Re: [PATCH] efi: READ_ONCE rng seed size before munmap

From: Ard Biesheuvel
Date: Mon Feb 17 2020 - 10:23:18 EST


On Mon, 17 Feb 2020 at 13:34, Jason A. Donenfeld <Jason@xxxxxxxxx> wrote:
>
> This function is consistent with using size instead of seed->size
> (except for one place that this patch fixes), but it reads seed->size
> without using READ_ONCE, which means the compiler might still do
> something unwanted. So, this commit simply adds the READ_ONCE
> wrapper.
>
> Signed-off-by: Jason A. Donenfeld <Jason@xxxxxxxxx>
> Cc: Ard Biesheuvel <ardb@xxxxxxxxxx>
> Cc: stable@xxxxxxxxxxxxxxx

Thanks Jason

I've queued this in efi/urgent with a fixes: tag rather than a cc:
stable, since it only applies clean to v5.4 and later. We'll need a
backport to 4.14 and 4.19 as well, which has a trivial conflict
(s/add_bootloader_randomness/add_device_randomness/) but we'll need to
wait for this patch to hit Linus's tree first.


> ---
> drivers/firmware/efi/efi.c | 4 ++--
> 1 file changed, 2 insertions(+), 2 deletions(-)
>
> diff --git a/drivers/firmware/efi/efi.c b/drivers/firmware/efi/efi.c
> index 621220ab3d0e..21ea99f65113 100644
> --- a/drivers/firmware/efi/efi.c
> +++ b/drivers/firmware/efi/efi.c
> @@ -552,7 +552,7 @@ int __init efi_config_parse_tables(void *config_tables, int count, int sz,
>
> seed = early_memremap(efi.rng_seed, sizeof(*seed));
> if (seed != NULL) {
> - size = seed->size;
> + size = READ_ONCE(seed->size);
> early_memunmap(seed, sizeof(*seed));
> } else {
> pr_err("Could not map UEFI random seed!\n");
> @@ -562,7 +562,7 @@ int __init efi_config_parse_tables(void *config_tables, int count, int sz,
> sizeof(*seed) + size);
> if (seed != NULL) {
> pr_notice("seeding entropy pool\n");
> - add_bootloader_randomness(seed->bits, seed->size);
> + add_bootloader_randomness(seed->bits, size);
> early_memunmap(seed, sizeof(*seed) + size);
> } else {
> pr_err("Could not map UEFI random seed!\n");
> --
> 2.25.0
>