(1) Add a whitelist for EHCI devices without SBRN registers.
(2) Add Kunpeng920's EHCI device to the whitelist.
Changes in v3:
- Fix some code style issues.
- Update struct name.
Changes in v2:
- Fix some code style issues.
- Update function name.
Longfang Liu (2):
USB:ehci:Add a whitelist for EHCI controllers
USB:ehci:fix Kunpeng920 ehci hardware problem
drivers/usb/host/ehci-pci.c | 29 +++++++++++++++++++++++++----
1 file changed, 25 insertions(+), 4 deletions(-)
--
2.8.1
Kunpeng920's EHCI controller does not have SBRN register.
Reading the SBRN register when the controller driver is
initialized will get 0.
When rebooting the EHCI driver, ehci_shutdown() will be called.
if the sbrn flag is 0, ehci_shutdown() will return directly.
The sbrn flag being 0 will cause the EHCI interrupt signal to
not be turned off after reboot. this interrupt that is not closed
will cause an exception to the device sharing the interrupt.
Therefore, the EHCI controller of kunpeng920 needs to be added
to the whitelist without SBRN register.
Signed-off-by: Longfang Liu <[email protected]>
---
drivers/usb/host/ehci-pci.c | 2 ++
1 file changed, 2 insertions(+)
diff --git a/drivers/usb/host/ehci-pci.c b/drivers/usb/host/ehci-pci.c
index f7f3156..3690e62 100644
--- a/drivers/usb/host/ehci-pci.c
+++ b/drivers/usb/host/ehci-pci.c
@@ -53,6 +53,8 @@ static const struct ehci_nosbrn {
} ehci_nosbrn_table[] = {
/* STMICRO ConneXT has no sbrn register */
{PCI_VENDOR_ID_STMICRO, PCI_DEVICE_ID_STMICRO_USB_HOST},
+ /* HUAWEI Kunpeng920 HiSilicon USB EHCI has no sbrn register */
+ {PCI_VENDOR_ID_HUAWEI, 0xa239},
{},
};
--
2.8.1
Some types of EHCI controllers do not have SBRN registers.
By comparing the white list, the operation of reading the SBRN
registers is skipped.
Subsequent EHCI controller types without SBRN registers can be
directly added to the white list.
The current patch does not affect the drive function.
Signed-off-by: Longfang Liu <[email protected]>
---
drivers/usb/host/ehci-pci.c | 27 +++++++++++++++++++++++----
1 file changed, 23 insertions(+), 4 deletions(-)
diff --git a/drivers/usb/host/ehci-pci.c b/drivers/usb/host/ehci-pci.c
index 3c3820a..f7f3156 100644
--- a/drivers/usb/host/ehci-pci.c
+++ b/drivers/usb/host/ehci-pci.c
@@ -47,6 +47,28 @@ static inline bool is_bypassed_id(struct pci_dev *pdev)
return !!pci_match_id(bypass_pci_id_table, pdev);
}
+static const struct ehci_nosbrn {
+ u16 vendor;
+ u16 device;
+} ehci_nosbrn_table[] = {
+ /* STMICRO ConneXT has no sbrn register */
+ {PCI_VENDOR_ID_STMICRO, PCI_DEVICE_ID_STMICRO_USB_HOST},
+ {},
+};
+
+static bool usb_forbid_sbrn_read(struct pci_dev *pdev)
+{
+ const struct ehci_nosbrn *entry;
+
+ for (entry = ehci_nosbrn_table; entry->vendor; entry++) {
+ if (pdev->vendor == entry->vendor &&
+ pdev->device == entry->device)
+ return true;
+ }
+
+ return false;
+}
+
/*
* 0x84 is the offset of in/out threshold register,
* and it is the same offset as the register of 'hostpc'.
@@ -288,10 +310,7 @@ static int ehci_pci_setup(struct usb_hcd *hcd)
}
/* Serial Bus Release Number is at PCI 0x60 offset */
- if (pdev->vendor == PCI_VENDOR_ID_STMICRO
- && pdev->device == PCI_DEVICE_ID_STMICRO_USB_HOST)
- ; /* ConneXT has no sbrn register */
- else
+ if (!usb_forbid_sbrn_read(pdev))
pci_read_config_byte(pdev, 0x60, &ehci->sbrn);
/* Keep this around for a while just in case some EHCI
--
2.8.1
On Fri, Apr 09, 2021 at 11:00:14AM +0800, Longfang Liu wrote:
> Some types of EHCI controllers do not have SBRN registers.
> By comparing the white list, the operation of reading the SBRN
> registers is skipped.
>
> Subsequent EHCI controller types without SBRN registers can be
> directly added to the white list.
Your subject line still contains "whitelist" when the code does not.
Please fix it up to reflect your recent changes.
>
> The current patch does not affect the drive function.
>
> Signed-off-by: Longfang Liu <[email protected]>
> ---
> drivers/usb/host/ehci-pci.c | 27 +++++++++++++++++++++++----
> 1 file changed, 23 insertions(+), 4 deletions(-)
>
> diff --git a/drivers/usb/host/ehci-pci.c b/drivers/usb/host/ehci-pci.c
> index 3c3820a..f7f3156 100644
> --- a/drivers/usb/host/ehci-pci.c
> +++ b/drivers/usb/host/ehci-pci.c
> @@ -47,6 +47,28 @@ static inline bool is_bypassed_id(struct pci_dev *pdev)
> return !!pci_match_id(bypass_pci_id_table, pdev);
> }
>
> +static const struct ehci_nosbrn {
> + u16 vendor;
> + u16 device;
> +} ehci_nosbrn_table[] = {
> + /* STMICRO ConneXT has no sbrn register */
> + {PCI_VENDOR_ID_STMICRO, PCI_DEVICE_ID_STMICRO_USB_HOST},
> + {},
> +};
> +
> +static bool usb_forbid_sbrn_read(struct pci_dev *pdev)
> +{
> + const struct ehci_nosbrn *entry;
> +
> + for (entry = ehci_nosbrn_table; entry->vendor; entry++) {
> + if (pdev->vendor == entry->vendor &&
> + pdev->device == entry->device)
> + return true;
There is a PCI-match-device function somewhere, right?
But as Alan said, why is all of this needed? You only have 2 devices
you are testing for, we have if statements that are MUCH bigger than
that for simple things like this.
And the odds that a new device will be added to this list in the future
is _VERY_ low given that I would think that no one is making new EHCI
controllers anymore.
So just make this one patch, and add your device id to that, like you
did in the first set of patches.
thanks,
greg k-h