"brian m. carlson" [off-list ref] writes:
That message comes from OpenSSH. I've seen it quite frequently in
various other (non-Git) cases. I think it's fair for us to exit
unsuccessfully if OpenSSH exits unsuccessfully in this case. For
example, an attacker could try to tamper with the connection and send
additional data, which OpenSSH would detect and exit unsuccessfully for.
We also in general need to detect truncation attacks, which OpenSSH will
do for us here.
It's possible that if there's an older version of OpenSSH being used,
that the problem happens to be related to a bug of some sort. There
were some versions which had various bugs that could be triggered by a
rekey, which, if the threshold is set low enough, could be the cause of
this particular problem.
I think the fact that it's not being seen with HTTPS is the ultimate
clue here.
This suspiously sounds familiar. Asking for "close ssh connection"
in the lore.kernel.org archive finds this:
https://lore.kernel.org/git/YKTg8nYjSGpKbq8W@coredump.intra.peff.net/
On Jul 26, 2021, at 22:20, Junio C Hamano [off-list ref] wrote:
This suspiously sounds familiar. Asking for "close ssh connection"
in the lore.kernel.org archive finds this:
https://lore.kernel.org/git/YKTg8nYjSGpKbq8W@coredump.intra.peff.net/
The described symptoms match what I am seeing. It makes sense that ancient hardware would bring out the bug – triggering the server timeout with a long local computation.
Switching to protocol 0 successfully works around the issue. This concludes my interest in the error. I suggest reviving a discussion of Jeff King's patch here:
https://lore.kernel.org/git/YKU4mbq%2Fam%2FliAN4@coredump.intra.peff.net/
Evan
On Mon, Jul 26, 2021 at 11:05:27PM -0400, Evan Miller wrote:
The described symptoms match what I am seeing. It makes sense that
ancient hardware would bring out the bug – triggering the server
timeout with a long local computation.
Switching to protocol 0 successfully works around the issue. This
concludes my interest in the error. I suggest reviving a discussion of
Jeff King's patch here:
https://lore.kernel.org/git/YKU4mbq%2Fam%2FliAN4@coredump.intra.peff.net/
Yeah, everything you've described is consistent with the problem being
fixed there. Good news, then: the patch is already in Git's "master"
branch, and will be part of the next release (which should be v2.33.0).
If you can reproduce the problem at will and building from master isn't
too hard, it would be nice to know if that indeed fixes it for you.
-Peff