[PATCH] net: fealnx: make driver work on architectures without I/O ports

[PATCH] net: fealnx: make driver work on architectures without I/O ports

[Ethan Nelson-Moore]

Devices supported by the fealnx driver support both MMIO and PIO access
(they have a PCI BAR for each). However, the driver always tries to use
the PIO BAR on architectures other than Alpha. This makes the driver
not work on architectures without I/O port mapping support. The comment
explaining why this was done explains that some x86 systems have issues
with MMIO. To enable the driver on all architectures while preventing
potential regressions, change the driver to only use PIO on x86.

Issue discovered by manual inspection.

Fixes: 1da177e4c3f4 ("Linux-2.6.12-rc2")
Cc: stable@vger.kernel.org
Signed-off-by: Ethan Nelson-Moore <enelsonmoore@gmail.com>
---
 drivers/net/ethernet/fealnx.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/net/ethernet/fealnx.c b/drivers/net/ethernet/fealnx.c
index 3c9961806f75..51dd09107242 100644
--- a/drivers/net/ethernet/fealnx.c
+++ b/drivers/net/ethernet/fealnx.c
@@ -91,7 +91,7 @@ static int full_duplex[MAX_UNITS] = { -1, -1, -1, -1, -1, -1, -1, -1 };
 
 /* This driver was written to use PCI memory space, however some x86 systems
    work only with I/O space accesses. */
-#ifndef __alpha__
+#ifdef CONFIG_X86
 #define USE_IO_OPS
 #endif
 
-- 
2.43.0

Re: [PATCH] net: fealnx: make driver work on architectures without I/O ports

[Andrew Lunn]

On Mon, Apr 27, 2026 at 07:11:36PM -0700, Ethan Nelson-Moore wrote:
> Devices supported by the fealnx driver support both MMIO and PIO access
> (they have a PCI BAR for each). However, the driver always tries to use
> the PIO BAR on architectures other than Alpha. This makes the driver
> not work on architectures without I/O port mapping support. The comment
> explaining why this was done explains that some x86 systems have issues
> with MMIO. To enable the driver on all architectures while preventing
> potential regressions, change the driver to only use PIO on x86.
> 
> Issue discovered by manual inspection.

Does that mean it is untested? Probably broken?

> Fixes: 1da177e4c3f4 ("Linux-2.6.12-rc2")

Given that nobody has cared about this for the last 20 years, why
change it now?

Please stop making changes to old drivers, unless you have the
hardware in your hands and you have a real problem with it. If those
conditions are not met, you are just wasting peoples time.

If you want to do something useful, please spend some time reviewing
other peoples patches. Or get some real hardware in your hands and add
new features to its driver.

    Andrew

---
pw-bot: cr