[PATCH net-next] ptp: ocp: add UPF_NO_THRE_TEST flag for serial ports

[PATCH net-next] ptp: ocp: add UPF_NO_THRE_TEST flag for serial ports

[Jonathan Lemon]

From: Jonathan Lemon <bsd@fb.com>

The serial port driver attempts to test for correct THRE behavior
on startup.  However, it does this by disabling interrupts, and
then intentionally trying to trigger an interrupt in order to see
if the IIR bit is set in the UART.

However, in this FPGA design, the UART interrupt is generated
through the MSI vector, so when interrupts are re-enabled after
the test, the DMAR-IR reports an unhandled IRTE entry, since
no irq handler is installed at this point - it is installed
after the test.

This only happens on the /second/ open of the UART, since on the
first open, the x86_vector has installed and activated by the
driver probe, and is correctly handled.  When the serial port is
closed for the first time, this vector is deactivated and removed,
leading to this error.

Signed-off-by: Jonathan Lemon <jonathan.lemon@gmail.com>
---
 drivers/ptp/ptp_ocp.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/ptp/ptp_ocp.c b/drivers/ptp/ptp_ocp.c
index 36c09269a12d..64cc92c2426b 100644
--- a/drivers/ptp/ptp_ocp.c
+++ b/drivers/ptp/ptp_ocp.c
@@ -1752,7 +1752,7 @@ ptp_ocp_serial_line(struct ptp_ocp *bp, struct ocp_resource *r)
 	uart.port.mapbase = pci_resource_start(pdev, 0) + r->offset;
 	uart.port.irq = pci_irq_vector(pdev, r->irq_vec);
 	uart.port.uartclk = 50000000;
-	uart.port.flags = UPF_FIXED_TYPE | UPF_IOREMAP;
+	uart.port.flags = UPF_FIXED_TYPE | UPF_IOREMAP | UPF_NO_THRE_TEST;
 	uart.port.type = PORT_16550A;
 
 	return serial8250_register_8250_port(&uart);
-- 
2.31.1

Re: [PATCH net-next] ptp: ocp: add UPF_NO_THRE_TEST flag for serial ports

[patchwork-bot+netdevbpf]

Hello:

This patch was applied to netdev/net-next.git (master)
by Jakub Kicinski <kuba@kernel.org>:

On Wed,  9 Mar 2022 14:34:27 -0800 you wrote:
> From: Jonathan Lemon <bsd@fb.com>
> 
> The serial port driver attempts to test for correct THRE behavior
> on startup.  However, it does this by disabling interrupts, and
> then intentionally trying to trigger an interrupt in order to see
> if the IIR bit is set in the UART.
> 
> [...]

Here is the summary with links:
  - [net-next] ptp: ocp: add UPF_NO_THRE_TEST flag for serial ports
    https://git.kernel.org/netdev/net-next/c/c17c4059df24

You are awesome, thank you!
-- 
Deet-doot-dot, I am a bot.
https://korg.docs.kernel.org/patchwork/pwbot.html