You are viewing a plain text version of this content. The canonical link for it is here.
Posted to commits@nuttx.apache.org by "pussuw (via GitHub)" <gi...@apache.org> on 2023/05/24 12:18:26 UTC
[GitHub] [nuttx] pussuw opened a new pull request, #9374: net/socket: Fix send() / recv() in BUILD_KERNEL
pussuw opened a new pull request, #9374:
URL: https://github.com/apache/nuttx/pull/9374
## Summary
The send() and recv() functions take in user data (as pointers) and these were passed -as is- to the underlying logic, which will fail miserably if the data is used in e.g. devif_poll().
Why? Because the callback comes directly from driver logic and at that point the wrong mappings are almost certainly active -> crash due to page fault. This fixes this behavior.
## Impact
Fix page fault issue with socket:send() / recv().
## Testing
icicle:knsh
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: commits-unsubscribe@nuttx.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [nuttx] mu578 commented on pull request #9374: net/socket: Fix send() / recv() in BUILD_KERNEL
Posted by "mu578 (via GitHub)" <gi...@apache.org>.
mu578 commented on PR #9374:
URL: https://github.com/apache/nuttx/pull/9374#issuecomment-1561239108
@pkarashchenko To me, that's elegant and simple, avoiding copying would require to implement a zero-copy model, using COW when necessary. Could be petitioned as a next-next-next major kernel-os-release. It would require first to design an appropriate model that fits actual + taking account of every single part it impacts; takes a lot of carrot cake juice.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: commits-unsubscribe@nuttx.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [nuttx] pussuw commented on pull request #9374: net/socket: Fix send() / recv() in BUILD_KERNEL
Posted by "pussuw (via GitHub)" <gi...@apache.org>.
pussuw commented on PR #9374:
URL: https://github.com/apache/nuttx/pull/9374#issuecomment-1561093961
Btw if you want a bit of reference for the iterative solution, iov_iter in Linux is basically the feature I'm talking about
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: commits-unsubscribe@nuttx.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [nuttx] xiaoxiang781216 commented on a diff in pull request #9374: net/socket: Fix send() / recv() in BUILD_KERNEL
Posted by "xiaoxiang781216 (via GitHub)" <gi...@apache.org>.
xiaoxiang781216 commented on code in PR #9374:
URL: https://github.com/apache/nuttx/pull/9374#discussion_r1204247966
##########
net/socket/send.c:
##########
@@ -156,30 +156,7 @@ ssize_t psock_send(FAR struct socket *psock, FAR const void *buf, size_t len,
ssize_t send(int sockfd, FAR const void *buf, size_t len, int flags)
Review Comment:
can move to libc and change to normal function?
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: commits-unsubscribe@nuttx.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [nuttx] patacongo commented on pull request #9374: net/socket: Fix send() / recv() in BUILD_KERNEL
Posted by "patacongo (via GitHub)" <gi...@apache.org>.
patacongo commented on PR #9374:
URL: https://github.com/apache/nuttx/pull/9374#issuecomment-1568748481
> Putting `send()`/`recv()` in libc and using `psock_send()`/`psock_recv()` as the system call is a fine idea, _provided_ that the table that maps socket descriptors to psock structures in accessible in user space.
A complication is that `send()`/`recv()` are also cancellation points and cancellation points cannot be implemented in user space.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: commits-unsubscribe@nuttx.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [nuttx] pkarashchenko commented on pull request #9374: net/socket: Fix send() / recv() in BUILD_KERNEL
Posted by "pkarashchenko (via GitHub)" <gi...@apache.org>.
pkarashchenko commented on PR #9374:
URL: https://github.com/apache/nuttx/pull/9374#issuecomment-1561067618
Just a general question: Can we "translate" user pointer to "kernel aware" address and bypass copying of data? Or that process is not trivial?
