpfc: Change default IRQ model on AMD architectures") Cc: <stable(a)vger.kernel.org> # v5.5+ Signed-off-by: Dick Kennedy <dick.kennedy(a)broadcom.com> Signed-off-by: James Smart <james.smart(a)broadcom.com> --- drivers/scsi/lpfc/lpfc_init.c | 1 - 1 file changed, 1 deletion(-) diff --git a/drivers/scsi/lpfc/lpfc_init.c b/drivers/scsi/lpfc/lpfc_init.c index 05ace6916b66..89c3ba0a0df9 100644 --- a/drivers/scsi/lpfc/lpfc_init.c +++ b/drivers/scsi/lpfc/lpfc_init.c @@ -11376,7 +11376,6 @@ lpfc_irq_clear_aff(struct lpfc_hba_eq_hdl *eqhdl) { cpumask_clear(&eqhdl->aff_mask); irq_clear_status_flags(eqhdl->irq, IRQ_NO_BALANCING); - irq_set_affinity_hint(eqhdl->irq, &eqhdl->aff_mask); } /** -- 2.26.2

4 years, 4 months

1
0
0 0

[PATCH 0/2] Backport uclamp static key to 5.4.y stable

by Qais Yousef

Hi Greg/Sasha The following 2 patches are backports of 46609ce22703: ("sched/uclamp: Protect uclamp fast path code with static key") e65855a52b47: ("sched/uclamp: Fix a deadlock when enabling uclamp static key") into 5.4.y stable tree. The conflict was trivial and due to: 1. uclamp_rq_util_with() was renamed from util_util_with() 2. 2 local variables were converted to unsigned long from unsigned int. I did do compile test on aarch64 and x86_64 and both looked fine. Beside I ran a quick and short mmtest to verify the functionality and got the following results which is inline with what's expected. 5.4.y 5.4.y-static-keys Hmean send-64 188.46 ( 0.00%) 189.95 * 0.79%* Hmean send-128 375.65 ( 0.00%) 379.75 * 1.09%* 7.32% -0.33% [kernel.kallsyms] [k] try_to_wake_up 0.58% -0.55% [kernel.kallsyms] [k] deactivate_task 0.50% -0.45% [kernel.kallsyms] [k] activate_task That said, it's Friday afternoon and I am off next week. If I did something stupid and didn't find me, please accept my apologies in advance and will fix it as soon as I am back online. Thanks -- Qais Yousef Qais Yousef (2): sched/uclamp: Protect uclamp fast path code with static key sched/uclamp: Fix a deadlock when enabling uclamp static key kernel/sched/core.c | 81 +++++++++++++++++++++++++++++++- kernel/sched/cpufreq_schedutil.c | 2 +- kernel/sched/sched.h | 47 +++++++++++++++++- 3 files changed, 126 insertions(+), 4 deletions(-) -- 2.17.1

4 years, 4 months

2
3
0 0

Stable inclusion request: Revert "ath10k: fix DMA related firmware crashes on multiple devices" for 4.4, 4.9, 4.14, 4.19 and 5.4

by Nick Lowe

Hi, Please can commit a1769bb68a850508a492e3674ab1e5e479b11254 be back merged to the 4.4, 4.9, 4.14, 4.19 and 5.4 LTS kernels to resolve a PCIe hang with Wave 2-generation 802.11ac QCA chips? commit a1769bb68a850508a492e3674ab1e5e479b11254 author Zhi Chen <zhichen(a)codeaurora.org> date 2020-01-14 12:35:21 +0800 https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?… The issue resolved is that a value of 0 has a different meaning and instruction depending on the specific chip. Wave 1 and Wave 2 QCA 802.11ac chips use this differently. This meant that the original commit introduced a regression for the Wave 2-generation 802.11ac QCA chips when seeking to resolve a firmware crash issue affecting some Wave 1-generation 802.11ac QCA chips. This subsequently has only been partially reverted via this commit, specifically for the 10_4 target only via TARGET_10_4_DMA_BURST_SIZE 1 The original commit that caused this regression from back in 2015 is 76d164f582150fd0259ec0fcbc485470bcd8033e https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?… With thanks, Nick

