Stephen Boyd [Tue, 21 Jun 2016 22:53:14 +0000 (15:53 -0700)]
clk: qcom: Remove gcc_aggre1_pnoc_ahb_clk from msm8996
This clk is critical to operation of the SoC and should never be
turned off. Furthermore, there are no consumers of this clk so
let's just delete it so things like eMMC work.
Reported-by: Srinivas Kandagatla <srinivas.kandagatla@linaro.org> Fixes: b1e010c0730a ("clk: qcom: Add MSM8996 Global Clock Control (GCC) driver") Signed-off-by: Stephen Boyd <sboyd@codeaurora.org>
The gxbb clock controller is the primary clock generation unit for the
AmLogic GXBB SoC. It is clocked by a fixed 24MHz xtal, contains several
PLLs and the usual post-dividers, muxes, dividers and leaf gates that
are fed into various IP blocks in the SoC.
Tested-by: Kevin Hilman <khilman@baylibre.com> Signed-off-by: Michael Turquette <mturquette@baylibre.com>
Fractional MPLLs are a superset of the existing AmLogic MPLLs. They add
in a couple of new bitfields for further dividing the clock rate to
achieve rates with fractional hertz.
Tested-by: Kevin Hilman <khilman@baylibre.com> Signed-off-by: Michael Turquette <mturquette@baylibre.com>
There are a series of peripheral and system gate clocks that fan out
from the clk81 signal. Add a helper macro to statically initialize these
gate clocks.
Tested-by: Kevin Hilman <khilman@baylibre.com> Signed-off-by: Michael Turquette <mturquette@baylibre.com>
Break the AmLogic clock code up so that only the necessary parts are
compiled and linked. The core code is selected by both arm and arm64
builds with COMMON_CLK_AMLOGIC. The individual drivers have their own
config options as well.
Tested-by: Kevin Hilman <khilman@baylibre.com> Signed-off-by: Michael Turquette <mturquette@baylibre.com>
This patch creates a proper platform_driver for the meson8b clock
controller. Use of CLK_OF_DECLARE is removed, and can be added back in
later if very early registration of some clocks is required.
Tested-by: Kevin Hilman <khilman@baylibre.com> Signed-off-by: Michael Turquette <mturquette@baylibre.com>
Remove the composite clock registration function and helpers. Replace
unnecessary configuration struct with static initialization of the
desired clock type.
To preserve git bisect this patch also flips the switch and starts using
of_clk_add_hw_provider instead of the deprecated meson_clk_register_clks
method. As a byproduct clk.c can be deleted.
Tested-by: Kevin Hilman <khilman@baylibre.com> Signed-off-by: Michael Turquette <mturquette@baylibre.com>
Remove the fixed factor registration function and helpers. Replace
unnecessary configuration struct with static initialization of the
desired clock type.
Tested-by: Kevin Hilman <khilman@baylibre.com> Signed-off-by: Michael Turquette <mturquette@baylibre.com>
Remove the fixed_rate registration function and helpers from clkc.[ch].
Replace unnecessary configuration struct with static initialization of
the desired clock type.
While we're here, begin the transition to a proper platform_driver and
call of_clk_add_hw_provider with a shiny new struct clk_hw_onecell_data.
Tested-by: Kevin Hilman <khilman@baylibre.com> Signed-off-by: Michael Turquette <mturquette@baylibre.com>
Stephen Boyd [Tue, 21 Jun 2016 01:00:16 +0000 (18:00 -0700)]
Merge tag 'clk-samsung-4.8' of git://linuxtv.org/snawrocki/samsung into clk-next
Merge changes from Sylwester Nawrocki for samsung clk drivers:
- a fix for exynos7 to prevent gating some critical CMU clocks,
- addition of CPU clocks for CPU frequency scaling on Exynos5433 SoCs,
- additions for exynos5410 SoC required for Odroid XU board support,
- register accessors fixes for kernels built for big endian operation
(mostly exynos4 SoCs),
- Exynos5433 clock definitions fixes required for suspend to RAM and
the audio subsystem operation,
- many cleanups changing attributes of the clock initializer data
* tag 'clk-samsung-4.8' of git://linuxtv.org/snawrocki/samsung: (41 commits)
clk: samsung: exynos5433: Add CLK_IGNORE_UNUSED flag to PCIE device
clk: samsung: exynos5433: Add CLK_IGNORE_UNUSED flags to avoid hang during S2R
clk: samsung: exynos5433: Add CLK_IGNORE_UNUSED flag for AUD UART
clk: samsung: exynos4: fixup reg access on be
clk: samsung: fixup endian in pll clk
clk: samsung: exynos5410: Add WDT, ACLK266 and SSS clocks
clk: samsung: exynos5433: add CPU clocks configuration data and instantiate CPU clocks
clk: samsung: cpu: prepare for adding Exynos5433 CPU clocks
clk: samsung: exynos5433: prepare for adding CPU clocks
clk: samsung: Suppress unbinding to prevent theoretical attacks
clk: samsung: exynos5420: Set ID for aclk333 gate clock
clk: samsung: exynos5410: Add TMU clock
clk: samsung: exynos5410: Add I2C, HSI2C and RTC clocks
clk: samsung: exynos5410: Add serial3, USB and PWM clocks
clk: samsung: exynos3250: Move PLL rates data to init section
clk: samsung: Fully constify mux parent names
clk: samsung: exynos5250: Move sleep init function to init section
clk: samsung: exynos5420: Move sleep init function and PLL data to init section
clk: samsung: exynos5433: Move PLL rates data to init section
clk: samsung: exynos5433: Constify all clock initializers
...
