linux/drivers
David Collins 1b18af40c1 spmi: spmi-pmic-arb: fix irq_set_type race condition
The qpnpint_irq_set_type() callback function configures the type
(edge vs level) and polarity (high, low, or both) of a particular
PMIC interrupt within a given peripheral.  To do this, it reads
the three consecutive IRQ configuration registers, modifies the
specified IRQ bit within the register values, and finally writes
the three modified register values back to the PMIC.  While a
spinlock is used to provide mutual exclusion on the SPMI bus
during the register read and write calls, there is no locking
around the overall read, modify, write sequence.  This opens up
the possibility of a race condition if two tasks set the type of
a PMIC IRQ within the same peripheral simultaneously.

When the race condition is encountered, both tasks will read the
old value of the registers and IRQ bits set by one of the tasks
will be dropped upon the register write of the other task.  This
then leads to PMIC IRQs being enabled with an incorrect type and
polarity configured.  Such misconfiguration can lead to an IRQ
storm that overwhelms the system and causes it to crash.

This race condition and IRQ storm have been observed when using
a pair of pm8941-pwrkey devices to handle PMK8350 pwrkey and
resin interrupts.  The independent devices probe asynchronously
in parallel and can simultaneously request and configure PMIC
IRQs in the same PMIC peripheral.

For a good case, the IRQ configuration calls end up serialized
due to timing deltas and the register read/write sequence looks
like this:

1. pwrkey probe: SPMI  read(0x1311): 0x00, 0x00, 0x00
2. pwrkey probe: SPMI write(0x1311): 0x80, 0x80, 0x80
3. resin probe:  SPMI  read(0x1311): 0x80, 0x80, 0x80
4. resin probe:  SPMI write(0x1311): 0xC0, 0xC0, 0xC0

The final register states after both devices have requested and
enabled their respective IRQs is thus:

0x1311: 0xC0
0x1312: 0xC0
0x1313: 0xC0
0x1314: 0x00
0x1315: 0xC0

For a bad case, the IRQ configuration calls end up occurring
simultaneously and the race condition is encountered.  The
register read/write sequence then looks like this:

1. pwrkey probe: SPMI  read(0x1311): 0x00, 0x00, 0x00
2. resin probe:  SPMI  read(0x1311): 0x00, 0x00, 0x00
3. pwrkey probe: SPMI write(0x1311): 0x80, 0x80, 0x80
4. resin probe:  SPMI write(0x1311): 0x40, 0x40, 0x40

In this case, the final register states after both devices have
requested and enabled their respective IRQs is thus:

0x1311: 0x40
0x1312: 0x40
0x1313: 0x40
0x1314: 0x00
0x1315: 0xC0

This corresponds to the resin IRQ being configured for both
rising and falling edges, as expected.  However, the pwrkey IRQ
is misconfigured as level type with both polarity high and low
set to disabled.  The PMIC IRQ triggering hardware treats this
particular register configuration as if level low triggering is
enabled.

The raw pwrkey IRQ signal is low when the power key is not being
pressed.  Thus, the pwrkey IRQ begins firing continuously in an
IRQ storm.

Fix the race condition by holding the spmi-pmic-arb spinlock for
the duration of the read, modify, write sequence performed in the
qpnpint_irq_set_type() function.  Split the pmic_arb_read_cmd()
and pmic_arb_write_cmd() functions each into three parts so that
hardware register IO is decoupled from spinlock locking.  This
allows a new function pmic_arb_masked_write() to be added which
locks the spinlock and then calls register IO functions to
perform SPMI read and write commands in a single atomic
operation.

