[ Upstream commit 5ca86eae55 ]
Afer commit 1fa5ae857b ("driver core: get rid of struct device's
bus_id string array"), the name of device is allocated dynamically. It
needs to be freed when of_device_register() fails. Call put_device() to
give up the reference that's taken in device_initialize(), so that it
can be freed in kobject_cleanup() when the refcount hits 0.
macio device is freed in macio_release_dev(), so the kfree() can be
removed.
Fixes: 1fa5ae857b ("driver core: get rid of struct device's bus_id string array")
Signed-off-by: Yang Yingliang <yangyingliang@huawei.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/20221104032551.1075335-1-yangyingliang@huawei.com
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit 01657bc14a ]
We currently have 3 different ways that __iommu_probe_device() may be
called, but no real guarantee that multiple callers can't tread on each
other, especially once asynchronous driver probe gets involved. It would
likely have taken a fair bit of luck to hit this previously, but commit
57365a04c9 ("iommu: Move bus setup to IOMMU device registration") ups
the odds since now it's not just omap-iommu that may trigger multiple
bus_iommu_probe() calls in parallel if probing asynchronously.
Add a lock to ensure we can't try to double-probe a device, and also
close some possible race windows to make sure we're truly robust against
trying to double-initialise a group via two different member devices.
Reported-by: Brian Norris <briannorris@chromium.org>
Signed-off-by: Robin Murphy <robin.murphy@arm.com>
Tested-by: Brian Norris <briannorris@chromium.org>
Fixes: 57365a04c9 ("iommu: Move bus setup to IOMMU device registration")
Link: https://lore.kernel.org/r/1946ef9f774851732eed78760a78ec40dbc6d178.1667591503.git.robin.murphy@arm.com
Signed-off-by: Joerg Roedel <jroedel@suse.de>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit 0462681e20 ]
On an iMX6ULL the following message appears when a wakealarm is set:
echo 0 > /sys/class/rtc/rtc1/wakealarm
rtc rtc1: Timeout trying to get valid LPSRT Counter read
This does not always happen but is reproducible quite often (7 out of 10
times). The problem appears because the iMX6ULL is not able to read the
registers within one 32kHz clock cycle which is the base clock of the
RTC. Therefore, this patch allows a difference of up to 320 cycles
(10ms). 10ms was chosen to be big enough even on systems with less cpu
power (e.g. iMX6ULL). According to the reference manual a difference is
fine:
- If the two consecutive reads are similar, the value is correct.
The values have to be similar, not equal.
Fixes: cd7f3a249d ("rtc: snvs: Add timeouts to avoid kernel lockups")
Reviewed-by: Francesco Dolcini <francesco.dolcini@toradex.com>
Signed-off-by: Stefan Eichenberger <stefan.eichenberger@toradex.com>
Signed-off-by: Francesco Dolcini <francesco@dolcini.it>
Link: https://lore.kernel.org/r/20221106115915.7930-1-francesco@dolcini.it
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit 508ccdfb86 ]
Notice that cmos_wake_setup() is the only user of acpi_rtc_info and it
can operate on the cmos_rtc variable directly, so it need not set the
platform_data pointer before cmos_do_probe() is called. Instead, it
can be called by cmos_do_probe() in the case when the platform_data
pointer is not set to implement the default behavior (which is to use
the FADT information as long as ACPI support is enabled).
Modify the code accordingly.
While at it, drop a comment that doesn't really match the code it is
supposed to be describing.
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Reviewed-by: Zhang Rui <rui.zhang@intel.com>
Tested-by: Zhang Rui <rui.zhang@intel.com>
Reviewed-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Link: https://lore.kernel.org/r/4803444.31r3eYUQgx@kreacher
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Stable-dep-of: 83ebb7b303 ("rtc: cmos: Disable ACPI RTC event on removal")
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit 11bf53a38c ]
The PCIe QMP PHY requires different programming sequences when being
used for the RC (Root Complex) or for the EP (End Point) modes. Allow
selecting the submode and thus selecting a set of PHY programming
tables.
Since the RC and EP modes share common some common init sequence, the
common sequence is kept in the main table and the sequence differences
are pushed to the extra tables.
Reviewed-by: Johan Hovold <johan+linaro@kernel.org>
Signed-off-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org>
Link: https://lore.kernel.org/r/20220927092207.161501-3-dmitry.baryshkov@linaro.org
Signed-off-by: Vinod Koul <vkoul@kernel.org>
Stable-dep-of: 9ddcd920f8 ("phy: qcom-qmp-pcie: Fix high latency with 4x2 PHY when ASPM is enabled")
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit 2584068a9e ]
Commit af6643242d ("phy: qcom-qmp-pcie: split pcs_misc region for ipq6018
pcie gen3") reworked the pcs regs values and removed the 0x400 offset
for each pcs_misc regs.
