[ Upstream commit 1833e327a5 ]
This is needed by LKDTM (crash dump test module), it calls
flush_icache_range(), which on RISC-V turns into flush_icache_all(). On
other architectures, the actual implementation is exported, so follow
that precedence and export it here too.
Fixes build of CONFIG_LKDTM that fails with:
ERROR: "flush_icache_all" [drivers/misc/lkdtm/lkdtm.ko] undefined!
Signed-off-by: Olof Johansson <olof@lixom.net>
Signed-off-by: Paul Walmsley <paul.walmsley@sifive.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit c65e41538b ]
firmware attempts to load test modules that require root access
and fail. Fix it to check for root uid and exit with skip code
instead.
Before this fix:
selftests: firmware: fw_run_tests.sh
modprobe: ERROR: could not insert 'test_firmware': Operation not permitted
You must have the following enabled in your kernel:
CONFIG_TEST_FIRMWARE=y
CONFIG_FW_LOADER=y
CONFIG_FW_LOADER_USER_HELPER=y
CONFIG_IKCONFIG=y
CONFIG_IKCONFIG_PROC=y
not ok 1 selftests: firmware: fw_run_tests.sh # SKIP
With this fix:
selftests: firmware: fw_run_tests.sh
skip all tests: must be run as root
not ok 1 selftests: firmware: fw_run_tests.sh # SKIP
Signed-off-by: Shuah Khan <skhan@linuxfoundation.org>
Reviwed-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Shuah Khan <skhan@linuxfoundation.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit e4dc9a4c31 ]
The sector size of the block layer is 512 bytes, but integrity interval
size might be different (in case of 4K block size of the media). At the
initiator side the virtual start sector is the one that was originally
submitted by the block layer (512 bytes) for the Reftag usage. The
initiator converts the Reftag to integrity interval units and sends it to
the target. So the target virtual start sector should be calculated at
integrity interval units. prepare_fn() and complete_fn() don't remap
correctly the Reftag when using incorrect units of the virtual start
sector, which leads to the following protection error at the device:
"blk_update_request: protection error, dev sdb, sector 2048 op 0x0:(READ)
flags 0x10000 phys_seg 1 prio class 0"
To fix that, set the seed in integrity interval units.
Link: https://lore.kernel.org/r/1576078562-15240-1-git-send-email-israelr@mellanox.com
Signed-off-by: Israel Rukshin <israelr@mellanox.com>
Reviewed-by: Max Gurtovoy <maxg@mellanox.com>
Reviewed-by: Sagi Grimberg <sagi@grimberg.me>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
commit 1f0d5c911b upstream.
We expect 64-bit calculation result from below statement, however
in 32-bit machine, looped left shift operation on pgoff_t type
variable may cause overflow issue, fix it by forcing type cast.
page->index << PAGE_SHIFT;
Fixes: 26de9b1171 ("f2fs: avoid unnecessary updating inode during fsync")
Fixes: 0a2aa8fbb9 ("f2fs: refactor __exchange_data_block for speed up")
Signed-off-by: Chao Yu <yuchao0@huawei.com>
Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit bb99ff9baa upstream.
When the single antenna diversity support was sent upstream, only some
definitions were sent, due to a bad revert.
Fix this by adding the actual code.
Fixes: 5952e0ec3f ("iwlwifi: mvm: add support for single antenna diversity")
Signed-off-by: Luca Coelho <luciano.coelho@intel.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 091c6e9c08 upstream.
When building with Clang + -Wtautological-pointer-compare:
drivers/net/wireless/realtek/rtlwifi/regd.c:389:33: warning: comparison
of address of 'rtlpriv->regd' equal to a null pointer is always false
[-Wtautological-pointer-compare]
if (wiphy == NULL || &rtlpriv->regd == NULL)
~~~~~~~~~^~~~ ~~~~
1 warning generated.
The address of an array member is never NULL unless it is the first
struct member so remove the unnecessary check. This was addressed in
the staging version of the driver in commit f986978b32 ("Staging:
rtlwifi: remove unnecessary NULL check").