Link: https://lore.kernel.org/r/20211118034719.28971-1-quic_collinsd@quicinc.com
Signed-off-by: David Collins <quic_collinsd@quicinc.com>
Signed-off-by: Stephen Boyd <sboyd@kernel.org>
Link: https://lore.kernel.org/r/20211216190812.1574801-7-sboyd@kernel.org
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2021-12-17 17:18:18 +01:00
..
accessibility speakup: remove redundant assignment of variable i 2021-11-26 17:01:53 +01:00
acpi Merge branch 'acpi-properties' 2021-11-26 19:45:31 +01:00
amba ARM: 9119/1: amba: Properly handle device probe without IRQ domain 2021-10-19 10:30:53 +01:00
android Merge v5.15-rc5 into char-misc-next 2021-12-13 10:17:10 +01:00
ata libata: add horkage for ASMedia 1092 2021-12-09 11:20:47 +09:00
atm
auxdisplay auxdisplay: cfag12864bfb: code indent should use tabs where possible 2021-10-22 00:13:16 +02:00
base arch_topology: Fix missing clear cluster_cpumask in remove_cpu_topology() 2021-11-11 13:09:33 +01:00
bcma pci-v5.16-changes 2021-11-06 14:36:12 -07:00
block block-5.16-2021-12-03 2021-12-04 08:38:25 -08:00
bluetooth TTY / Serial driver update for 5.16-rc1 2021-11-04 09:09:37 -07:00
bus bus: mhi: pci_generic: Introduce Sierra EM919X support 2021-12-17 17:17:15 +01:00
cdrom for-5.16/cdrom-2021-10-29 2021-11-01 10:09:14 -07:00
char Merge v5.15-rc5 into char-misc-next 2021-12-13 10:17:10 +01:00
clk clk: versatile: clk-icst: use after free on error path 2021-12-07 12:25:29 -08:00
clocksource clocksource/drivers/arm_arch_timer: Force inlining of erratum_set_next_event_generic() 2021-12-10 17:47:00 +01:00
comedi comedi: Move "comedi_isadma.h" to <linux/comedi/comedi_isadma.h> 2021-11-26 16:48:59 +01:00
connector
counter counter: Fix use-after-free race condition for events_queue_size write 2021-10-21 13:02:47 +02:00
cpufreq cpufreq: Fix a comment in cpufreq_policy_free 2021-12-01 20:02:11 +01:00
cpuidle ARM: SoC drivers for 5.16 2021-11-03 17:00:52 -07:00
crypto pci-v5.16-changes 2021-11-06 14:36:12 -07:00
cxl cxl for v5.16 2021-11-08 11:49:48 -08:00
dax
dca
devfreq Merge branches 'pm-opp' and 'pm-cpufreq' 2021-11-10 14:06:51 +01:00
dio
dma dmaengine updates for v5.16-rc1 2021-11-10 11:47:55 -08:00
dma-buf dma-buf: system_heap: Use 'for_each_sgtable_sg' in pages free flow 2021-12-01 15:30:10 +05:30
edac - amd64_edac: Add support for three-rank interleaving mode which is 2021-11-01 15:02:49 -07:00
eisa
extcon extcon: usbc-tusb320: Add support for TUSB320L 2021-10-27 14:13:39 +09:00
firewire SCSI misc on 20211105 2021-11-05 08:42:02 -07:00
firmware firmware: xilinx: instantiate xilinx event manager driver 2021-12-03 16:08:46 +01:00
fpga
fsi fsi: sbefifo: Use interruptible mutex locking 2021-10-22 09:54:33 +10:30
gnss
gpio gpio: rockchip: needs GENERIC_IRQ_CHIP to fix build errors 2021-11-16 09:41:44 +01:00
gpu drm fixes for 5.16-rc5 2021-12-10 11:29:53 -08:00
greybus
hid HID: Ignore battery for Elan touchscreen on Asus UX550VE 2021-12-08 18:16:07 +01:00
hsi HSI changes for the 5.16 series 2021-11-04 13:56:55 -07:00
hv Drivers: hv: balloon: Use VMBUS_RING_SIZE() wrapper for dm_ring_size 2021-11-15 12:35:56 +00:00
hwmon hwmon: (pwm-fan) Ensure the fan going on in .probe() 2021-11-30 06:44:18 -08:00
hwspinlock
hwtracing coresight: trbe: Work around write to out of range 2021-10-27 11:46:01 -06:00
i2c i2c: mpc: Use atomic read and fix break condition 2021-12-10 22:27:30 +01:00
i3c
idle
iio iio: trigger: stm32-timer: fix MODULE_ALIAS 2021-12-04 15:37:02 +00:00
infiniband RDMA/irdma: Don't arm the CQ more than two times if no CE for this CQ 2021-12-07 13:53:01 -04:00
input xen: add "not_essential" flag to struct xenbus_driver 2021-11-23 13:41:29 -06:00
interconnect
iommu iommu/vt-d: Fix unmap_pages support 2021-11-26 22:54:47 +01:00
ipack
irqchip irqchip/irq-bcm7120-l2: Add put_device() after of_find_device_by_node() 2021-12-10 13:23:13 +00:00
isdn mISDN: Fix return values of the probe function 2021-10-19 13:09:28 +01:00
leds
macintosh Merge branch 'akpm' (patches from Andrew) 2021-11-06 14:08:17 -07:00
mailbox mailbox: imx: support i.MX8ULP S4 MU 2021-10-29 23:03:09 -05:00
mcb
md md: fix double free of mddev->private in autorun_array() 2021-12-10 09:11:07 -08:00
media media fixes for v5.16-rc3 2021-11-22 14:58:57 -08:00
