[PATCH] Fix for handling bad IPMI ACPI data

[PATCH] Fix for handling bad IPMI ACPI data

[Corey Minyard]

[-- Attachment #1: Type: text/plain, Size: 1 bytes --]



[-- Attachment #2: ipmi_broken_acpi_register_bits.diff --]
[-- Type: text/x-patch, Size: 1305 bytes --]

If the ACPI register bit width is zero (an invalid value) assume
it is the default spacing.  This avoids some coredumps on invalid
data and makes some systems work that have broken ACPI data.

Signed-off-by: Corey Minyard <minyard@acm.org>

Index: linux-2.6.12-rc2/drivers/char/ipmi/ipmi_si_intf.c
===================================================================
--- linux-2.6.12-rc2.orig/drivers/char/ipmi/ipmi_si_intf.c
+++ linux-2.6.12-rc2/drivers/char/ipmi/ipmi_si_intf.c
@@ -1557,8 +1557,17 @@
 		info->irq_setup = NULL;
 	}
 
-	regspacings[intf_num] = spmi->addr.register_bit_width / 8;
-	info->io.regspacing = spmi->addr.register_bit_width / 8;
+	if (spmi->addr.register_bit_width) {
+		/* A (hopefully) properly formed register bit width. */
+		regspacings[intf_num] = spmi->addr.register_bit_width / 8;
+		info->io.regspacing = spmi->addr.register_bit_width / 8;
+	} else {
+		/* Some broken systems get this wrong and set the value
+		 * to zero.  Assume it is the default spacing.  If that
+		 * is wrong, too bad, the vendor should fix the tables. */
+		regspacings[intf_num] = DEFAULT_REGSPACING;
+		info->io.regspacing = DEFAULT_REGSPACING;
+	}
 	regsizes[intf_num] = regspacings[intf_num];
 	info->io.regsize = regsizes[intf_num];
 	regshifts[intf_num] = spmi->addr.register_bit_offset;