While we are here, fix the following checkpatch warning:
CHECK: Comparison to NULL could be written "!wiphy"
35: FILE: drivers/net/wireless/realtek/rtlwifi/regd.c:389:
+ if (wiphy == NULL)
Fixes: 0c8173385e ("rtl8192ce: Add new driver")
Link:https://github.com/ClangBuiltLinux/linux/issues/750
Signed-off-by: Nathan Chancellor <natechancellor@gmail.com>
Acked-by: Ping-Ke Shih <pkshih@realtek.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit ee4cce9b9d upstream.
debugfs amsdu_len sets only the max_amsdu_len for ieee80211 station
so take it into consideration while getting max amsdu
Fixes: af2984e9e6 ("iwlwifi: mvm: add a debugfs entry to set a fixed size AMSDU for all TX packets")
Signed-off-by: Mordechay Goodstein <mordechay.goodstein@intel.com>
Signed-off-by: Luca Coelho <luciano.coelho@intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 2de860b4a7 upstream.
As platform_get_irq_byname() now prints an error when the interrupt
does not exist, scary warnings may be printed for optional interrupts:
renesas_spi e6b10000.spi: IRQ rx not found
renesas_spi e6b10000.spi: IRQ mux not found
Fix this by calling platform_get_irq_byname_optional() instead.
Remove the no longer needed printing of platform_get_irq errors, as the
remaining calls to platform_get_irq() and platform_get_irq_byname() take
care of that.
Fixes: 7723f4c5ec ("driver core: platform: Add an error message to platform_get_irq*()")
Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be>
Reviewed-by: Stephen Boyd <swboyd@chromium.org>
Link: https://lore.kernel.org/r/20191016143101.28738-1-geert+renesas@glider.be
Signed-off-by: Mark Brown <broonie@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit fed8d8c7a6 upstream.
The driver does the wrong thing when cs_change is set on a non-last
xfer in a message. When cs_change is set, the driver deactivates the
CS and leaves it off until a later xfer again has cs_change set whereas
it should be briefly toggling CS off and on again.
This patch brings the behaviour of the driver back in line with the
documentation and common sense. The delay of 10 us is the same as is
used by the default spi_transfer_one_message() function in spi.c.
[gregory: rebased on for-5.5 from spi tree]
Fixes: 8090d6d1a4 ("spi: atmel: Refactor spi-atmel to use SPI framework queue")
Signed-off-by: Mans Rullgard <mans@mansr.com>
Acked-by: Nicolas Ferre <nicolas.ferre@atmel.com>
Signed-off-by: Gregory CLEMENT <gregory.clement@bootlin.com>
Link: https://lore.kernel.org/r/20191018153504.4249-1-gregory.clement@bootlin.com
Signed-off-by: Mark Brown <broonie@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 3d63ee5deb upstream.
spi_nor_read_raw() assigns the result of 'ssize_t spi_nor_read_data()'
to the 'int ret' variable, while 'ssize_t' is a 64-bit type and *int*
is a 32-bit type on the 64-bit machines. This silent truncation isn't
really valid, so fix up the variable's type.
Fixes: f384b352cb ("mtd: spi-nor: parse Serial Flash Discoverable Parameters (SFDP) tables")
Signed-off-by: Sergei Shtylyov <sergei.shtylyov@cogentembedded.com>
Signed-off-by: Tudor Ambarus <tudor.ambarus@microchip.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit a719a75a77 upstream.
spi_nor_read() assigns the result of 'ssize_t spi_nor_read_data()'
to the 'int ret' variable, while 'ssize_t' is a 64-bit type and *int*
is a 32-bit type on the 64-bit machines. This silent truncation isn't
really valid, so fix up the variable's type.
Fixes: 59451e1233 ("mtd: spi-nor: change return value of read/write")
Signed-off-by: Sergei Shtylyov <sergei.shtylyov@cogentembedded.com>
Signed-off-by: Tudor Ambarus <tudor.ambarus@microchip.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 10256f0009 upstream.
If there are more than one valid snod on the sleb->nodes list,
do_kill_orphans will malloc ino more than once without releasing
previous ino's memory. Finally, it will trigger memory leak.
Fixes: ee1438ce5d ("ubifs: Check link count of inodes when...")
Signed-off-by: Zhihao Cheng <chengzhihao1@huawei.com>
Signed-off-by: zhangyi (F) <yi.zhang@huawei.com>
Signed-off-by: Richard Weinberger <richard@nod.at>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit df22b5b3ec upstream.
In the ubifs_jnl_write_inode() functon, it calls ubifs_iget()
with xent->inum. The xent->inum is __le64, but the ubifs_iget()
takes native cpu endian.
