Re: [PATCH v3 08/10] PCI: apple: Implement MSI support

From
Marc Zyngier <maz@kernel.org>
Date
2021-09-17 09:10:48
Also in
linux-pci, lkml
Reply to
Alyssa Rosenzweig, 2021-09-13
On Mon, 13 Sep 2021 21:43:23 +0100,
Alyssa Rosenzweig [off-list ref] wrote:
quoted
+static void apple_msi_compose_msg(struct irq_data *data, struct msi_msg *msg)
+{
+	BUILD_BUG_ON(upper_32_bits(DOORBELL_ADDR));
+
+	msg->address_hi = upper_32_bits(DOORBELL_ADDR);
+	msg->address_lo = lower_32_bits(DOORBELL_ADDR);
+	msg->data = data->hwirq;
+}
...
quoted
@@ -269,6 +378,14 @@ static int apple_pcie_port_setup_irq(struct apple_pcie_port *port)
 
 	irq_set_chained_handler_and_data(irq, apple_port_irq_handler, port);
 
+	/* Configure MSI base address */
+	writel_relaxed(lower_32_bits(DOORBELL_ADDR), port->base + PORT_MSIADDR);
+
+	/* Enable MSIs, shared between all ports */
+	writel_relaxed(0, port->base + PORT_MSIBASE);
+	writel_relaxed((ilog2(port->pcie->nvecs) << PORT_MSICFG_L2MSINUM_SHIFT) |
+		       PORT_MSICFG_EN, port->base + PORT_MSICFG);
+
 	return 0;
 }
I think the BUILD_BUG_ON makes more sense next to configuring the base
address (which only has a 32-bit register, the BUILD_BUG_ON justifies
using writel and not writeq), rather than configuring the message (which
specifies the full 64-bits).
Indeed, this makes a bit more sense. Thanks for pointing this out.

	M.

-- 
Without deviation from the norm, progress is not possible.
Thread (27 messages) 27 messages, 6 authors, last reply 2021-09-19
Keyboard shortcuts
hback out one level
jnext message in thread / next series-diff position
kprevious message in thread / previous series-diff position
ldrill in
Escclose help / fold thread tree
?toggle this help