WSS over HTTP/2 returns invalid RST_STREAM frame #1752
Comments
lws-team
added a commit
that referenced
this issue
Nov 4, 2019
rst_stream is almost always called from the h2 parser... in the case a ws-over-h2 is closing though, it can't reference the h2n parser sid context to get the sid to send... it's unrelated at that time. https://bugzilla.mozilla.org/show_bug.cgi?id=1590299 #1752
lws-team
added a commit
that referenced
this issue
Nov 4, 2019
rst_stream is almost always called from the h2 parser... in the case a ws-over-h2 is closing though, it can't reference the h2n parser sid context to get the sid to send... it's unrelated at that time. https://bugzilla.mozilla.org/show_bug.cgi?id=1590299 #1752
|
I pushed a patch on master and v3.2-stable that should solve the wrong sid on RST_STREAM when it's ws-over-h2 (only). I'm not sure that's the root cause of the problem, since it's already decided to close the wsi at that point and the bugzilla bug is complaining the ws link is closing unexpectedly. Also I opened a firefox bugzilla 9 months ago about ws-over-h2 behaviour I can still see from the lws testserver even with this patch, it may be related |
|
Btw https://libwebsockets.org/testserver is updated to current master and should demonstrate the RST_STREAM fix. It still passes all h2spec tests in CI as well. |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
This issue was first reported to Mozilla in bug 1590299.
The text was updated successfully, but these errors were encountered: