The ni_65xx driver has an "interrupt" subdevice that supports Comedi
asynchronous commands, placing a value in the Comedi buffer for each
interrupt. The subdevice uses Comedi's 16-bit sample format but the
interrupt handler is calling `comedi_buf_write_samples()` with the
address of a 32-bit integer `&s->state`. On bigendian machines, this
will copy 2 bytes from the wrong end of the 32-bit integer. This isn't
really a problem since `s->state` will always be 0 for this subdevice,
but clean it up by using a 16-bit variable initialized to 0 to pass the
value.
Signed-off-by: Ian Abbott <abbo...@mev.co.uk>
---
drivers/staging/comedi/drivers/ni_65xx.c | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)
diff --git a/drivers/staging/comedi/drivers/ni_65xx.c
b/drivers/staging/comedi/drivers/ni_65xx.c
index eb3f9f7109da..7cd8497420f2 100644
--- a/drivers/staging/comedi/drivers/ni_65xx.c
+++ b/drivers/staging/comedi/drivers/ni_65xx.c
@@ -472,6 +472,7 @@ static irqreturn_t ni_65xx_interrupt(int irq, void *d)
struct comedi_device *dev = d;
struct comedi_subdevice *s = dev->read_subdev;
unsigned int status;
+ unsigned short val = 0;
status = readb(dev->mmio + NI_65XX_STATUS_REG);
if ((status & NI_65XX_STATUS_INT) == 0)
@@ -482,7 +483,7 @@ static irqreturn_t ni_65xx_interrupt(int irq, void *d)
writeb(NI_65XX_CLR_EDGE_INT | NI_65XX_CLR_OVERFLOW_INT,
dev->mmio + NI_65XX_CLR_REG);
- comedi_buf_write_samples(s, &s->state, 1);
+ comedi_buf_write_samples(s, &val, 1);
comedi_handle_events(dev, s);
return IRQ_HANDLED;
--
2.30.0
_______________________________________________
devel mailing list
de...@linuxdriverproject.org
http://driverdev.linuxdriverproject.org/mailman/listinfo/driverdev-devel
