Skip to contentSkip to conversation

ide-cs (CompactFlash) driver, rm irq warning

PATCH In review David Brownell 2 participants 5 messages Oct 2, 2006

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 ide-cs (CompactFlash) driver, rm irq warning

Git rid of the runtime warning about pcmcia not supporting exclusive IRQs, so "the driver needs updating".

drivers/ide/legacy/ide-cs.c +1 -1
@@ -120,7 +120,7 @@ static int ide_probe(struct pcmcia_devic
     link->io.Attributes1 = IO_DATA_PATH_WIDTH_AUTO;
     link->io.Attributes2 = IO_DATA_PATH_WIDTH_8;
     link->io.IOAddrLines = 3;
-    link->irq.Attributes = IRQ_TYPE_EXCLUSIVE;
+    link->irq.Attributes = IRQ_TYPE_DYNAMIC_SHARING;
     link->irq.IRQInfo1 = IRQ_LEVEL_ID;
     link->conf.Attributes = CONF_ENABLE_IRQ;
     link->conf.IntType = INT_MEMORY_AND_IO;
#1 lore raw

Ar Llu, 2006-10-02 am 09:02 -0700, ysgrifennodd David Brownell:

Quoted 4 lines

Git rid of the runtime warning about pcmcia not supporting exclusive IRQs, so "the driver needs updating".

You've audited the code to check this is safely handled ?

#2 lore raw
Quoted 7 lines · Alan Cox

Ar Llu, 2006-10-02 am 09:02 -0700, ysgrifennodd David Brownell:

Quoted 4 lines

Git rid of the runtime warning about pcmcia not supporting exclusive IRQs, so "the driver needs updating".

You've audited the code to check this is safely handled ?

As best I can, yes. IDE being somewhat black-boxish to most of us. The patch literally does no more than shut up the warning found in drivers/pcmcia/pcmcia_resource.c ... it doesn't change any other behavior in the PCMCIA layer. And the IDE layer never appeared to have a problem.

The IRQ handler seems to be drivers/ide/ide-io.c::ide_intr() and comments there reflect the expectation that it handle shared IRQs.

- Dave

Acked-by
#3 lore raw

Ar Llu, 2006-10-02 am 11:58 -0700, ysgrifennodd David Brownell:

Quoted 2 lines

The IRQ handler seems to be drivers/ide/ide-io.c::ide_intr() and comments there reflect the expectation that it handle shared IRQs.

I was more worried what the pcmcia side may be up to but yes this should get dealt with and -mm will find any horrors fast.

#4 lore raw
Quoted 5 lines · Alan Cox

Ar Llu, 2006-10-02 am 11:58 -0700, ysgrifennodd David Brownell:

Quoted 2 lines

The IRQ handler seems to be drivers/ide/ide-io.c::ide_intr() and comments there reflect the expectation that it handle shared IRQs.

I was more worried what the pcmcia side may be up to

The relevant flag appears to be ignored inside drivers/pcmcia except to trigger the annoying message ... leaving the remaining potential for side effects inside the old IDE stack, which would already be showing bug symptoms if there were any.

Quoted 5 lines

but yes this should
get dealt with and -mm will find any horrors fast.