- 13 Mar, 2023 36 commits
-
-
AngeloGioacchino Del Regno authored
Convert the MT8516 clock drivers to be platform drivers and use the common probe mechanism. Thanks to the conversion, more error handling was added to the clocks registration. Signed-off-by:
AngeloGioacchino Del Regno <angelogioacchino.delregno@collabora.com> Reviewed-by:
Chen-Yu Tsai <wenst@chromium.org> Link: https://lore.kernel.org/r/20230306140543.1813621-30-angelogioacchino.delregno@collabora.comSigned-off-by:
Stephen Boyd <sboyd@kernel.org>
-
AngeloGioacchino Del Regno authored
In preparation for migrating mt8516 clocks to the common simple probe mechanism, convert the apmixedsys to be a separated platform driver and move it to clk-mt8516-apmixedsys.c. While at it, also fix some indentation issues. During the conversion, error handling was added to the apmixedsys probe function. Signed-off-by:
-
AngeloGioacchino Del Regno authored
Convert the MT7622 topckgen and pericfg clock drivers to platform drivers and use the simple probe mechanism. This also allows to build these clocks as modules. Thanks to the conversion, more error handling was added to the clocks registration. Signed-off-by:
-
AngeloGioacchino Del Regno authored
The infracfg driver cannot be converted to clk_mtk_simple_probe() as it registers cpumuxes, which is not supported on the common probing mechanism: for this reason, move it to its own file. While at it, also convert it to be a platform driver instead; to do so, also add a .remove() callback for this driver. During the conversion, error handling was added to the infracfg probe function. Signed-off-by:
-
AngeloGioacchino Del Regno authored
Add a .remove() callback to the apmixedsys driver to allow full module build; while at it, also change the usage of builtin_platform_driver() to module_platform_driver() to actually make use of the new callback. Signed-off-by:
-
AngeloGioacchino Del Regno authored
In preparation for migrating mt7622 clocks to the common simple probe mechanism, move apmixedsys clocks to a different file. While at it, use the builtin_platform_driver() macro for it. During the conversion, error handling was added to the apmixedsys probe function. Signed-off-by:
-
AngeloGioacchino Del Regno authored
Instead of calling clk_prepare_enable() for clocks that shall stay enabled, use the CLK_IS_CRITICAL flag, which purpose is exactly that. Fixes: 2fc0a509 ("clk: mediatek: add clock support for MT7622 SoC") Signed-off-by:
-
AngeloGioacchino Del Regno authored
All the various MediaTek clock drivers are, in a way or another, redefining the GATE_MTK() macro with different names: while some are doing that by actually using GATE_MTK(), others are copying it entirely (hence, entirely redefining it). Change all clock drivers to always and consistently use this macro. Signed-off-by:
-
AngeloGioacchino Del Regno authored
Switch to mtk_clk_simple_{probe,remove}() for infracfg and topckgen clocks on MT8183 to allow full module build for clock drivers. Differently from other MediaTek clock drivers, it was necessary to change the name of the `clk13m` clock, as that is already declared in the SoC's devicetree as a "fixed-factor-clock" (with the same name) and redeclaring it here would obviously fail to register the entire clock controller; this clock wasn't dropped only to retain compatibility with older devicetrees As a note, the `clk13m` clock is not mentioned in any parent names array(s) as the correct one (csw_f26m_d2) is already used in place of that. Thanks to the conversion, more error handling was added to the clocks registration. Signed-off-by: -
AngeloGioacchino Del Regno authored
Increase human readability and decrease number of lines by compressing the clock array entries where possible, to a maximum of ~95 columns. Signed-off-by:
-
AngeloGioacchino Del Regno authored
In preparation for migrating all other mt8183 clocks to the common mtk_clk_simple_probe(), move apmixedsys clocks to a different file. While at it, use the builtin_platform_driver() macro for it and fix some indentation issues in the PLLs table. During the conversion, error handling was added to the apmixedsys probe function. Signed-off-by:
-
AngeloGioacchino Del Regno authored
Convert topckgen and infracfg clock drivers to use the common mtk_clk_simple_probe() mechanism and change this from the old "static" CLK_OF_DECLARE to be a platform driver, allowing it to eventually be built as a module. Thanks to the conversion, more error handling was added to the clocks registration. Signed-off-by:
-
AngeloGioacchino Del Regno authored
In preparation for converting the MT8167 clock drivers to be proper platform_driver(s), drop the __initconst annotation from all of the clock arrays since they will be used not only during init but also during runtime. Signed-off-by:
-
AngeloGioacchino Del Regno authored
In preparation for migrating all other MT8167 clocks to the common mtk_clk_simple_probe(), move apmixedsys clocks to a different file. While at it, also migrate away from the legacy CLK_OF_DECLARE and convert this clock driver to be a platform_driver instead. During the conversion, error handling was added to the apmixedsys probe function. Signed-off-by:
-
AngeloGioacchino Del Regno authored
Use the GATE_MTK macro to compress the GATE_TOP{0..5} macros. No functional changes. Signed-off-by: -
AngeloGioacchino Del Regno authored
