I just had the chance to talk with David at OLS ... I think we're in
broad agreement on a number of issues. Try this one (against a recent
git tree ... possibly not the most recent).
commit f2d0a280d5d8067e55be930f2bc89ed19587e569
Author: Matthew Wilcox <matthew@xxxxxx>
Date: Thu Jul 24 14:07:25 2008 -0400
MSI: Don't disable MSIs if the mask bit isn't supported
David Vrabel has a device which generates an interrupt storm on the
INTx pin if we disable MSI interrupts altogether. Masking interrupts is
only a performance optimisation, so we can ignore the request to mask
the interrupt.
Signed-off-by: Matthew Wilcox <willy@xxxxxxxxxxxxxxx>
diff --git a/drivers/pci/msi.c b/drivers/pci/msi.c
index 15af618..1835481 100644
--- a/drivers/pci/msi.c
+++ b/drivers/pci/msi.c
@@ -126,7 +126,16 @@ static void msix_flush_writes(unsigned int irq)
}
}
-static void msi_set_mask_bits(unsigned int irq, u32 mask, u32 flag)
+/*
+ * PCI 2.3 does not specify mask bits for each MSI interrupt. Attempting to
+ * mask all MSI interrupts by clearing the MSI enable bit does not work
+ * reliably as devices without an INTx disable bit will then generate a
+ * level IRQ which will never be cleared.
+ *
+ * Returns 1 if it succeeded in masking the interrupt and 0 if the device
+ * doesn't support MSI masking.
+ */
+static int msi_set_mask_bits(unsigned int irq, u32 mask, u32 flag)
{
struct msi_desc *entry;
@@ -144,8 +153,7 @@ static void msi_set_mask_bits(unsigned int irq, u32 mask, u32 flag)
mask_bits |= flag & mask;
pci_write_config_dword(entry->dev, pos, mask_bits);
} else {
- __msi_set_enable(entry->dev, entry->msi_attrib.pos,
- !flag);
+ return 0;
}
break;
case PCI_CAP_ID_MSIX:
@@ -161,6 +169,7 @@ static void msi_set_mask_bits(unsigned int irq, u32 mask, u32 flag)
break;
}
entry->msi_attrib.masked = !!flag;
+ return 1;
}
void read_msi_msg(unsigned int irq, struct msi_msg *msg)
----- End forwarded message -----
--
Intel are signing my paycheques ... these opinions are still mine
"Bill, look, we understand that you're interested in selling us this
operating system, but compare it to ours. We can't possibly take such
a retrograde step."
--
To unsubscribe from this list: send the line "unsubscribe linux-pci" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
