This is a note to let you know that I've just added the patch titled mmc: mediatek: Fix system suspend/resume support for CQHCI to the 5.9-stable tree which can be found at: http://www.kernel.org/git/?p=linux/kernel/git/stable/stable-queue.git;a=summary The filename of the patch is: mmc-mediatek-fix-system-suspend-resume-support-for-cqhci.patch and it can be found in the queue-5.9 subdirectory. If you, or anyone else, feels it should not be added to the stable tree, please let <stable@xxxxxxxxxxxxxxx> know about it. >From c0a2074ac575fff2848c8ef804bdc8590466c36c Mon Sep 17 00:00:00 2001 From: Wenbin Mei <wenbin.mei@xxxxxxxxxxxx> Date: Wed, 18 Nov 2020 14:34:05 +0800 Subject: mmc: mediatek: Fix system suspend/resume support for CQHCI From: Wenbin Mei <wenbin.mei@xxxxxxxxxxxx> commit c0a2074ac575fff2848c8ef804bdc8590466c36c upstream. Before we got these errors on MT8192 platform: [ 59.153891] Restarting tasks ... [ 59.154540] done. [ 59.159175] PM: suspend exit [ 59.218724] mtk-msdc 11f60000.mmc: phase: [map:fffffffe] [maxlen:31] [final:16] [ 119.776083] mmc0: cqhci: timeout for tag 9 [ 119.780196] mmc0: cqhci: ============ CQHCI REGISTER DUMP =========== [ 119.786709] mmc0: cqhci: Caps: 0x100020b6 | Version: 0x00000510 [ 119.793225] mmc0: cqhci: Config: 0x00000101 | Control: 0x00000000 [ 119.799706] mmc0: cqhci: Int stat: 0x00000000 | Int enab: 0x00000000 [ 119.806177] mmc0: cqhci: Int sig: 0x00000000 | Int Coal: 0x00000000 [ 119.812670] mmc0: cqhci: TDL base: 0x00000000 | TDL up32: 0x00000000 [ 119.819149] mmc0: cqhci: Doorbell: 0x003ffc00 | TCN: 0x00000200 [ 119.825656] mmc0: cqhci: Dev queue: 0x00000000 | Dev Pend: 0x00000000 [ 119.832155] mmc0: cqhci: Task clr: 0x00000000 | SSC1: 0x00001000 [ 119.838627] mmc0: cqhci: SSC2: 0x00000000 | DCMD rsp: 0x00000000 [ 119.845174] mmc0: cqhci: RED mask: 0xfdf9a080 | TERRI: 0x0000891c [ 119.851654] mmc0: cqhci: Resp idx: 0x00000000 | Resp arg: 0x00000000 [ 119.865773] mmc0: cqhci: : =========================================== [ 119.872358] mmc0: running CQE recovery >From these logs, we found TDL base was back to the default value. After suspend, the mmc host is powered off by HW, and bring CQE register to the default value, so we add system suspend/resume interface, then bring CQE to deactivated state before suspend, it will be enabled by CQE first request after resume. Signed-off-by: Wenbin Mei <wenbin.mei@xxxxxxxxxxxx> Link: https://lore.kernel.org/r/20201118063405.24906-1-wenbin.mei@xxxxxxxxxxxx Fixes: 88bd652b3c74 ("mmc: mediatek: command queue support") Cc: stable@xxxxxxxxxxxxxxx [Ulf: Renamed functions] Signed-off-by: Ulf Hansson <ulf.hansson@xxxxxxxxxx> Signed-off-by: Greg Kroah-Hartman <gregkh@xxxxxxxxxxxxxxxxxxx> --- drivers/mmc/host/mtk-sd.c | 22 ++++++++++++++++++++-- 1 file changed, 20 insertions(+), 2 deletions(-) --- a/drivers/mmc/host/mtk-sd.c +++ b/drivers/mmc/host/mtk-sd.c @@ -2654,11 +2654,29 @@ static int msdc_runtime_resume(struct de msdc_restore_reg(host); return 0; } + +static int msdc_suspend(struct device *dev) +{ + struct mmc_host *mmc = dev_get_drvdata(dev); + int ret; + + if (mmc->caps2 & MMC_CAP2_CQE) { + ret = cqhci_suspend(mmc); + if (ret) + return ret; + } + + return pm_runtime_force_suspend(dev); +} + +static int msdc_resume(struct device *dev) +{ + return pm_runtime_force_resume(dev); +} #endif static const struct dev_pm_ops msdc_dev_pm_ops = { - SET_SYSTEM_SLEEP_PM_OPS(pm_runtime_force_suspend, - pm_runtime_force_resume) + SET_SYSTEM_SLEEP_PM_OPS(msdc_suspend, msdc_resume) SET_RUNTIME_PM_OPS(msdc_runtime_suspend, msdc_runtime_resume, NULL) }; Patches currently in stable-queue which might be from wenbin.mei@xxxxxxxxxxxx are queue-5.9/mmc-mediatek-fix-system-suspend-resume-support-for-cqhci.patch