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: commits-unsubscribe@nuttx.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [nuttx] xiaoxiang781216 commented on pull request #9374: net/socket: Fix send() / recv() in BUILD_KERNEL
Posted by "xiaoxiang781216 (via GitHub)" <gi...@apache.org>.
xiaoxiang781216 commented on PR #9374:
URL: https://github.com/apache/nuttx/pull/9374#issuecomment-1565222686
Do you mean the binary compatibility or the source compatibility?
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: commits-unsubscribe@nuttx.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [nuttx] pkarashchenko commented on pull request #9374: net/socket: Fix send() / recv() in BUILD_KERNEL
Posted by "pkarashchenko (via GitHub)" <gi...@apache.org>.
pkarashchenko commented on PR #9374:
URL: https://github.com/apache/nuttx/pull/9374#issuecomment-1561170315
Let's go with copy solution until we find something better.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: commits-unsubscribe@nuttx.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [nuttx] pussuw commented on a diff in pull request #9374: net/socket: Fix send() / recv() in BUILD_KERNEL
Posted by "pussuw (via GitHub)" <gi...@apache.org>.
pussuw commented on code in PR #9374:
URL: https://github.com/apache/nuttx/pull/9374#discussion_r1206286161
##########
net/socket/send.c:
##########
@@ -156,30 +156,7 @@ ssize_t psock_send(FAR struct socket *psock, FAR const void *buf, size_t len,
ssize_t send(int sockfd, FAR const void *buf, size_t len, int flags)
Review Comment:
Ok, what about recv() then, should that be removed too ? Maybe we can do another patch for that because it is not related to this fix ?
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: commits-unsubscribe@nuttx.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [nuttx] xiaoxiang781216 commented on a diff in pull request #9374: net/socket: Fix send() / recv() in BUILD_KERNEL
Posted by "xiaoxiang781216 (via GitHub)" <gi...@apache.org>.
xiaoxiang781216 commented on code in PR #9374:
URL: https://github.com/apache/nuttx/pull/9374#discussion_r1206284580
##########
net/socket/send.c:
##########
@@ -156,30 +156,7 @@ ssize_t psock_send(FAR struct socket *psock, FAR const void *buf, size_t len,
ssize_t send(int sockfd, FAR const void *buf, size_t len, int flags)
Review Comment:
> You mean remove the system call for send() ?
Yes.
> I guess it could be done but I think we should keep it as a system call.
Normally, if the function can be simply forwarded to system call, it should be move to libs/libc as normal function.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: commits-unsubscribe@nuttx.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [nuttx] patacongo commented on pull request #9374: net/socket: Fix send() / recv() in BUILD_KERNEL
Posted by "patacongo (via GitHub)" <gi...@apache.org>.
patacongo commented on PR #9374:
URL: https://github.com/apache/nuttx/pull/9374#issuecomment-1568734516
> @mu578 if it is moved into libc it will still be available to kernel, as libc is built for the kernel as well.
Which is not to say that `send()`/`recv()` _should_ be used within the OS. In general, application APIs with socket descriptor arguments are not permitted within the OS because the socket descriptor is only valid in the context of the application task. It often is valid, but good programming practice is to always avoid the use of socket descriptors within the OS.
Within the OS, there are a special set of socket APIs that are usable and do not depend upon anything from the task context. These are prototyped in `include/nuttx/net/net.h` and begin with` psock_` like `psock_send()` and `psock_recv().`
There is a similar set of interfaces for application APIs that use file descriptors.
Putting `send()`/`recv()` in libc and using `psock_send()`/`psock_recv()` as the system call is a fine idea, _provided_ that the table that maps socket descriptors to psock structures in accessible in user space. It was not user accessible in the distant past and hence `send()`/`recv() `were the system calls.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: commits-unsubscribe@nuttx.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [nuttx] pussuw commented on pull request #9374: net/socket: Fix send() / recv() in BUILD_KERNEL
Posted by "pussuw (via GitHub)" <gi...@apache.org>.
pussuw commented on PR #9374:
URL: https://github.com/apache/nuttx/pull/9374#issuecomment-1568755006
> psock_send()/psock_recv() are not cancellation points.