4 years, 4 months

2
1
0 0

[PATCH] s390/numa: set node distance to LOCAL_DISTANCE

by Alexander Gordeev

commit 535e4fc623fab2e09a0653fc3a3e17f382ad0251 upstream. The original upstream commit applies only to 5.7 and 5.8 stable trees. This is backport to 4.4, 4.9, 4.14, 4.19 and 5.4 trees. The node distance is hardcoded to 0, which causes a trouble for some user-level applications. In particular, "libnuma" expects the distance of a node to itself as LOCAL_DISTANCE. This update removes the offending node distance override. Cc: Heiko Carstens <hca(a)linux.ibm.com> Fixes: 3a368f742da1 ("s390/numa: add core infrastructure") Signed-off-by: Alexander Gordeev <agordeev(a)linux.ibm.com> --- arch/s390/numa/numa.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/arch/s390/numa/numa.c b/arch/s390/numa/numa.c index d2910fa..8f95db6 100644 --- a/arch/s390/numa/numa.c +++ b/arch/s390/numa/numa.c @@ -51,7 +51,7 @@ void numa_update_cpu_topology(void) int __node_distance(int a, int b) { - return mode->distance ? mode->distance(a, b) : 0; + return mode->distance ? mode->distance(a, b) : LOCAL_DISTANCE; } EXPORT_SYMBOL(__node_distance); -- 1.8.3.1

4 years, 4 months

1
0
0 0

Stable inclusion request, 5.7

by Jens Axboe

Hi, Can you cherry-pick this one: commit e697deed834de15d2322d0619d51893022c90ea2 Author: Jiufei Xue <jiufei.xue(a)linux.alibaba.com> Date: Wed Jun 10 13:41:59 2020 +0800 io_uring: check file O_NONBLOCK state for accept into 5.7-stable? Thanks! -- Jens Axboe

4 years, 4 months

2
2
0 0

[git:media_tree/master] media: rc: do not access device via sysfs after rc_unregister_device()

by Mauro Carvalho Chehab

This is an automatic generated email to let you know that the following patch were queued: Subject: media: rc: do not access device via sysfs after rc_unregister_device() Author: Sean Young <sean(a)mess.org> Date: Sat Aug 8 13:38:02 2020 +0200 Device drivers do not expect to have change_protocol or wakeup re-programming to be accesed after rc_unregister_device(). This can cause the device driver to access deallocated resources. Cc: <stable(a)vger.kernel.org> # 4.16+ Signed-off-by: Sean Young <sean(a)mess.org> Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei(a)kernel.org> drivers/media/rc/rc-main.c | 12 ++++++++++++ 1 file changed, 12 insertions(+) --- diff --git a/drivers/media/rc/rc-main.c b/drivers/media/rc/rc-main.c index e1cda80a4b25..dee8a9f3d80a 100644 --- a/drivers/media/rc/rc-main.c +++ b/drivers/media/rc/rc-main.c @@ -1292,6 +1292,10 @@ static ssize_t store_protocols(struct device *device, } mutex_lock(&dev->lock); + if (!dev->registered) { + mutex_unlock(&dev->lock); + return -ENODEV; + } old_protocols = *current_protocols; new_protocols = old_protocols; @@ -1430,6 +1434,10 @@ static ssize_t store_filter(struct device *device, return -EINVAL; mutex_lock(&dev->lock); + if (!dev->registered) { + mutex_unlock(&dev->lock); + return -ENODEV; + } new_filter = *filter; if (fattr->mask) @@ -1544,6 +1552,10 @@ static ssize_t store_wakeup_protocols(struct device *device, int i; mutex_lock(&dev->lock); + if (!dev->registered) { + mutex_unlock(&dev->lock); + return -ENODEV; + } allowed = dev->allowed_wakeup_protocols;

4 years, 4 months

1
0
0 0

[git:media_tree/fixes] media: rc: do not access device via sysfs after rc_unregister_device()

by Mauro Carvalho Chehab

This is an automatic generated email to let you know that the following patch were queued: Subject: media: rc: do not access device via sysfs after rc_unregister_device() Author: Sean Young <sean(a)mess.org> Date: Sat Aug 8 13:38:02 2020 +0200 Device drivers do not expect to have change_protocol or wakeup re-programming to be accesed after rc_unregister_device(). This can cause the device driver to access deallocated resources. Cc: <stable(a)vger.kernel.org> # 4.16+ Signed-off-by: Sean Young <sean(a)mess.org> Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei(a)kernel.org> drivers/media/rc/rc-main.c | 12 ++++++++++++ 1 file changed, 12 insertions(+) --- diff --git a/drivers/media/rc/rc-main.c b/drivers/media/rc/rc-main.c index e1cda80a4b25..dee8a9f3d80a 100644 --- a/drivers/media/rc/rc-main.c +++ b/drivers/media/rc/rc-main.c @@ -1292,6 +1292,10 @@ static ssize_t store_protocols(struct device *device, } mutex_lock(&dev->lock); + if (!dev->registered) { + mutex_unlock(&dev->lock); + return -ENODEV; + } old_protocols = *current_protocols; new_protocols = old_protocols; @@ -1430,6 +1434,10 @@ static ssize_t store_filter(struct device *device, return -EINVAL; mutex_lock(&dev->lock); + if (!dev->registered) { + mutex_unlock(&dev->lock); + return -ENODEV; + } new_filter = *filter; if (fattr->mask) @@ -1544,6 +1552,10 @@ static ssize_t store_wakeup_protocols(struct device *device, int i; mutex_lock(&dev->lock); + if (!dev->registered) { + mutex_unlock(&dev->lock); + return -ENODEV; + } allowed = dev->allowed_wakeup_protocols;

