IF YOU WOULD LIKE TO GET AN ACCOUNT, please write an
email to Administrator. User accounts are meant only to access repo
and report issues and/or generate pull requests.
This is a purpose-specific Git hosting for
BaseALT
projects. Thank you for your understanding!
Только зарегистрированные пользователи имеют доступ к сервису!
Для получения аккаунта, обратитесь к администратору.
commit 2a93c6cbd5a703d44c414a3c3945a87ce11430ba upstream.
Commit 'e3e56c050ab6 ("soc: qcom: rpmhpd: Make power_on actually enable
the domain")' aimed to make sure that a power-domain that is being
enabled without any particular performance-state requested will at least
turn the rail on, to avoid filling DeviceTree with otherwise unnecessary
required-opps properties.
But in the event that aggregation happens on a disabled power-domain, with
an enabled peer without performance-state, both the local and peer
corner are 0. The peer's enabled_corner is not considered, with the
result that the underlying (shared) resource is disabled.
One case where this can be observed is when the display stack keeps mmcx
enabled (but without a particular performance-state vote) in order to
access registers and sync_state happens in the rpmhpd driver. As mmcx_ao
is flushed the state of the peer (mmcx) is not considered and mmcx_ao
ends up turning off "mmcx.lvl" underneath mmcx. This has been observed
several times, but has been painted over in DeviceTree by adding an
explicit vote for the lowest non-disabled performance-state.
Fixes: e3e56c050ab6 ("soc: qcom: rpmhpd: Make power_on actually enable the domain")
Reported-by: Johan Hovold <johan@kernel.org>
Closes: https://lore.kernel.org/linux-arm-msm/ZdMwZa98L23mu3u6@hovoldconsulting.com/
Cc: <stable@vger.kernel.org>
Signed-off-by: Bjorn Andersson <quic_bjorande@quicinc.com>
Reviewed-by: Konrad Dybcio <konrad.dybcio@linaro.org>
Reviewed-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org>
Tested-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org>
Reviewed-by: Abhinav Kumar <quic_abhinavk@quicinc.com>
Reviewed-by: Stephen Boyd <swboyd@chromium.org>
Tested-by: Johan Hovold <johan+linaro@kernel.org>
Link: https://lore.kernel.org/r/20240226-rpmhpd-enable-corner-fix-v1-1-68c004cec48c@quicinc.com
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit f1a1bc8775b26345aba2be278118999e7f661d3d ]
Usually there is only one llcc device. But if there were a second, even
a failed probe call would modify the global drv_data pointer. So check
if drv_data is valid before overwriting it.
Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Fixes: a3134fb09e0b ("drivers: soc: Add LLCC driver")
Link: https://lore.kernel.org/r/20230926083229.2073890-1-u.kleine-koenig@pengutronix.de
Signed-off-by: Bjorn Andersson <andersson@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
commit 8d207400fd6b79c92aeb2f33bb79f62dff904ea2 upstream.
The QMI TLV value for strings in a lot of qmi element info structures
account for null terminated strings with MAX_LEN + 1. If a string is
actually MAX_LEN + 1 length, this will cause an out of bounds access
when the NULL character is appended in decoding.
Fixes: 9b8a11e82615 ("soc: qcom: Introduce QMI encoder/decoder")
Cc: stable@vger.kernel.org
Signed-off-by: Chris Lew <quic_clew@quicinc.com>
Signed-off-by: Praveenkumar I <quic_ipkumar@quicinc.com>
Link: https://lore.kernel.org/r/20230801064712.3590128-1-quic_ipkumar@quicinc.com
Signed-off-by: Bjorn Andersson <andersson@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit a7b484b1c9332a1ee12e8799d62a11ee3f8e0801 ]
Since we're using these two macros to read a value from a register, we
need to use the FIELD_GET instead of the FIELD_PREP macro, otherwise
we're getting wrong values.
So instead of:
[ 3.111779] ocmem fdd00000.sram: 2 ports, 1 regions, 512 macros, not interleaved
we now get the correct value of:
[ 3.129672] ocmem fdd00000.sram: 2 ports, 1 regions, 2 macros, not interleaved
Fixes: 88c1e9404f1d ("soc: qcom: add OCMEM driver")
Reviewed-by: Caleb Connolly <caleb.connolly@linaro.org>
Reviewed-by: Konrad Dybcio <konrad.dybcio@linaro.org>
Signed-off-by: Luca Weiss <luca@z3ntu.xyz>
Link: https://lore.kernel.org/r/20230506-msm8226-ocmem-v3-1-79da95a2581f@z3ntu.xyz
Signed-off-by: Bjorn Andersson <andersson@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit e81a16e77259294cd4ff0a9c1fbe5aa0e311a47d ]
It might be useful to know what hardware version of the OCMEM block the
SoC contains. Add a debug print for that.
Signed-off-by: Luca Weiss <luca@z3ntu.xyz>
Reviewed-by: Konrad Dybcio <konrad.dybcio@linaro.org>
Signed-off-by: Bjorn Andersson <andersson@kernel.org>
Link: https://lore.kernel.org/r/20230509-ocmem-hwver-v3-1-e51f3488e0f4@z3ntu.xyz
Stable-dep-of: a7b484b1c933 ("soc: qcom: ocmem: Fix NUM_PORTS & NUM_MACROS macros")
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit 6049aae52392539e505bfb8ccbcff3c26f1d2f0b ]
If an error occurs after a successful pm_genpd_init() call, it should be
undone by a corresponding pm_genpd_remove().
Add the missing call in the error handling path, as already done in the
remove function.