Peng Fan [Mon, 13 Jun 2016 11:34:21 +0000 (19:34 +0800)]
clk: correct comments for __clk_determine_rate
Correct comments for __clk_determine_rate.
Signed-off-by: Peng Fan <van.freenix@gmail.com> Cc: Michael Turquette <mturquette@baylibre.com> Cc: Stephen Boyd <sboyd@codeaurora.org> Signed-off-by: Stephen Boyd <sboyd@codeaurora.org>
Roman Volkov [Tue, 7 Jun 2016 21:56:10 +0000 (00:56 +0300)]
clk: vt8500: rework wm8650_find_pll_bits()
PLL clock on WM8650 is calculated in the following way:
M * parent [O1] => / P [O2] => / D [O3]
Where O2 is 600MHz >= (M * parent) / P >= 300MHz.
Current algorithm does not met this requirement, so that the
function may return rates which are not supported by the hardware.
This patch fixes the algorithm and simplifies the code, reducing
the calculation time by ~10000 times (according to usermode app) by
removing the nested loops.
Signed-off-by: Roman Volkov <rvolkov@v1ros.org> Signed-off-by: Stephen Boyd <sboyd@codeaurora.org>
Arnd Bergmann [Tue, 7 Jun 2016 21:56:09 +0000 (00:56 +0300)]
clk: vt8500: fix gcc-4.9 warnings
This fixes some false positive warnings we get with older compiler
versions:
clk-vt8500.c: In function ‘wm8650_find_pll_bits’:
clk-vt8500.c:430:12: ‘best_div2’ may be used uninitialized in this function
clk-vt8500.c:429:12: ‘best_div1’ may be used uninitialized in this function
clk-vt8500.c:428:14: ‘best_mul’ may be used uninitialized in this function
clk-vt8500.c: In function ‘wm8750_find_pll_bits’:
clk-vt8500.c:509:12: ‘best_div2’ may be used uninitialized in this function
clk-vt8500.c:508:12: ‘best_div1’ may be used uninitialized in this function
clk-vt8500.c:507:14: ‘best_mul’ may be used uninitialized in this function
clk-vt8500.c: In function ‘wm8850_find_pll_bits’:
clk-vt8500.c:560:12: ‘best_div2’ may be used uninitialized in this function
clk-vt8500.c:559:12: ‘best_div1’ may be used uninitialized in this function
clk-vt8500.c:558:14: ‘best_mul’ may be used uninitialized in this function
As the local variables are only use for temporaries, we can just
as well assign the final values directly, which also makes the
code slightly shorter.
Signed-off-by: Arnd Bergmann <arnd@arndb.de> Signed-off-by: Roman Volkov <rvolkov@v1ros.org> Signed-off-by: Stephen Boyd <sboyd@codeaurora.org>
Ben Dooks [Tue, 7 Jun 2016 16:38:09 +0000 (17:38 +0100)]
clk: at91: make of_sama5d2_clk_generated_setup() static
The of_sama5d2_clk_generated_setup() is not exported outside
of the driver, so make it static to fix the warning about it
being not static:
drivers/clk/at91/clk-generated.c:270:13: warning: symbol 'of_sama5d2_clk_generated_setup' was not declared. Should it be static?
Signed-off-by: Ben Dooks <ben.dooks@codethink.co.uk> Acked-by: Boris Brezillon <boris.brezillon@free-electrons.com> Signed-off-by: Stephen Boyd <sboyd@codeaurora.org>
Stephen Boyd [Tue, 21 Jun 2016 00:01:45 +0000 (17:01 -0700)]
Merge tag 'v4.7-rockchip-clk-fixes1' of git://git.kernel.org/pub/scm/linux/kernel/git/mmind/linux-rockchip into clk-fixes
A bunch of fixes. Some for the newly added rk3399 clock tree, some
concerning error handling and initialization and a revert of the
mmc-phase clock initialization, as this could conflict with the
bootloader setting of this clock and a real solution to initing
the phase correctly from dw_mmc went in as fix for 4.7 through
the mmc tree.
* tag 'v4.7-rockchip-clk-fixes1' of git://git.kernel.org/pub/scm/linux/kernel/git/mmind/linux-rockchip:
clk: rockchip: release io resource when failing to init clk on rk3399
clk: rockchip: fix cpuclk registration error handling
clk: rockchip: Revert "clk: rockchip: reset init state before mmc card initialization"
clk: rockchip: fix incorrect parent for rk3399's {c,g}pll_aclk_perihp_src
clk: rockchip: mark rk3399 GIC clocks as critical
clk: rockchip: initialize flags of clk_init_data in mmc-phase clock
Wei Yongjun [Fri, 17 Jun 2016 17:24:23 +0000 (17:24 +0000)]
clk: Fix return value check in oxnas_stdclk_probe()
In case of error, the function syscon_node_to_regmap() returns
ERR_PTR() and never returns NULL. The NULL test in the return
value check should be replaced with IS_ERR().
