On 08/25/2016 02:22 PM, Steven J. Hill wrote:
We clear the OF_POPULATED flag for the GPIO controller node, otherwise
the GPIO lines used by the MMC driver are never probed.
Fixes: 15cc2ed6dcf9 ("of/irq: Mark initialised interrupt controllers as populated")
Signed-off-by: Steven J. Hill <Steven.Hill@xxxxxxxxxx>
---
arch/mips/cavium-octeon/octeon-irq.c | 7 +++++++
1 file changed, 7 insertions(+)
diff --git a/arch/mips/cavium-octeon/octeon-irq.c b/arch/mips/cavium-octeon/octeon-irq.c
index 5a9b87b..41d12d4 100644
--- a/arch/mips/cavium-octeon/octeon-irq.c
+++ b/arch/mips/cavium-octeon/octeon-irq.c
@@ -1619,6 +1619,13 @@ static int __init octeon_irq_init_gpio(
return -ENOMEM;
}
+ /*
+ * Clear the OF_POPULATED flag that was set above for the
Can we s/above/in of_irq_init()/ to be less ambiguous?
+ * GPIO controller so that the lines used by the MMC driver
I suspect that it is not just MMC that was broken by commit
15cc2ed6dcf9. Can we get a real description of exactly which kernel
facilities are impacted? Is it all GPIO, or what?
+ * will not be skipped.
+ */
+ of_node_clear_flag(gpio_node, OF_POPULATED);
+
return 0;
}
/*
