2013-03-12 20:17:29

by Artem Savkov

[permalink] [raw]
Subject: [PATCH] dmi_scan: proper buf type in dmi_present()

buf needs to be unsigned in dmi_present(), otherwise dmi_base calculation
doesn't work properly leading to a wraparound during dmi_ioremap:

[ 0.000000] ------------[ cut here ]------------
[ 0.000000] WARNING: at arch/x86/mm/ioremap.c:524 __early_ioremap+0x85/0x18a()
[ 0.000000] Modules linked in:
[ 0.000000] Pid: 0, comm: swapper Not tainted 3.9.0-rc2-next-20130312 #283
[ 0.000000] Call Trace:
[ 0.000000] [<c1036992>] warn_slowpath_common+0x72/0xa0
[ 0.000000] [<c19cc176>] ? __early_ioremap+0x85/0x18a
[ 0.000000] [<c19cc176>] ? __early_ioremap+0x85/0x18a
[ 0.000000] [<c10369e2>] warn_slowpath_null+0x22/0x30
[ 0.000000] [<c19cc176>] __early_ioremap+0x85/0x18a
[ 0.000000] [<c19cc1ed>] ? __early_ioremap+0xfc/0x18a
[ 0.000000] [<c19cc412>] early_ioremap+0xd/0xf
[ 0.000000] [<c19fbba9>] dmi_present+0xf5/0x22a
[ 0.000000] [<c19fbd9d>] dmi_scan_machine+0xbf/0x11c
[ 0.000000] [<c19bd781>] setup_arch+0x4db/0xca0
[ 0.000000] [<c168105c>] ? printk+0x3d/0x3f
[ 0.000000] [<c19ba763>] start_kernel+0x7b/0x329
[ 0.000000] [<c19ba361>] i386_start_kernel+0x137/0x13a
[ 0.000000] ---[ end trace 4eaa2a86a8e2da22 ]---
[ 0.000000] DMI not present or invalid.

Introduced in "dmi_scan: refactor dmi_scan_machine(), {smbios,dmi}_present()"

Signed-off-by: Artem Savkov <[email protected]>
---
drivers/firmware/dmi_scan.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/firmware/dmi_scan.c b/drivers/firmware/dmi_scan.c
index 3439f59..40e940d 100644
--- a/drivers/firmware/dmi_scan.c
+++ b/drivers/firmware/dmi_scan.c
@@ -410,7 +410,7 @@ static void __init dmi_dump_ids(void)
printk(KERN_CONT "\n");
}

-static int __init dmi_present(const char *buf)
+static int __init dmi_present(const u8 *buf)
{
int smbios_ver;

--
1.8.1.5


2013-03-14 00:32:05

by Ben Hutchings

[permalink] [raw]
Subject: Re: [PATCH] dmi_scan: proper buf type in dmi_present()

On Wed, 2013-03-13 at 00:17 +0400, Artem Savkov wrote:
> buf needs to be unsigned in dmi_present(), otherwise dmi_base calculation
> doesn't work properly leading to a wraparound during dmi_ioremap:
>
> [ 0.000000] ------------[ cut here ]------------
> [ 0.000000] WARNING: at arch/x86/mm/ioremap.c:524 __early_ioremap+0x85/0x18a()
> [ 0.000000] Modules linked in:
> [ 0.000000] Pid: 0, comm: swapper Not tainted 3.9.0-rc2-next-20130312 #283
> [ 0.000000] Call Trace:
> [ 0.000000] [<c1036992>] warn_slowpath_common+0x72/0xa0
> [ 0.000000] [<c19cc176>] ? __early_ioremap+0x85/0x18a
> [ 0.000000] [<c19cc176>] ? __early_ioremap+0x85/0x18a
> [ 0.000000] [<c10369e2>] warn_slowpath_null+0x22/0x30
> [ 0.000000] [<c19cc176>] __early_ioremap+0x85/0x18a
> [ 0.000000] [<c19cc1ed>] ? __early_ioremap+0xfc/0x18a
> [ 0.000000] [<c19cc412>] early_ioremap+0xd/0xf
> [ 0.000000] [<c19fbba9>] dmi_present+0xf5/0x22a
> [ 0.000000] [<c19fbd9d>] dmi_scan_machine+0xbf/0x11c
> [ 0.000000] [<c19bd781>] setup_arch+0x4db/0xca0
> [ 0.000000] [<c168105c>] ? printk+0x3d/0x3f
> [ 0.000000] [<c19ba763>] start_kernel+0x7b/0x329
> [ 0.000000] [<c19ba361>] i386_start_kernel+0x137/0x13a
> [ 0.000000] ---[ end trace 4eaa2a86a8e2da22 ]---
> [ 0.000000] DMI not present or invalid.
>
> Introduced in "dmi_scan: refactor dmi_scan_machine(), {smbios,dmi}_present()"
>
> Signed-off-by: Artem Savkov <[email protected]>
Acked-by: Ben Hutchings <[email protected]>

Wow, this is a pretty obvious fix. Somehow this code did work for at
least two of us, so I suppose we were (un)lucky to have bit 7 clear on
all the critical bytes!

Ben.

> ---
> drivers/firmware/dmi_scan.c | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/drivers/firmware/dmi_scan.c b/drivers/firmware/dmi_scan.c
> index 3439f59..40e940d 100644
> --- a/drivers/firmware/dmi_scan.c
> +++ b/drivers/firmware/dmi_scan.c
> @@ -410,7 +410,7 @@ static void __init dmi_dump_ids(void)
> printk(KERN_CONT "\n");
> }
>
> -static int __init dmi_present(const char *buf)
> +static int __init dmi_present(const u8 *buf)
> {
> int smbios_ver;
>

--
Ben Hutchings
Humans are not rational beings; they are rationalising beings.


Attachments:
signature.asc (828.00 B)
This is a digitally signed message part