All clocks in this driver are supported by the common simple probe mechanism and it's now possible to migrate to it. While at it, also switch to using the module_platform_driver() macro. Signed-off-by:
-
AngeloGioacchino Del Regno authored
In preparation for migrating this driver to the simple probe mechanism, join the audio gates to the top_misc_mux_gates array of mtk_composite clocks in one top_misc_muxes array. While at it, since the `apll_i2s0_parents` array is for all i2s clocks, rename that to `apll_i2s_parents`. Signed-off-by:
-
AngeloGioacchino Del Regno authored
On this SoC some clocks apparently don't have different offsets for set/clr/sta registers hence they can be set, cleared and status-read on one register: this means that it was possible to use simpler gate clocks instead of custom mtk_gate ones. In preparation for converting this clock driver to the common probe mechanism for MediaTek clocks, perform a conversion from simple_gate to mtk_gate clocks since the latter does provide implicit support for simple gate clocks as well. Signed-off-by:
-
AngeloGioacchino Del Regno authored
In preparation for migrating all other mt8365 clocks to the common mtk_clk_simple_probe(), move apmixedsys clocks to a different file. While at it, use the builtin_platform_driver() macro for it. During the conversion, error handling was added to the apmixedsys probe function. Signed-off-by:
-
AngeloGioacchino Del Regno authored
Now that all of the clocks in clk-mt2712.c are using the common mtk_clk_simple_{probe,remove}() callbacks we can safely migrate to module_platform_driver. While at it, also drop all references to `simple` in the specific context of mt2712 as that was used in the past only to allow us to have two platform_driver(s) in one file. Signed-off-by: -
AngeloGioacchino Del Regno authored
Add a .remove() callback to the apmixedsys driver to allow full module build; while at it, also change the usage of builtin_platform_driver() to module_platform_driver() to actually make use of the new callback. Signed-off-by:
-
AngeloGioacchino Del Regno authored
The only clock driver that does not support mtk_clk_simple_probe() is apmixedsys: in preparation for enabling module build of non-critical mt2712 clocks, move this to its own file. While at it, also fix some indentation issues in the PLLs table. Signed-off-by:
-
AngeloGioacchino Del Regno authored
This function was completely missing error handling: add it. Fixes: e2f744a8 ("clk: mediatek: Add MT2712 clock support") Signed-off-by:
-
AngeloGioacchino Del Regno authored
Compress the clock arrays entries to allow a maximum of 90 columns: this greatly increases readability and also generously reduces the amount of lines. While at it, also fix some indentation here and there. This is a cosmetic change. No functional changes. Signed-off-by:
-
AngeloGioacchino Del Regno authored
Now that the common mtk_clk_simple_{probe,remove}() functions can deal with divider clocks it is possible to migrate more clock drivers to it: in this case, it's about topckgen. While at it, also perform a fast migration for mcucfg. Thanks to the conversion, more error handling was added to the clocks registration. Signed-off-by: -
AngeloGioacchino Del Regno authored
Add support for divider clocks register/unregister in the common mtk_clk_simple_probe() and mtk_clk_simple_remove() functions. Signed-off-by:
Miles Chen <miles.chen@mediatek.com> Tested-by:
-
AngeloGioacchino Del Regno authored
Reduce duplication and simplify all MediaTek multimedia clock drivers by migrating away from defining custom probe functions for each driver and instead use mtk_clk_pdev_probe(). While at it, also add a .remove() callback to all of the multimedia clock drivers where missing. Signed-off-by:
-
AngeloGioacchino Del Regno authored
Introduce functions clk_mtk_pdev_probe() and clk_mtk_pdev_remove(): these will be useful to commonize the probe and remove handlers for multimedia (clk-mtxxxx-mm) drivers as these are registered by the mtk-mmsys driver instead of having their own devicetree compatible. In order to do this, the main logic of clk_mtk_simple{probe,remove}() was moved to new static __clk_mtk_simple_{probe,remove}() functions that take as parameter a pointer to struct device_node because when registering the clocks from mtk-mmsys we want to pass a pointer to the clock driver's parent (which is, obviously, mtk-mmsys) struct device_node instead. As for the clock driver's platform data: for the devicetree case, we keep using the standard match_data mechanism, else we retrieve it from an id_table. Signed-off-by: -
AngeloGioacchino Del Regno authored
Instead of using of_device_get_match_data(), switch to the generic device_get_match_data(). Signed-off-by:
-
AngeloGioacchino Del Regno authored
Add FHCTL parameters and register PLLs through FHCTL to add support for frequency hopping and SSC. FHCTL will be enabled only on PLLs specified in devicetree. This commit brings functional changes only upon addition of devicetree configuration. Signed-off-by:
-
AngeloGioacchino Del Regno authored
Add FHCTL parameters and register PLLs through FHCTL to add support for frequency hopping and SSC. FHCTL will be enabled only on PLLs specified in devicetree. This commit brings functional changes only upon addition of devicetree configuration. Signed-off-by:
-
AngeloGioacchino Del Regno authored
