From: Leo Unglaub Subject: dbupdate(8): pass dbupdates.tgz to signify -C To: tech@openbsd.org Date: Tue, 10 Nov 2026 12:37:30 +0100 Hey, while reading the dbupdate command to understand how it works, I noticed that signify(1) is called without specifying the file to verify: signify -C -q -p "${PUBKEY}" -x out-SHA256.sig As far as i understand it, when called with no file arguments, signify -C only checks the files listed in out-SHA256.sig. If out-SHA256.sig has a valid signature for an empty checksum list (or does not list dbupdates.tgz), signify -C exits 0 without verifying dbupdates.tgz. This is more of a theoretical issue than a real-world one, and I am not 100% sure if this is the preferred way to handle it, but passing dbupdates.tgz explicitly would ensure signify -C fails unless dbupdates.tgz is present in out-SHA256.sig and matches its checksum. It would also keep the verification logic consistent with syspatch(8) and sysupgrade(8), which explicitly pass the downloaded filenames to sha256(1) -C. diff --git a/usr.sbin/dbupdate/dbupdate.sh b/usr.sbin/dbupdate/dbupdate.sh index 10ee908b35c..debe2f63643 100644 --- a/usr.sbin/dbupdate/dbupdate.sh +++ b/usr.sbin/dbupdate/dbupdate.sh @@ -65,7 +65,7 @@ fi # Get the new file & verify the signing ftp https://ftp.openbsd.org/pub/OpenBSD/dbupdates/dbupdates.tgz > /dev/null -signify -C -q -p "${PUBKEY}" -x out-SHA256.sig +signify -C -q -p "${PUBKEY}" -x out-SHA256.sig dbupdates.tgz # extract tar xfz dbupdates.tgz If I misunderstood the signify(1) manual and specially the -C part, please ignore my email. Greetings Leo