When the initialization code in hpet founds a memory resource and does
not found IRQ, it does not unmap the memory resource previously
mapped.
There are buggy BIOSes which report resources exactly like this and
what is worse the memory region bases point to normal RAM. This
normally would not matter since the space is not touched. But when PAT
is turned on, ioremap causes the page to be uncached and sets this bit
in page->flags.
Then when the page is about to be used by the allocator, it is
reported as:
BUG: Bad page state in process md5sum pfn:3ed00
page:ffffea0000dbd800 count:0 mapcount:0 mapping:(null) index:0x0
page flags: 0x20000001000000(uncached)
Pid: 7956, comm: md5sum Not tainted 2.6.34-12-desktop #1
Call Trace:
[<ffffffff810df851>] bad_page+0xb1/0x100
[<ffffffff810dfa45>] prep_new_page+0x1a5/0x1c0
[<ffffffff810dfe01>] get_page_from_freelist+0x3a1/0x640
[<ffffffff810e01af>] __alloc_pages_nodemask+0x10f/0x6b0
...
In this particular case:
1) HPET returns 3ed00000 as memory region base, but it is not in
reserved ranges reported by the BIOS (excerpt):
BIOS-e820: 0000000000100000 - 00000000af6cf000 (usable)
BIOS-e820: 00000000af6cf000 - 00000000afdcf000 (reserved)
2) there is no IRQ resource reported by HPET method. On the other
hand, the Intel HPET specs (1.0a) says (3.2.5.1):
_CRS (
// Report 1K of memory consumed by this Timer Block
memory range consumed
// Optional: only used if BIOS allocates Interrupts [1]
IRQs consumed
)
[1] For case where Timer Block is configured to consume IRQ0/IRQ8 AND
Legacy 8254/Legacy RTC hardware still exists, the device objects
associated with 8254 & RTC devices should not report IRQ0/IRQ8 as
“consumed resources.”
So in theory we should check whether if it is the case and use those
interrupts instead.
Anyway the address reported by the BIOS here is bogus, so non-presence
of IRQ doesn't mean the "optional" part in point 2).
Since I got no reply previously, fix this by simply unmapping the
space when IRQ is not found and memory region was mapped previously.
It would be probably more safe to walk the resources again and unmap
appropriately depending on type. But as we now use only ioremap for
both 2 memory resource types, it is not necessarily needed right now.
References: https://bugzilla.novell.com/show_bug.cgi?id=629908
Signed-off-by: Jiri Slaby <[email protected]>
Cc: Clemens Ladisch <[email protected]>
---
drivers/char/hpet.c | 2 ++
1 files changed, 2 insertions(+), 0 deletions(-)
diff --git a/drivers/char/hpet.c b/drivers/char/hpet.c
index 8372a34..e3e184b 100644
--- a/drivers/char/hpet.c
+++ b/drivers/char/hpet.c
@@ -1017,6 +1017,8 @@ static int hpet_acpi_add(struct acpi_device *device)
return -ENODEV;
if (!data.hd_address || !data.hd_nirqs) {
+ if (data.hd_address)
+ iounmap(data.hd_address);
printk("%s: no address or irqs in _CRS\n", __func__);
return -ENODEV;
}
--
1.7.2.1
Jiri Slaby wrote:
> When the initialization code in hpet founds a memory resource and does
> not found IRQ, it does not unmap the memory resource previously
> mapped.
> [...]
> Since I got no reply previously, fix this by simply unmapping the
> space when IRQ is not found and memory region was mapped previously.
>
> Signed-off-by: Jiri Slaby <[email protected]>
> Cc: Clemens Ladisch <[email protected]>
Acked-by: Clemens Ladisch <[email protected]>
> ---
> drivers/char/hpet.c | 2 ++
> 1 files changed, 2 insertions(+), 0 deletions(-)
>
> --- a/drivers/char/hpet.c
> +++ b/drivers/char/hpet.c
> @@ -1017,6 +1017,8 @@ static int hpet_acpi_add(struct acpi_device *device)
> return -ENODEV;
>
> if (!data.hd_address || !data.hd_nirqs) {
> + if (data.hd_address)
> + iounmap(data.hd_address);
> printk("%s: no address or irqs in _CRS\n", __func__);
> return -ENODEV;
> }