Fixes: bf6910abf548 ("power: avs: Add support for CPR (Core Power Reduction)")
Signed-off-by: Christophe JAILLET <christophe.jaillet@wanadoo.fr>
Reviewed-by: Ulf Hansson <ulf.hansson@linaro.org>
Signed-off-by: Bjorn Andersson <andersson@kernel.org>
Link: https://lore.kernel.org/r/0f520597dbad89ab99c217c8986912fa53eaf5f9.1671293108.git.christophe.jaillet@wanadoo.fr
Signed-off-by: Sasha Levin <sashal@kernel.org>
commit 599d41fb8ea8bd2a99ca9525dd69405020e43dda upstream.
APR should not fail if the service device tree node does not have
the qcom,protection-domain property, since this functionality does
not exist on older platforms such as MSM8916 and MSM8996.
Ignore -EINVAL (returned when the property does not exist) to fix
a regression on 6.2-rc1 that prevents audio from working:
qcom,apr remoteproc0:smd-edge.apr_audio_svc.-1.-1:
Failed to read second value of qcom,protection-domain
qcom,apr remoteproc0:smd-edge.apr_audio_svc.-1.-1:
Failed to add apr 3 svc
Fixes: 6d7860f5750d ("soc: qcom: apr: Add check for idr_alloc and of_property_read_string_index")
Signed-off-by: Stephan Gerhold <stephan@gerhold.net>
Reviewed-by: Bjorn Andersson <andersson@kernel.org>
Signed-off-by: Bjorn Andersson <andersson@kernel.org>
Link: https://lore.kernel.org/r/20221229151648.19839-3-stephan@gerhold.net
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit 6d7860f5750d73da2fa1a1f6c9405058a593fa32 ]
As idr_alloc() and of_property_read_string_index() can return negative
numbers, it should be better to check the return value and deal with
the exception.
Therefore, it should be better to use goto statement to stop and return
error.
Fixes: 6adba21eb434 ("soc: qcom: Add APR bus driver")
Signed-off-by: Jiasheng Jiang <jiasheng@iscas.ac.cn>
Reviewed-by: Bjorn Andersson <andersson@kernel.org>
Signed-off-by: Bjorn Andersson <andersson@kernel.org>
Link: https://lore.kernel.org/r/20221107014403.3606-1-jiasheng@iscas.ac.cn
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit 99139b80c1b3d73026ed8be2de42c52e2976ab64 ]
APR and other packet routers like GPR are pretty much same and
interact with other drivers in similar way.
Ex: GPR ports can be considered as APR services, only difference
is they are allocated dynamically.
Other difference is packet layout, which should not matter
with the apis abstracted. Apart from this the rest of the
functionality is pretty much identical across APR and GPR.
Make the apr code more reusable by abstracting it service level,
rather than device level so that we do not need to write
new drivers for other new packet routers like GPR.
This patch is in preparation to add GPR support to this driver.
Signed-off-by: Srinivas Kandagatla <srinivas.kandagatla@linaro.org>
Reviewed-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Signed-off-by: Bjorn Andersson <bjorn.andersson@linaro.org>
Link: https://lore.kernel.org/r/20210927135559.738-4-srinivas.kandagatla@linaro.org
Stable-dep-of: 6d7860f5750d ("soc: qcom: apr: Add check for idr_alloc and of_property_read_string_index")
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit c882c899ead3545102a4d71b5fbe73b9e4bc2657 ]
The function platform_get_irq prints an error message into the kernel
log when the irq isn't found.
Since the interrupt is actually optional and not provided by some SoCs,
use platform_get_irq_optional which does not print an error message.
Fixes: c081f3060fab ("soc: qcom: Add support to register LLCC EDAC driver")
Signed-off-by: Luca Weiss <luca.weiss@fairphone.com>
Signed-off-by: Bjorn Andersson <andersson@kernel.org>
Link: https://lore.kernel.org/r/20221104153041.412020-1-luca.weiss@fairphone.com
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit 90681f53b9381c23ff7762a3b13826d620c272de ]
In qcom_smem_state_register() and qcom_smem_state_release(), we
should better use of_node_get() and of_node_put() for the reference
creation and destruction of 'device_node'.
Fixes: 9460ae2ff308 ("soc: qcom: Introduce common SMEM state machine code")
Signed-off-by: Liang He <windhl@126.com>
Signed-off-by: Bjorn Andersson <bjorn.andersson@linaro.org>
Link: https://lore.kernel.org/r/20220721135217.1301039-2-windhl@126.com
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit af8f6f39b8afd772fda4f8e61823ef8c021bf382 ]
There are two refcount leak bugs in qcom_smsm_probe():
(1) The 'local_node' is escaped out from for_each_child_of_node() as
the break of iteration, we should call of_node_put() for it in error
path or when it is not used anymore.
(2) The 'node' is escaped out from for_each_available_child_of_node()
as the 'goto', we should call of_node_put() for it in goto target.
Fixes: c97c4090ff72 ("soc: qcom: smsm: Add driver for Qualcomm SMSM")
Signed-off-by: Liang He <windhl@126.com>
Signed-off-by: Bjorn Andersson <bjorn.andersson@linaro.org>
Link: https://lore.kernel.org/r/20220721135217.1301039-1-windhl@126.com
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit a6232f2aa99ce470799992e99e0012945bb5308f ]
QCOM_RPMPD requires PM_GENERIC_DOMAINS/_OF, which in turns requires
CONFIG_PM. I forgot about the latter in my earlier patch (it's still
in -next as of the time of committing, hence no Fixes: tag). Fix it.
