Re: [PATCH v5 1/2] PCI: dwc: Implement general suspend/resume functionality for L2/L3 transitionse

On Fri, Jul 28, 2023 at 11:49:28AM -0400, Frank Li wrote:

On Fri, Jul 28, 2023 at 09:02:38PM +0530, Manivannan Sadhasivam wrote:

quoted

On Mon, Jul 24, 2023 at 05:58:29PM -0400, Frank Li wrote:

quoted

Introduced helper function dw_pcie_get_ltssm to retrieve SMLH_LTSS_STATE.
Added API pme_turn_off and exit_from_l2 for managing L2/L3 state transitions.
...

quoted

quoted

+static void dw_pcie_set_dstate(struct dw_pcie *pci, pci_power_t dstate)
+{
+	u8 offset = dw_pcie_find_capability(pci, PCI_CAP_ID_PM);
+	u16 val;
+
+	val = dw_pcie_readw_dbi(pci, offset + PCI_PM_CTRL);
+	val &= ~PCI_PM_CTRL_STATE_MASK;
+	val |= ((u16 __force)dstate) & PCI_PM_CTRL_STATE_MASK;

Why can't just,

val |= dstate;

fixed a build warning.

Closes: https://lore.kernel.org/oe-kbuild-all/202307211904.zExw4Q8H-lkp@intel.com/ (local)

This is the sparse warning from the lkp URL:

  sparse warnings: (new ones prefixed by >>)
  >> drivers/pci/controller/dwc/pcie-designware-host.c:824:13: sparse: sparse: invalid assignment: |=
  >> drivers/pci/controller/dwc/pcie-designware-host.c:824:13: sparse:    left side has type unsigned int
  >> drivers/pci/controller/dwc/pcie-designware-host.c:824:13: sparse:    right side has type restricted pci_power_t

We have a zillion of those warnings already (try "make C=2 drivers/pci/").

Personally I think it's better to omit the ugly cast and live with the
warning for now.  Someday somebody will figure a better way to use
pci_power_t that will fix all these warnings at once.  I'd rather wait
for that fix than clutter all the uses with casts like this.

Bjorn