I think the suggestion was to use sendto() and recvfrom() as the system call for send() and recv(). I combined send() into sendto() as it is IMO just a specific flavor of it. I think this optimization is not strictly necessary (although possible) and that is why I suggested it be left out of this PR / patch.
Using the kernel APIs psock_send()/recv() will obviously not do.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: commits-unsubscribe@nuttx.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [nuttx] patacongo commented on pull request #9374: net/socket: Fix send() / recv() in BUILD_KERNEL
Posted by "patacongo (via GitHub)" <gi...@apache.org>.
patacongo commented on PR #9374:
URL: https://github.com/apache/nuttx/pull/9374#issuecomment-1569365119
> @patacongo theoretically then, it should fail with bad_descriptor access, if any access performed within kern,
Two C libraries are built in the KERNEL build: One for applications and one for the kernel. If we want to prohibit kernel access to C library functions that use file or socket descriptors, then we can always exclude them from the kernel C library. build That is really not a bad idea.
That would be done by testing __KERNEL__ which is defined when the kernel C library is built.
[This really needs to be its own issue]
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: commits-unsubscribe@nuttx.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [nuttx] pussuw commented on a diff in pull request #9374: net/socket: Fix send() / recv() in BUILD_KERNEL
Posted by "pussuw (via GitHub)" <gi...@apache.org>.
pussuw commented on code in PR #9374:
URL: https://github.com/apache/nuttx/pull/9374#discussion_r1206286161
##########
net/socket/send.c:
##########
@@ -156,30 +156,7 @@ ssize_t psock_send(FAR struct socket *psock, FAR const void *buf, size_t len,
ssize_t send(int sockfd, FAR const void *buf, size_t len, int flags)
Review Comment:
Ok, what about recv() then, should that be removed too ? Maybe we can do another patch and PR for that because it is not related to this fix ?
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: commits-unsubscribe@nuttx.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [nuttx] mu578 commented on pull request #9374: net/socket: Fix send() / recv() in BUILD_KERNEL
Posted by "mu578 (via GitHub)" <gi...@apache.org>.
mu578 commented on PR #9374:
URL: https://github.com/apache/nuttx/pull/9374#issuecomment-1569317233
@pussuw ok I understand now, personally, in such case, I do prefer redundancy with a comment about pairing concerns.
I didn't not elaborate but concerns could be lower down at file_descriptor read and write (it is context aware), not their frontends.
@patacongo theoretically then, it should fail with bad_descriptor access, if any access performed within kern, I am ignorant of that part of NuttX, I need to follow the stream.
For historical reason, there are in unix-like models a "kern-sock" API vs a user-bsd-sock API. There is nothing preventing to rethink that model in 2023; could think about a model using access permissions.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: commits-unsubscribe@nuttx.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [nuttx] pussuw commented on pull request #9374: net/socket: Fix send() / recv() in BUILD_KERNEL
Posted by "pussuw (via GitHub)" <gi...@apache.org>.
pussuw commented on PR #9374:
URL: https://github.com/apache/nuttx/pull/9374#issuecomment-1561091201
> Just a general question: Can we "translate" user pointer to "kernel aware" address and bypass copying of data? Or that process is not trivial?
Yes it is doable, in fact the infra is here: https://github.com/apache/nuttx/pull/9368
Simply obtaining a kernel addressable address (once) for the user data is not enough, because if the data goes across multiple pages you would need to do it again and again by iterating the user data page-by-page, or do a large continuous mapping (and make sure the mapping is released once it is not needed any more).
Both are doable, but the work amount is not trivial by any means, basically it would be a brand new, and very big feature. It would of course be much better than this but the amount of work is not trivial.
