On Fri, May 29, 2026 at 02:16:55AM -0500, Simon Glass wrote: > Hi Philippe, > > On 2026-05-28T08:19:01, Philippe Reynes <[email protected]> > wrote: > > ecdsa: fix support of secp521r1 > > > > Current implementation of ecdsa only supports key len aligned on > > 8 bits. But the curve secp521r1 uses a key of 521 bits which is not > > aligned on 8 bits. In this commit, we update the keys management > > for ecdsa to support keys that are not aligned on 8 bits. > > > > Reviewed-by: Raymond Mao <[email protected]> > > Signed-off-by: Philippe Reynes <[email protected]> > > > > lib/ecdsa/ecdsa-libcrypto.c | 65 > > +++++++++++++++++++++++++++++++++++++++++++-- > > lib/ecdsa/ecdsa-verify.c | 65 > > ++++++++++++++++++++++++++++++++++++++++++--- > > lib/fdt-libcrypto.c | 2 +- > > tools/image-sig-host.c | 7 +++++ > > 4 files changed, 132 insertions(+), 7 deletions(-) > > Reviewed-by: Simon Glass <[email protected]> > > questions / nits below > > > diff --git a/lib/ecdsa/ecdsa-libcrypto.c b/lib/ecdsa/ecdsa-libcrypto.c > > @@ -41,10 +43,26 @@ struct ecdsa_public_key { > > +static char *memdup(char *buf, size_t size) > > +{ > > + char *dup; > > + > > + dup = malloc(size); > > + if (dup) > > + memcpy(dup, buf, size); > > + > > + return dup; > > +} > > Please match the U-Boot signature: void *memdup(const void *src, > size_t len) (see include/linux/string.h). Making buf const lets the > call sites lose the (char *) casts. Also note that fdt_get_key() now > hands back malloc'd buffers - please spell out the caller's free > responsibility in a function comment.
Simon, I've noticed of late that you're providing a lot of Reviewed-by tags in the same email that you provide a lot of substantial and non-trivial feedback. This would normally negate the Reviewed-by tag, so can you please refrain from providing both the tag and feedback in the same email? Thanks. -- Tom
signature.asc
Description: PGP signature