Signed-off-by: Wei Yongjun <yongjun_wei@trendmicro.com.cn> Acked-by: Neil Armstrong <narmstrong@baylibre.com> Fixes: 0bbd72b4c64f ("clk: Add Oxford Semiconductor OXNAS Standard Clocks") Signed-off-by: Stephen Boyd <sboyd@codeaurora.org>
Maxime Ripard [Mon, 16 May 2016 12:47:02 +0000 (14:47 +0200)]
clk: multiplier: Prevent the multiplier from under / over flowing
In the current multiplier base clock implementation, if the
CLK_SET_RATE_PARENT flag isn't set, the code will not make sure that the
multiplier computed remains within the boundaries of our clock.
This means that if the clock we want to reach is below the parent rate,
or if the multiplier is above the maximum that we can reach, we will end up
with a completely bogus one that the clock cannot achieve.
Fixes: f2e0a53271a4 ("clk: Add a basic multiplier clock") Signed-off-by: Maxime Ripard <maxime.ripard@free-electrons.com> Signed-off-by: Michael Turquette <mturquette@baylibre.com>
Link: lkml.kernel.org/r/1463402840-17062-3-git-send-email-maxime.ripard@free-electrons.com
Paul E. McKenney [Tue, 26 Apr 2016 21:02:23 +0000 (14:02 -0700)]
clk: Use _rcuidle suffix to allow clk_core_enable() to used from idle
This commit fixes the RCU use-from-idle bug corresponding the following
splat:
> [ INFO: suspicious RCU usage. ]
> 4.6.0-rc5-next-20160426+ #1127 Not tainted
> -------------------------------
> include/trace/events/clk.h:45 suspicious rcu_dereference_check() usage!
>
> other info that might help us debug this:
>
>
> RCU used illegally from idle CPU!
> rcu_scheduler_active = 1, debug_locks = 0
> RCU used illegally from extended quiescent state!
> 2 locks held by swapper/0/0:
> #0: (&oh->hwmod_key#30){......}, at: [<c0121afc>] omap_hwmod_enable+0x18/0x44
> #1: (enable_lock){......}, at: [<c0630684>] clk_enable_lock+0x18/0x124
>
> stack backtrace:
> CPU: 0 PID: 0 Comm: swapper/0 Not tainted 4.6.0-rc5-next-20160426+ #1127
> Hardware name: Generic OMAP36xx (Flattened Device Tree)
> [<c0110290>] (unwind_backtrace) from [<c010c3a8>] (show_stack+0x10/0x14)
> [<c010c3a8>] (show_stack) from [<c047fd68>] (dump_stack+0xb0/0xe4)
> [<c047fd68>] (dump_stack) from [<c06315c0>] (clk_core_enable+0x1e0/0x36c)
> [<c06315c0>] (clk_core_enable) from [<c0632298>] (clk_enable+0x1c/0x38)
> [<c0632298>] (clk_enable) from [<c01204e0>] (_enable_clocks+0x18/0x7c)
> [<c01204e0>] (_enable_clocks) from [<c012137c>] (_enable+0x114/0x2ec)
> [<c012137c>] (_enable) from [<c0121b08>] (omap_hwmod_enable+0x24/0x44)
> [<c0121b08>] (omap_hwmod_enable) from [<c0122ad0>] (omap_device_enable+0x3c/0x90)
> [<c0122ad0>] (omap_device_enable) from [<c0122b34>] (_od_runtime_resume+0x10/0x38)
> [<c0122b34>] (_od_runtime_resume) from [<c052cc00>] (__rpm_callback+0x2c/0x60)
> [<c052cc00>] (__rpm_callback) from [<c052cc54>] (rpm_callback+0x20/0x80)
> [<c052cc54>] (rpm_callback) from [<c052df7c>] (rpm_resume+0x3d0/0x6f0)
> [<c052df7c>] (rpm_resume) from [<c052e2e8>] (__pm_runtime_resume+0x4c/0x64)
> [<c052e2e8>] (__pm_runtime_resume) from [<c04bf2c4>] (omap2_gpio_resume_after_idle+0x54/0x68)
> [<c04bf2c4>] (omap2_gpio_resume_after_idle) from [<c01269dc>] (omap3_enter_idle_bm+0xfc/0x1ec)
> [<c01269dc>] (omap3_enter_idle_bm) from [<c0601888>] (cpuidle_enter_state+0x80/0x3d4)
> [<c0601888>] (cpuidle_enter_state) from [<c0183b08>] (cpu_startup_entry+0x198/0x3a0)
> [<c0183b08>] (cpu_startup_entry) from [<c0b00c0c>] (start_kernel+0x354/0x3c8)
> [<c0b00c0c>] (start_kernel) from [<8000807c>] (0x8000807c)
Reported-by: Tony Lindgren <tony@atomide.com> Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com> Tested-by: Tony Lindgren <tony@atomide.com> Tested-by: Guenter Roeck <linux@roeck-us.net> Cc: Russell King <linux@arm.linux.org.uk> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Michael Turquette <mturquette@baylibre.com> Cc: Stephen Boyd <sboyd@codeaurora.org> Cc: <linux-omap@vger.kernel.org> Cc: <linux-arm-kernel@lists.infradead.org> Cc: <linux-clk@vger.kernel.org>
Paul E. McKenney [Tue, 26 Apr 2016 19:43:57 +0000 (12:43 -0700)]
clk: Add _rcuidle tracepoints to allow clk_core_disable() use from idle
This commit adds an _rcuidle suffix to a pair of trace events to
prevent the following splat:
> ===============================
> [ INFO: suspicious RCU usage. ]
> 4.6.0-rc5-next-20160426+ #1114 Not tainted
> -------------------------------
> include/trace/events/clk.h:59 suspicious rcu_dereference_check() usage!