I think that this should be changed to passing le64_to_cpu(xent->inum)
to fix the following sparse warning:
fs/ubifs/journal.c:902:58: warning: incorrect type in argument 2 (different base types)
fs/ubifs/journal.c:902:58: expected unsigned long inum
fs/ubifs/journal.c:902:58: got restricted __le64 [usertype] inum
Fixes: 7959cf3a75 ("ubifs: journal: Handle xattrs like files")
Signed-off-by: Ben Dooks <ben.dooks@codethink.co.uk>
Signed-off-by: Richard Weinberger <richard@nod.at>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit fb03082a54 upstream.
In the commit 4f0a1a1ae3 ("memory: mtk-smi: Invoke pm runtime_callback
to enable clocks"), we use pm_runtime callback to enable/disable the smi
larb clocks. It will cause the larb's clock may not be disabled when
suspend. That is because device_prepare will call pm_runtime_get_noresume
which will keep the larb's PM runtime status still is active when suspend,
then it won't enter our pm_runtime suspend callback to disable the
corresponding clocks.
This patch adds suspend pm_ops to force disable the clocks, Use "LATE" to
make sure it disable the larb's clocks after the multimedia devices.
Fixes: 4f0a1a1ae3 ("memory: mtk-smi: Invoke pm runtime_callback to enable clocks")
Signed-off-by: Anan Sun <anan.sun@mediatek.com>
Signed-off-by: Yong Wu <yong.wu@mediatek.com>
Signed-off-by: Joerg Roedel <jroedel@suse.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit da3cc91b8d upstream.
The commit 4d689b6194 ("iommu/io-pgtable-arm-v7s: Convert to IOMMU API
TLB sync") help move the tlb_sync of unmap from v7s into the iommu
framework. It helps add a new function "mtk_iommu_iotlb_sync", But it
lacked the lock, then it will cause the variable "tlb_flush_active"
may be changed unexpectedly, we could see this warning log randomly:
mtk-iommu 10205000.iommu: Partial TLB flush timed out, falling back to
full flush
The HW requires tlb_flush/tlb_sync in pairs strictly, this patch adds
a new tlb_lock for tlb operations to fix this issue.
Fixes: 4d689b6194 ("iommu/io-pgtable-arm-v7s: Convert to IOMMU API TLB sync")
Signed-off-by: Yong Wu <yong.wu@mediatek.com>
Signed-off-by: Joerg Roedel <jroedel@suse.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit a2cbf80a84 upstream.
The FIELDPIC_FLAG_E bit should be set when field_pic_flag exists in stream,
it is currently set based on field_pic_flag of current frame.
The PIC_FIELDMODE_E bit is correctly set based on the field_pic_flag.
Fix this by setting the FIELDPIC_FLAG_E bit when frame_mbs_only is not set.
Fixes: dea0a82f3d ("media: hantro: Add support for H264 decoding on G1")
Signed-off-by: Jonas Karlman <jonas@kwiboo.se>
Reviewed-by: Boris Brezillon <boris.brezillon@collabora.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit c3df30a01d upstream.
In the implementation of aspeed_video_probe() the allocated memory for
video should be released if either devm_ioremap_resource()
or aspeed_video_init() or aspeed_video_setup_video() fails. Replace
kzalloc() with devm_kzalloc to avoid explicit release for video.
Fixes: d2b4387f3b ("media: platform: Add Aspeed Video Engine driver")
Signed-off-by: Navid Emamdoost <navid.emamdoost@gmail.com>
Reviewed-by: Jae Hyun Yoo <jae.hyun.yoo@linux.intel.com>
Reviewed-by: Eddie James <eajames@linux.ibm.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit e17f08e316 upstream.
Scaling list supplied from userspace should be in matrix order
and can be used without applying the inverse scanning process.
The HW also only support 8x8 scaling list for the Y component, indices 0
and 1 in the scaling list supplied from userspace.
Remove reordering and write the scaling matrix in an order expected by
the VPU, also only allocate memory for the two 8x8 lists supported.
Fixes: a9471e2562 ("media: hantro: Add core bits to support H264 decoding")
Signed-off-by: Jonas Karlman <jonas@kwiboo.se>
Reviewed-by: Philipp Zabel <p.zabel@pengutronix.de>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit a3fd80198d upstream.
The BIT decoder picture run temporarily locks the bitstream mutex while
the coda device mutex is locked, to refill the bitstream ring buffer.
Consequently, the decoder start command, which locks both mutexes when
flushing the bitstream ring buffer, must lock the coda device mutex
first as well, to avoid an ABBA deadlock.
