Now that all of the mmsys routing tables have been fixed,
migrate all of them to use the MMSYS_ROUTE() macro: this
will make sure that future additions to any of the tables
for the currently supported SoCs are compile-time sanity
checked, greatly reducing room for (way too common) mistakes.

Signed-off-by: AngeloGioacchino Del Regno <angelogioacchino.delregno@collabora.com>

The mmsys driver reads the routing table and writes to the
hardware `val & mask`, but multiple entries in the mmsys
routing table for the MT8365 SoC are setting a 0x0 mask:
this effectively writes .. nothing .. to the hardware.

That would never work, and if the display controller was
actually working with the mmsys doing no routing at all,
that was only because the bootloader was correctly setting
the display controller routing registers before booting the
kernel, and the mmsys was never reset.

Make this table to actually set the routing by adding the
correct register masks to it.

While at it, also change MOUT val definitions to BIT(x), as
the MOUT registers are effectively checking for each bit to
enable output to the corresponding HW.
Please note that, for this SoC, only the MOUT registers are
checking bits (as those can enable multiple outputs), while
the others are purely reading a number to select an input.

Fixes: bc3fc5c0 ("soc: mediatek: mmsys: add MT8365 support")
Signed-off-by: AngeloGioacchino Del Regno <angelogioacchino.delregno@collabora.com>

The mmsys routing table for this SoC was effectively missing
initialization of the val variable of struct mtk_mmsys_routes:
this means that `val` was incorrectly initialized to zero,
hence the registers were wrongly initialized.

Add the required regval to all of the entries of the routing
table for this SoC to fix display controller functionality.

Fixes: 060f7875 ("soc: mediatek: mmsys: Add support for MT8167 SoC")
Signed-off-by: AngeloGioacchino Del Regno <angelogioacchino.delregno@collabora.com>

Add routing paths to support Display Stream Compression on the
VDOSYS0 pipelines ending with DSI or DisplayPort (DP_INTF).

Signed-off-by: AngeloGioacchino Del Regno <angelogioacchino.delregno@collabora.com>

MT8188 uses DPI1 to output to the HDMI controller: add the
Start of Frame and End of Frame configuration for the DPI1
IP to the tables to unblock generation and sending of these
signals to the GCE.

Signed-off-by: AngeloGioacchino Del Regno <angelogioacchino.delregno@collabora.com>

Change the initialization data in the arrays of structure
mtk_mmsys_routes to make use of the MMSYS_ROUTE() macro:
this will make sure that each array entry's SEL value fits
in its corresponding register mask with a compile time check.

Signed-off-by: AngeloGioacchino Del Regno <angelogioacchino.delregno@collabora.com>

Every MediaTek SoC with multimedia capabilities has an array of
structure mtk_mmsys_routes that defines a multimedia connection
between hardware components.

This connection is activated by writing a (masked) value in each
specific register, and the association between from->to path and
value to write is expressed as an entry in that array.

Failing to set the right path does not give any meaningful error
and makes things to simply not work as the data will either not
be retrieved from the right input port, or will be written to
the wrong output port (or both): since a misconfiguration may
effectively still be a possibly correct configuration at the HW
level, this may be only giving side effects in terms of simply
getting no functionality but, again, no errors.

In order to reduce room for mistakes in declarations of the
mmsys routes, add a macro that compile-time checks that the
provided value does at least fit in the register mask.

Signed-off-by: AngeloGioacchino Del Regno <angelogioacchino.delregno@collabora.com>

The VDO1_MERGE4 hardware (merge5 software component) should be
set to enable output to DPI1_SEL by setting BIT(2) but, despite
the intention being exactly that, this won't work because the
declared register mask is wrong as it is set as GENMASK(1, 0).

Register MERGE4_MOUT_EN in VDO1 has four used bits [3, 0] so
fix the mask to reflect that.
That, in turn, allows the mmsys driver to actually set BIT(2)
in this register, fixing the MERGE4 output to DPI1 selection.

Fixes: c0349314 ("soc: mediatek: Support MT8188 VDOSYS1 in mtk-mmsys")
Signed-off-by: AngeloGioacchino Del Regno <angelogioacchino.delregno@collabora.com>

Commit e9f826b0 ("soc: qcom: pmic_glink: simplify locking with
guard()") was on top of a fix [1] which was moving the 'pg =
__pmic_glink' assignment into the critical section.  Unfortunately the
actual fix was not applied and instead rebased version of the next patch
got in.  The resulting code is in general correct, but now there is a
duplicated assignment 'pg = __pmic_glink'.