>
> other info that might help us debug this:
>
>
> RCU used illegally from idle CPU!
> rcu_scheduler_active = 1, debug_locks = 0
> RCU used illegally from extended quiescent state!
> 2 locks held by swapper/0/0:
> #0: (&oh->hwmod_key#30){......}, at: [<c0121b40>] omap_hwmod_idle+0x18/0x44
> #1: (enable_lock){......}, at: [<c0630998>] clk_enable_lock+0x18/0x124
>
> stack backtrace:
> CPU: 0 PID: 0 Comm: swapper/0 Not tainted 4.6.0-rc5-next-20160426+ #1114
> Hardware name: Generic OMAP36xx (Flattened Device Tree)
> [<c0110290>] (unwind_backtrace) from [<c010c3a8>] (show_stack+0x10/0x14)
> [<c010c3a8>] (show_stack) from [<c047fd68>] (dump_stack+0xb0/0xe4)
> [<c047fd68>] (dump_stack) from [<c0631618>] (clk_core_disable+0x17c/0x348)
> [<c0631618>] (clk_core_disable) from [<c0632774>] (clk_disable+0x24/0x30)
> [<c0632774>] (clk_disable) from [<c0120590>] (_disable_clocks+0x18/0x7c)
> [<c0120590>] (_disable_clocks) from [<c0121680>] (_idle+0x12c/0x230)
> [<c0121680>] (_idle) from [<c0121b4c>] (omap_hwmod_idle+0x24/0x44)
> [<c0121b4c>] (omap_hwmod_idle) from [<c0122c24>] (omap_device_idle+0x3c/0x90)
> [<c0122c24>] (omap_device_idle) from [<c052cc00>] (__rpm_callback+0x2c/0x60)
> [<c052cc00>] (__rpm_callback) from [<c052cc54>] (rpm_callback+0x20/0x80)
> [<c052cc54>] (rpm_callback) from [<c052d150>] (rpm_suspend+0x100/0x768)
> [<c052d150>] (rpm_suspend) from [<c052ec58>] (__pm_runtime_suspend+0x64/0x84)
> [<c052ec58>] (__pm_runtime_suspend) from [<c04bf25c>] (omap2_gpio_prepare_for_idle+0x5c/0x70)
> [<c04bf25c>] (omap2_gpio_prepare_for_idle) from [<c0125568>] (omap_sram_idle+0x140/0x244)
> [<c0125568>] (omap_sram_idle) from [<c01269dc>] (omap3_enter_idle_bm+0xfc/0x1ec)
> [<c01269dc>] (omap3_enter_idle_bm) from [<c0601bdc>] (cpuidle_enter_state+0x80/0x3d4)
> [<c0601bdc>] (cpuidle_enter_state) from [<c0183b08>] (cpu_startup_entry+0x198/0x3a0)
> [<c0183b08>] (cpu_startup_entry) from [<c0b00c0c>] (start_kernel+0x354/0x3c8)
> [<c0b00c0c>] (start_kernel) from [<8000807c>] (0x8000807c)
Reported-by: Tony Lindgren <tony@atomide.com> Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com> Tested-by: Tony Lindgren <tony@atomide.com> Tested-by: Guenter Roeck <linux@roeck-us.net> Cc: Russell King <linux@arm.linux.org.uk> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Michael Turquette <mturquette@baylibre.com> Cc: Stephen Boyd <sboyd@codeaurora.org> Cc: <linux-omap@vger.kernel.org> Cc: <linux-arm-kernel@lists.infradead.org> Cc: <linux-clk@vger.kernel.org>
clk: samsung: cpu: prepare for adding Exynos5433 CPU clocks
Exynos5433 uses different register layout for CPU clock registers
than earlier SoCs so add new code for handling this layout. Also
add new CLK_CPU_HAS_E5433_REGS_LAYOUT flag to request using it.
There should be no functional change resulting from this patch.
Cc: Kukjin Kim <kgene@kernel.org> CC: Krzysztof Kozlowski <k.kozlowski@samsung.com> Signed-off-by: Bartlomiej Zolnierkiewicz <b.zolnierkie@samsung.com> Signed-off-by: Sylwester Nawrocki <s.nawrocki@samsung.com>
clk: samsung: Suppress unbinding to prevent theoretical attacks
Although unbinding a driver requires root privileges but it still might
be used theoretically in certain attacks (by triggering NULL pointer
exception or memory corruption if driver does not provide proper remove
callbacks or core does not handle it).
