Bug 3298

Summary:	Conversion of SSH Ed25519 private key associated to a passphrase into PEM format \| 'unsupported key type ED25519'
Product:	Portable OpenSSH	Reporter:	ricky.tigg
Component:	ssh-keygen	Assignee:	Assigned to nobody <unassigned-bugs>
Status:	CLOSED DUPLICATE
Severity:	normal	CC:	djm
Priority:	P5
Version:	8.5p1
Hardware:	amd64
OS:	Linux

Description ricky.tigg 2021-04-13 22:35:57 AEST

OS: Fedora 34; Component: openssh-8.5p1-2.fc34.x86_64

Failed: conversion of SSH private Ed25519 key associated to a passphrase into PEM format in order to be compatible for use with OpenSSL.

Prerequisite: Creation of an SSH Ed25519 key pair – using 'ssh-keygen' – associated to a passphrase.

Task: export file to PEM format:

$ ssh-keygen -e -f id_ed25519 -m pem -out id_ed25519.pem
do_convert_to_pem: unsupported key type ED25519

Reference: Patch providing basic support of Ed25519 keys via PKCS#11 tokens: https://github.com/openssh/openssh-portable/pull/230

No mention in manual 'man ssh-keygen' regards non-compatibility of OpenSSH-specific format – RFC4716 – with OpenSSL, i could reasonably assume that OpenSSL might support it.

Comment 1 Damien Miller 2021-04-30 14:04:50 AEST


*** This bug has been marked as a duplicate of bug 3195 ***

Comment 2 Damien Miller 2022-02-25 13:58:33 AEDT

closing bugs resolved before openssh-8.9

Comment 3 ricky.tigg 2022-02-25 20:20:24 AEDT

Sent (chronological order): "marked as a duplicate of bug", "closing bugs resolved before openssh-8.9".

I have trouble to extract a coherence of such chronology. Well, the report, whose present report is a duplicate of, is in state open. Both reports were published against Openssh versions older than version 8.9.

Therefore how can the present report be closed as resolved as version previous to 8.9, while the other is not.

Coherence would lead 
- either to close the present as duplicate for unique reason, therefore without being associated to a tier reason such as resolved as version previous to version 8.9
- or to close the present as resolved as version previous to version 8.9 for unique reason, therefore without being associated to a tier reason such as duplicate, and closing as well those reports which it is a duplicate of.