On Sat, Jul 29, 2023 at 10:58:46AM +0900, Damien Le Moal wrote:
> On 7/29/23 10:55, Damien Le Moal wrote:
> > On 7/29/23 10:35, Serge Semin wrote:
> >> On Mon, Jul 24, 2023 at 01:02:11PM +0530, Manivannan Sadhasivam wrote:
> >>> On Fri, Jul 21, 2023 at 05:10:27PM +0900, Damien Le Moal wrote:
> >>>> On 7/21/23 16:44, Yoshihiro Shimoda wrote:
> >>>>> Using "INTx" instead of "legacy" is more specific. So, rename
> >>>>> PCI_EPC_IRQ_LEGACY to PCI_EPC_IRQ_INTX.
> >>>>>
> >>>>> Suggested-by: Bjorn Helgaas <helgaas@xxxxxxxxxx>
> >>>>> Signed-off-by: Yoshihiro Shimoda <yoshihiro.shimoda.uh@xxxxxxxxxxx>
> >>>>> Reviewed-by: Manivannan Sadhasivam <mani@xxxxxxxxxx>
> >>>>> Acked-by: Jesper Nilsson <jesper.nilsson@xxxxxxxx> # ARTPEC
> >>>>> Reviewed-by: Serge Semin <fancer.lancer@xxxxxxxxx>
> >>>>
> >>
> >>>> I would rather drop completely the PCI_EPC_IRQ_XXX enum and simply use the
> >>>> PCI_IRQ_XXX macros used everywhere. Less definitions :)
> >>>>
> >>>> See attached patch that I have in my queue (about to send that).
> >>>>
> >>>
> >>> It looks better! This patch should be dropped.
> >>
> >> Back then Bjorn specifically asked to change the names suffix in a
> >> preparation patch before adding the INTx support to the DW PCIe core
> >> driver (see the Sb tag in the patch log). Damien, seeing you cleanup
> >> the names anyway what about fixing the macro suffix too: INTx instead
> >> of LEGACY)?
> >
> > Sure, I can do that. That is going to be a gigantic patch though given that
> > PCI_IRQ_LEGACY is used well beyond the ep/pcie controller drivers.
> > While I agree it would be nice to do, not sure it is worth such code churn.
> >
> >> Mani, Damien, what do you suggest to Yoshihiro to do with the
> >> LEGACY/INTx names in the following up patches of this series?
> >
> > If everyone is OK with the patch I proposed (the PCI_IRQ_LEGACY -> PCI_IRQ_INTx
> > change can go on top), then I can rebase it and send it next week (the remaining
> > of my EP cleanup series needs some more testing & rebasing). Yoshihiro can
> > either include it in his series or rebase on it if the patch is added to
> > pci-next quickly.
>
> Note that we could start by simply defining an alias:
>
> #define PCI_IRQ_INTx PCI_IRQ_LEGACY
>
> and gradually convert all drivers using it until we can get rid of PCI_IRQ_LEGACY.
I try to catch additions of "legacy," e.g., in new drivers, but I
agree this patch looks like it might be more churn than it's worth.
But I like your idea of an alias, Damien. Maybe something like the
below to make it more obvious that the preferred usage is the "INTX"
form.
diff --git a/include/linux/pci-epc.h b/include/linux/pci-epc.h
index 5cb694031072..6c0bb4c5d12e 100644
--- a/include/linux/pci-epc.h
+++ b/include/linux/pci-epc.h
@@ -21,11 +21,13 @@ enum pci_epc_interface_type {
enum pci_epc_irq_type {
PCI_EPC_IRQ_UNKNOWN,
- PCI_EPC_IRQ_LEGACY,
+ PCI_EPC_IRQ_INTX,
PCI_EPC_IRQ_MSI,
PCI_EPC_IRQ_MSIX,
};
+#define PCI_EPC_IRQ_LEGACY PCI_EPC_IRQ_INTX
+
static inline const char *
pci_epc_interface_string(enum pci_epc_interface_type type)
{
diff --git a/include/linux/pci.h b/include/linux/pci.h
index c69a2cc1f412..6638e0cd487f 100644
--- a/include/linux/pci.h
+++ b/include/linux/pci.h
@@ -1048,11 +1048,13 @@ enum {
PCI_SCAN_ALL_PCIE_DEVS = 0x00000040, /* Scan all, not just dev 0 */
};
-#define PCI_IRQ_LEGACY (1 << 0) /* Allow legacy interrupts */
+#define PCI_IRQ_INTX (1 << 0) /* Allow INTx interrupts */
#define PCI_IRQ_MSI (1 << 1) /* Allow MSI interrupts */
#define PCI_IRQ_MSIX (1 << 2) /* Allow MSI-X interrupts */
#define PCI_IRQ_AFFINITY (1 << 3) /* Auto-assign affinity */
+#define PCI_IRQ_LEGACY PCI_IRQ_INTX /* prefer PCI_IRQ_INTX */
+
/* These external functions are only available when PCI support is enabled */
#ifdef CONFIG_PCI