Samsung clock drivers are essential for system operation so their
removal is not expected. More over, the Exynos3250 ISP clock driver does
not implement remove() driver callback and it is not buildable as
modules.
Suppress the unbind interface for Exynos3250 ISP and S3C2410 DCLK clock
drivers.
Suggested-by: Marek Szyprowski <m.szyprowski@samsung.com> Signed-off-by: Krzysztof Kozlowski <k.kozlowski@samsung.com> Signed-off-by: Sylwester Nawrocki <s.nawrocki@samsung.com>
clk: samsung: exynos5420: Set ID for aclk333 gate clock
The aclk333 clock needs to be ungated during the MFC power domain switch,
so set the clock ID to allow the Exynos power domain logic to lookup this
clock if is defined in the MFC PD device tree node.
Signed-off-by: Javier Martinez Canillas <javier@osg.samsung.com> Reviewed-by: Krzysztof Kozlowski <k.kozlowski@samsung.com> Tested-by: Marek Szyprowski <m.szyprowski@samsung.com> Signed-off-by: Sylwester Nawrocki <s.nawrocki@samsung.com>
The PNAME macro defines array of strings for names of mux parents.
Although the strings itself were const but pointers to them were not thus
this data resided in initdata. Make this an array of const pointers to
const strings and move to initconst section.
Signed-off-by: Krzysztof Kozlowski <k.kozlowski@samsung.com> Signed-off-by: Sylwester Nawrocki <s.nawrocki@samsung.com>
clk: samsung: exynos5420: Move sleep init function and PLL data to init section
The exynos5420_clk_sleep_init() function and arrays with initialization
data of PLLs can be moved to init section because they are referenced
only from other init-level symbols.
Signed-off-by: Krzysztof Kozlowski <k.kozlowski@samsung.com> Signed-off-by: Sylwester Nawrocki <s.nawrocki@samsung.com>
clk: samsung: Constify clock init data with clock arrays
samsung_cmu_register_one() can accept pointer to const initialization
data: struct samsung_cmu_info. The members of the latter can also be
pointers to const data.
Signed-off-by: Krzysztof Kozlowski <k.kozlowski@samsung.com> Signed-off-by: Sylwester Nawrocki <s.nawrocki@samsung.com>
This patch adds CLK_IS_CRITICAL flag to ACLK_CCORE_133 and ACLK_FSYS0_200
clocks. These clocks are critical for accessing CMU_CCORE and CMU_FSYS0
blocks registers. Let these clocks to be enabled all the time.
Signed-off-by: Alim Akhtar <alim.akhtar@samsung.com> Signed-off-by: Sylwester Nawrocki <s.nawrocki@samsung.com>
Douglas Anderson [Thu, 12 May 2016 18:03:16 +0000 (11:03 -0700)]
clk: rockchip: Revert "clk: rockchip: reset init state before mmc card initialization"
This reverts commit 7a03fe6f48f3 ("clk: rockchip: reset init state
before mmc card initialization").
Though not totally obvious from the commit message nor from the source
code, that commit appears to be trying to reset the "_drv" MMC clocks to
90 degrees (note that the "_sample" MMC clocks have a shift of 0 so are
not touched).
The major problem here is that it doesn't properly reset things. The
phase is a two bit field and the commit only touches one of the two
bits. Thus the commit had the following affect:
- phase 0 => phase 90
- phase 90 => phase 90
- phase 180 => phase 270
- phase 270 => phase 270
Things get even weirder if you happen to have a bootloader that was
actually using delay elements (should be no reason to, but you never
know), since those are additional bits that weren't touched by the
original patch.
This is unlikely to be what we actually want. Checking on rk3288-veyron
devices, I can see that the bootloader leaves these clocks as:
- emmc: phase 180
- sdmmc: phase 90
- sdio0: phase 90
Thus on rk3288-veyron devices the commit we're reverting had the effect
of changing the eMMC clock to phase 270. This probably explains the
scattered reports I've heard of eMMC devices not working on some veyron
devices when using the upstream kernel.
The original commit was presumably made because previously the kernel
didn't touch the "_drv" phase at all and relied on whatever value was
there when the kernel started. If someone was using a bootloader that
touched the "_drv" phase then, indeed, we should have code in the kernel
to fix that. ...and also, to get ideal timings, we should also have the
kernel change the phase depending on the speed mode. In fact, that's
the subject of a recent patch I posted at
<https://patchwork.kernel.org/patch/9075141/>.
Ideally, we should take both the patch posted to dw_mmc and this
revert. Since those will likely go through different trees, here I
describe behavior with the combos:
1. Just this revert: likely will fix rk3288-veyron eMMC on some devices
+ other cases; might break someone with a strange bootloader that
sets the phase to 0 or one that uses delay elements (pretty
unpredicable what would happen in that case).
2. Just dw_mmc patch: fixes everyone. Effectly the dw_mmc patch will
totally override the broken patch and fix everything.
3. Both patches: fixes everyone. Once dw_mmc is initting properly then
any defaults from the clock code doesn't mattery.