This patches a full system crash I get with UDP sockets, so it does the trick for me.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: commits-unsubscribe@nuttx.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [nuttx] mu578 commented on pull request #9374: net/socket: Fix send() / recv() in BUILD_KERNEL
Posted by "mu578 (via GitHub)" <gi...@apache.org>.
mu578 commented on PR #9374:
URL: https://github.com/apache/nuttx/pull/9374#issuecomment-1564841976
@xiaoxiang781216 @pussuw for clarity, what's the near future plan? keep it as is, or moving it at user-space only availability? As now, it is kernel available and accepted, meanwhile if tomorrow moved, it will break code compatibility relaying on such feature.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: commits-unsubscribe@nuttx.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [nuttx] mu578 commented on pull request #9374: net/socket: Fix send() / recv() in BUILD_KERNEL
Posted by "mu578 (via GitHub)" <gi...@apache.org>.
mu578 commented on PR #9374:
URL: https://github.com/apache/nuttx/pull/9374#issuecomment-1561990119
@acassis zero copy memory model @future
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: commits-unsubscribe@nuttx.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [nuttx] pussuw commented on a diff in pull request #9374: net/socket: Fix send() / recv() in BUILD_KERNEL
Posted by "pussuw (via GitHub)" <gi...@apache.org>.
pussuw commented on code in PR #9374:
URL: https://github.com/apache/nuttx/pull/9374#discussion_r1206277597
##########
net/socket/send.c:
##########
@@ -156,30 +156,7 @@ ssize_t psock_send(FAR struct socket *psock, FAR const void *buf, size_t len,
ssize_t send(int sockfd, FAR const void *buf, size_t len, int flags)
Review Comment:
You mean remove the system call for send() ? I guess it could be done but I think we should keep it as a system call.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: commits-unsubscribe@nuttx.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [nuttx] pussuw commented on pull request #9374: net/socket: Fix send() / recv() in BUILD_KERNEL
Posted by "pussuw (via GitHub)" <gi...@apache.org>.
pussuw commented on PR #9374:
URL: https://github.com/apache/nuttx/pull/9374#issuecomment-1568529336
@mu578 if it is moved into libc it will still be available to kernel, as libc is built for the kernel as well. To me keeping send() / recv() as system call is fine, I don't have any opinion on that. The kernel/user API separation is quite vague, I have cleaned up some of the APIs from kernel that _definitely_ don't belong there, due to being a security risk, but not others.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: commits-unsubscribe@nuttx.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [nuttx] patacongo commented on pull request #9374: net/socket: Fix send() / recv() in BUILD_KERNEL
Posted by "patacongo (via GitHub)" <gi...@apache.org>.
patacongo commented on PR #9374:
URL: https://github.com/apache/nuttx/pull/9374#issuecomment-1568777811
> I think the suggestion was to use sendto() and recvfrom() as the system call for send() and recv().
It already works that way:
- https://github.com/apache/nuttx/blob/master/net/socket/send.c#L161
- https://github.com/apache/nuttx/blob/master/net/socket/recv.c#L61
`sendto()` and `recvfrom()` are the system calls. They implement the cancellation points and call:
- psock_sendto: https://github.com/apache/nuttx/blob/master/net/socket/sendto.c#L244
- psock_recvfrom: https://github.com/apache/nuttx/blob/master/net/socket/recvfrom.c#L198
`psock_send()` /` psock_recv()` are only used inside the OS.
And these all, eventually boil down to `psock_sendmsg()` and `psock_recvmsg()`.
