Greg Kroah-Hartman [off-list ref] wrote:

On Thu, Jul 01, 2021 at 10:47:54PM -0700, Greg Thelen wrote:

quoted

Commit a66d21d7dba8 ("usb: xhci: Add support for Renesas controller with
memory") added renesas_usb_fw.mem firmware reference to xhci-pci.  Thus
modinfo indicates xhci-pci.ko has "firmware: renesas_usb_fw.mem".  But
the firmware is only actually used with CONFIG_USB_XHCI_PCI_RENESAS.  An
unusable firmware reference can trigger safety checkers which look for
drivers with unmet firmware dependencies.

Avoid referring to renesas_usb_fw.mem in circumstances when it cannot be
loaded (when CONFIG_USB_XHCI_PCI_RENESAS isn't set).

Signed-off-by: Greg Thelen <redacted>
---
 drivers/usb/host/xhci-pci.c | 8 ++++++++
 1 file changed, 8 insertions(+)

diff --git a/drivers/usb/host/xhci-pci.c b/drivers/usb/host/xhci-pci.c
index 18c2bbddf080..cb148da7a789 100644
--- a/drivers/usb/host/xhci-pci.c
+++ b/drivers/usb/host/xhci-pci.c

@@ -618,8 +618,10 @@ static void xhci_pci_shutdown(struct usb_hcd *hcd)
 /*-------------------------------------------------------------------------*/
 
 static const struct xhci_driver_data reneses_data = {
+#if IS_ENABLED(CONFIG_USB_XHCI_PCI_RENESAS)

If this is not enabled, why is the whole structure needed?

Removing the struct meant 2 more #if guards needed below to initialize
pci_ids, which exceeded my taste.

quoted

 	.quirks  = XHCI_RENESAS_FW_QUIRK,
 	.firmware = "renesas_usb_fw.mem",
+#endif
 };
 
 /* PCI driver selection metadata; PCI hotplugging uses this */

@@ -636,7 +638,13 @@ static const struct pci_device_id pci_ids[] = {
 	{ /* end: all zeroes */ }
 };
 MODULE_DEVICE_TABLE(pci, pci_ids);
+/*
+ * Without CONFIG_USB_XHCI_PCI_RENESAS renesas_xhci_check_request_fw() won't
+ * load firmware, so don't encumber the xhci-pci driver with it.
+ */
+#if IS_ENABLED(CONFIG_USB_XHCI_PCI_RENESAS)
 MODULE_FIRMWARE("renesas_usb_fw.mem");
+#endif

Why not just have this one chunk, why is the first chunk needed?

It seemed cleaner to remove both renesas_usb_fw.mem references rather
than leave one unusable one in reneses_data above.  But a simpler patch
is appealing, especially given the number of #if guards needed to remove
reneses_data.  Patch V2 will just go with this 1-hunk approach.

And a blank line please before your comment?

Will do.

And can you add a Fixes: line to the changelog text when you resend?

Will do.

thanks,

greg k-h