4 years, 4 months

1
0
0 0

[git:media_tree/master] media: rc: uevent sysfs file races with rc_unregister_device()

by Mauro Carvalho Chehab

This is an automatic generated email to let you know that the following patch were queued: Subject: media: rc: uevent sysfs file races with rc_unregister_device() Author: Sean Young <sean(a)mess.org> Date: Sat Aug 8 13:19:12 2020 +0200 Only report uevent file contents if device still registered, else we might read freed memory. Reported-by: syzbot+ceef16277388d6f24898(a)syzkaller.appspotmail.com Cc: Hillf Danton <hdanton(a)sina.com> Cc: <stable(a)vger.kernel.org> # 4.16+ Signed-off-by: Sean Young <sean(a)mess.org> Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei(a)kernel.org> drivers/media/rc/rc-main.c | 32 ++++++++++++++++---------------- 1 file changed, 16 insertions(+), 16 deletions(-) --- diff --git a/drivers/media/rc/rc-main.c b/drivers/media/rc/rc-main.c index 7b53066d9d07..e1cda80a4b25 100644 --- a/drivers/media/rc/rc-main.c +++ b/drivers/media/rc/rc-main.c @@ -1601,25 +1601,25 @@ static void rc_dev_release(struct device *device) kfree(dev); } -#define ADD_HOTPLUG_VAR(fmt, val...) \ - do { \ - int err = add_uevent_var(env, fmt, val); \ - if (err) \ - return err; \ - } while (0) - static int rc_dev_uevent(struct device *device, struct kobj_uevent_env *env) { struct rc_dev *dev = to_rc_dev(device); + int ret = 0; - if (dev->rc_map.name) - ADD_HOTPLUG_VAR("NAME=%s", dev->rc_map.name); - if (dev->driver_name) - ADD_HOTPLUG_VAR("DRV_NAME=%s", dev->driver_name); - if (dev->device_name) - ADD_HOTPLUG_VAR("DEV_NAME=%s", dev->device_name); + mutex_lock(&dev->lock); - return 0; + if (!dev->registered) + ret = -ENODEV; + if (ret == 0 && dev->rc_map.name) + ret = add_uevent_var(env, "NAME=%s", dev->rc_map.name); + if (ret == 0 && dev->driver_name) + ret = add_uevent_var(env, "DRV_NAME=%s", dev->driver_name); + if (ret == 0 && dev->device_name) + ret = add_uevent_var(env, "DEV_NAME=%s", dev->device_name); + + mutex_unlock(&dev->lock); + + return ret; } /* @@ -2011,14 +2011,14 @@ void rc_unregister_device(struct rc_dev *dev) del_timer_sync(&dev->timer_keyup); del_timer_sync(&dev->timer_repeat); - rc_free_rx_device(dev); - mutex_lock(&dev->lock); if (dev->users && dev->close) dev->close(dev); dev->registered = false; mutex_unlock(&dev->lock); + rc_free_rx_device(dev); + /* * lirc device should be freed with dev->registered = false, so * that userspace polling will get notified.

4 years, 4 months

1
0
0 0

[git:media_tree/fixes] media: rc: uevent sysfs file races with rc_unregister_device()