Reported-by: kernel test robot <lkp@intel.com>
Signed-off-by: Konrad Dybcio <konrad.dybcio@somainline.org>
Signed-off-by: Bjorn Andersson <bjorn.andersson@linaro.org>
Link: https://lore.kernel.org/r/20220707212158.32684-1-konrad.dybcio@somainline.org
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit e6e0951414a314e7db3e9e24fd924b3e15515288 ]
Every iteration of for_each_available_child_of_node() decrements
the reference count of the previous node.
When breaking early from a for_each_available_child_of_node() loop,
we need to explicitly call of_node_put() on the child node.
Add missing of_node_put() to avoid refcount leak.
Fixes: 05589b30b21a ("soc: qcom: Extend AOSS QMP driver to support resources that are used to wake up the SoC.")
Signed-off-by: Miaoqian Lin <linmq006@gmail.com>
Signed-off-by: Bjorn Andersson <bjorn.andersson@linaro.org>
Link: https://lore.kernel.org/r/20220606064252.42595-1-linmq006@gmail.com
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit 92a563fcf14b3093226fb36f12e9b5cf630c5a5d ]
of_parse_phandle() returns a node pointer with refcount
incremented, we should use of_node_put() on it when not need anymore.
Add missing of_node_put() to avoid refcount leak.
of_node_put() will check NULL pointer.
Fixes: 88c1e9404f1d ("soc: qcom: add OCMEM driver")
Signed-off-by: Miaoqian Lin <linmq006@gmail.com>
Reviewed-by: Brian Masney <masneyb@onstation.org>
Signed-off-by: Bjorn Andersson <bjorn.andersson@linaro.org>
Link: https://lore.kernel.org/r/20220602042430.1114-1-linmq006@gmail.com
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit 5334a3b12a7233b31788de60d61bfd890059d783 ]
The llcc-qcom driver can be compiled as a module, but lacks
MODULE_DEVICE_TABLE() and will therefore not be loaded automatically.
Fix this.
Fixes: a3134fb09e0b ("drivers: soc: Add LLCC driver")
Signed-off-by: Bjorn Andersson <bjorn.andersson@linaro.org>
Reviewed-by: Sai Prakash Ranjan <quic_saipraka@quicinc.com>
Reviewed-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org>
Link: https://lore.kernel.org/r/20220408213336.581661-3-bjorn.andersson@linaro.org
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit aad66a3c78da668f4506356c2fdb70b7a19ecc76 ]
The device_node pointer is returned by of_parse_phandle() with refcount
incremented. We should use of_node_put() on it when done.
Fixes: c97c4090ff72 ("soc: qcom: smsm: Add driver for Qualcomm SMSM")
Signed-off-by: Miaoqian Lin <linmq006@gmail.com>
Signed-off-by: Bjorn Andersson <bjorn.andersson@linaro.org>
Link: https://lore.kernel.org/r/20220308073648.24634-1-linmq006@gmail.com
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit 8fd3f18ea31a398ecce4a6d3804433658678b0a3 ]
The device_node pointer is returned by of_parse_phandle() with refcount
incremented. We should use of_node_put() on it when done.
Fixes: 50e99641413e ("soc: qcom: smp2p: Qualcomm Shared Memory Point to Point")
Signed-off-by: Miaoqian Lin <linmq006@gmail.com>
Signed-off-by: Bjorn Andersson <bjorn.andersson@linaro.org>
Link: https://lore.kernel.org/r/20220308071942.22942-1-linmq006@gmail.com
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit 8030cb9a55688c1339edd284d9d6ce5f9fc75160 ]
Quoting the header comments, IRQF_ONESHOT is "Used by threaded interrupts
which need to keep the irq line disabled until the threaded handler has
been run.". When applied to an interrupt that doesn't request a threaded
irq then IRQF_ONESHOT has a lesser known (undocumented?) side effect,
which it to disable the forced threading of the irq. For "normal" kernels
(without forced threading) then, if there is no thread_fn, then
IRQF_ONESHOT is a nop.
In this case disabling forced threading is not appropriate for this driver
because it calls wake_up_all() and this API cannot be called from
no-thread interrupt handlers on PREEMPT_RT systems (deadlock risk, triggers
sleeping-while-atomic warnings).
Fix this by removing IRQF_ONESHOT.
Fixes: 2209481409b7 ("soc: qcom: Add AOSS QMP driver")
Signed-off-by: Daniel Thompson <daniel.thompson@linaro.org>
[bjorn: Added Fixes tag]
Signed-off-by: Bjorn Andersson <bjorn.andersson@linaro.org>
Link: https://lore.kernel.org/r/20220127173554.158111-1-daniel.thompson@linaro.org
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit 0ff027027e05a866491bbb53494f0e2a61354c85 ]
The reference taken by 'of_find_device_by_node()' must be released when
not needed anymore.
Add the corresponding 'put_device()' in the error handling path.
Fixes: 01f937ffc468 ("soc: qcom: ocmem: don't return NULL in of_get_ocmem")
Signed-off-by: Miaoqian Lin <linmq006@gmail.com>
Signed-off-by: Bjorn Andersson <bjorn.andersson@linaro.org>
Link: https://lore.kernel.org/r/20220107073126.2335-1-linmq006@gmail.com
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit 5a811126d38f9767a20cc271b34db7c8efc5a46c ]
Because of the possible failure of the allocation, data->domains might
be NULL pointer and will cause the dereference of the NULL pointer
later.
Therefore, it might be better to check it and directly return -ENOMEM
without releasing data manually if fails, because the comment of the
devm_kmalloc() says "Memory allocated with this function is
automatically freed on driver detach.".