This change caused the malfunction of ipq8074 downstream since it still
has the legacy pcs table where pcs_misc are not placed on a different
table and instead put together assuming the offset of 0x400 for the
related pcs_misc regs.
Split pcs_misc init cfg from the ipq8074 pcs init table to be handled
correctly to prepare for actual support for gen3 pcie for ipq8074.
Fixes: af6643242d ("phy: qcom-qmp-pcie: split pcs_misc region for ipq6018 pcie gen3")
Reported-by: Robert Marko <robimarko@gmail.com>
Tested-by: Robert Marko <robimarko@gmail.com>
Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
Link: https://lore.kernel.org/r/20221103212125.17156-1-ansuelsmth@gmail.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit 2566ad8ec4 ]
SM8250 configuration tables are split into two parts: the common one and
the PHY-specific tables. Make this split more formal. Rather than having
a blind renamed copy of all QMP table fields, add separate struct
qmp_phy_cfg_tables and add two instances of this structure to the struct
qmp_phy_cfg. Later on this will be used to support different PHY modes
(RC vs EP).
Reviewed-by: Johan Hovold <johan+linaro@kernel.org>
Signed-off-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org>
Link: https://lore.kernel.org/r/20220927092207.161501-2-dmitry.baryshkov@linaro.org
Signed-off-by: Vinod Koul <vkoul@kernel.org>
Stable-dep-of: 2584068a9e ("phy: qcom-qmp-pcie: split pcs_misc init cfg for ipq8074 pcs table")
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit 5eccd0d9fa ]
When dynamically scaling the PWM clock, the function
dev_pm_opp_set_rate() may set the PWM clock to a rate that is lower than
what is required. The clock rate requested when calling
dev_pm_opp_set_rate() is the minimum clock rate that is needed to drive
the PWM to achieve the required period. Hence, if the actual clock
rate is less than the requested clock rate, then the required period
cannot be achieved and configuring the PWM fails. Fix this by
calling clk_round_rate() to check if the clock rate that will be provided
is sufficient and if not, double the required clock rate to ensure the
required period can be attained.
Fixes: 8c193f4714 ("pwm: tegra: Optimize period calculation")
Signed-off-by: Jon Hunter <jonathanh@nvidia.com>
Acked-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Thierry Reding <thierry.reding@gmail.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit f271946117 ]
For the case where dev_pm_opp_set_rate() is called to set the PWM clock
rate, the requested rate is calculated as ...
required_clk_rate = (NSEC_PER_SEC / period_ns) << PWM_DUTY_WIDTH;
The above calculation may lead to rounding errors because the
NSEC_PER_SEC is divided by 'period_ns' before applying the
PWM_DUTY_WIDTH multiplication factor. For example, if the period is
45334ns, the above calculation yields a rate of 5646848Hz instead of
5646976Hz. Fix this by applying the multiplication factor before
dividing and using the DIV_ROUND_UP macro which yields the expected
result of 5646976Hz.
Fixes: 1d7796bdb6 ("pwm: tegra: Support dynamic clock frequency configuration")
Signed-off-by: Jon Hunter <jonathanh@nvidia.com>
Reviewed-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Thierry Reding <thierry.reding@gmail.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit defbab270d ]
Commit bc27fb68aa ("include/uapi/linux/byteorder, swab: force inlining
of some byteswap operations") added __always_inline to swab functions
and commit 283d757378 ("uapi/linux/stddef.h: Provide __always_inline to
userspace headers") added a definition of __always_inline for use in
exported headers when the kernel's compiler.h is not available.
However, since swab.h does not include stddef.h, if the header soup does
not indirectly include it, the definition of __always_inline is missing,
resulting in a compilation failure, which was observed compiling the
perf tool using exported headers containing this commit:
In file included from /usr/include/linux/byteorder/little_endian.h:12:0,
from /usr/include/asm/byteorder.h:14,
from tools/include/uapi/linux/perf_event.h:20,
from perf.h:8,
from builtin-bench.c:18:
/usr/include/linux/swab.h:160:8: error: unknown type name `__always_inline'
static __always_inline __u16 __swab16p(const __u16 *p)
Fix this by replacing the inclusion of linux/compiler.h with
linux/stddef.h to ensure that we pick up that definition if required,
without relying on it's indirect inclusion. compiler.h is then included
indirectly, via stddef.h.
Fixes: 283d757378 ("uapi/linux/stddef.h: Provide __always_inline to userspace headers")
Signed-off-by: Matt Redfearn <matt.redfearn@mips.com>
Signed-off-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Tested-by: Nathan Chancellor <nathan@kernel.org>
Reviewed-by: Petr Vaněk <arkamar@atlas.cz>
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit 4c6b3af390 ]
The PCS_USB register block lives at an offset of 0x1000 from the PCS
region on SC8280XP so add the missing offset to avoid corrupting
unrelated registers on runtime suspend.
Note that the current binding is broken as it does not describe the
PCS_USB region and the PCS register size does not cover PCS_USB and the
regions in between. As Linux currently maps full pages, simply adding
the offset to driver works until the binding has been fixed.
Fixes: c0c7769cda ("phy: qcom-qmp: Add SC8280XP USB3 UNI phy")
Reviewed-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org>
Signed-off-by: Johan Hovold <johan+linaro@kernel.org>
Link: https://lore.kernel.org/r/20221028160435.26948-2-johan+linaro@kernel.org
Signed-off-by: Vinod Koul <vkoul@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit 9a8ddb35a9 ]
In current code, the following sysfs attributes are exposed to user to
show or update the values:
max_read_buffers (max_tokens)
read_buffer_limit (token_limit)
group/read_buffers_allowed (group/tokens_allowed)
group/read_buffers_reserved (group/tokens_reserved)
group/use_read_buffer_limit (group/use_token_limit)
>From Intel IAA spec [1], Intel IAA does not support Read Buffer
allocation control. So these sysfs attributes should not be supported on
IAA device.
Fix this issue by making these sysfs attributes invisible through
is_visible() filter when the device is IAA.
Add description in the ABI documentation to mention that these
attributes are not visible when the device does not support Read Buffer
allocation control.
[1]: https://cdrdv2.intel.com/v1/dl/getContent/721858
Fixes: fde212e44f ("dmaengine: idxd: deprecate token sysfs attributes for read buffers")
Fixes: c52ca47823 ("dmaengine: idxd: add configuration component of driver")
Signed-off-by: Xiaochen Shen <xiaochen.shen@intel.com>
Reviewed-by: Dave Jiang <dave.jiang@intel.com>
Link: https://lore.kernel.org/r/20221022074949.11719-1-xiaochen.shen@intel.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit e563cc0c78 ]
Allocated iova ranges need to be invalidated immediately or otherwise
they might or might not work when used by master or CPU. This was
discovered when running video decoder conformity test with Cedrus. Some
videos were now and then decoded incorrectly and generated page faults.
According to vendor driver, it's enough to invalidate just start and end
TLB and PTW cache lines. Documentation says that neighbouring lines must
be invalidated too. Finally, when page fault occurs, that iova must be
invalidated the same way, according to documentation.
Fixes: 4100b8c229 ("iommu: Add Allwinner H6 IOMMU driver")
Signed-off-by: Jernej Skrabec <jernej.skrabec@gmail.com>
Link: https://lore.kernel.org/r/20221025165415.307591-6-jernej.skrabec@gmail.com
Signed-off-by: Joerg Roedel <jroedel@suse.de>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit eac0104dc6 ]
Because driver has enum type permissions and iommu subsystem has bitmap
type, we have to be careful how check for combined read and write
permissions is done. In such case, we have to mask both permissions and
check that both are set at the same time.
Current code just masks both flags but doesn't check that both are set.
In short, it always sets R/W permission, regardles if requested
permissions were RO, WO or RW. Fix that.
Fixes: 4100b8c229 ("iommu: Add Allwinner H6 IOMMU driver")
Signed-off-by: Jernej Skrabec <jernej.skrabec@gmail.com>
Link: https://lore.kernel.org/r/20221025165415.307591-4-jernej.skrabec@gmail.com
Signed-off-by: Joerg Roedel <jroedel@suse.de>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit 9ad0c1252e ]
Reset signal is asserted by writing 0 to the corresponding locations of
masters we want to reset. So in order to deassert all reset signals, we
should write 1's to all locations.
Current code writes 1's to locations of masters which were just reset
which is good. However, at the same time it also writes 0's to other
locations and thus asserts reset signals of remaining masters. Fix code
by writing all 1's when we want to deassert all reset signals.
This bug was discovered when working with Cedrus (video decoder). When
it faulted, display went blank due to reset signal assertion.
