]> git.ipfire.org Git - thirdparty/kernel/stable.git/commitdiff
usb: dwc2: use a longer AHB idle timeout in dwc2_core_reset()
authorMartin Blumenstingl <martin.blumenstingl@googlemail.com>
Thu, 20 Jun 2019 17:50:22 +0000 (19:50 +0200)
committerGreg Kroah-Hartman <gregkh@linuxfoundation.org>
Sun, 14 Jul 2019 06:09:49 +0000 (08:09 +0200)
commit dfc4fdebc5d62ac4e2fe5428e59b273675515fb2 upstream.

Use a 10000us AHB idle timeout in dwc2_core_reset() and make it
consistent with the other "wait for AHB master IDLE state" ocurrences.

This fixes a problem for me where dwc2 would not want to initialize when
updating to 4.19 on a MIPS Lantiq VRX200 SoC. dwc2 worked fine with
4.14.
Testing on my board shows that it takes 180us until AHB master IDLE
state is signalled. The very old vendor driver for this SoC (ifxhcd)
used a 1 second timeout.
Use the same timeout that is used everywhere when polling for
GRSTCTL_AHBIDLE instead of using a timeout that "works for one board"
(180us in my case) to have consistent behavior across the dwc2 driver.

Cc: linux-stable <stable@vger.kernel.org> # 4.19+
Acked-by: Minas Harutyunyan <hminas@synopsys.com>
Signed-off-by: Martin Blumenstingl <martin.blumenstingl@googlemail.com>
Signed-off-by: Felipe Balbi <felipe.balbi@linux.intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
drivers/usb/dwc2/core.c

index 55d5ae2a7ec7bf805d66502f842f739926e7995b..51d83f77dc04ac447789eebb602c80f3a17f2bab 100644 (file)
@@ -531,7 +531,7 @@ int dwc2_core_reset(struct dwc2_hsotg *hsotg, bool skip_wait)
        }
 
        /* Wait for AHB master IDLE state */
-       if (dwc2_hsotg_wait_bit_set(hsotg, GRSTCTL, GRSTCTL_AHBIDLE, 50)) {
+       if (dwc2_hsotg_wait_bit_set(hsotg, GRSTCTL, GRSTCTL_AHBIDLE, 10000)) {
                dev_warn(hsotg->dev, "%s: HANG! AHB Idle timeout GRSTCTL GRSTCTL_AHBIDLE\n",
                         __func__);
                return -EBUSY;