| Age | Commit message (Collapse) | Author | Files | Lines |
|---|
|
These classes accidentally overlapped in naming with the nghttp2
classes. This is because this class, unlike most nghttp2 classes
doesn't end in _ptr for a type. This changes the class name to add a
_ex to differentiate the two classes, and avoid a warning in clang.
Tested: Unit tests pass. Code only used in unit test.
Change-Id: I91a6982264df69bc65166ab38feddc21f72cd223
Signed-off-by: Ed Tanous <ed@tanous.net>
|
|
This allows http2 connections to now host authenticated endpoints.
Note, this work exposed that the http2 path was not calling
preparePayload() and responses were therefore missing the
Content-Length header. preparePayload is now called, and Content-Length
is added to the unit tests.
This commit also allows a full Redfish Service Validator test to pass
entirely using HTTP2.
Tested: Unit tests pass.
Curl /redfish/v1/Managers/bmc/LogServices/Journal/Entries
(which returns a payload larger than 16kB) succeeds and returns the
data.
Manually logging in with both basic and session authentication succeeds
over http2.
A modified Redfish-Service-Validator, changed to use httpx as its
backend, (thus using http2) succeeds.
Change-Id: I956f3ff8f442e9826312c6147d7599ab136a8e7c
Signed-off-by: Ed Tanous <ed@tanous.net>
|
|
This unit test currently only tests a simple connect and settings frame
transfer, but should form the basis for more complex testing in the
future.
Tested: Unit tests pass
Change-Id: Ieb803dbe490129ec5fe99fb3d4505a06202e282e
Signed-off-by: Ed Tanous <ed@tanous.net>
|
|
Using the mem_send methods of nghttp2 can reduce the amount of buffering
we need to do. This is recommended by the nghttp2 docs.
Tested: Enabled experimental-http. Curl succeeds on /redfish/v1, and
shows:
* using HTTP/2
* [HTTP/2] [1] OPENED stream for https://localhost:18080/redfish/v1
Change-Id: I287d8c956f064d244116fac853055a17fca915a2
Signed-off-by: Ed Tanous <ed@tanous.net>
|
|
std::format is a much more modern logging solution, and gives us a lot
more flexibility, and better compile times when doing logging.
Unfortunately, given its level of compile time checks, it needs to be a
method, instead of the stream style logging we had before. This
requires a pretty substantial change. Fortunately, this change can be
largely automated, via the script included in this commit under
scripts/replace_logs.py. This is to aid people in moving their
patchsets over to the new form in the short period where old patches
will be based on the old logging. The intention is that this script
eventually goes away.
The old style logging (stream based) looked like.
BMCWEB_LOG_DEBUG << "Foo " << foo;
The new equivalent of the above would be:
BMCWEB_LOG_DEBUG("Foo {}", foo);
In the course of doing this, this also cleans up several ignored linter
errors, including macro usage, and array to pointer deconstruction.
Note, This patchset does remove the timestamp from the log message. In
practice, this was duplicated between journald and bmcweb, and there's
no need for both to exist.
One design decision of note is the addition of logPtr. Because the
compiler can't disambiguate between const char* and const MyThing*, it's
necessary to add an explicit cast to void*. This is identical to how
fmt handled it.
Tested: compiled with logging meson_option enabled, and launched bmcweb
Saw the usual logging, similar to what was present before:
```
[Error include/webassets.hpp:60] Unable to find or open /usr/share/www/ static file hosting disabled
[Debug include/persistent_data.hpp:133] Restored Session Timeout: 1800
[Debug redfish-core/include/event_service_manager.hpp:671] Old eventService config not exist
[Info src/webserver_main.cpp:59] Starting webserver on port 18080
[Error redfish-core/include/event_service_manager.hpp:1301] inotify_add_watch failed for redfish log file.
[Info src/webserver_main.cpp:137] Start Hostname Monitor Service...
```
Signed-off-by: Ed Tanous <ed@tanous.net>
Change-Id: I86a46aa2454be7fe80df608cb7e5573ca4029ec8
|
|
HTTP/2 gives a number of optimizations, while keeping support for the
protocol. HTTP/2 support was recently added to the Redfish
specification. The largest performance increase in bmc usage is likely
header compression. Almost all requests reuse the same header values,
so the hpack based compression scheme in HTTP/2 allows OpenBMC to be
more efficient as a transport, and has the potential to significantly
reduce the number of bytes we're sending on the wire.
This commit adds HTTP2 support to bmcweb through nghttp2 library. When
static linked into bmcweb, this support adds 53.4KB to the bmcweb binary
size. nghttp2 is available in meta-oe already.
Given the experimental nature of this option, it is added under the
meson option "experimental-http2" and disabled by default. The hope is
to enable it at some point in the future.
To accomplish the above, there a new class, HTTP2Connection is created.
This is intended to isolate HTTP/2 connections code from HttpConnection
such that it is far less likely to cause bugs, although it does
duplicate about 20 lines of code (async_read_some, async_write_some,
buffers, etc). This seems worth it for the moment.
In a similar way to Websockets, when an HTTP/2 connection is detected
through ALPN, the HTTP2Connection class will be instantiated, and the
socket object passed to it, thus allowing the Connection class to be
destroyed, and the HTTP2Connection to take over for the user.
Tested: Redfish service validator passes with option enabled
With option disabled
GET /redfish/v1 in curl shows ALPN non negotiation, and fallback to
http1.1
With the option enable
GET /redfish/v1 in curl shows ALPN negotiates to HTTP2
Change-Id: I7839e457e0ba918b0695e04babddd0925ed3383c
Signed-off-by: Ed Tanous <edtanous@google.com>
|
