Buy commercial curl support from WolfSSL. We help you work
out your issues, debug your libcurl applications, use the API, port to new
platforms, add new features and more. With a team lead by the curl founder
himself.
Re: Crash libCurl 7.76.1 32-bit build in H3 transfers
- Contemporary messages sorted: [ by date ] [ by thread ] [ by subject ] [ by author ] [ by messages with attachments ]
From: Daniel Stenberg via curl-library <curl-library_at_cool.haxx.se>
Date: Thu, 6 May 2021 23:04:48 +0200 (CEST)
On Thu, 6 May 2021, Dmitry Karpov via curl-library wrote:
> So "size_t datalen" in libCurl is different from "uint64_t datalen" in the
> ngtcp2 library.
Fix pending: https://github.com/curl/curl/pull/7027
> The crash was so obvious, so it was a bit surprising how it was able to
> escape pre-release testing. Maybe H3, as experimental feature, wasn't tested
> enough on 32-builds?
We don't test h3 at all yet on any platform, I'm sorry to say! :-O
We still lack a test server/proxy for this purpose and we have at least three
rather nasty-sounding HTTP/3 related bugs in KNOWN_BUGS:
https://curl.se/docs/knownbugs.html#If_the_HTTP_3_server_closes_conn
I would prefer doing h3 tests with a h3 <=> h1 proxy front, a little like
stunnel but for h3, so that we could still serve the core stuff with our
existing h1 test server. But it's not a hard requirement.
I would love to get this going better and further but I've not had time for it
lately and no one has stepped up wanting to sponsor my work on h3.
> 32-bit builds are still widely used on embedded devices, so it will help a
> lot if you guys look at this issue and look for other potential pitfalls
> like this when switching from 64-bit builds to 32-bit builds.
Once we get a few h3-tests landed, I think it would make sense and be good to
have at least one set of them running on a 32 bit system, yes. The availablity
of convenient CI builds on 32 bit systems might be a bit limited though.
Date: Thu, 6 May 2021 23:04:48 +0200 (CEST)
On Thu, 6 May 2021, Dmitry Karpov via curl-library wrote:
> So "size_t datalen" in libCurl is different from "uint64_t datalen" in the
> ngtcp2 library.
Fix pending: https://github.com/curl/curl/pull/7027
> The crash was so obvious, so it was a bit surprising how it was able to
> escape pre-release testing. Maybe H3, as experimental feature, wasn't tested
> enough on 32-builds?
We don't test h3 at all yet on any platform, I'm sorry to say! :-O
We still lack a test server/proxy for this purpose and we have at least three
rather nasty-sounding HTTP/3 related bugs in KNOWN_BUGS:
https://curl.se/docs/knownbugs.html#If_the_HTTP_3_server_closes_conn
I would prefer doing h3 tests with a h3 <=> h1 proxy front, a little like
stunnel but for h3, so that we could still serve the core stuff with our
existing h1 test server. But it's not a hard requirement.
I would love to get this going better and further but I've not had time for it
lately and no one has stepped up wanting to sponsor my work on h3.
> 32-bit builds are still widely used on embedded devices, so it will help a
> lot if you guys look at this issue and look for other potential pitfalls
> like this when switching from 64-bit builds to 32-bit builds.
Once we get a few h3-tests landed, I think it would make sense and be good to
have at least one set of them running on a 32 bit system, yes. The availablity
of convenient CI builds on 32 bit systems might be a bit limited though.
-- / daniel.haxx.se | Commercial curl support up to 24x7 is available! | Private help, bug fixes, support, ports, new features | https://www.wolfssl.com/contact/ ------------------------------------------------------------------- Unsubscribe: https://cool.haxx.se/list/listinfo/curl-library Etiquette: https://curl.se/mail/etiquette.htmlReceived on 2021-05-06