Fixes: bbe3a66c3f5a ("soc: qcom: rpmpd: Add a Power domain driver to model corners")
Signed-off-by: Jiasheng Jiang <jiasheng@iscas.ac.cn>
Signed-off-by: Bjorn Andersson <bjorn.andersson@linaro.org>
Link: https://lore.kernel.org/r/20211231094419.1941054-1-jiasheng@iscas.ac.cn
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit 92c550f9ffd2884bb5def52b5c0485a35e452784 ]
do_div() does a 64-by-32 division. Here the divisor is an unsigned long
which on some platforms is 64 bit wide. So use div64_ul instead of do_div
to avoid a possible truncation.
Reported-by: Zeal Robot <zealci@zte.com.cn>
Signed-off-by: Changcheng Deng <deng.changcheng@zte.com.cn>
Signed-off-by: Bjorn Andersson <bjorn.andersson@linaro.org>
Link: https://lore.kernel.org/r/20211125014311.45942-1-deng.changcheng@zte.com.cn
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit e3e56c050ab6e3f1bd811f0787f50709017543e4 ]
The general expectation is that powering on a power-domain should make
the power domain deliver some power, and if a specific performance state
is needed further requests has to be made.
But in contrast with other power-domain implementations (e.g. rpmpd) the
RPMh does not have an interface to enable the power, so the driver has
to vote for a particular corner (performance level) in rpmh_power_on().
But the corner is never initialized, so a typical request to simply
enable the power domain would not actually turn on the hardware. Further
more, when no more clients vote for a performance state (i.e. the
aggregated vote is 0) the power domain would be turned off.
Fix both of these issues by always voting for a corner with non-zero
value, when the power domain is enabled.
The tracking of the lowest non-zero corner is performed to handle the
corner case if there's ever a domain with a non-zero lowest corner, in
which case both rpmh_power_on() and rpmh_rpmhpd_set_performance_state()
would be allowed to use this lowest corner.
Fixes: 279b7e8a62cc ("soc: qcom: rpmhpd: Add RPMh power domain driver")
Signed-off-by: Bjorn Andersson <bjorn.andersson@linaro.org>
Reviewed-by: Stephen Boyd <swboyd@chromium.org>
Link: https://lore.kernel.org/r/20211005033732.2284447-1-bjorn.andersson@linaro.org
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit 5d16af6a921f5a4e7038671be5478cba4b7cfe81 ]
Fixes the following W=1 kernel build warning(s):
drivers/soc/qcom/rpmhpd.c:52: warning: Function parameter or member 'parent' not described in 'rpmhpd'
drivers/soc/qcom/rpmhpd.c:52: warning: Function parameter or member 'corner' not described in 'rpmhpd'
drivers/soc/qcom/rpmhpd.c:52: warning: Function parameter or member 'active_corner' not described in 'rpmhpd'
Cc: Andy Gross <agross@kernel.org>
Cc: Bjorn Andersson <bjorn.andersson@linaro.org>
Cc: linux-arm-msm@vger.kernel.org
Reviewed-by: Bjorn Andersson <bjorn.andersson@linaro.org>
Signed-off-by: Lee Jones <lee.jones@linaro.org>
Link: https://lore.kernel.org/r/20201103152838.1290217-22-lee.jones@linaro.org
Signed-off-by: Bjorn Andersson <bjorn.andersson@linaro.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit 833d51d7c66d6708abbc02398892b96b950167b9 ]
PT_LOAD type denotes that the segment should be loaded into the final
firmware memory region. Hash segment is not one such, because it's only
needed for PAS init and shouldn't be in the final firmware memory region.
That's why mdt_phdr_valid() explicitly reject non PT_LOAD segment and
hash segment. This actually makes the hash segment type check in
qcom_mdt_read_metadata() unnecessary and redundant. For a hash segment,
it won't be loaded into firmware memory region anyway, due to the
QCOM_MDT_TYPE_HASH check in mdt_phdr_valid(), even if it has a PT_LOAD
type for some reason (misusing or abusing?).
Some firmware files on Sony phones are such examples, e.g WCNSS firmware
of Sony Xperia M4 Aqua phone. The type of hash segment is just PT_LOAD.
Drop the unnecessary hash segment type check in qcom_mdt_read_metadata()
to fix firmware loading failure on these phones, while hash segment is
still kept away from the final firmware memory region.
Fixes: 498b98e93900 ("soc: qcom: mdt_loader: Support loading non-split images")
Signed-off-by: Shawn Guo <shawn.guo@linaro.org>
Reviewed-by: Marijn Suijten <marijn.suijten@somainline.org>
Signed-off-by: Bjorn Andersson <bjorn.andersson@linaro.org>
Link: https://lore.kernel.org/r/20210828070202.7033-1-shawn.guo@linaro.org
Signed-off-by: Sasha Levin <sashal@kernel.org>
commit a89f355e469dcda129c2522be4fdba00c1c74c83 upstream.
In "qmp_cooling_devices_register", the count value is initially
QMP_NUM_COOLING_RESOURCES, which is 2. Based on the initial count value,
the memory for cooling_devs is allocated. Then while calling the
"qmp_cooling_device_add" function, count value is post-incremented for
each child node.
This makes the out of bound access to the cooling_dev array. Fix it by
passing the QMP_NUM_COOLING_RESOURCES definition to devm_kzalloc() and
initializing the count to 0.
While at it, let's also free the memory allocated to cooling_dev if no
cooling device is found in DT and during unroll phase.
Cc: stable@vger.kernel.org # 5.4
Fixes: 05589b30b21a ("soc: qcom: Extend AOSS QMP driver to support resources that are used to wake up the SoC.")
