On Friday 13 January 2006 17:24, Bjorn Helgaas wrote:
> ... the
> DMI stuff crashes HP sx2000 (and probably sx1000) boxes, probably
> because of some memory attribute problem. So I'll have more
> feedback after I debug that ;-)
It *is* a memory attribute problem. The current code always calls
ioremap() on efi.smbios. The first problem is that this is a
physical address on x86, but a virtual address on ia64.
The second problem is that we don't check the supported attributes
for the SMBIOS table. On HP sx1000/sx2000, these tables are in system
memory, which doesn't support uncacheable access, so ioremap() does
the wrong thing.
The patch below addresses both problems (but I can't test the x86 EFI
change). I don't really like it, because the memory attribute checking
is not complete (it only checks the first page, not the whole range),
and there's already very similar code in acpi_os_map_memory(),
acpi_os_read_memory(), acpi_os_write_memory(), and efi_range_is_wc().
But it's a start, and maybe the consolidation could be done later.
Index: work-mm3/arch/i386/kernel/dmi_scan.c
===================================================================
--- work-mm3.orig/arch/i386/kernel/dmi_scan.c 2006-01-17 15:18:42.000000000 -0700
+++ work-mm3/arch/i386/kernel/dmi_scan.c 2006-01-17 16:58:11.000000000 -0700
@@ -39,9 +39,18 @@
void (*decode)(struct dmi_header *))
{
u8 *buf, *data;
- int i = 0;
+ int iomem = 1, i = 0;
- buf = dmi_ioremap(base, len);
+ if (efi_enabled) {
+ if (efi_mem_attributes(base & EFI_MEMORY_WB)) {
+ iomem = 0;
+ buf = (u8 *) phys_to_virt(base);
+ } else if (efi_mem_attributes(base & EFI_MEMORY_UC))
+ buf = dmi_ioremap(base, len);
+ else
+ buf = NULL;
+ } else
+ buf = dmi_ioremap(base, len);
if (buf == NULL)
return -1;
@@ -66,7 +75,8 @@
data += 2;
i++;
}
- dmi_iounmap(buf, len);
+ if (iomem)
+ dmi_iounmap(buf, len);
return 0;
}
@@ -216,19 +226,30 @@
int rc;
if (efi_enabled) {
- if (!efi.smbios)
+ unsigned long phys_addr = __pa(efi.smbios);
+ int iomem = 0;
+
+ if (!phys_addr)
goto out;
/* This is called as a core_initcall() because it isn't
* needed during early boot. This also means we can
* iounmap the space when we're done with it.
*/
- p = ioremap((unsigned long)efi.smbios, 0x10000);
+ if (efi_mem_attributes(phys_addr & EFI_MEMORY_WB))
+ p = (char *) phys_to_virt(phys_addr);
+ else if (efi_mem_attributes(phys_addr & EFI_MEMORY_UC)) {
+ iomem = 1;
+ p = ioremap(phys_addr, 0x10000);
+ } else
+ p = NULL;
+
if (p == NULL)
goto out;
rc = dmi_present(p + 0x10); /* offset of _DMI_ string */
- iounmap(p);
+ if (iomem)
+ iounmap(p);
if (!rc)
return;
}
Index: work-mm3/arch/i386/kernel/efi.c
===================================================================
--- work-mm3.orig/arch/i386/kernel/efi.c 2005-10-27 18:02:08.000000000 -0600
+++ work-mm3/arch/i386/kernel/efi.c 2006-01-17 17:10:20.000000000 -0700
@@ -391,7 +391,7 @@
printk(KERN_INFO " ACPI=0x%lx ", config_tables[i].table);
} else
if (efi_guidcmp(config_tables[i].guid, SMBIOS_TABLE_GUID) == 0) {
- efi.smbios = (void *) config_tables[i].table;
+ efi.smbios = __va(config_tables[i].table);
printk(KERN_INFO " SMBIOS=0x%lx ", config_tables[i].table);
} else
if (efi_guidcmp(config_tables[i].guid, HCDP_TABLE_GUID) == 0) {
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
[Index of Archives]
[Kernel Newbies]
[Netfilter]
[Bugtraq]
[Photo]
[Stuff]
[Gimp]
[Yosemite News]
[MIPS Linux]
[ARM Linux]
[Linux Security]
[Linux RAID]
[Video 4 Linux]
[Linux for the blind]
[Linux Resources]
