kernel/drivers/mmc/host/mtk-sd.c, branch linux-6.9.y

mmc: mtk-sd: Extend number of tuning steps

2023-12-07T14:13:08Z

Previously, during the MSDC calibration process, a full clock cycle actually not be covered, which in some cases didn't yield the best results and could cause CRC errors. This problem is particularly evident when MSDC is used as an SDIO host. In fact, MSDC support tuning up to a maximum of 64 steps, but by default, the step number is 32. By increase the tuning step, we are more likely to cover more parts of a clock cycle, and get better calibration result. To illustrate, when tuning 32 steps, if the obtained window has a hole near the middle, like this: 0xffc07ff (hex), then the selected delay will be the 6 (counting from right to left). (32 <- 1) 1111 1111 1100 0000 0000 0111 11(1)1 1111 However, if we tune 64 steps, the window obtained may look like this: 0xfffffffffffc07ff. The final selected delay will be 44, which is safer as it is further away from the hole: (64 <- 1) 1111 ... (1)111 1111 1111 1111 1111 1100 0000 0000 0111 1111 1111 In this case, delay 6 selected through 32 steps tuning is obviously not optimal, and this delay is closer to the hole, using it would easily cause CRC problems. As per mesaurements taken on mediatek SoC platform, the tuning phase will take: eMMC - 32 steps: ~3ms - 64 steps: ~6ms SDIO - 32 steps: ~4ms - 64 steos: ~7ms Tuning more steps won't prolong boot times by any meaningful amount of time, so for SD/SDIO the default tuning steps will be adjust to 64. But for eMMC, it is still preferred to use 32 steps tuning as otherwise there would be performance lose when accessing the RPMB partition(requiring retuning each time). You can configure property "mediatek,tuning-step" in MSDC dts node to adjust the step number. Signed-off-by: Axe Yang Reviewed-by: AngeloGioacchino Del Regno Link: https://lore.kernel.org/r/20231207063535.29546-3-axe.yang@mediatek.com Signed-off-by: Ulf Hansson

mmc: mtk-sd: Increase the verbosity of msdc_track_cmd_data

2023-12-07T13:52:52Z

This log message is necessary for debugging, so enable it by default to debug issues that are hard to reproduce locally. Signed-off-by: Pin-yen Lin Reviewed-by: Wenbin Mei Reviewed-by: AngeloGioacchino Del Regno Link: https://lore.kernel.org/r/20231201102747.3854573-1-treapking@chromium.org Signed-off-by: Ulf Hansson

mmc: mtk-sd: Use readl_poll_timeout_atomic in msdc_reset_hw

2023-09-26T14:47:44Z

Use atomic readl_poll_timeout_atomic, because msdc_reset_hw may be invoked in IRQ handler in the following context: msdc_irq() -> msdc_cmd_done() -> msdc_reset_hw() The following kernel BUG stack trace can be observed on Genio 1200 EVK after initializing MSDC1 hardware during kernel boot: [ 1.187441] BUG: scheduling while atomic: swapper/0/0/0x00010002 [ 1.189157] Modules linked in: [ 1.204633] CPU: 0 PID: 0 Comm: swapper/0 Tainted: G W 5.15.42-mtk+modified #1 [ 1.205713] Hardware name: MediaTek Genio 1200 EVK-P1V2-EMMC (DT) [ 1.206484] Call trace: [ 1.206796] dump_backtrace+0x0/0x1ac [ 1.207266] show_stack+0x24/0x30 [ 1.207692] dump_stack_lvl+0x68/0x84 [ 1.208162] dump_stack+0x1c/0x38 [ 1.208587] __schedule_bug+0x68/0x80 [ 1.209056] __schedule+0x6ec/0x7c0 [ 1.209502] schedule+0x7c/0x110 [ 1.209915] schedule_hrtimeout_range_clock+0xc4/0x1f0 [ 1.210569] schedule_hrtimeout_range+0x20/0x30 [ 1.211148] usleep_range_state+0x84/0xc0 [ 1.211661] msdc_reset_hw+0xc8/0x1b0 [ 1.212134] msdc_cmd_done.isra.0+0x4ac/0x5f0 [ 1.212693] msdc_irq+0x104/0x2d4 [ 1.213121] __handle_irq_event_percpu+0x68/0x280 [ 1.213725] handle_irq_event+0x70/0x15c [ 1.214230] handle_fasteoi_irq+0xb0/0x1a4 [ 1.214755] handle_domain_irq+0x6c/0x9c [ 1.215260] gic_handle_irq+0xc4/0x180 [ 1.215741] call_on_irq_stack+0x2c/0x54 [ 1.216245] do_interrupt_handler+0x5c/0x70 [ 1.216782] el1_interrupt+0x30/0x80 [ 1.217242] el1h_64_irq_handler+0x1c/0x2c [ 1.217769] el1h_64_irq+0x78/0x7c [ 1.218206] cpuidle_enter_state+0xc8/0x600 [ 1.218744] cpuidle_enter+0x44/0x5c [ 1.219205] do_idle+0x224/0x2d0 [ 1.219624] cpu_startup_entry+0x30/0x80 [ 1.220129] rest_init+0x108/0x134 [ 1.220568] arch_call_rest_init+0x1c/0x28 [ 1.221094] start_kernel+0x6c0/0x700 [ 1.221564] __primary_switched+0xc0/0xc8 Fixes: ffaea6ebfe9c ("mmc: mtk-sd: Use readl_poll_timeout instead of open-coded polling") Signed-off-by: Pablo Sun Reviewed-by: Chen-Yu Tsai Reviewed-by: AngeloGioacchino Del Regno Cc: stable@vger.kernel.org Link: https://lore.kernel.org/r/20230922095348.22182-1-pablo.sun@mediatek.com Signed-off-by: Ulf Hansson

