[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20171214102513.GA2371@dhcp-128-65.nay.redhat.com>
Date: Thu, 14 Dec 2017 18:25:13 +0800
From: Dave Young <dyoung@...hat.com>
To: David Howells <dhowells@...hat.com>
Cc: keyrings@...r.kernel.org, linux-kernel@...r.kernel.org,
kexec@...ts.infradead.org, David Woodhouse <dwmw2@...radead.org>
Subject: Re: [PATCH] certs: always use secondary keyring first if possible
On 11/18/17 at 12:47pm, Dave Young wrote:
> Commit d3bfe84129f6 introduced secondary_trusted_keys keyring, current
> users of verify_pkcs7_signature are below:
> net/wireless/reg.c : uses its own trusted_keys
> kernel/module_signing.c : pass NULL trusted_keys
> crypto/asymmetric_keys/verify_pefile.c : pass NULL trusted_keys
>
> For both module and pefile verification, there is no reason to use builtin
> keys only. Actually in Fedora kernel module signing code passes 1UL, but
> kexec code does not pass 1UL for pefile verification thus we have below bug
> https://bugzilla.redhat.com/show_bug.cgi?id=1470995
>
> Drop the hard code 1UL checking so that pefile verification can use
> secondary keyring as well.
>
> Signed-off-by: Dave Young <dyoung@...hat.com>
> ---
> certs/system_keyring.c | 2 --
> 1 file changed, 2 deletions(-)
>
> --- linux-x86.orig/certs/system_keyring.c
> +++ linux-x86/certs/system_keyring.c
> @@ -229,8 +229,6 @@ int verify_pkcs7_signature(const void *d
> goto error;
>
> if (!trusted_keys) {
> - trusted_keys = builtin_trusted_keys;
> - } else if (trusted_keys == (void *)1UL) {
> #ifdef CONFIG_SECONDARY_TRUSTED_KEYRING
> trusted_keys = secondary_trusted_keys;
> #else
Another ping.
If the (-1UL) is really needed, below file need update to use it
But I think it is ugly..
crypto/asymmetric_keys/verify_pefile.c
Thanks
Dave
Powered by blists - more mailing lists