Fixes: 7a03fe6f48f3 ("clk: rockchip: reset init state before mmc card initialization") Signed-off-by: Douglas Anderson <dianders@chromium.org> Reviewed-by: Shawn Lin <shawn.lin@rock-chips.com>
[emmc and sdmmc still work on all current boards in mainline after this
revert, so they should take precedence over any out-of-tree board that
will hopefully again get fixed with the better upcoming dw_mmc change.] Signed-off-by: Heiko Stuebner <heiko@sntech.de>
Signed-off-by: Brian Norris <briannorris@chromium.org> Reviewed-by: Douglas Anderson <dianders@chromium.org> Signed-off-by: Heiko Stuebner <heiko@sntech.de>
Heiko Stuebner [Tue, 17 May 2016 18:57:50 +0000 (20:57 +0200)]
clk: rockchip: initialize flags of clk_init_data in mmc-phase clock
The flags element of clk_init_data was never initialized for mmc-
phase-clocks resulting in the element containing a random value
and thus possibly enabling unwanted clock flags.
Fixes: 89bf26cbc1a0 ("clk: rockchip: Add support for the mmc clock phases using the framework") Cc: stable@vger.kernel.org Signed-off-by: Heiko Stuebner <heiko@sntech.de>
George Spelvin [Sun, 29 May 2016 05:26:41 +0000 (01:26 -0400)]
Rename other copy of hash_string to hashlen_string
The original name was simply hash_string(), but that conflicted with a
function with that name in drivers/base/power/trace.c, and I decided
that calling it "hashlen_" was better anyway.
But you have to do it in two places.
[ This caused build errors for architectures that don't define
CONFIG_DCACHE_WORD_ACCESS - Linus ]
Mikulas Patocka [Tue, 24 May 2016 20:49:18 +0000 (22:49 +0200)]
hpfs: implement the show_options method
The HPFS filesystem used generic_show_options to produce string that is
displayed in /proc/mounts. However, there is a problem that the options
may disappear after remount. If we mount the filesystem with option1
and then remount it with option2, /proc/mounts should show both option1
and option2, however it only shows option2 because the whole option
string is replaced with replace_mount_options in hpfs_remount_fs.
To fix this bug, implement the hpfs_show_options function that prints
options that are currently selected.
Mikulas Patocka [Tue, 24 May 2016 20:48:33 +0000 (22:48 +0200)]
affs: fix remount failure when there are no options changed
Commit c8f33d0bec99 ("affs: kstrdup() memory handling") checks if the
kstrdup function returns NULL due to out-of-memory condition.
However, if we are remounting a filesystem with no change to
filesystem-specific options, the parameter data is NULL. In this case,
kstrdup returns NULL (because it was passed NULL parameter), although no
out of memory condition exists. The mount syscall then fails with
ENOMEM.
This patch fixes the bug. We fail with ENOMEM only if data is non-NULL.
The patch also changes the call to replace_mount_options - if we didn't
pass any filesystem-specific options, we don't call
replace_mount_options (thus we don't erase existing reported options).
Mikulas Patocka [Tue, 24 May 2016 20:47:00 +0000 (22:47 +0200)]
hpfs: fix remount failure when there are no options changed
Commit ce657611baf9 ("hpfs: kstrdup() out of memory handling") checks if
the kstrdup function returns NULL due to out-of-memory condition.
However, if we are remounting a filesystem with no change to
filesystem-specific options, the parameter data is NULL. In this case,
kstrdup returns NULL (because it was passed NULL parameter), although no
out of memory condition exists. The mount syscall then fails with
ENOMEM.
This patch fixes the bug. We fail with ENOMEM only if data is non-NULL.
The patch also changes the call to replace_mount_options - if we didn't
pass any filesystem-specific options, we don't call
replace_mount_options (thus we don't erase existing reported options).
Fixes: ce657611baf9 ("hpfs: kstrdup() out of memory handling") Signed-off-by: Mikulas Patocka <mpatocka@redhat.com> Cc: stable@vger.kernel.org Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
VDSO:
- Build microMIPS VDSO for microMIPS kernels.
- Fix aliasing warning by building with `-fno-strict-aliasing' for
debugging but also tracing them might result in recursion.
Misc:
- Add missing FROZEN hotplug notifier transitions.
- Fix clk binding example for varioius PIC32 devices.
- Fix cpu interrupt controller node-names in the DT files.
- Fix XPA CPU feature separation.
- Fix write_gc0_* macros when writing zero.
- Add inline asm encoding helpers.
- Add missing VZ accessor microMIPS encodings.
- Fix little endian microMIPS MSA encodings.
- Add 64-bit HTW fields and fix its configuration.
- Fix sigreturn via VDSO on microMIPS kernel.
- Lots of typo fixes.
- Add definitions of SegCtl registers and use them"
Linus Torvalds [Sat, 28 May 2016 23:15:25 +0000 (16:15 -0700)]
Merge branch 'hash' of git://ftp.sciencehorizons.net/linux
Pull string hash improvements from George Spelvin:
"This series does several related things:
- Makes the dcache hash (fs/namei.c) useful for general kernel use.
