USB: EHCI: remove PORT_RWC_BITS when clearing USB_PORT_FEAT_ENABLE

In the ClearPortFeature/USB_PORT_FEAT_ENABLE case, ehci_hub_control()
would read from status_reg, clear PORT_PE, and write the result back to
status_reg. This would clear any bits in PORT_RWC_BITS that were set in
the registers. Fix this by masking these bits off before the write.

Since this masking is common across all ClearPortFeature cases, move it
into a single early location to avoid duplicating it.

Remove the same bugfix from ehci-tegra.c's tegra_ehci_hub_control(), now
that this case is correctly handled by the core.

Signed-off-by: Stephen Warren <swarren@nvidia.com>
Acked-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

1 files changed, 1 insertions, 12 deletions

diff --git a/drivers/usb/host/ehci-tegra.c b/drivers/usb/host/ehci-tegra.c
index 4c8bef615cec..9f817680ff04 100644
--- a/drivers/usb/host/ehci-tegra.c
+++ b/drivers/usb/host/ehci-tegra.c
@@ -148,18 +148,7 @@ static int tegra_ehci_hub_control(
 
         spin_lock_irqsave(&ehci->lock, flags);
 
-        /*
-         * In ehci_hub_control() for USB_PORT_FEAT_ENABLE clears the other bits
-         * that are write on clear, by writing back the register read value, so
-         * USB_PORT_FEAT_ENABLE is handled by masking the set on clear bits
-         */
-        if (typeReq == ClearPortFeature && wValue == USB_PORT_FEAT_ENABLE) {
-                temp = ehci_readl(ehci, status_reg) & ~PORT_RWC_BITS;
-                ehci_writel(ehci, temp & ~PORT_PE, status_reg);
-                goto done;
-        }
-
-        else if (typeReq == GetPortStatus) {
+        if (typeReq == GetPortStatus) {
                 temp = ehci_readl(ehci, status_reg);
                 if (tegra->port_resuming && !(temp & PORT_SUSPEND)) {
                         /* Resume completed, re-enable disconnect detection */