-
Notifications
You must be signed in to change notification settings - Fork 706
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
ktls: add sendfile #4186
ktls: add sendfile #4186
Conversation
0ecc5f1
to
10cb2b9
Compare
-Werror-implicit-function-declaration | ||
-Werror-implicit-function-declaration -Wno-unused-variable |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Related to #4187
My new S2N_LINUX_SENDFILE feature flag was failing in make (but not in cmake) with "unused variable", and so was S2N_KTLS_SUPPORTED. Probably best just to ignore unused variables, since they have nothing to do with feature support.
But if we think that's too risky, I can just fix both feature tests to not have unused variables. However, I suspect we'll keep making this mistake.
10cb2b9
to
d0c1e14
Compare
Resolved issues:
Resolves #4171
Description of changes:
Add a wrapper method for sendfile.
There are actually multiple definitions of sendfile:
ssize_t sendfile(int out_fd, int in_fd, off_t *_Nullable offset, size_t count);
int sendfile(int fd, int s, off_t offset, size_t nbytes, struct sf_hdtr *hdtr, off_t *sbytes, int flags);
int sendfile(int fd, int s, off_t offset, off_t *len, struct sf_hdtr *hdtr, int flags);
We only need to support Linux for now, since we only support ktls in Linux. However, we need to ensure that the signature we choose can work for all implementations. So I dropped any feature only supported by one implementation, like Linux's special handling of offset==NULL and the sf_hdtr field. Our signature should be as minimal as possible.
Call-outs:
Why not just let customers call sendfile directly? This would prevent us from ever updating the connection in response to a send (like if we track bytes sent) and from sending control messages like post-handshake messages.
Testing:
A little unit test that tests sendfile without ktls actually enabled. It tries to send itself :)
By submitting this pull request, I confirm that my contribution is made under the terms of the Apache 2.0 license.