Signed-off-by: Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org>
Link: https://lore.kernel.org/r/20210629153249.73428-1-manivannan.sadhasivam@linaro.org
Signed-off-by: Bjorn Andersson <bjorn.andersson@linaro.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit e3d4571955050736bbf3eda0a9538a09d9fcfce8 ]
The SMSM driver detects interrupt edges by tracking the last state
it has seen (and has triggered the interrupt handler for). This works
fine, but only if the interrupt does not change state while masked.
For example, if an interrupt is unmasked while the state is HIGH,
the stored last_value for that interrupt might still be LOW. Then,
when the remote processor triggers smsm_intr() we assume that nothing
has changed, even though the state might have changed from HIGH to LOW.
Attempt to fix this by checking the current remote state before
unmasking an IRQ. Use atomic operations to avoid the interrupt handler
from interfering with the unmask function.
This fixes modem crashes in some edge cases with the BAM-DMUX driver.
Specifically, the BAM-DMUX interrupt handler is not called for the
HIGH -> LOW smsm state transition if the BAM-DMUX driver is loaded
(and therefore unmasks the interrupt) after the modem was already started:
qcom-q6v5-mss 4080000.remoteproc: fatal error received: a2_task.c:3188:
Assert FALSE failed: A2 DL PER deadlock timer expired waiting for Apps ACK
Fixes: c97c4090ff72 ("soc: qcom: smsm: Add driver for Qualcomm SMSM")
Signed-off-by: Stephan Gerhold <stephan@gerhold.net>
Link: https://lore.kernel.org/r/20210712135703.324748-2-stephan@gerhold.net
Signed-off-by: Bjorn Andersson <bjorn.andersson@linaro.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit d43b3a989bc8c06fd4bbb69a7500d180db2d68e8 ]
rpmhpd_aggregate_corner() takes a corner as parameter, but in
rpmhpd_power_off() the code requests the level of the first corner
instead.
In all (known) current cases the first corner has level 0, so this
change should be a nop, but in case that there's a power domain with a
non-zero lowest level this makes sure that rpmhpd_power_off() actually
requests the lowest level - which is the closest to "power off" we can
get.
While touching the code, also skip the unnecessary zero-initialization
of "ret".
Fixes: 279b7e8a62cc ("soc: qcom: rpmhpd: Add RPMh power domain driver")
Reviewed-by: Rajendra Nayak <rnayak@codeaurora.org>
Reviewed-by: Stephen Boyd <swboyd@chromium.org>
Reviewed-by: Sibi Sankar <sibis@codeaurora.org>
Tested-by: Sibi Sankar <sibis@codeaurora.org>
Link: https://lore.kernel.org/r/20210703005416.2668319-2-bjorn.andersson@linaro.org
Signed-off-by: Bjorn Andersson <bjorn.andersson@linaro.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit 0648c55e3a21ccd816e99b6600d6199fbf39d23a ]
Given that no validation of how much data the firmware loader read in
for a given segment truncated segment files would best case result in a
hash verification failure, without any indication of what went wrong.
Improve this by validating that the firmware loader did return the
amount of data requested.
Fixes: 445c2410a449 ("soc: qcom: mdt_loader: Use request_firmware_into_buf()")
Reviewed-by: Sibi Sankar <sibis@codeaurora.org>
Link: https://lore.kernel.org/r/20210107232526.716989-1-bjorn.andersson@linaro.org
Signed-off-by: Bjorn Andersson <bjorn.andersson@linaro.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit 84168d1b54e76a1bcb5192991adde5176abe02e3 ]
The code validates that segments of p_memsz bytes of a segment will fit
in the provided memory region, but does not validate that p_filesz bytes
will, which means that an incorrectly crafted ELF header might write
beyond the provided memory region.
Fixes: 051fb70fd4ea ("remoteproc: qcom: Driver for the self-authenticating Hexagon v5")
Reviewed-by: Sibi Sankar <sibis@codeaurora.org>
Link: https://lore.kernel.org/r/20210107233119.717173-1-bjorn.andersson@linaro.org
Signed-off-by: Bjorn Andersson <bjorn.andersson@linaro.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit 0c9fdcdba68208270ae85d39600ea97da1718344 ]
Currently, GENI devices like i2c-qcom-geni fails to probe in ACPI boot,
if interconnect support is enabled. That's because interconnect driver
only supports DT right now. As interconnect is not necessarily required
for basic function of GENI devices, let's shield geni_icc_get() call,
and then all other ICC calls become nop due to NULL icc_path, so that
GENI devices keep working for ACPI boot.
Reviewed-by: Bjorn Andersson <bjorn.andersson@linaro.org>
Signed-off-by: Shawn Guo <shawn.guo@linaro.org>
Link: https://lore.kernel.org/r/20210114112928.11368-1-shawn.guo@linaro.org
Signed-off-by: Bjorn Andersson <bjorn.andersson@linaro.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
commit 29d96eb261345c8d888e248ae79484e681be2faa upstream.
This reverts commit 048eb908a1f2 ("soc: qcom-geni-se: Add interconnect
support to fix earlycon crash")
ICC core and platforms drivers supports sync_state feature, which
ensures that the default ICC BW votes from the bootloader is not
removed until all it's consumers are probes.
The proxy votes were needed in case other QUP child drivers
I2C, SPI probes before UART, they can turn off the QUP-CORE clock
which is shared resources for all QUP driver, this causes unclocked
access to HW from earlycon.
Given above support from ICC there is no longer need to maintain
proxy votes on QUP-CORE ICC node from QUP wrapper driver for early
console usecase, the default votes won't be removed until real
console is probed.
