2013-04-10 08:59:51

by Richard Weinberger

Subject: [PATCH 1/2] x86,efi: Check max_size only if it is non-zero.

Some EFI implementations return always a MaximumVariableSize of 0,
check against max_size only if it is non-zero.
My Intel DQ67SW desktop board has such an implementation.

Signed-off-by: Richard Weinberger <[email protected]>
---
arch/x86/platform/efi/efi.c | 7 ++++++-
1 file changed, 6 insertions(+), 1 deletion(-)

diff --git a/arch/x86/platform/efi/efi.c b/arch/x86/platform/efi/efi.c
index c89c245..3f96a48 100644
--- a/arch/x86/platform/efi/efi.c
+++ b/arch/x86/platform/efi/efi.c
@@ -1018,7 +1018,12 @@ efi_status_t efi_query_variable_store(u32 attributes, unsigned long size)
if (status != EFI_SUCCESS)
return status;

- if (!storage_size || size > remaining_size || size > max_size ||
+ if (!max_size && remaining_size > size)
+ printk_once(KERN_ERR FW_BUG "Broken EFI implementation"
+ " is returning MaxVariableSize=0\n");
+
+ if (!storage_size || size > remaining_size ||
+ (max_size && size > max_size) ||
(remaining_size - size) < (storage_size / 2))
return EFI_OUT_OF_RESOURCES;

--
1.8.1.4

2013-04-10 09:00:35

by Richard Weinberger

Subject: [PATCH 2/2] x86,efi: Implement no_storage_paranoia parameter

Using this parameter one can disable the storage_size/2 check if
he is really sure that the UEFI does sane gc and fulfills the spec.

This parameter is useful if a devices uses more than 50% of the
storage by default.
The Intel DQSW67 desktop board is such an sucker for exmaple.

Signed-off-by: Richard Weinberger <[email protected]>
---
arch/x86/platform/efi/efi.c | 13 ++++++++++++-
1 file changed, 12 insertions(+), 1 deletion(-)

diff --git a/arch/x86/platform/efi/efi.c b/arch/x86/platform/efi/efi.c
index 3f96a48..1b0efb6 100644
--- a/arch/x86/platform/efi/efi.c
+++ b/arch/x86/platform/efi/efi.c
@@ -41,6 +41,7 @@
#include <linux/io.h>
#include <linux/reboot.h>
#include <linux/bcd.h>
+#include <linux/module.h>

#include <asm/setup.h>
#include <asm/efi.h>
@@ -71,6 +72,13 @@ static efi_system_table_t efi_systab __initdata;

unsigned long x86_efi_facility;

+static bool efivars_no_storage_paranoia;
+module_param_named(no_storage_paranoia, efivars_no_storage_paranoia, bool, 0644);
+MODULE_PARM_DESC(no_storage_paranoia, "Use this parameter only if you are very"
+ " sure that your EFI implemenation does sane garbage"
+ " collection and fulfills the UEFI spec."
+ " Otherwise your board may brick."
+ " See: http://mjg59.dreamwidth.org/#entry-22855");
/*
* Returns 1 if 'facility' is enabled, 0 otherwise.
*/
@@ -1023,7 +1031,10 @@ efi_status_t efi_query_variable_store(u32 attributes, unsigned long size)
" is returning MaxVariableSize=0\n");

if (!storage_size || size > remaining_size ||
- (max_size && size > max_size) ||
+ (max_size && size > max_size))
+ return EFI_OUT_OF_RESOURCES;
+
+ if (!efivars_no_storage_paranoia &&
(remaining_size - size) < (storage_size / 2))
return EFI_OUT_OF_RESOURCES;

--
1.8.1.4

2013-04-11 15:13:23

by Matt Fleming

Subject: Re: [PATCH 1/2] x86,efi: Check max_size only if it is non-zero.

On 10/04/13 09:59, Richard Weinberger wrote:
> Some EFI implementations return always a MaximumVariableSize of 0,
> check against max_size only if it is non-zero.
> My Intel DQ67SW desktop board has such an implementation.
>
> Signed-off-by: Richard Weinberger <[email protected]>
> ---
> arch/x86/platform/efi/efi.c | 7 ++++++-
> 1 file changed, 6 insertions(+), 1 deletion(-)

Applied, thanks!

2013-04-11 15:19:08

by Matt Fleming

Subject: Re: [PATCH 2/2] x86,efi: Implement no_storage_paranoia parameter

On 10/04/13 09:59, Richard Weinberger wrote:
> Using this parameter one can disable the storage_size/2 check if
> he is really sure that the UEFI does sane gc and fulfills the spec.
>
> This parameter is useful if a devices uses more than 50% of the
> storage by default.
> The Intel DQSW67 desktop board is such an sucker for exmaple.
>
> Signed-off-by: Richard Weinberger <[email protected]>
> ---
> arch/x86/platform/efi/efi.c | 13 ++++++++++++-
> 1 file changed, 12 insertions(+), 1 deletion(-)
>
> diff --git a/arch/x86/platform/efi/efi.c b/arch/x86/platform/efi/efi.c
> index 3f96a48..1b0efb6 100644
> --- a/arch/x86/platform/efi/efi.c
> +++ b/arch/x86/platform/efi/efi.c
> @@ -41,6 +41,7 @@
> #include <linux/io.h>
> #include <linux/reboot.h>
> #include <linux/bcd.h>
> +#include <linux/module.h>
>
> #include <asm/setup.h>
> #include <asm/efi.h>
> @@ -71,6 +72,13 @@ static efi_system_table_t efi_systab __initdata;
>
> unsigned long x86_efi_facility;
>
> +static bool efivars_no_storage_paranoia;
> +module_param_named(no_storage_paranoia, efivars_no_storage_paranoia, bool, 0644);

Could you rename this to 'efi_no_storage_paranoia' so that it isn't
confused with drivers/firmware/efivars.c? Also, it would be a good idea
to document this new kernel parameter in Documentation/kernel-parameters.

> +MODULE_PARM_DESC(no_storage_paranoia, "Use this parameter only if you are very"
> + " sure that your EFI implemenation does sane garbage"
> + " collection and fulfills the UEFI spec."
> + " Otherwise your board may brick."
> + " See: http://mjg59.dreamwidth.org/#entry-22855");

I think we should avoid including urls in the kernel source. The rest of
the description is fine.