Fixes: 4100b8c229 ("iommu: Add Allwinner H6 IOMMU driver")
Signed-off-by: Jernej Skrabec <jernej.skrabec@gmail.com>
Link: https://lore.kernel.org/r/20221025165415.307591-2-jernej.skrabec@gmail.com
Signed-off-by: Joerg Roedel <jroedel@suse.de>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit bf8d2dd2ed ]
Since commit fa7e9ecc5e ("iommu/s390: Tolerate repeat attach_dev
calls") we can end up with duplicates in the list of devices attached to
a domain. This is inefficient and confusing since only one domain can
actually be in control of the IOMMU translations for a device. Fix this
by detaching the device from the previous domain, if any, on attach.
Add a WARN_ON() in case we still have attached devices on freeing the
domain. While here remove the re-attach on failure dance as it was
determined to be unlikely to help and may confuse debug and recovery.
Fixes: fa7e9ecc5e ("iommu/s390: Tolerate repeat attach_dev calls")
Reviewed-by: Matthew Rosato <mjrosato@linux.ibm.com>
Reviewed-by: Jason Gunthorpe <jgg@nvidia.com>
Signed-off-by: Niklas Schnelle <schnelle@linux.ibm.com>
Link: https://lore.kernel.org/r/20221025115657.1666860-2-schnelle@linux.ibm.com
Signed-off-by: Joerg Roedel <jroedel@suse.de>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit 38cd167d1f ]
The power-down delay was included in the first version of the QMP driver
as an optional delay after powering on the PHY (using
POWER_DOWN_CONTROL) and just before starting it. Later changes modified
this sequence by powering on before initialising the PHY, but the
optional delay stayed where it was (i.e. before starting the PHY).
The vendor driver does not use a delay before starting the PHY and this
is likely not needed on any platform unless there is a corresponding
delay in the vendor kernel init sequence tables (i.e. in devicetree).
Let's keep the delay for now, but drop the redundant delay period
configuration while increasing the unnecessarily low timer slack
somewhat.
Reviewed-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org>
Signed-off-by: Johan Hovold <johan+linaro@kernel.org>
Link: https://lore.kernel.org/r/20221012081241.18273-15-johan+linaro@kernel.org
Signed-off-by: Vinod Koul <vkoul@kernel.org>
Stable-dep-of: 922adfd59e ("phy: qcom-qmp-usb: correct registers layout for IPQ8074 USB3 PHY")
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit 51bd33069f ]
The power-down delay was included in the first version of the QMP driver
as an optional delay after powering on the PHY (using
POWER_DOWN_CONTROL) and just before starting it. Later changes modified
this sequence by powering on before initialising the PHY, but the
optional delay stayed where it was (i.e. before starting the PHY).
The vendor driver does not use a delay before starting the PHY and this
is likely not needed on any platform unless there is a corresponding
delay in the vendor kernel init sequence tables (i.e. in devicetree).
But as the vendor kernel do have a 1 ms delay *after* starting the PHY
and before starting to poll the status it is possible that later
contributors have simply not noticed that the mainline power-down delay
is not equivalent.
As the current delay before even starting the PHY is pretty much
pointless and likely a mistake, move the delay after starting the PHY
which avoids a few iterations of polling and speeds up startup by 1 ms
(the poll loop otherwise takes about 1.8 ms).
Note that MSM8998 has never used a power-down delay so add a flag to
skip the delay in case starting the PHY is faster on MSM8998. This can
be removed after someone takes a measurement.
Signed-off-by: Johan Hovold <johan+linaro@kernel.org>
Reviewed-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org>
Link: https://lore.kernel.org/r/20221012081241.18273-10-johan+linaro@kernel.org
Signed-off-by: Vinod Koul <vkoul@kernel.org>
Stable-dep-of: 4a9eac5ae2 ("phy: qcom-qmp-pcie: fix sc8180x initialisation")
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit e71906144b ]
The power-down delay was included in the first version of the QMP driver
as an optional delay after powering on the PHY (using
POWER_DOWN_CONTROL) and just before starting it. Later changes modified
this sequence by powering on before initialising the PHY, but the
optional delay stayed where it was (i.e. before starting the PHY).
The vendor driver does not use a delay before starting the PHY and this
is likely not needed on any platform unless there is a corresponding
delay in the vendor kernel init sequence tables (i.e. in devicetree).
Let's keep the delay for now, but drop the redundant delay period
configuration while increasing the unnecessarily low timer slack
somewhat.
Reviewed-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org>
Signed-off-by: Johan Hovold <johan+linaro@kernel.org>
Link: https://lore.kernel.org/r/20221012081241.18273-9-johan+linaro@kernel.org
Signed-off-by: Vinod Koul <vkoul@kernel.org>
Stable-dep-of: 4a9eac5ae2 ("phy: qcom-qmp-pcie: fix sc8180x initialisation")
Signed-off-by: Sasha Levin <sashal@kernel.org>
