In ufs_qcom_dev_ref_clk_ctrl(), it was noted that the ref_clk needs to be stable for at least 1us. Even though there is wmb() to make sure the write gets "completed", there is no guarantee that the write actually reached the UFS device. There is a good chance that the write could be stored in a Write Buffer (WB). In that case, even though the CPU waits for 1us, the ref_clk might not be stable for that period.
So lets do a readl() to make sure that the previous write has reached the UFS device before udelay().
Also, the wmb() after writel_relaxed is not really needed. Both writel and readl are ordered on all architectures and the CPU won't speculate instructions after readl() due to the in-built control dependency with read value on weakly ordered architectures. So it can be safely removed.
Cc: stable@vger.kernel.org Fixes: f06fcc7155dc ("scsi: ufs-qcom: add QUniPro hardware support and power optimizations") Signed-off-by: Manivannan Sadhasivam manivannan.sadhasivam@linaro.org --- drivers/scsi/ufs/ufs-qcom.c | 7 +++++-- 1 file changed, 5 insertions(+), 2 deletions(-)
diff --git a/drivers/scsi/ufs/ufs-qcom.c b/drivers/scsi/ufs/ufs-qcom.c index 6ee33cc0ad09..f47a16b7cff5 100644 --- a/drivers/scsi/ufs/ufs-qcom.c +++ b/drivers/scsi/ufs/ufs-qcom.c @@ -687,8 +687,11 @@ static void ufs_qcom_dev_ref_clk_ctrl(struct ufs_qcom_host *host, bool enable)
writel_relaxed(temp, host->dev_ref_clk_ctrl_mmio);
- /* ensure that ref_clk is enabled/disabled before we return */ - wmb(); + /* + * Make sure the write to ref_clk reaches the destination and + * not stored in a Write Buffer (WB). + */ + readl(host->dev_ref_clk_ctrl_mmio);
/* * If we call hibern8 exit after this, we need to make sure that
On Sat 23 Apr 07:02 PDT 2022, Manivannan Sadhasivam wrote:
In ufs_qcom_dev_ref_clk_ctrl(), it was noted that the ref_clk needs to be stable for at least 1us. Even though there is wmb() to make sure the write gets "completed", there is no guarantee that the write actually reached the UFS device. There is a good chance that the write could be stored in a Write Buffer (WB). In that case, even though the CPU waits for 1us, the ref_clk might not be stable for that period.
So lets do a readl() to make sure that the previous write has reached the UFS device before udelay().
Also, the wmb() after writel_relaxed is not really needed. Both writel and readl are ordered on all architectures and the CPU won't speculate instructions after readl() due to the in-built control dependency with read value on weakly ordered architectures. So it can be safely removed.
Cc: stable@vger.kernel.org Fixes: f06fcc7155dc ("scsi: ufs-qcom: add QUniPro hardware support and power optimizations") Signed-off-by: Manivannan Sadhasivam manivannan.sadhasivam@linaro.org
Reviewed-by: Bjorn Andersson bjorn.andersson@linaro.org
drivers/scsi/ufs/ufs-qcom.c | 7 +++++-- 1 file changed, 5 insertions(+), 2 deletions(-)
diff --git a/drivers/scsi/ufs/ufs-qcom.c b/drivers/scsi/ufs/ufs-qcom.c index 6ee33cc0ad09..f47a16b7cff5 100644 --- a/drivers/scsi/ufs/ufs-qcom.c +++ b/drivers/scsi/ufs/ufs-qcom.c @@ -687,8 +687,11 @@ static void ufs_qcom_dev_ref_clk_ctrl(struct ufs_qcom_host *host, bool enable) writel_relaxed(temp, host->dev_ref_clk_ctrl_mmio);
/* ensure that ref_clk is enabled/disabled before we return */wmb();
/** Make sure the write to ref_clk reaches the destination and* not stored in a Write Buffer (WB).*/readl(host->dev_ref_clk_ctrl_mmio);/* * If we call hibern8 exit after this, we need to make sure that -- 2.25.1
linux-stable-mirror@lists.linaro.org
-
Bjorn Andersson -
Manivannan Sadhasivam