Cc: stable@vger.kernel.org
Fixes: 266cd33b5913 ("interconnect: qcom: Ensure that the floor bandwidth value is enforced")
Fixes: 7d3b0b0d8184 ("interconnect: qcom: Use icc_sync_state")
Signed-off-by: Roja Rani Yarubandi <rojay@codeaurora.org>
Signed-off-by: Akash Asthana <akashast@codeaurora.org>
Reviewed-by: Matthias Kaehlcke <mka@chromium.org>
Link: https://lore.kernel.org/r/20210324101836.25272-2-rojay@codeaurora.org
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit 01f937ffc4686837d6c43dea80c6ade6cbd2940a ]
If ocmem probe fails for whatever reason, of_get_ocmem returned NULL.
Without this, users must check for both NULL and IS_ERR on the returned
pointer - which didn't happen in drivers/gpu/drm/msm/adreno/adreno_gpu.c
leading to a NULL pointer dereference.
Reviewed-by: Brian Masney <masneyb@onstation.org>
Fixes: 88c1e9404f1d ("soc: qcom: add OCMEM driver")
Signed-off-by: Luca Weiss <luca@z3ntu.xyz>
Link: https://lore.kernel.org/r/20210130142349.53335-1-luca@z3ntu.xyz
Signed-off-by: Bjorn Andersson <bjorn.andersson@linaro.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit 5fb33d8960dc7abdabc6fe599a30c2c99b082ef6 ]
These need to be < ARRAY_SIZE() instead of <= ARRAY_SIZE() to prevent
accessing one element beyond the end of the array.
Acked-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org>
Reviewed-by: Douglas Anderson <dianders@chromium.org>
Reviewed-by: Stephen Boyd <swboyd@chromium.org>
Fixes: e9247e2ce577 ("soc: qcom: socinfo: fix printing of pmic_model")
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Link: https://lore.kernel.org/r/YAf+o85Z9lgkq3Nw@mwanda
Signed-off-by: Bjorn Andersson <bjorn.andersson@linaro.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
commit fc3e62e25c3896855b7c3d72df19ca6be3459c9f upstream.
smp2p_update_bits() should disable interrupts when it acquires its
spinlock. This is important because without the _irqsave, a priority
inversion can occur.
This function is called both with interrupts enabled in
qcom_q6v5_request_stop(), and with interrupts disabled in
ipa_smp2p_panic_notifier(). IRQ handling of spinlocks should be
consistent to avoid the panic notifier deadlocking because it's
sitting on the thread that's already got the lock via _request_stop().
Found via lockdep.
Cc: stable@vger.kernel.org
Fixes: 50e99641413e7 ("soc: qcom: smp2p: Qualcomm Shared Memory Point to Point")
Reviewed-by: Bjorn Andersson <bjorn.andersson@linaro.org>
Reviewed-by: Stephen Boyd <swboyd@chromium.org>
Signed-off-by: Evan Green <evgreen@chromium.org>
Link: https://lore.kernel.org/r/20200929133040.RESEND.1.Ideabf6dcdfc577cf39ce3d95b0e4aa1ac8b38f0c@changeid
Signed-off-by: Bjorn Andersson <bjorn.andersson@linaro.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit a161ffe4b877721d8917e18e70461d255a090f19 ]
clang static analysis reports this problem
pdr_interface.c:596:6: warning: Branch condition evaluates
to a garbage value
if (!req.service_path[0])
^~~~~~~~~~~~~~~~~~~~
This check that req.service_path was set in an earlier loop.
However req is a stack variable and its initial value
is undefined.
So initialize req to 0.
Fixes: fbe639b44a82 ("soc: qcom: Introduce Protection Domain Restart helpers")
Reviewed-by: Sibi Sankar <sibis@codeaurora.org>
Signed-off-by: Tom Rix <trix@redhat.com>
Link: https://lore.kernel.org/r/20200819184637.15648-1-trix@redhat.com
Signed-off-by: Bjorn Andersson <bjorn.andersson@linaro.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit 4b6ea87be44ef34732846fc71e44c41125f0c4fa ]
On geni-i2c transfers using DMA, it was seen that if you program the
command (I2C_READ) before calling geni_se_rx_dma_prep() that it could
cause interrupts to fire. If we get unlucky, these interrupts can
just keep firing (and not be handled) blocking further progress and
hanging the system.
In commit 02b9aec59243 ("i2c: i2c-qcom-geni: Fix DMA transfer race")
we avoided that by making sure we didn't program the command until
after geni_se_rx_dma_prep() was called. While that avoided the
problems, it also turns out to be invalid. At least in the TX case we
started seeing sporadic corrupted transfers. This is easily seen by
adding an msleep() between the DMA prep and the writing of the
command, which makes the problem worse. That means we need to revert
that commit and find another way to fix the bogus IRQs.
Specifically, after reverting commit 02b9aec59243 ("i2c:
i2c-qcom-geni: Fix DMA transfer race"), I put some traces in. I found
that the when the interrupts were firing like crazy:
- "m_stat" had bits for M_RX_IRQ_EN, M_RX_FIFO_WATERMARK_EN set.
- "dma" was set.
Further debugging showed that I could make the problem happen more
reliably by adding an "msleep(1)" any time after geni_se_setup_m_cmd()
ran up until geni_se_rx_dma_prep() programmed the length.
A rather simple fix is to change geni_se_select_dma_mode() so it's a
true inverse of geni_se_select_fifo_mode() and disables all the FIFO
related interrupts. Now the problematic interrupts can't fire and we
can program things in the correct order without worrying.
As part of this, let's also change the writel_relaxed() in the prepare
function to a writel() so that our DMA is guaranteed to be prepared
now that we can't rely on geni_se_setup_m_cmd()'s writel().
NOTE: the only current user of GENI_SE_DMA in mainline is i2c.