Sorry, I should not talk off the top of my head without checking things. A lot has changed and I don't always keep up with the changes.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: commits-unsubscribe@nuttx.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [nuttx] mu578 commented on pull request #9374: net/socket: Fix send() / recv() in BUILD_KERNEL
Posted by "mu578 (via GitHub)" <gi...@apache.org>.
mu578 commented on PR #9374:
URL: https://github.com/apache/nuttx/pull/9374#issuecomment-1569399631
@patacongo yes I do understand the principle of shared vs by passed or split, now how one can make understand such concerns in design? yes it is two models colliding issues. To know, I don't argue here, it is just from good smart approach; it lets room for unwanted imagination and interpretation, it happens; I don't pretend to have good solutions to address it either.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: commits-unsubscribe@nuttx.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [nuttx] patacongo commented on pull request #9374: net/socket: Fix send() / recv() in BUILD_KERNEL
Posted by "patacongo (via GitHub)" <gi...@apache.org>.
patacongo commented on PR #9374:
URL: https://github.com/apache/nuttx/pull/9374#issuecomment-1568781633
> The system call change / optimization is not strictly necessary (although possible) and that is why I suggested it be left out of this PR / patch.
That is reasonable too. The is the tip a of much larger issue: All APIs that receive socket descriptors as inputs should be implemented in libc and call an a lower level OS function via a call gate. It might be better handled as one larger effort.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: commits-unsubscribe@nuttx.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [nuttx] xiaoxiang781216 merged pull request #9374: net/socket: Fix send() / recv() in BUILD_KERNEL
Posted by "xiaoxiang781216 (via GitHub)" <gi...@apache.org>.
xiaoxiang781216 merged PR #9374:
URL: https://github.com/apache/nuttx/pull/9374
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: commits-unsubscribe@nuttx.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [nuttx] xiaoxiang781216 commented on a diff in pull request #9374: net/socket: Fix send() / recv() in BUILD_KERNEL
Posted by "xiaoxiang781216 (via GitHub)" <gi...@apache.org>.
xiaoxiang781216 commented on code in PR #9374:
URL: https://github.com/apache/nuttx/pull/9374#discussion_r1205643544
##########
net/socket/recvfrom.c:
##########
@@ -152,11 +154,39 @@ ssize_t recvfrom(int sockfd, FAR void *buf, size_t len, int flags,
{
FAR struct socket *psock;
ssize_t ret;
+#ifdef CONFIG_BUILD_KERNEL
+ struct sockaddr kaddr;
Review Comment:
need change to sockaddr_storage to support IPv6 or other transport
##########
net/socket/sendto.c:
##########
@@ -198,11 +200,39 @@ ssize_t sendto(int sockfd, FAR const void *buf, size_t len, int flags,
{
FAR struct socket *psock;
ssize_t ret;
+#ifdef CONFIG_BUILD_KERNEL
+ struct sockaddr kaddr;
Review Comment:
ditto
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: commits-unsubscribe@nuttx.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [nuttx] pussuw commented on a diff in pull request #9374: net/socket: Fix send() / recv() in BUILD_KERNEL
Posted by "pussuw (via GitHub)" <gi...@apache.org>.
pussuw commented on code in PR #9374:
URL: https://github.com/apache/nuttx/pull/9374#discussion_r1206277761
##########
net/socket/recvfrom.c:
##########
@@ -152,11 +154,39 @@ ssize_t recvfrom(int sockfd, FAR void *buf, size_t len, int flags,
{
FAR struct socket *psock;
ssize_t ret;
+#ifdef CONFIG_BUILD_KERNEL
+ struct sockaddr kaddr;
Review Comment:
Thanks for pointing this out, will fix!
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: commits-unsubscribe@nuttx.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [nuttx] xiaoxiang781216 commented on a diff in pull request #9374: net/socket: Fix send() / recv() in BUILD_KERNEL
Posted by "xiaoxiang781216 (via GitHub)" <gi...@apache.org>.
xiaoxiang781216 commented on code in PR #9374:
URL: https://github.com/apache/nuttx/pull/9374#discussion_r1206319255
##########
net/socket/send.c:
##########
@@ -156,30 +156,7 @@ ssize_t psock_send(FAR struct socket *psock, FAR const void *buf, size_t len,
ssize_t send(int sockfd, FAR const void *buf, size_t len, int flags)
Review Comment:
could be.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: commits-unsubscribe@nuttx.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org