memory memory: mtk-smi: Fix a null dereference for the ostd 2021-11-25 14:46:00 +01:00
memstick memstick: r592: Fix a UAF bug when removing the driver 2021-10-19 13:04:42 +02:00
message pci-v5.16-changes 2021-11-06 14:36:12 -07:00
mfd chrome platform changes for 5.16 2021-11-10 11:36:43 -08:00
misc lkdtm updates for v5.17-rc1 2021-12-17 10:06:21 +01:00
mmc mmc: mediatek: free the ext_csd when mmc_get_ext_csd success 2021-12-09 10:30:11 +01:00
most most: usb: replace snprintf in show functions with sysfs_emit 2021-11-26 17:03:47 +01:00
mtd Revert "mtd_blkdevs: don't scan partitions for plain mtdblock" 2021-12-10 11:52:34 -07:00
mux mux: add support for delay after muxing 2021-10-21 20:02:42 +01:00
net Char/misc driver fixes for 5.16-rc5 2021-12-12 10:16:34 -08:00
nfc nfc: virtual_ncidev: change default device permissions 2021-11-26 11:14:31 -08:00
ntb
nubus
nvdimm libnvdimm for v5.16 2021-11-10 10:56:02 -08:00
nvme nvmet-tcp: fix possible list corruption for unexpected command failure 2021-12-08 16:36:58 +01:00
nvmem nvmem: core: set size for sysfs bin file 2021-12-03 16:03:37 +01:00
of of/irq: Add a quirk for controllers with their own definition of interrupt-map 2021-12-03 11:30:22 -06:00
opp
parisc
parport
pci pci-v5.16-fixes-2 2021-12-10 11:56:05 -08:00
pcmcia Core: 2021-11-02 06:20:58 -07:00
perf ACPI updates for 5.16-rc1 2021-11-02 15:58:39 -07:00
phy phy: HiSilicon: Fix copy and paste bug in error handling 2021-11-23 10:42:13 +05:30
pinctrl pinctrl: qcom: sm8350: Correct UFS and SDC offsets 2021-11-16 02:19:15 +01:00
platform platform-drivers-x86 for v5.16-3 2021-12-07 10:10:20 -08:00
pnp
power power: supply: bq25890: Fix initial setting of the F_CONV_RATE field 2021-11-02 16:48:47 +01:00
powercap powercap: DTPM: Drop unused local variable from init_dtpm() 2021-12-03 17:51:59 +01:00
pps
ps3
ptp ptp: ocp: Fix a couple NULL vs IS_ERR() checks 2021-11-18 12:12:55 +00:00
pwm pwm: vt8500: Rename pwm_busy_wait() to make it obviously driver-specific 2021-11-05 11:57:13 +01:00
rapidio rapidio: avoid bogus __alloc_size warning 2021-11-06 13:30:33 -07:00
ras
regulator - Remove Drivers 2021-11-08 12:07:52 -08:00
remoteproc
reset ARM: SoC drivers for 5.16 2021-11-03 17:00:52 -07:00
rpmsg remoteproc updates for v5.16 2021-11-10 09:07:26 -08:00
rtc RTC for 5.16 2021-11-12 11:44:31 -08:00
s390 s390: replace snprintf in show functions with sysfs_emit 2021-11-16 12:29:19 +01:00
sbus
scsi SCSI fixes on 20211211 2021-12-11 16:28:27 -08:00
sh
siox
slimbus
soc driver: soc: xilinx: register for power events in zynqmp power driver 2021-12-03 16:08:46 +01:00
soundwire soundwire: qcom: add debugfs entry for soundwire register dump 2021-10-20 20:54:59 +05:30
spi spi: Fixes for v5.16 2021-11-18 14:35:41 -08:00
spmi spmi: spmi-pmic-arb: fix irq_set_type race condition 2021-12-17 17:18:18 +01:00
ssb
staging staging: rtl8192e: Fix use after free in _rtl92e_pci_disconnect() 2021-11-17 14:08:57 +01:00
target scsi: target: configfs: Delete unnecessary checks for NULL 2021-11-18 23:07:02 -05:00
tc
tee optee: fix kfree NULL pointer 2021-11-16 14:41:23 +01:00
thermal thermal: int340x: Fix VCoRefLow MMIO bit offset for TGL 2021-12-08 15:29:22 +01:00
thunderbolt thunderbolt: Changes for v5.16 merge window 2021-10-25 13:17:29 +02:00
tty TTY/Serial fixes for 5.16-rc4 2021-12-05 09:13:20 -08:00
uio uio: uio_dmem_genirq: Catch the Exception 2021-12-05 14:25:00 +01:00
usb USB fixes for 5.16-rc5 2021-12-12 10:20:57 -08:00
vdpa vdpa_sim: avoid putting an uninitialized iova_domain 2021-11-24 19:00:29 -05:00
vfio vfio/pci: Fix OpRegion read 2021-11-30 11:41:49 -07:00
vhost vhost-vdpa: clean irqs before reseting vdpa device 2021-11-24 19:00:28 -05:00
video TTY/Serial fixes for 5.16-rc4 2021-12-05 09:13:20 -08:00
virt nitro_enclaves: Add KUnit tests for contiguous physical memory regions merging 2021-11-17 16:11:09 +01:00
virtio Revert "virtio_ring: validate used buffer length" 2021-11-24 18:47:27 -05:00
visorbus
vlynq
vme
w1 w1: Misuse of get_user()/put_user() reported by sparse 2021-12-03 14:30:05 +01:00
watchdog linux-watchdog 5.16-rc1 tag 2021-11-10 09:41:22 -08:00
xen xen: detect uninitialized xenbus in xenbus_init 2021-11-24 08:55:15 -06:00
zorro
Kconfig
Makefile