Fixes: 37692de5d523 ("i2c: i2c-qcom-geni: Add bus driver for the Qualcomm GENI I2C controller")
Fixes: 02b9aec59243 ("i2c: i2c-qcom-geni: Fix DMA transfer race")
Signed-off-by: Douglas Anderson <dianders@chromium.org>
Reviewed-by: Stephen Boyd <swboyd@chromium.org>
Reviewed-by: Akash Asthana <akashast@codeaurora.org>
Tested-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org>
Link: https://lore.kernel.org/r/20201013142448.v2.1.Ifdb1b69fa3367b81118e16e9e4e63299980ca798@changeid
Signed-off-by: Bjorn Andersson <bjorn.andersson@linaro.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
Various driver updates for platforms. A bulk of this is smaller fixes or
cleanups, but some of the new material this time around is:
- Support for Nvidia Tegra234 SoC
- Ring accelerator support for TI AM65x
- PRUSS driver for TI platforms
- Renesas support for R-Car V3U SoC
- Reset support for Cortex-M4 processor on i.MX8MQ
There are also new socinfo entries for a handful of different SoCs
and platforms.
-----BEGIN PGP SIGNATURE-----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=tfPb
-----END PGP SIGNATURE-----
Merge tag 'armsoc-drivers' of git://git.kernel.org/pub/scm/linux/kernel/git/soc/soc
Pull ARM SoC-related driver updates from Olof Johansson:
"Various driver updates for platforms. A bulk of this is smaller fixes
or cleanups, but some of the new material this time around is:
- Support for Nvidia Tegra234 SoC
- Ring accelerator support for TI AM65x
- PRUSS driver for TI platforms
- Renesas support for R-Car V3U SoC
- Reset support for Cortex-M4 processor on i.MX8MQ
There are also new socinfo entries for a handful of different SoCs and
platforms"
* tag 'armsoc-drivers' of git://git.kernel.org/pub/scm/linux/kernel/git/soc/soc: (131 commits)
drm/mediatek: reduce clear event
soc: mediatek: cmdq: add clear option in cmdq_pkt_wfe api
soc: mediatek: cmdq: add jump function
soc: mediatek: cmdq: add write_s_mask value function
soc: mediatek: cmdq: add write_s value function
soc: mediatek: cmdq: add read_s function
soc: mediatek: cmdq: add write_s_mask function
soc: mediatek: cmdq: add write_s function
soc: mediatek: cmdq: add address shift in jump
soc: mediatek: mtk-infracfg: Fix kerneldoc
soc: amlogic: pm-domains: use always-on flag
reset: sti: reset-syscfg: fix struct description warnings
reset: imx7: add the cm4 reset for i.MX8MQ
dt-bindings: reset: imx8mq: add m4 reset
reset: Fix and extend kerneldoc
reset: reset-zynqmp: Added support for Versal platform
dt-bindings: reset: Updated binding for Versal reset driver
reset: imx7: Support module build
soc: fsl: qe: Remove unnessesary check in ucc_set_tdm_rxtx_clk
soc: fsl: qman: convert to use be32_add_cpu()
...
I had queued up a batch of fixes that got a bit close to the release for
sending in before the merge window opened, so I'm including them in the
batch of pull requests instead. They're mostly smaller DT tweaks and
fixes, the usual mix that we tend to have through the releases.
-----BEGIN PGP SIGNATURE-----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=tvbv
-----END PGP SIGNATURE-----
Merge tag 'armsoc-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/soc/soc
Pull ARM SoC fixes from Olof Johansson:
"I had queued up a batch of fixes that got a bit close to the release
for sending in before the merge window opened, so I'm including them
in the merge window batch instead.
Mostly smaller DT tweaks and fixes, the usual mix that we tend to have
through the releases"
* tag 'armsoc-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/soc/soc:
ARM: dts: iwg20d-q7-common: Fix touch controller probe failure
ARM: OMAP2+: Restore MPU power domain if cpu_cluster_pm_enter() fails
ARM: dts: am33xx: modify AM33XX_IOPAD for #pinctrl-cells = 2
soc: actions: include header to fix missing prototype
arm64: dts: ti: k3-j721e: Rename mux header and update macro names
soc: qcom: pdr: Fixup array type of get_domain_list_resp message
arm64: dts: qcom: pm660: Fix missing pound sign in interrupt-cells
arm64: dts: qcom: kitakami: Temporarily disable SDHCI1
arm64: dts: sdm630: Temporarily disable SMMUs by default
arm64: dts: sdm845: Fixup OPP table for all qup devices
arm64: dts: allwinner: h5: remove Mali GPU PMU module
ARM: dts: sun8i: r40: bananapi-m2-ultra: Fix dcdc1 regulator
soc: xilinx: Fix error code in zynqmp_pm_probe()
The avs drivers are all SoC specific drivers that doesn't share any code.
Instead they are located in a directory, mostly to keep similar
functionality together. From a maintenance point of view, it makes better
sense to collect SoC specific drivers like these, into the SoC specific
directories.
Therefore, let's move the qcom-cpr driver to the qcom directory.
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
Acked-by: Bjorn Andersson <bjorn.andersson@linaro.org>
Acked-by: Niklas Cassel <nks@flawful.org>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Fix the array type of the domain_list QMI response in PDR.
-----BEGIN PGP SIGNATURE-----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=RBzc
-----END PGP SIGNATURE-----
Merge tag 'qcom-drivers-fixes-for-5.9' of https://git.kernel.org/pub/scm/linux/kernel/git/qcom/linux into arm/fixes
Qualcomm driver fixes for v5.9
Fix the array type of the domain_list QMI response in PDR.
