On Mon, Jul 22, 2024 at 08:19:41AM -0400, Stefan Berger wrote:
> On 7/17/24 12:17, Lukas Wunner wrote:
> > On Tue, Mar 16, 2021 at 05:07:32PM -0400, Stefan Berger wrote:
> > > +/*
> > > + * Get the r and s components of a signature from the X509 certificate.
> > > + */
> > > +static int ecdsa_get_signature_rs(u64 *dest, size_t hdrlen, unsigned char tag,
> > > + const void *value, size_t vlen, unsigned int ndigits)
> > > +{
> > > + size_t keylen = ndigits * sizeof(u64);
> > > + ssize_t diff = vlen - keylen;
> > > + const char *d = value;
> > > + u8 rs[ECC_MAX_BYTES];
> > > +
> > > + if (!value || !vlen)
> > > + return -EINVAL;
> > > +
> > > + /* diff = 0: 'value' has exacly the right size
> > > + * diff > 0: 'value' has too many bytes; one leading zero is allowed that
> > > + * makes the value a positive integer; error on more
> > > + * diff < 0: 'value' is missing leading zeros, which we add
> > > + */
> > > + if (diff > 0) {
> > > + /* skip over leading zeros that make 'value' a positive int */
> > > + if (*d == 0) {
> > > + vlen -= 1;
> > > + diff--;
> > > + d++;
> > > + }
> > > + if (diff)
> > > + return -EINVAL;
> > > + }
> > > + if (-diff >= keylen)
> > > + return -EINVAL;
> >
> > There's an oddity in the above-quoted function. The check ...
> >
> > + if (-diff >= keylen)
> > + return -EINVAL;
> >
> > ... seems superfluous.
>
> You're right, this check is not necessary.
After staring at the code a little longer I've realized that
the purpose of this if-clause is likely to check for a signed
integer overflow. So it *does* seem to have a purpose,
but it's quite subtle and not very obvious.
I've provisionally added the (untested) commit below to my
development branch to make it more obvious what's going on.
Using check_sub_overflow() might be an alternative.
I want to ask mips maintainers first whether signed integer
overflows can really cause an exception on their arch
as commit 36ccf1c0e391 suggests, despite -fno-strict-overflow...
-- >8 --
Subject: [PATCH] crypto: ecdsa - Avoid signed integer overflow on signature
decoding
When extracting a signature component R or S from an ASN.1-encoded
integer, ecdsa_get_signature_rs() subtracts the expected length
"bufsize" from the ASN.1 length "vlen" (both of unsigned type size_t)
and stores the result in "diff" (of signed type ssize_t).
This results in a signed integer overflow if vlen > SSIZE_MAX + bufsize.
The kernel is compiled with -fno-strict-overflow, which implies -fwrapv,
meaning signed integer overflow is not undefined behavior. And the
function does check for overflow:
if (-diff >= bufsize)
return -EINVAL;
However that's not very readable and may trigger a false-positive with
CONFIG_UBSAN_SIGNED_WRAP=y. It also seems that certain Mips CPUs may
raise an exception regardless of -fno-strict-overflow (see do_ov() in
arch/mips/kernel/traps.c).
Avoid by comparing the two unsigned variables directly and erroring out
if "vlen" is too large.
Signed-off-by: Lukas Wunner <lukas@xxxxxxxxx>
---
crypto/ecdsa.c | 17 ++++-------------
1 file changed, 4 insertions(+), 13 deletions(-)
diff --git a/crypto/ecdsa.c b/crypto/ecdsa.c
index 08c2c76..0cead9b 100644
--- a/crypto/ecdsa.c
+++ b/crypto/ecdsa.c
@@ -36,29 +36,20 @@ static int ecdsa_get_signature_rs(u64 *dest, size_t hdrlen, unsigned char tag,
const void *value, size_t vlen, unsigned int ndigits)
{
size_t bufsize = ndigits * sizeof(u64);
- ssize_t diff = vlen - bufsize;
const char *d = value;
- if (!value || !vlen)
+ if (!value || !vlen || vlen > bufsize + 1)
return -EINVAL;
- /* diff = 0: 'value' has exacly the right size
- * diff > 0: 'value' has too many bytes; one leading zero is allowed that
- * makes the value a positive integer; error on more
- * diff < 0: 'value' is missing leading zeros
- */
- if (diff > 0) {
+ if (vlen > bufsize) {
/* skip over leading zeros that make 'value' a positive int */
if (*d == 0) {
vlen -= 1;
- diff--;
d++;
- }
- if (diff)
+ } else {
return -EINVAL;
+ }
}
- if (-diff >= bufsize)
- return -EINVAL;
ecc_digits_from_bytes(d, vlen, dest, ndigits);
--
2.43.0