mmc: mtk-sd: Convert to platform remove callback returning void

2023-08-15T10:45:04Z

The .remove() callback for a platform driver returns an int which makes many driver authors wrongly assume it's possible to do error handling by returning an error code. However the value returned is (mostly) ignored and this typically results in resource leaks. To improve here there is a quest to make the remove callback return void. In the first step of this quest all drivers are converted to .remove_new() which already returns void. Trivially convert this driver from always returning zero in the remove callback to the void returning variant. Cc: Uwe Kleine-König Signed-off-by: Yangtao Li Reviewed-by: AngeloGioacchino Del Regno Link: https://lore.kernel.org/r/20230727070051.17778-5-frank.li@vivo.com Signed-off-by: Ulf Hansson

mmc: Explicitly include correct DT includes

2023-08-09T11:21:48Z

The DT of_device.h and of_platform.h date back to the separate of_platform_bus_type before it as merged into the regular platform bus. As part of that merge prepping Arm DT support 13 years ago, they "temporarily" include each other. They also include platform_device.h and of.h. As a result, there's a pretty much random mix of those include files used throughout the tree. In order to detangle these headers and replace the implicit includes with struct declarations, users need to explicitly include the correct includes. Signed-off-by: Rob Herring Acked-by: Adrian Hunter Link: https://lore.kernel.org/r/20230718143054.1065288-1-robh@kernel.org Signed-off-by: Ulf Hansson

mmc: Merge branch fixes into next

2023-06-19T11:41:58Z

Merge the mmc fixes for v6.4-rc[n] into the next branch, to allow them to get tested together with the new mmc changes that are targeted for v6.5. Signed-off-by: Ulf Hansson

mmc: mtk-sd: fix deferred probing

2023-06-19T11:31:12Z

The driver overrides the error codes returned by platform_get_irq() to -EINVAL, so if it returns -EPROBE_DEFER, the driver will fail the probe permanently instead of the deferred probing. Switch to propagating the error codes upstream. Fixes: 208489032bdd ("mmc: mediatek: Add Mediatek MMC driver") Signed-off-by: Sergey Shtylyov Link: https://lore.kernel.org/r/20230617203622.6812-4-s.shtylyov@omp.ru Signed-off-by: Ulf Hansson

mmc: mtk-sd: reduce CIT for better performance

2023-06-12T13:20:08Z

CQHCI_SSC1 indicates to CQE the polling period to use when using periodic SEND_QUEUE_STATUS(CMD13) polling. Since MSDC CQE uses msdc_hclk as ITCFVAL, so driver should use hclk frequency to get the actual time. The default value 0x1000 that corresponds to 150us for MediaTek SoCs, let's decrease it to 0x40 that corresponds to 2.35us, which can improve the performance of some eMMC devices. Reported-by: kernel test robot Signed-off-by: Wenbin Mei Reviewed-by: AngeloGioacchino Del Regno Acked-by: Adrian Hunter Reviewed-by: Alexandre Mergnat Link: https://lore.kernel.org/r/20230609101355.5220-2-wenbin.mei@mediatek.com Signed-off-by: Ulf Hansson

mmc: mediatek: Avoid ugly error message when SDIO wakeup IRQ isn't used

2023-06-12T13:16:18Z

When I boot a kukui-kodama board, I see an ugly warning in my kernel log: mtk-msdc 11240000.mmc: error -ENXIO: IRQ sdio_wakeup not found It's pretty normal not to have an "sdio_wakeup" IRQ defined. In fact, no device trees in mainline seem to have it. Let's use the platform_get_irq_byname_optional() to avoid the error message. Fixes: 527f36f5efa4 ("mmc: mediatek: add support for SDIO eint wakup IRQ") Signed-off-by: Douglas Anderson Reviewed-by: Matthias Brugger Link: https://lore.kernel.org/r/20230510064434.1.I935404c5396e6bf952e99bb7ffb744c6f7fd430b@changeid Signed-off-by: Ulf Hansson

mmc: Avoid open coding by using mmc_op_tuning()

2022-12-07T12:29:13Z

Replace code with the already defined function. No functional changes. Signed-off-by: ChanWoo Lee Reviewed-by: Adrian Hunter Link: https://lore.kernel.org/r/20221124080031.14690-1-cw9316.lee@samsung.com Signed-off-by: Ulf Hansson