* tag 'qcom-drivers-fixes-for-5.9' of https://git.kernel.org/pub/scm/linux/kernel/git/qcom/linux:
soc: qcom: pdr: Fixup array type of get_domain_list_resp message
Link: https://lore.kernel.org/r/20200921235241.36463-1-bjorn.andersson@linaro.org
Signed-off-by: Olof Johansson <olof@lixom.net>
APR client incorrectly prints out "ret" variable on pdr_add_lookup failure,
it should be printing the error value returned by the lookup instead.
Fixes: 8347356626028 ("soc: qcom: apr: Add avs/audio tracking functionality")
Signed-off-by: Sibi Sankar <sibis@codeaurora.org>
Link: https://lore.kernel.org/r/20200915154232.27523-1-sibis@codeaurora.org
Signed-off-by: Bjorn Andersson <bjorn.andersson@linaro.org>
The array type of get_domain_list_resp is incorrectly marked as NO_ARRAY.
Due to which the following error was observed when using pdr helpers with
the downstream proprietary pd-mapper. Fix this up by marking it as
VAR_LEN_ARRAY instead.
Err logs:
qmi_decode_struct_elem: Fault in decoding: dl(2), db(27), tl(160), i(1), el(1)
failed to decode incoming message
PDR: tms/servreg get domain list txn wait failed: -14
PDR: service lookup for tms/servreg failed: -14
Tested-by: Rishabh Bhatnagar <rishabhb@codeaurora.org>
Fixes: fbe639b44a82 ("soc: qcom: Introduce Protection Domain Restart helpers")
Reported-by: Rishabh Bhatnagar <rishabhb@codeaurora.org>
Signed-off-by: Sibi Sankar <sibis@codeaurora.org>
Link: https://lore.kernel.org/r/20200914145807.1224-1-sibis@codeaurora.org
Signed-off-by: Bjorn Andersson <bjorn.andersson@linaro.org>
The busy loop in rpmh_rsc_send_data() is written with the assumption
that the udelay will be preempted by the tcs_tx_done() irq handler when
the TCS slots are all full. This doesn't hold true when the calling
thread is an irqthread and the tcs_tx_done() irq is also an irqthread.
That's because kernel irqthreads are SCHED_FIFO and thus need to
voluntarily give up priority by calling into the scheduler so that other
threads can run.
I see RCU stalls when I boot with irqthreads on the kernel commandline
because the modem remoteproc driver is trying to send an rpmh async
message from an irqthread that needs to give up the CPU for the rpmh
irqthread to run and clear out tcs slots.
rcu: INFO: rcu_preempt self-detected stall on CPU
rcu: 0-....: (1 GPs behind) idle=402/1/0x4000000000000002 softirq=2108/2109 fqs=4920
(t=21016 jiffies g=2933 q=590)
Task dump for CPU 0:
irq/11-smp2p R running task 0 148 2 0x00000028
Call trace:
dump_backtrace+0x0/0x154
show_stack+0x20/0x2c
sched_show_task+0xfc/0x108
dump_cpu_task+0x44/0x50
rcu_dump_cpu_stacks+0xa4/0xf8
rcu_sched_clock_irq+0x7dc/0xaa8
update_process_times+0x30/0x54
tick_sched_handle+0x50/0x64
tick_sched_timer+0x4c/0x8c
__hrtimer_run_queues+0x21c/0x36c
hrtimer_interrupt+0xf0/0x22c
arch_timer_handler_phys+0x40/0x50
handle_percpu_devid_irq+0x114/0x25c
__handle_domain_irq+0x84/0xc4
gic_handle_irq+0xd0/0x178
el1_irq+0xbc/0x180
save_return_addr+0x18/0x28
return_address+0x54/0x88
preempt_count_sub+0x40/0x88
_raw_spin_unlock_irqrestore+0x4c/0x6c
___ratelimit+0xd0/0x128
rpmh_rsc_send_data+0x24c/0x378
__rpmh_write+0x1b0/0x208
rpmh_write_async+0x90/0xbc
rpmhpd_send_corner+0x60/0x8c
rpmhpd_aggregate_corner+0x8c/0x124
rpmhpd_set_performance_state+0x8c/0xbc
_genpd_set_performance_state+0xdc/0x1b8
dev_pm_genpd_set_performance_state+0xb8/0xf8
q6v5_pds_disable+0x34/0x60 [qcom_q6v5_mss]
qcom_msa_handover+0x38/0x44 [qcom_q6v5_mss]
q6v5_handover_interrupt+0x24/0x3c [qcom_q6v5]
handle_nested_irq+0xd0/0x138
qcom_smp2p_intr+0x188/0x200
irq_thread_fn+0x2c/0x70
irq_thread+0xfc/0x14c
kthread+0x11c/0x12c
ret_from_fork+0x10/0x18
This busy loop naturally lends itself to using a wait queue so that each
thread that tries to send a message will sleep waiting on the waitqueue
and only be woken up when a free slot is available. This should make
things more predictable too because the scheduler will be able to sleep
tasks that are waiting on a free tcs instead of the busy loop we
currently have today.
Reviewed-by: Maulik Shah <mkshah@codeaurora.org>
Reviewed-by: Douglas Anderson <dianders@chromium.org>
Tested-by: Stanimir Varbanov <stanimir.varbanov@linaro.org>
Cc: Douglas Anderson <dianders@chromium.org>
Cc: Maulik Shah <mkshah@codeaurora.org>
Cc: Lina Iyer <ilina@codeaurora.org>
Signed-off-by: Stephen Boyd <swboyd@chromium.org>
Link: https://lore.kernel.org/r/20200724211711.810009-1-sboyd@kernel.org
Signed-off-by: Bjorn Andersson <bjorn.andersson@linaro.org>
