Feat unix streams #24
Conversation
Add flow_info and scope_id to IPv6 address translation
| addr6.sin6_flowinfo = 0; | ||
| addr6.sin6_scope_id = 0; | ||
| addr6.sin6_flowinfo = need_revert?htonl(peer_addr->u.inet6.addr.flow_info):peer_addr->u.inet6.addr.flow_info; | ||
| addr6.sin6_scope_id = need_revert?htonl(peer_addr->u.inet6.addr.scope_id):peer_addr->u.inet6.addr.scope_id;; |
There was a problem hiding this comment.
You are sure the need_revert test also apply to those new field?
There was a problem hiding this comment.
I'm 99% sure, but let me make 100% sure all the same.
| @@ -4,6 +4,8 @@ | |||
| #define FIONREAD 1 | |||
| #define FIONBIO 2 | |||
|
|
|||
| #define TIOCGWINSZ 0x101 | |||
There was a problem hiding this comment.
using random constant can be risky in the future...
There was a problem hiding this comment.
Definitely. I did this since the FIONREAD and FIONBIO constants where already redeclared as 1 and 2, which seems random enough; however, I went for an extra 0x100 to keep FIO and TIO kind of distinct. The usual implementation of this constant takes into account things such as the size of the result struct, which I suppose is relevant when implementing the syscall in the kernel. For our usecase, we don't really need the additional data, so I thought a simple number would suffice. I can change it to something else if you feel it's inappropriate.
Arshia001
force-pushed
the
feat-unix-streams
branch
2 times, most recently
from
30b2ff9
to
3c194fb
Compare
Arshia001
force-pushed
the
feat-unix-streams
branch
from
3c194fb
to
c7cfb2d
Compare
|
CI is red, but that seems to be an issue with the CI test iotself? |
|
Yes, the red pipelines are from the original wasi-libc. The one we merged today is green. |
A few changes related to sockets, unix streams and tty (all needed for
crosstermto work):sockaddr_un