Fixes: e7fd95849b ("media: coda: flush bitstream ring buffer on decoder restart")
Signed-off-by: Philipp Zabel <p.zabel@pengutronix.de>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 704c6c80fb upstream.
>From isp_video_release(), &isp->video_lock is held and subsequent
vb2_fop_release() tries to lock vdev->lock which is same with the
previous one. Replace vb2_fop_release() with _vb2_fop_release() to
fix the recursive locking.
Fixes: 1380f5754c ("[media] videobuf2: Add missing lock held on vb2_fop_release")
Signed-off-by: Seung-Woo Kim <sw0312.kim@samsung.com>
Reviewed-by: Sylwester Nawrocki <s.nawrocki@samsung.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 5439fa9263 upstream.
It is not clear what pixel format is actually configured in hardware on
reset. MEDIA_BUS_FMT_YUYV8_2X8, assumed on device probe since the
driver was intiially submitted, is for sure not the one.
Fix it by explicitly applying a known, driver default frame format just
after initial device reset.
Fixes: 2f6e240479 ("[media] SoC Camera: add driver for OV6650 sensor")
Signed-off-by: Janusz Krzysztofik <jmkrzyszt@gmail.com>
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 39034bb0c2 upstream.
Commit da298c6d98 ("[media] v4l2: replace video op g_mbus_fmt by pad
op get_fmt") converted a former ov6650_g_fmt() video operation callback
to an ov6650_get_fmt() pad operation callback. However, the converted
function disregards a format->which flag that pad operations should
obey and always returns active frame format settings.
That can be fixed by always responding to V4L2_SUBDEV_FORMAT_TRY with
-EINVAL, or providing the response from a pad config argument, likely
updated by a former user call to V4L2_SUBDEV_FORMAT_TRY .set_fmt().
Since implementation of the latter is trivial, go for it.
Fixes: da298c6d98 ("[media] v4l2: replace video op g_mbus_fmt by pad op get_fmt")
Signed-off-by: Janusz Krzysztofik <jmkrzyszt@gmail.com>
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 1c6a2b6309 upstream.
User arguments passed to .get/set_fmt() pad operation callbacks may
contain unsupported values. The driver takes control over frame size
and pixel code as well as colorspace and field attributes but has never
cared for remainig format attributes, i.e., ycbcr_enc, quantization
and xfer_func, introduced by commit 11ff030c73 ("[media]
v4l2-mediabus: improve colorspace support"). Fix it.
Set up a static v4l2_mbus_framefmt structure with attributes
initialized to reasonable defaults and use it for updating content of
user provided arguments. In case of V4L2_SUBDEV_FORMAT_ACTIVE,
postpone frame size update, now performed from inside ov6650_s_fmt()
helper, util the user argument is first updated in ov6650_set_fmt() with
default frame format content. For V4L2_SUBDEV_FORMAT_TRY, don't copy
all attributes to pad config, only those handled by the driver, then
fill the response with the default frame format updated with resulting
pad config format code and frame size.
Fixes: 11ff030c73 ("[media] v4l2-mediabus: improve colorspace support")
Signed-off-by: Janusz Krzysztofik <jmkrzyszt@gmail.com>
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 1250073189 upstream.
Since its initial submission, the driver selects V4L2_COLORSPACE_JPEG
for supported formats other than V4L2_MBUS_FMT_SBGGR8_1X8. According
to v4l2-compliance test program, V4L2_COLORSPACE_JPEG applies
exclusively to V4L2_PIX_FMT_JPEG. Since the sensor does not support
JPEG format, fix it to always select V4L2_COLORSPACE_SRGB.
Fixes: 2f6e240479 ("[media] SoC Camera: add driver for OV6650 sensor")
Signed-off-by: Janusz Krzysztofik <jmkrzyszt@gmail.com>
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit ff98a5f624 upstream.
Commit ca74b316df ("arm: Use common cpu_topology structure and
functions.") changed cpu_coregroup_mask() from the ARM32 specific
implementation in arch/arm/include/asm/topology.h to the one shared
with ARM64 and RISCV in drivers/base/arch_topology.c.
Currently on ARM32 (TC2 w/ CONFIG_SCHED_MC) the task scheduler setup
code (w/ CONFIG_SCHED_DEBUG) shows this during CPU hotplug:
ERROR: groups don't span domain->span
It happens to CPUs of the cluster of the CPU which gets hot-plugged
out on scheduler domain MC.
