[PATCH 4.14] efi: random: Properly limit the size of the random seed

[PATCH 4.14] efi: random: Properly limit the size of the random seed

[Ben Hutchings]

[-- Attachment #1: Type: text/plain, Size: 1514 bytes --]

Commit be36f9e7517e ("efi: READ_ONCE rng seed size before munmap")
added a READ_ONCE() and also changed the call to
add_bootloader_randomness() to use the local size variable.  Neither
of these changes was actually needed and this was not backported to
the 4.14 stable branch.

Commit 161a438d730d ("efi: random: reduce seed size to 32 bytes")
reverted the addition of READ_ONCE() and added a limit to the value of
size.  This depends on the earlier commit, because size can now differ
from seed->size, but it was wrongly backported to the 4.14 stable
branch by itself.

Apply the missing change to the add_bootloader_randomness() parameter
(except that here we are still using add_device_randomness()).

Fixes: 700485f70e50 ("efi: random: reduce seed size to 32 bytes")
Signed-off-by: Ben Hutchings <ben@decadent.org.uk>
---
 drivers/firmware/efi/efi.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/firmware/efi/efi.c b/drivers/firmware/efi/efi.c
index ed981f5e29ae..cc64869d8420 100644
--- a/drivers/firmware/efi/efi.c
+++ b/drivers/firmware/efi/efi.c
@@ -541,7 +541,7 @@ int __init efi_config_parse_tables(void *config_tables, int count, int sz,
 			seed = early_memremap(efi.rng_seed,
 					      sizeof(*seed) + size);
 			if (seed != NULL) {
-				add_device_randomness(seed->bits, seed->size);
+				add_device_randomness(seed->bits, size);
 				early_memunmap(seed, sizeof(*seed) + size);
 				pr_notice("seeding entropy pool\n");
 			} else {

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 833 bytes --]

Re: [PATCH 4.14] efi: random: Properly limit the size of the random seed

[Ben Hutchings]

[-- Attachment #1: Type: text/plain, Size: 1135 bytes --]

On Thu, 2022-12-01 at 00:47 +0100, Ben Hutchings wrote:
> Commit be36f9e7517e ("efi: READ_ONCE rng seed size before munmap")
> added a READ_ONCE() and also changed the call to
> add_bootloader_randomness() to use the local size variable.  Neither
> of these changes was actually needed and this was not backported to
> the 4.14 stable branch.
> 
> Commit 161a438d730d ("efi: random: reduce seed size to 32 bytes")
> reverted the addition of READ_ONCE() and added a limit to the value of
> size.  This depends on the earlier commit, because size can now differ
> from seed->size, but it was wrongly backported to the 4.14 stable
> branch by itself.
> 
> Apply the missing change to the add_bootloader_randomness() parameter
> (except that here we are still using add_device_randomness()).
[...]

This made me wonder: shouldn't commit 18b915ac6b0a ("efi/random: Treat
EFI_RNG_PROTOCOL output as bootloader randomness") be applied to these
older stable branches?  Without that, the EFI RNG can't be distrusted
if necessary.

Ben.

-- 
Ben Hutchings
compatible: Gracefully accepts erroneous data from any source

[-- Attachment #2: This is a digitally signed message part --]
[-- Type: application/pgp-signature, Size: 833 bytes --]

Re: [PATCH 4.14] efi: random: Properly limit the size of the random seed

[Greg KH]

On Thu, Dec 01, 2022 at 12:53:03AM +0100, Ben Hutchings wrote:
> On Thu, 2022-12-01 at 00:47 +0100, Ben Hutchings wrote:
> > Commit be36f9e7517e ("efi: READ_ONCE rng seed size before munmap")
> > added a READ_ONCE() and also changed the call to
> > add_bootloader_randomness() to use the local size variable.  Neither
> > of these changes was actually needed and this was not backported to
> > the 4.14 stable branch.
> > 
> > Commit 161a438d730d ("efi: random: reduce seed size to 32 bytes")
> > reverted the addition of READ_ONCE() and added a limit to the value of
> > size.  This depends on the earlier commit, because size can now differ
> > from seed->size, but it was wrongly backported to the 4.14 stable
> > branch by itself.
> > 
> > Apply the missing change to the add_bootloader_randomness() parameter
> > (except that here we are still using add_device_randomness()).
> [...]
> 
> This made me wonder: shouldn't commit 18b915ac6b0a ("efi/random: Treat
> EFI_RNG_PROTOCOL output as bootloader randomness") be applied to these
> older stable branches?  Without that, the EFI RNG can't be distrusted
> if necessary.

Makes sense, want to send a backport on top of this one as the original
will not work?

greg k-h

Re: [PATCH 4.14] efi: random: Properly limit the size of the random seed

[Greg KH]

On Thu, Dec 01, 2022 at 12:47:22AM +0100, Ben Hutchings wrote:
> Commit be36f9e7517e ("efi: READ_ONCE rng seed size before munmap")
> added a READ_ONCE() and also changed the call to
> add_bootloader_randomness() to use the local size variable.  Neither
> of these changes was actually needed and this was not backported to
> the 4.14 stable branch.
> 
> Commit 161a438d730d ("efi: random: reduce seed size to 32 bytes")
> reverted the addition of READ_ONCE() and added a limit to the value of
> size.  This depends on the earlier commit, because size can now differ
> from seed->size, but it was wrongly backported to the 4.14 stable
> branch by itself.
> 
> Apply the missing change to the add_bootloader_randomness() parameter
> (except that here we are still using add_device_randomness()).
> 
> Fixes: 700485f70e50 ("efi: random: reduce seed size to 32 bytes")
> Signed-off-by: Ben Hutchings <ben@decadent.org.uk>
> ---
>  drivers/firmware/efi/efi.c | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)

Both now queued up, thanks,

greg k-h