[1] https://lore.kernel.org/all/20240822164815.230167-1-krzysztof.kozlowski@linaro.org/



Fixes: e9f826b0 ("soc: qcom: pmic_glink: simplify locking with guard()")
Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
Link: https://lore.kernel.org/r/20250111172326.101779-1-krzysztof.kozlowski@linaro.org


Signed-off-by: Bjorn Andersson <andersson@kernel.org>

Various Nvidia userspace applications and tests access following fuse
via Fuse nvmem interface:

	* odmid
	* odminfo
	* boot_security_info
	* public_key_hash
	* reserved_odm0
	* reserved_odm1
	* reserved_odm2
	* reserved_odm3
	* reserved_odm4
	* reserved_odm5
	* reserved_odm6
	* reserved_odm7
	* odm_lock
	* pk_h1
	* pk_h2
	* revoke_pk_h0
	* revoke_pk_h1
	* security_mode
	* system_fw_field_ratchet0
	* system_fw_field_ratchet1
	* system_fw_field_ratchet2
	* system_fw_field_ratchet3
	* optin_enable

Update tegra234_fuse_keepouts list to allow reading these fuse from
nvmem sysfs interface.

Signed-off-by: Kartik Rajput <kkartik@nvidia.com>
Link: https://lore.kernel.org/r/20241127061053.16775-1-kkartik@nvidia.com


Signed-off-by: Thierry Reding <treding@nvidia.com>

Fix spelling error in tegra234_lookup_slave_timeout().

Signed-off-by: liujing <liujing@cmss.chinamobile.com>
Link: https://lore.kernel.org/r/20241209055148.3749-1-liujing@cmss.chinamobile.com


Signed-off-by: Thierry Reding <treding@nvidia.com>

Kernel coding style expects all drivers to ignore debugfs errors.
Partially because it is purely for debugging, not for important user
interfaces.  Simplify the code by dropping unnecessary probe failuring
and error message on debugfs failures, which also fixes incorrect usage
IS_ERR_OR_NULL() and Smatch warning:

  drivers/soc/tegra/cbb/tegra-cbb.c:80 tegra_cbb_err_debugfs_init() warn: passing zero to 'PTR_ERR'

Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
Link: https://lore.kernel.org/r/20250104141958.115911-1-krzysztof.kozlowski@linaro.org


Signed-off-by: Thierry Reding <treding@nvidia.com>

The Qualcomm SM7225 is practically identical to SM6350, so add an entry
using that data.

Signed-off-by: Luca Weiss <luca.weiss@fairphone.com>
Link: https://lore.kernel.org/r/20241220-pd-mapper-sm7225-v1-1-68f85a87288d@fairphone.com


Signed-off-by: Bjorn Andersson <andersson@kernel.org>

The 'broadcast' register space is present only in chipsets that
have multiple instances of LLCC IP. Since IPQ5424 has only one
instance, both the LLCC and LLCC_BROADCAST points to the same
register space.

Signed-off-by: Varadarajan Narayanan <quic_varada@quicinc.com>
Reviewed-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org>
Reviewed-by: Konrad Dybcio <konrad.dybcio@oss.qualcomm.com>
Link: https://lore.kernel.org/r/20241121051935.1055222-3-quic_varada@quicinc.com


Signed-off-by: Bjorn Andersson <andersson@kernel.org>

Driver removal should fully clean up - unmap the memory.

Fixes: 0890beb2 ("soc: mediatek: add mt6779 devapc driver")
Cc: stable@vger.kernel.org
Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
Link: https://lore.kernel.org/r/20250104142012.115974-2-krzysztof.kozlowski@linaro.org


Signed-off-by: AngeloGioacchino Del Regno <angelogioacchino.delregno@collabora.com>

Error paths of mtk_devapc_probe() should unmap the memory.  Reported by
Smatch:

  drivers/soc/mediatek/mtk-devapc.c:292 mtk_devapc_probe() warn: 'ctx->infra_base' from of_iomap() not released on lines: 277,281,286.

Fixes: 0890beb2 ("soc: mediatek: add mt6779 devapc driver")
Cc: stable@vger.kernel.org
Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
Link: https://lore.kernel.org/r/20250104142012.115974-1-krzysztof.kozlowski@linaro.org


Signed-off-by: AngeloGioacchino Del Regno <angelogioacchino.delregno@collabora.com>

Update soc_id table for the Qualcomm QCS9075 SoC
to represent qcs9075 machine.