Add FHCTL parameters and register PLLs through FHCTL to add support for frequency hopping and SSC. FHCTL will be enabled only on PLLs specified in devicetree. This commit brings functional changes only upon addition of devicetree configuration. Signed-off-by:
-
AngeloGioacchino Del Regno authored
Add FHCTL parameters and register PLLs through FHCTL to add support for frequency hopping and SSC. FHCTL will be enabled only on PLLs specified in devicetree. This commit brings functional changes only upon addition of devicetree configuration. Signed-off-by:
-
AngeloGioacchino Del Regno authored
MT6795, MT8173, MT8192 and MT8195 support Frequency Hopping: add the compatibles for them. Signed-off-by:
Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org> Reviewed-by:
-
AngeloGioacchino Del Regno authored
These functions are used by the various MediaTek apmixed clock drivers that may be built as modules: export the common functions used to parse related devicetree properties, register and unregister the PLLFH clocks. Signed-off-by:
-
AngeloGioacchino Del Regno authored
The Frequency Hopping Controller (FHCTL) seems to have different versions, as it has a slightly different register layout on some older SoCs like MT6795, MT8173, MT8183 (and others). This driver is indeed compatible with at least some of those older IP revisions, so all we need to do is to add a way to select the right register layout at registration time. Signed-off-by:
-
- 05 Mar, 2023 4 commits
-
-
Linus Torvalds authored
-
Linus Torvalds authored
Commit aa47a7c2 ("lib/cpumask: deprecate nr_cpumask_bits") resulted in the cpumask operations potentially becoming hugely less efficient, because suddenly the cpumask was always considered to be variable-sized. The optimization was then later added back in a limited form by commit 6f9c07be ("lib/cpumask: add FORCE_NR_CPUS config option"), but that FORCE_NR_CPUS option is not useful in a generic kernel and more of a special case for embedded situations with fixed hardware. Instead, just re-introduce the optimization, with some changes. Instead of depending on CPUMASK_OFFSTACK being false, and then always using the full constant cpumask width, this introduces three different cpumask "sizes": - the exact size (nr_cpumask_bits) remains identical to nr_cpu_ids. This is used for situations where we should use the exact size. - the "small" size (small_cpumask_bits) is the NR_CPUS constant if it fits in a single word and the bitmap operations thus end up able to trigger the "small_const_nbits()" optimizations. This is used for the operations that have optimized single-word cases that get inlined, notably the bit find and scanning functions. - the "large" size (large_cpumask_bits) is the NR_CPUS constant if it is an sufficiently small constant that makes simple "copy" and "clear" operations more efficient. This is arbitrarily set at four words or less. As a an example of this situation, without this fixed size optimization, cpumask_clear() will generate code like movl nr_cpu_ids(%rip), %edx addq $63, %rdx shrq $3, %rdx andl $-8, %edx callq memset@PLT on x86-64, because it would calculate the "exact" number of longwords that need to be cleared. In contrast, with this patch, using a MAX_CPU of 64 (which is quite a reasonable value to use), the above becomes a single movq $0,cpumask instruction instead, because instead of caring to figure out exactly how many CPU's the system has, it just knows that the cpumask will be a single word and can just clear it all. Note that this does end up tightening the rules a bit from the original version in another way: operations that set bits in the cpumask are now limited to the actual nr_cpu_ids limit, whereas we used to do the nr_cpumask_bits thing almost everywhere in the cpumask code. But if you just clear bits, or scan for bits, we can use the simpler compile-time constants. In the process, remove 'cpumask_complement()' and 'for_each_cpu_not()' which were not useful, and which fundamentally have to be limited to 'nr_cpu_ids'. Better remove them now than have somebody introduce use of them later. Of course, on x86-64 with MAXSMP there is no sane small compile-time constant for the cpumask sizes, and we end up using the actual CPU bits, and will generate the above kind of horrors regardless. Please don't use MAXSMP unless you really expect to have machines with thousands of cores. Signed-off-by:
Linus Torvalds <torvalds@linux-foundation.org>
-
git://git.kernel.org/pub/scm/linux/kernel/git/herbert/crypto-2.6Linus Torvalds authored
Pull crypto fix from Herbert Xu: "Fix a regression in the caam driver" * tag 'v6.3-p2' of git://git.kernel.org/pub/scm/linux/kernel/git/herbert/crypto-2.6: crypto: caam - Fix edesc/iv ordering mixup
-
git://git.kernel.org/pub/scm/linux/kernel/git/tip/tipLinus Torvalds authored
Pull x86 updates from Thomas Gleixner: "A small set of updates for x86: - Return -EIO instead of success when the certificate buffer for SEV guests is not large enough - Allow STIPB to be enabled with legacy IBSR. Legacy IBRS is cleared on return to userspace for performance reasons, but the leaves user space vulnerable to cross-thread attacks which STIBP prevents. Update the documentation accordingly" * tag 'x86-urgent-2023-03-05' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip: virt/sev-guest: Return -EIO if certificate buffer is not large enough Documentation/hw-vuln: Document the interaction between IBRS and STIBP x86/speculation: Allow enabling STIBP with legacy IBRS
-
