* [PATCH] PNP: don't fail device init if no DMA channel available
@ 2007-08-13 11:24 Jan Beulich
0 siblings, 0 replies; only message in thread
From: Jan Beulich @ 2007-08-13 11:24 UTC (permalink / raw)
To: linux-kernel; +Cc: ambx1
Most drivers for devices supporting ISA DMA can operate without DMA as
well (falling back zo PIO). Thus it seems inappropriate for PNP to fail
device initialization in case none of the possible DMA channels are
available. Instead, it should be left to the driver to decide what to
do if request_dma() fails.
The patch at once adjusts the code to account for the fact that
pnp_assign_dma() now doesn't need to report failure anymore.
Signed-off-by: Jan Beulich <jbeulich@novell.com>
Cc: Adam Belay <ambx1@neo.rr.com>
drivers/pnp/manager.c | 28 ++++++++++------------------
1 files changed, 10 insertions(+), 18 deletions(-)
--- linux-2.6.23-rc3/drivers/pnp/manager.c 2007-08-13 09:00:32.000000000 +0200
+++ 2.6.23-rc3-pnp-no-dma/drivers/pnp/manager.c 2007-08-07 10:40:43.000000000 +0200
@@ -170,7 +170,7 @@ static int pnp_assign_irq(struct pnp_dev
return 0;
}
-static int pnp_assign_dma(struct pnp_dev *dev, struct pnp_dma *rule, int idx)
+static void pnp_assign_dma(struct pnp_dev *dev, struct pnp_dma *rule, int idx)
{
resource_size_t *start, *end;
unsigned long *flags;
@@ -181,19 +181,15 @@ static int pnp_assign_dma(struct pnp_dev
1, 3, 5, 6, 7, 0, 2, 4
};
- if (!dev || !rule)
- return -EINVAL;
-
if (idx >= PNP_MAX_DMA) {
pnp_err
("More than 2 dmas is incompatible with pnp specifications.");
- /* pretend we were successful so at least the manager won't try again */
- return 1;
+ return;
}
/* check if this resource has been manually set, if so skip */
if (!(dev->res.dma_resource[idx].flags & IORESOURCE_AUTO))
- return 1;
+ return;
start = &dev->res.dma_resource[idx].start;
end = &dev->res.dma_resource[idx].end;
@@ -203,19 +199,17 @@ static int pnp_assign_dma(struct pnp_dev
*flags |= rule->flags | IORESOURCE_DMA;
*flags &= ~IORESOURCE_UNSET;
- if (!rule->map) {
- *flags |= IORESOURCE_DISABLED;
- return 1; /* skip disabled resource requests */
- }
-
for (i = 0; i < 8; i++) {
if (rule->map & (1 << xtab[i])) {
*start = *end = xtab[i];
if (pnp_check_dma(dev, idx))
- return 1;
+ return;
}
}
- return 0;
+#ifdef MAX_DMA_CHANNELS
+ *start = *end = MAX_DMA_CHANNELS;
+#endif
+ *flags |= IORESOURCE_UNSET | IORESOURCE_DISABLED;
}
/**
@@ -342,8 +336,7 @@ static int pnp_assign_resources(struct p
irq = irq->next;
}
while (dma) {
- if (!pnp_assign_dma(dev, dma, ndma))
- goto fail;
+ pnp_assign_dma(dev, dma, ndma);
ndma++;
dma = dma->next;
}
@@ -379,8 +372,7 @@ static int pnp_assign_resources(struct p
irq = irq->next;
}
while (dma) {
- if (!pnp_assign_dma(dev, dma, ndma))
- goto fail;
+ pnp_assign_dma(dev, dma, ndma);
ndma++;
dma = dma->next;
}
^ permalink raw reply [flat|nested] only message in thread
only message in thread, other threads:[~2007-08-13 14:22 UTC | newest]
Thread overview: (only message) (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2007-08-13 11:24 [PATCH] PNP: don't fail device init if no DMA channel available Jan Beulich
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox
Powered by JetHome