Reviewed-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org>
Signed-off-by: Wasim Nazir <quic_wasimn@quicinc.com>
Link: https://lore.kernel.org/r/20241229152332.3068172-3-quic_wasimn@quicinc.com


Signed-off-by: Bjorn Andersson <andersson@kernel.org>

On MSM8916 devices, the serial number exposed in sysfs is constant and does
not change across individual devices. It's always:

  db410c:/sys/devices/soc0$ cat serial_number
  2644893864

The firmware used on MSM8916 exposes SOCINFO_VERSION(0, 8), which does not
have support for the serial_num field in the socinfo struct. There is an
existing check to avoid exposing the serial number in that case, but it's
not correct: When checking the item_size returned by SMEM, we need to make
sure the *end* of the serial_num is within bounds, instead of comparing
with the *start* offset. The serial_number currently exposed on MSM8916
devices is just an out of bounds read of whatever comes after the socinfo
struct in SMEM.

Fix this by changing offsetof() to offsetofend(), so that the size of the
field is also taken into account.

Cc: stable@vger.kernel.org
Fixes: efb448d0 ("soc: qcom: Add socinfo driver")
Signed-off-by: Stephan Gerhold <stephan.gerhold@linaro.org>
Reviewed-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org>
Link: https://lore.kernel.org/r/20241230-qcom-socinfo-serialno-oob-v1-1-9b7a890da3da@linaro.org


Signed-off-by: Bjorn Andersson <andersson@kernel.org>

If tensor_set_bits_atomic() is called with a mask of 0 the function will
just iterate over its bit, not perform any updates and return stack
value of 'ret'.

Also reported by smatch:

  drivers/soc/samsung/exynos-pmu.c:129 tensor_set_bits_atomic() error: uninitialized symbol 'ret'.

Fixes: 0b7c6075 ("soc: samsung: exynos-pmu: Add regmap support for SoCs that protect PMU regs")
Cc: stable@vger.kernel.org
Link: https://lore.kernel.org/r/20250104135605.109209-1-krzysztof.kozlowski@linaro.org


Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>

Use the proper API instead of open coding it.

However it looks like kthreads here could be replaced by the use of a
per-cpu workqueue instead.

Signed-off-by: Frederic Weisbecker <frederic@kernel.org>

The rmtfs class object is never modified and can be made constant.

Reviewed-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org>
Signed-off-by: Bartosz Golaszewski <bartosz.golaszewski@linaro.org>
Link: https://lore.kernel.org/r/20241202094903.18388-2-brgl@bgdev.pl


Signed-off-by: Bjorn Andersson <andersson@kernel.org>

Make it possible to build the module when COMPILE_TEST is enabled for
better build coverage.

Reviewed-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org>
Signed-off-by: Bartosz Golaszewski <bartosz.golaszewski@linaro.org>
Link: https://lore.kernel.org/r/20241202094903.18388-1-brgl@bgdev.pl


Signed-off-by: Bjorn Andersson <andersson@kernel.org>

Simplify error handling (less gotos) over locks with guard().

Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
Link: https://lore.kernel.org/r/20240822164853.231087-1-krzysztof.kozlowski@linaro.org


Signed-off-by: Bjorn Andersson <andersson@kernel.org>

If of_parse_phandle_with_args() succeeds, the OF node reference should
be dropped, regardless of number of phandle arguments.

Cc: stable@vger.kernel.org
Fixes: 9460ae2f ("soc: qcom: Introduce common SMEM state machine code")
Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
Reviewed-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org>
Link: https://lore.kernel.org/r/20240822164853.231087-2-krzysztof.kozlowski@linaro.org


Signed-off-by: Bjorn Andersson <andersson@kernel.org>

The Last Level Cache is split into many slices, each one of which can
be toggled on or off.

Only certain slices are recommended to be turned on unconditionally,
in order to reach optimal performance/latency/power levels.

Enable WRCACHE on X1 at boot, in accordance with internal
recommendations.

No significant performance difference is expected.

Fixes: b3cf69a4 ("soc: qcom: llcc: Add configuration data for X1E80100")
Cc: stable@vger.kernel.org
Reviewed-by: Rajendra Nayak <quic_rjendra@quicinc.com>
Signed-off-by: Konrad Dybcio <konrad.dybcio@oss.qualcomm.com>
Reviewed-by: Johan Hovold <johan+linaro@kernel.org>
Tested-by: Johan Hovold <johan+linaro@kernel.org>
Link: https://lore.kernel.org/r/20241219-topic-llcc_x1e_wrcache-v3-1-b9848d9c3d63@oss.qualcomm.com


Signed-off-by: Bjorn Andersson <andersson@kernel.org>

X1P42100 is a cousin of X1E80100, and hence can make use of the
latter's configuration. Do so.

