Project

General

Profile

Activity

From 2021-08-05 to 2021-08-11

2021-08-11

21:11 Bug #3093: Chrome 92, HTTP/2, fcgi, mutiple puts no response
If you do, please post a link to the issue, so that I can track it, too.
RFC 7540 (on which Google participated) s...
gstrauss
20:40 Bug #3093: Chrome 92, HTTP/2, fcgi, mutiple puts no response
This have to be a bug in Chrome. In h2_recv_headers, when the 9th header frame is received, h2c->sent_settings is set... stenvaag
20:00 Bug #3093: Chrome 92, HTTP/2, fcgi, mutiple puts no response
lighttpd HTTP/2 support limits SETTINGS_MAX_CONCURRENT_STREAMS to 8 and communicates this in the initial SETTINGS sen... gstrauss
14:18 Bug #3093: Chrome 92, HTTP/2, fcgi, mutiple puts no response
I have tried to locate the error today. My lastest finding is that when the number of put requests passes 8, gw_write... stenvaag
20:52 Bug #3084: Memory fragmentation with HTTP/2 enabled
For the PROPFIND/POST issue, could you try with @debug.log-request-header = "enable"@ and the patch further above to ... gstrauss
20:47 Bug #3084: Memory fragmentation with HTTP/2 enabled
I can not reproduce what you are seeing with @h2spec@. What is the target of @h2spec@ that you are using?
If I reme...
gstrauss
13:14 Bug #3084: Memory fragmentation with HTTP/2 enabled
Running h2spec on version 1.4.59, I get 2 errors.
Running h2spec on current git version, I get 6 errors.
Can yo...
flynn
12:35 Bug #3084: Memory fragmentation with HTTP/2 enabled
> Any hints how to debug this?
Firefox has some nice debugging tools in the Web Developer Tools, but I am by no mean...
gstrauss
12:18 Bug #3084: Memory fragmentation with HTTP/2 enabled
Using @debug.log-request-header = "enable"@ the biggest difference between firefox and chromium is te http-header use... flynn
11:55 Bug #3084: Memory fragmentation with HTTP/2 enabled
It works with chromium, it seems a pure firefox problem, maybe a detail in the HTTP/2 headers, flags or handling.
...
flynn
11:49 Bug #3084: Memory fragmentation with HTTP/2 enabled
If it helps, take those 84 chars. Add @\r\n@ after each @>@, including the final line. Change @<@ to ... gstrauss
11:06 Bug #3084: Memory fragmentation with HTTP/2 enabled
More progress:
- curl on cmdline works, if I remove the Content-Length header
- firefox seems to cause the problem,...
flynn
10:17 Bug #3084: Memory fragmentation with HTTP/2 enabled
I tried your request, but did not reproduce the issue. Things worked fine for me after changing @Content-Length: 122... gstrauss
09:54 Bug #3084: Memory fragmentation with HTTP/2 enabled
Request body is:
<?xml version="1.0" encoding="utf-8"?><D:propfind xmlns:D="DAV:"><D:prop><D:current-user-principal/...
flynn
09:40 Bug #3084: Memory fragmentation with HTTP/2 enabled
Most lighttpd backends use @src/http_cgi.c:http_cgi_headers()@, including mod_cgi, mod_fastcgi, and mod_scgi, so if o... gstrauss
09:24 Bug #3084: Memory fragmentation with HTTP/2 enabled
Partial success: with minimzed setup I see PROPFIND as request method with your php example above.
I have to searc...
flynn
09:18 Bug #3084: Memory fragmentation with HTTP/2 enabled
At the moment I try to reduce/minimize my setup ...
The PROPFIND request is produced by infcloud, a pure javascrip...
flynn
09:16 Bug #3084: Memory fragmentation with HTTP/2 enabled
Let's try a different approach. Maybe the bug is elsewhere and PROPFIND/POST is how you are observing it.
In your r...
gstrauss
08:57 Bug #3084: Memory fragmentation with HTTP/2 enabled
By any chance are you running multiple lighttpd instances, maybe proxying one to another, and maybe one instance is r... gstrauss
08:49 Bug #3084: Memory fragmentation with HTTP/2 enabled
I am not seeing an issue with HTTP/2 and PHP served via FastCGI.
foo.php...
gstrauss
08:36 Bug #3084: Memory fragmentation with HTTP/2 enabled
Thanks for continuing to look and narrowing things down. The backends all share a substantial amount of code with sl... gstrauss
08:24 Bug #3084: Memory fragmentation with HTTP/2 enabled
I tested more backend protocols
* scgi is ok, PROPFIND is reported as REQUEST_METHOD.
* reverse proxy is ok too
...
flynn
07:57 Bug #3084: Memory fragmentation with HTTP/2 enabled
In tests like this I always run lighttpd in foreground (option -D), to get all log log messages immeditely.
I can ...
flynn

2021-08-10

20:27 Bug #3093 (Invalid): Chrome 92, HTTP/2, fcgi, mutiple puts no response
Once you track the issue down, if the issue might be something others run into, please post a summary of what the pro... gstrauss
20:21 Bug #3093: Chrome 92, HTTP/2, fcgi, mutiple puts no response
Sorry about this. Setting fastcgi.debug = 3 suggests that fcgi is called. The bug have to be in my system. stenvaag
18:53 Bug #3093: Chrome 92, HTTP/2, fcgi, mutiple puts no response
> When using Chrome 92 my fcgi backend is not called.
The logs suggest otherwise. The logs you shared show
@2021...
gstrauss
12:13 Bug #3093: Chrome 92, HTTP/2, fcgi, mutiple puts no response
When disabling HTTP/2 there is no problem with Chrome. stenvaag
12:03 Bug #3093 (Fixed): Chrome 92, HTTP/2, fcgi, mutiple puts no response
I use Lighttpd (self compiled) with a self developed fcgi backend. I have used this in a production environment for 1... stenvaag

2021-08-08

20:59 Feature #2975: safe_memcmp new function proposal
FYI: a digest-specific constant time comparison was added in commit commit:0e749c1c and included in lighttpd 1.4.55 (... gstrauss

2021-08-06

18:33 Bug #3084: Memory fragmentation with HTTP/2 enabled
Did you install the clean build of lighttpd and restart lighttpd? Or is @lighttpd -tt@ failing and lighttpd not actu... gstrauss
18:32 Bug #3084: Memory fragmentation with HTTP/2 enabled
Yes, the CGI script gets the PROPFIND... gstrauss
18:19 Bug #3084: Memory fragmentation with HTTP/2 enabled
From my knowledge of the problem your test will succeed, because you do not check whether the PROPFIND method reache... flynn
17:44 Bug #3084: Memory fragmentation with HTTP/2 enabled
FYI: I have tested with a simple CGI script which prints the environment, and forced HTTP/2 and request method PROPFI... gstrauss
17:31 Bug #3084: Memory fragmentation with HTTP/2 enabled
On PROPFIND requests I still get 400 with the git version, I tested this morning, I try to debug later ... flynn
12:38 Bug #3084: Memory fragmentation with HTTP/2 enabled
> But I cannot test until the PROPFIND issue is solved with HTTP/2 ... (the code around the suggested workaround abov... gstrauss
07:29 Bug #3084: Memory fragmentation with HTTP/2 enabled
Sorry for the delay...
With a clean build of the current git version I do not get anymore a segfault on testing th...
flynn
 

Also available in: Atom