by Mauro Carvalho Chehab

This is an automatic generated email to let you know that the following patch were queued: Subject: media: rc: uevent sysfs file races with rc_unregister_device() Author: Sean Young <sean(a)mess.org> Date: Sat Aug 8 13:19:12 2020 +0200 Only report uevent file contents if device still registered, else we might read freed memory. Reported-by: syzbot+ceef16277388d6f24898(a)syzkaller.appspotmail.com Cc: Hillf Danton <hdanton(a)sina.com> Cc: <stable(a)vger.kernel.org> # 4.16+ Signed-off-by: Sean Young <sean(a)mess.org> Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei(a)kernel.org> drivers/media/rc/rc-main.c | 32 ++++++++++++++++---------------- 1 file changed, 16 insertions(+), 16 deletions(-) --- diff --git a/drivers/media/rc/rc-main.c b/drivers/media/rc/rc-main.c index 7b53066d9d07..e1cda80a4b25 100644 --- a/drivers/media/rc/rc-main.c +++ b/drivers/media/rc/rc-main.c @@ -1601,25 +1601,25 @@ static void rc_dev_release(struct device *device) kfree(dev); } -#define ADD_HOTPLUG_VAR(fmt, val...) \ - do { \ - int err = add_uevent_var(env, fmt, val); \ - if (err) \ - return err; \ - } while (0) - static int rc_dev_uevent(struct device *device, struct kobj_uevent_env *env) { struct rc_dev *dev = to_rc_dev(device); + int ret = 0; - if (dev->rc_map.name) - ADD_HOTPLUG_VAR("NAME=%s", dev->rc_map.name); - if (dev->driver_name) - ADD_HOTPLUG_VAR("DRV_NAME=%s", dev->driver_name); - if (dev->device_name) - ADD_HOTPLUG_VAR("DEV_NAME=%s", dev->device_name); + mutex_lock(&dev->lock); - return 0; + if (!dev->registered) + ret = -ENODEV; + if (ret == 0 && dev->rc_map.name) + ret = add_uevent_var(env, "NAME=%s", dev->rc_map.name); + if (ret == 0 && dev->driver_name) + ret = add_uevent_var(env, "DRV_NAME=%s", dev->driver_name); + if (ret == 0 && dev->device_name) + ret = add_uevent_var(env, "DEV_NAME=%s", dev->device_name); + + mutex_unlock(&dev->lock); + + return ret; } /* @@ -2011,14 +2011,14 @@ void rc_unregister_device(struct rc_dev *dev) del_timer_sync(&dev->timer_keyup); del_timer_sync(&dev->timer_repeat); - rc_free_rx_device(dev); - mutex_lock(&dev->lock); if (dev->users && dev->close) dev->close(dev); dev->registered = false; mutex_unlock(&dev->lock); + rc_free_rx_device(dev); + /* * lirc device should be freed with dev->registered = false, so * that userspace polling will get notified.

4 years, 4 months

1
0
0 0

patch "kobject: Restore old behaviour of kobject_del(NULL)" added to driver-core-linus

by gregkh＠linuxfoundation.org

This is a note to let you know that I've just added the patch titled kobject: Restore old behaviour of kobject_del(NULL) to my driver-core git tree which can be found at git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/driver-core.git in the driver-core-linus branch. The patch will show up in the next release of the linux-next tree (usually sometime within the next 24 hours during the week.) The patch will hopefully also be merged in Linus's tree for the next -rc kernel release. If you have any questions about this process, please let me know. >From 40b8b826a6998639dd1c26f0e127f18371e1058d Mon Sep 17 00:00:00 2001 From: Andy Shevchenko <andriy.shevchenko(a)linux.intel.com> Date: Mon, 3 Aug 2020 11:27:06 +0300 Subject: kobject: Restore old behaviour of kobject_del(NULL) The commit 079ad2fb4bf9 ("kobject: Avoid premature parent object freeing in kobject_cleanup()") inadvertently dropped a possibility to call kobject_del() with NULL pointer. Restore the old behaviour. Fixes: 079ad2fb4bf9 ("kobject: Avoid premature parent object freeing in kobject_cleanup()") Cc: stable <stable(a)vger.kernel.org> Reported-by: Qu Wenruo <quwenruo.btrfs(a)gmx.com> Cc: Heikki Krogerus <heikki.krogerus(a)linux.intel.com> Signed-off-by: Andy Shevchenko <andriy.shevchenko(a)linux.intel.com> Reviewed-by: Qu Wenruo <wqu(a)suse.com> Link: https://lore.kernel.org/r/20200803082706.65347-1-andriy.shevchenko@linux.in… Signed-off-by: Greg Kroah-Hartman <gregkh(a)linuxfoundation.org> --- lib/kobject.c | 6 +++++- 1 file changed, 5 insertions(+), 1 deletion(-) diff --git a/lib/kobject.c b/lib/kobject.c index 3afb939f2a1c..9dce68c378e6 100644 --- a/lib/kobject.c +++ b/lib/kobject.c @@ -637,8 +637,12 @@ static void __kobject_del(struct kobject *kobj) */ void kobject_del(struct kobject *kobj) { - struct kobject *parent = kobj->parent; + struct kobject *parent; + + if (!kobj) + return; + parent = kobj->parent; __kobject_del(kobj); kobject_put(parent); } -- 2.28.0

4 years, 4 months

1
0
0 0

2024

2023

2022

2021

2020

2019

2018

2017

Linux-stable-mirror August 2020