Signed-off-by: Konrad Dybcio <konrad.dybcio@oss.qualcomm.com>
Reviewed-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org>
Link: https://lore.kernel.org/r/20241221-topic-x1p4_soc-v1-3-55347831d73c@oss.qualcomm.com


Signed-off-by: Bjorn Andersson <andersson@kernel.org>

The helper devm_clk_bulk_get_all_enable() missed to return the number of
clocks stored in the clk_bulk_data table referenced by the clks
argument and, therefore, will be dropped.

Use the newly introduced devm_clk_bulk_get_all_enabled() variant
instead, which is consistent with devm_clk_bulk_get_all() in terms of
the returned value:

 > 0 if one or more clocks have been stored
 = 0 if there are no clocks
 < 0 if an error occurred

Reviewed-by: AngeloGioacchino Del Regno <angelogioacchino.delregno@collabora.com>
Reviewed-by: Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org>
Signed-off-by: Cristian Ciocaltea <cristian.ciocaltea@collabora.com>
Link: https://lore.kernel.org/r/20241217-clk_bulk_ena_fix-v5-1-aafbbb245155@collabora.com


Acked-by: AngeloGioacchino Del Regno <angelogioacchino.delregno@collabora.com>
Signed-off-by: Stephen Boyd <sboyd@kernel.org>

In order to have fine-grained control, use cmdq_pkt_eoc() and
cmdq_pkt_jump_rel() to replace cmdq_pkt_finalize().

Signed-off-by: Chun-Kuang Hu <chunkuang.hu@kernel.org>
Acked-by: Matthias Brugger <matthias.bgg@gmail.com>
Reviewed-by: AngeloGioacchino Del Regno <angelogioacchino.delregno@collabora.com>
Signed-off-by: Sebastian Fricke <sebastian.fricke@collabora.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

Add a configuration option for the RZ/G3E SoC.

Signed-off-by: Biju Das <biju.das.jz@bp.renesas.com>
Reviewed-by: Geert Uytterhoeven <geert+renesas@glider.be>
Link: https://lore.kernel.org/20241203105005.103927-6-biju.das.jz@bp.renesas.com


Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be>

Simplify error handling over locks with guard().  In few places this
elimiates error gotos and local variables.  Switch to guard() everywhere
(except when jumps would go over scoped guard) for consistency, even if
it does not bring benefit in such places.

Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
Link: https://lore.kernel.org/r/20240822164815.230167-2-krzysztof.kozlowski@linaro.org


[bjorn: Rebased ontop of v6.13-rc1]
Signed-off-by: Bjorn Andersson <andersson@kernel.org>

i.MX9 SoCs have SoC ID, SoC revision number and chip unique identifier
which are provided by the corresponding ARM trusted firmware API. This
patch intends to use SMC call to obtain these information and then
register i.MX9 SoC as a device.

Signed-off-by: Alice Guo <alice.guo@nxp.com>
Tested-by: Alexander Stein <alexander.stein@ew.tq-group.com>
Reviewed-by: Stefan Wahren <wahrenst@gmx.net>
Signed-off-by: Shawn Guo <shawnguo@kernel.org>

A device_node acquired via of_find_node_by_path() requires explicit
calls to of_node_put() when it is no longer needed to avoid leaking the
resource.

Instead of adding the missing calls to of_node_put() in all execution
paths, use the cleanup attribute for 'np' by means of the __free()
macro, which automatically calls of_node_put() when the variable goes
out of scope.

Fixes: 960ddf70 ("drivers: soc: atmel: Avoid calling at91_soc_init on non AT91 SoCs")
Signed-off-by: Javier Carrasco <javier.carrasco.cruz@gmail.com>
Link: https://lore.kernel.org/r/20241031-soc-atmel-soc-cleanup-v2-1-73f2d235fd98@gmail.com


Signed-off-by: Claudiu Beznea <claudiu.beznea@tuxon.dev>

A kernel test robot detected a missing error code:
   qmc.c:1942 qmc_probe() warn: missing error code 'ret'

Indeed, the error returned by platform_get_irq() is checked and the
operation is aborted in case of failure but the ret error code is
not set in that case.

Set the ret error code.