Turns out that the shared cpu_coregroup_mask() requires that the
hot-plugged CPU is removed from the core_sibling mask via
remove_cpu_topology(). Otherwise the 'is core_sibling subset of
cpumask_of_node()' doesn't work. In this case the task scheduler has to
deal with cpumask_of_node instead of core_sibling which is wrong on
scheduler domain MC.
e.g. CPU3 hot-plugged out on TC2 [cluster0: 0,3-4 cluster1: 1-2]:
cpu_coregroup_mask(): CPU3 cpumask_of_node=0-2,4 core_sibling=0,3-4
^
should be:
cpu_coregroup_mask(): CPU3 cpumask_of_node=0-2,4 core_sibling=0,4
Add remove_cpu_topology() to __cpu_disable() to remove the CPU from the
topology masks in case of a CPU hotplug out operation.
At the same time tweak store_cpu_topology() slightly so it will call
update_siblings_masks() in case of CPU hotplug in operation via
secondary_start_kernel()->smp_store_cpu_info().
This aligns the ARM32 implementation with the ARM64 one.
Guarding remove_cpu_topology() with CONFIG_GENERIC_ARCH_TOPOLOGY is
necessary since some Arm32 defconfigs (aspeed_g5_defconfig,
milbeaut_m10v_defconfig, spear13xx_defconfig) specify an explicit
# CONFIG_ARM_CPU_TOPOLOGY is not set
w/ ./arch/arm/Kconfig: select GENERIC_ARCH_TOPOLOGY if ARM_CPU_TOPOLOGY
Fixes: ca74b316df ("arm: Use common cpu_topology structure and functions")
Reviewed-by: Sudeep Holla <sudeep.holla@arm.com>
Reviewed-by: Lukasz Luba <lukasz.luba@arm.com>
Tested-by: Lukasz Luba <lukasz.luba@arm.com>
Tested-by: Ondrej Jirman <megous@megous.com>
Signed-off-by: Dietmar Eggemann <dietmar.eggemann@arm.com>
Signed-off-by: Russell King <rmk+kernel@armlinux.org.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 74887542fd upstream.
Per Documentation/DMA-API-HOWTO.txt,
To unmap a scatterlist, just call:
dma_unmap_sg(dev, sglist, nents, direction);
.. note::
The 'nents' argument to the dma_unmap_sg call must be
the _same_ one you passed into the dma_map_sg call,
it should _NOT_ be the 'count' value _returned_ from the
dma_map_sg call.
However in the driver, priv->nent is directly assigned with value
returned from dma_map_sg, and dma_unmap_sg use priv->nent for unmap,
this breaks the API usage.
So introduce a new entry orig_nent to remember 'nents'.
Fixes: da3564ee02 ("pch_uart: add multi-scatter processing")
Signed-off-by: Peng Fan <peng.fan@nxp.com>
Link: https://lore.kernel.org/r/1573623259-6339-1-git-send-email-peng.fan@nxp.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 9d72dcef89 upstream.
On PowerNV the PCIe topology is (currently) managed by the powernv platform
code in Linux in cooperation with the platform firmware. Linux's native
PCIe port service drivers operate independently of both and this can cause
problems.
The main issue is that the portbus driver will conflict with the platform
specific hotplug driver (pnv_php) over ownership of the MSI used to notify
the host when a hotplug event occurs. The portbus driver claims this MSI on
behalf of the individual port services because the same interrupt is used
for hotplug events, PMEs (on root ports), and link bandwidth change
notifications. The portbus driver will always claim the interrupt even if
the individual port service drivers, such as pciehp, are compiled out.
The second, bigger, problem is that the hotplug port service driver
fundamentally does not work on PowerNV. The platform assumes that all
PCI devices have a corresponding arch-specific handle derived from the DT
node for the device (pci_dn) and without one the platform will not allow
a PCI device to be enabled. This problem is largely due to historical
baggage, but it can't be resolved without significant re-factoring of the
platform PCI support.
We can fix these problems in the interim by setting the
"pcie_ports_disabled" flag during platform initialisation. The flag
indicates the platform owns the PCIe ports which stops the portbus driver
from being registered.
This does have the side effect of disabling all port services drivers
that is: AER, PME, BW notifications, hotplug, and DPC. However, this is
not a huge disadvantage on PowerNV since these services are either unused
or handled through other means.
Fixes: 66725152fb ("PCI/hotplug: PowerPC PowerNV PCI hotplug driver")
Signed-off-by: Oliver O'Halloran <oohall@gmail.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/20191118065553.30362-1-oohall@gmail.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