(Thanks to Bruce for noticing the zero-length corner case)
- Converts the string hashes in <linux/sunrpc/svcauth.h> to use the
above.
- Avoids 64-bit multiplies in hash_64() on 32-bit platforms. Two
32-bit multiplies will do well enough.
- Rids the world of the bad hash multipliers in hash_32.
This finishes the job started in commit 689de1d6ca95 ("Minimal
fix-up of bad hashing behavior of hash_64()")
The vast majority of Linux architectures have hardware support for
32x32-bit multiply and so derive no benefit from "simplified"
multipliers.
The few processors that do not (68000, h8/300 and some models of
Microblaze) have arch-specific implementations added. Those
patches are last in the series.
- Overhauls the dcache hash mixing.
The patch in commit 0fed3ac866ea ("namei: Improve hash mixing if
CONFIG_DCACHE_WORD_ACCESS") was an off-the-cuff suggestion.
Replaced with a much more careful design that's simultaneously
faster and better. (My own invention, as there was noting suitable
in the literature I could find. Comments welcome!)
- Modify the hash_name() loop to skip the initial HASH_MIX(). This
would let us salt the hash if we ever wanted to.
- Sort out partial_name_hash().
The hash function is declared as using a long state, even though
it's truncated to 32 bits at the end and the extra internal state
contributes nothing to the result. And some callers do odd things:
- fs/hfs/string.c only allocates 32 bits of state
- fs/hfsplus/unicode.c uses it to hash 16-bit unicode symbols not bytes
- Modify bytemask_from_count to handle inputs of 1..sizeof(long)
rather than 0..sizeof(long)-1. This would simplify users other
than full_name_hash"
Special thanks to Bruce Fields for testing and finding bugs in v1. (I
learned some humbling lessons about "obviously correct" code.)
On the arch-specific front, the m68k assembly has been tested in a
standalone test harness, I've been in contact with the Microblaze
maintainers who mostly don't care, as the hardware multiplier is never
omitted in real-world applications, and I haven't heard anything from
the H8/300 world"
* 'hash' of git://ftp.sciencehorizons.net/linux:
h8300: Add <asm/hash.h>
microblaze: Add <asm/hash.h>
m68k: Add <asm/hash.h>
<linux/hash.h>: Add support for architecture-specific functions
fs/namei.c: Improve dcache hash function
Eliminate bad hash multipliers from hash_32() and hash_64()
Change hash_64() return value to 32 bits
<linux/sunrpc/svcauth.h>: Define hash_str() in terms of hashlen_string()
fs/namei.c: Add hashlen_string() function
Pull out string hash to <linux/stringhash.h>
George Spelvin [Wed, 25 May 2016 18:19:49 +0000 (14:19 -0400)]
h8300: Add <asm/hash.h>
This will improve the performance of hash_32() and hash_64(), but due
to complete lack of multi-bit shift instructions on H8, performance will
still be bad in surrounding code.
Designing H8-specific hash algorithms to work around that is a separate
project. (But if the maintainers would like to get in touch...)
Signed-off-by: George Spelvin <linux@sciencehorizons.net> Cc: Yoshinori Sato <ysato@users.sourceforge.jp> Cc: uclinux-h8-devel@lists.sourceforge.jp
George Spelvin [Wed, 25 May 2016 15:06:09 +0000 (11:06 -0400)]
microblaze: Add <asm/hash.h>
Microblaze is an FPGA soft core that can be configured various ways.
If it is configured without a multiplier, the standard __hash_32()
will require a call to __mulsi3, which is a slow software loop.
Instead, use a shift-and-add sequence for the constant multiply.
GCC knows how to do this, but it's not as clever as some.
Signed-off-by: George Spelvin <linux@sciencehorizons.net> Cc: Alistair Francis <alistair.francis@xilinx.com> Cc: Michal Simek <michal.simek@xilinx.com>
George Spelvin [Fri, 27 May 2016 02:11:51 +0000 (22:11 -0400)]
<linux/hash.h>: Add support for architecture-specific functions
This is just the infrastructure; there are no users yet.
This is modelled on CONFIG_ARCH_RANDOM; a CONFIG_ symbol declares
the existence of <asm/hash.h>.
That file may define its own versions of various functions, and define
HAVE_* symbols (no CONFIG_ prefix!) to suppress the generic ones.
Included is a self-test (in lib/test_hash.c) that verifies the basics.
It is NOT in general required that the arch-specific functions compute
the same thing as the generic, but if a HAVE_* symbol is defined with
the value 1, then equality is tested.
Signed-off-by: George Spelvin <linux@sciencehorizons.net> Cc: Geert Uytterhoeven <geert@linux-m68k.org> Cc: Greg Ungerer <gerg@linux-m68k.org> Cc: Andreas Schwab <schwab@linux-m68k.org> Cc: Philippe De Muyter <phdm@macq.eu> Cc: linux-m68k@lists.linux-m68k.org Cc: Alistair Francis <alistai@xilinx.com> Cc: Michal Simek <michal.simek@xilinx.com> Cc: Yoshinori Sato <ysato@users.sourceforge.jp> Cc: uclinux-h8-devel@lists.sourceforge.jp
George Spelvin [Mon, 23 May 2016 11:43:58 +0000 (07:43 -0400)]
fs/namei.c: Improve dcache hash function
Patch 0fed3ac866 improved the hash mixing, but the function is slower
than necessary; there's a 7-instruction dependency chain (10 on x86)
each loop iteration.