Reported-by: kernel test robot <lkp@intel.com>
Reported-by: Dan Carpenter <dan.carpenter@linaro.org>
Closes: https://lore.kernel.org/r/202411051350.KNy6ZIWA-lkp@intel.com/


Fixes: 3178d58e ("soc: fsl: cpm1: Add support for QMC")
Cc: stable@vger.kernel.org
Signed-off-by: Herve Codina <herve.codina@bootlin.com>
Link: https://lore.kernel.org/r/20241105145623.401528-1-herve.codina@bootlin.com


Signed-off-by: Christophe Leroy <christophe.leroy@csgroup.eu>

of_find_compatible_node() requires a call to of_node_put() when the
pointer to the node is not required anymore to decrement its refcount
and avoid leaking memory.

Add the missing call to of_node_put() after the node has been used.

Cc: stable@vger.kernel.org
Fixes: e95f287d ("soc: fsl: handle RCPM errata A-008646 on SoC LS1021A")
Signed-off-by: Javier Carrasco <javier.carrasco.cruz@gmail.com>
Link: https://lore.kernel.org/r/20241013-rcpm-of_node_put-v1-1-9a8e55a01eae@gmail.com


Signed-off-by: Christophe Leroy <christophe.leroy@csgroup.eu>

The non-scoped variant of this macro turns error-prone as soon as error
paths are included, because explicit calls to of_node_put() are required
to avoid leaking memory.

Using its scoped counterpart simplifies the code by removing the need of
explicit calls to of_node_put(), as they are automatically triggered as
soon as the child node goes out of scope. Moreover, it is more robust as
it accounts for new error paths without having to worry about
decrementing the object's refcount.

Note that the device_node is declared within the macro, and its explicit
declaration can be dropped as well if it is not used anywhere else.

Signed-off-by: Javier Carrasco <javier.carrasco.cruz@gmail.com>
Acked-by: Herve Codina <herve.codina@bootlin.com>
Link: https://lore.kernel.org/r/20241002-tsa-scoped-v1-1-ba6a6d657f82@gmail.com


Signed-off-by: Christophe Leroy <christophe.leroy@csgroup.eu>

The struct device_link *link field in struct qcom_ice is only used to
store the result of a device_link_add call with the
DL_FLAG_AUTOREMOVE_SUPPLIER flag. With this flag, the resulting value
can only be used to check whether the link is present or not, as per the
device_link_add description, hence this commit removes the field.

Signed-off-by: Joe Hattori <joe@pf.is.s.u-tokyo.ac.jp>
Link: https://lore.kernel.org/r/20241030025046.303342-1-joe@pf.is.s.u-tokyo.ac.jp


Signed-off-by: Bjorn Andersson <andersson@kernel.org>

Add SoC ID for Qualcomm IPQ5424/IPQ5404.

Signed-off-by: Manikanta Mylavarapu <quic_mmanikan@quicinc.com>
Reviewed-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
Link: https://lore.kernel.org/r/20241016151528.2893599-3-quic_mmanikan@quicinc.com


Signed-off-by: Bjorn Andersson <andersson@kernel.org>

Commit c14e64b4 ("soc: qcom: llcc: Support chipsets that can write
to llcc") made the code not configure certain registers on SDM845 due
to firmware security policies. That turned out only to concern SDM845,
but the condition was chosen such that all other entries (for SoCs that
didnot need it) were required to set .need_llcc_cfg = true.

Flip the condition, so the default is "doesn't need the workaround".

Signed-off-by: Konrad Dybcio <konrad.dybcio@oss.qualcomm.com>
Link: https://lore.kernel.org/r/20241104-topic-llcc_flip-v1-1-3003c846d131@oss.qualcomm.com


[bjorn: Dropped a few newly added need_llcc_cfg uses]
Signed-off-by: Bjorn Andersson <andersson@kernel.org>

Add LLCC configuration for the QCS8300 platform. There is an errata on
LB_CNT information on QCS8300 platform, hardcode num_banks to get the
correct value.

Signed-off-by: Jingyi Wang <quic_jingyw@quicinc.com>
Reviewed-by: Konrad Dybcio <konrad.dybcio@oss.qualcomm.com>
Link: https://lore.kernel.org/r/20241031-qcs8300_llcc-v3-2-bb56952cb83b@quicinc.com


Signed-off-by: Bjorn Andersson <andersson@kernel.org>

Add LLCC configuration support for the QCS615 platform.

Signed-off-by: Song Xue <quic_songxue@quicinc.com>
Link: https://lore.kernel.org/r/20241010-add_llcc_support_for_qcs615-v2-2-044432450a75@quicinc.com


Signed-off-by: Bjorn Andersson <andersson@kernel.org>