Word-at-a-time access is a very tight loop (which is good, because
link_path_walk() is one of the hottest code paths in the entire kernel),
and the hash mixing function must not have a longer latency to avoid
slowing it down.
There do not appear to be any published fast hash functions that:
1) Operate on the input a word at a time, and
2) Don't need to know the length of the input beforehand, and
3) Have a single iterated mixing function, not needing conditional
branches or unrolling to distinguish different loop iterations.
One of the algorithms which comes closest is Yann Collet's xxHash, but
that's two dependent multiplies per word, which is too much.
The key insights in this design are:
1) Barring expensive ops like multiplies, to diffuse one input bit
across 64 bits of hash state takes at least log2(64) = 6 sequentially
dependent instructions. That is more cycles than we'd like.
2) An operation like "hash ^= hash << 13" requires a second temporary
register anyway, and on a 2-operand machine like x86, it's three
instructions.
3) A better use of a second register is to hold a two-word hash state.
With careful design, no temporaries are needed at all, so it doesn't
increase register pressure. And this gets rid of register copying
on 2-operand machines, so the code is smaller and faster.
4) Using two words of state weakens the requirement for one-round mixing;
we now have two rounds of mixing before cancellation is possible.
5) A two-word hash state also allows operations on both halves to be
done in parallel, so on a superscalar processor we get more mixing
in fewer cycles.
I ended up using a mixing function inspired by the ChaCha and Speck
round functions. It is 6 simple instructions and 3 cycles per iteration
(assuming multiply by 9 can be done by an "lea" instruction):
x ^= *input++;
y ^= x; x = ROL(x, K1);
x += y; y = ROL(y, K2);
y *= 9;
Not only is this reversible, two consecutive rounds are reversible:
if you are given the initial and final states, but not the intermediate
state, it is possible to compute both input words. This means that at
least 3 words of input are required to create a collision.
(It also has the property, used by hash_name() to avoid a branch, that
it hashes all-zero to all-zero.)
The rotate constants K1 and K2 were found by experiment. The search took
a sample of random initial states (I used 1023) and considered the effect
of flipping each of the 64 input bits on each of the 128 output bits two
rounds later. Each of the 8192 pairs can be considered a biased coin, and
adding up the Shannon entropy of all of them produces a score.
The best-scoring shifts also did well in other tests (flipping bits in y,
trying 3 or 4 rounds of mixing, flipping all 64*63/2 pairs of input bits),
so the choice was made with the additional constraint that the sum of the
shifts is odd and not too close to the word size.
The final state is then folded into a 32-bit hash value by a less carefully
optimized multiply-based scheme. This also has to be fast, as pathname
components tend to be short (the most common case is one iteration!), but
there's some room for latency, as there is a fair bit of intervening logic
before the hash value is used for anything.
(Performance verified with "bonnie++ -s 0 -n 1536:-2" on tmpfs. I need
a better benchmark; the numbers seem to show a slight dip in performance
between 4.6.0 and this patch, but they're too noisy to quote.)
Special thanks to Bruce fields for diligent testing which uncovered a
nasty fencepost error in an earlier version of this patch.
[checkpatch.pl formatting complaints noted and respectfully disagreed with.]
Signed-off-by: George Spelvin <linux@sciencehorizons.net> Tested-by: J. Bruce Fields <bfields@redhat.com>
George Spelvin [Fri, 27 May 2016 03:00:23 +0000 (23:00 -0400)]
Eliminate bad hash multipliers from hash_32() and hash_64()
The "simplified" prime multipliers made very bad hash functions, so get rid
of them. This completes the work of 689de1d6ca.
To avoid the inefficiency which was the motivation for the "simplified"
multipliers, hash_64() on 32-bit systems is changed to use a different
algorithm. It makes two calls to hash_32() instead.
drivers/media/usb/dvb-usb-v2/af9015.c uses the old GOLDEN_RATIO_PRIME_32
for some horrible reason, so it inherits a copy of the old definition.
Signed-off-by: George Spelvin <linux@sciencehorizons.net> Cc: Antti Palosaari <crope@iki.fi> Cc: Mauro Carvalho Chehab <m.chehab@samsung.com>
George Spelvin [Fri, 27 May 2016 02:22:01 +0000 (22:22 -0400)]
Change hash_64() return value to 32 bits
That's all that's ever asked for, and it makes the return
type of hash_long() consistent.
It also allows (upcoming patch) an optimized implementation
of hash_64 on 32-bit machines.
I tried adding a BUILD_BUG_ON to ensure the number of bits requested
was never more than 32 (most callers use a compile-time constant), but
adding <linux/bug.h> to <linux/hash.h> breaks the tools/perf compiler
unless tools/perf/MANIFEST is updated, and understanding that code base
well enough to update it is too much trouble. I did the rest of an
allyesconfig build with such a check, and nothing tripped.
Signed-off-by: George Spelvin <linux@sciencehorizons.net>