sendmsg control headers, do they work in android ver 26, how do i find this out? #1929
-
|
Greetings, My question has two parts:
Notes: |
Beta Was this translation helpful? Give feedback.
Replies: 4 comments 3 replies
-
|
this is really a kernel question, and will most likely depend on what
kernel the device(s) in question have rather than the OS version, and
probably also on things like what their maximum msg_controllen is
configured to. (if you can read /proc/sys/net/core/optmem_max on the
device(s) in question, it might be worth a look at that.)
…On Sun, Sep 10, 2023 at 2:38 PM timprepscius ***@***.***> wrote:
Greetings,
My question has two parts:
1. do sendmsg msghdr msg_controllen msg_control work in android ver 26
2. how can I definitively determine what is and what is not supported?
Notes:
I'm trying to determine why a quic implementation is failing on android,
while working ios/linux/mac/windows.
the sendmsg function, when msg_controllen > 0, return -1 with errno EINVAL.
—
Reply to this email directly, view it on GitHub
<#1929>, or unsubscribe
<https://github.com/notifications/unsubscribe-auth/AMVLEWGA3EKBBDBS5FROEATXZYXNJANCNFSM6AAAAAA4SN3VPA>
.
You are receiving this because you are subscribed to this thread.Message
ID: ***@***.***>
|
Beta Was this translation helpful? Give feedback.
-
|
Is there a place I can look for this type of information? I feel like there should be some document that I can look at instead of asking on random github discussion forums. |
Beta Was this translation helpful? Give feedback.
-
depends on the device. i'd start by looking at the Android Common Kernel source (or just looking at "what's the kernel version(s) on the devices that are failing for me, and what's the lowest kernel version i see on a device that does work"). but if you're only seeing it on devices from a particular OEM or SoC vendor, it's possible that the behavior difference is in that specific kernel (and may not even be reflected upstream).
the linux kernel is quite a large API surface. when there's an actual bug that's caused widespread trouble, we'll typically add a CTS test. but where there's no test (and there's certainly no direct test for sendmsg() control messages) there's pretty wide latitude. in this specific case, there are some parts of the system that use control messages.
no, because they'll just say the same thing --- "what exact kernel(s)?", and then tell you to talk to the maintainers of those kernels.
not that i'm aware of, no. |
Beta Was this translation helpful? Give feedback.
-
|
I know the kernel, it is "Linux version 3.10.84-perf-g51b663cc723-04742-g5bcfaca9e61" So- just to verify, you are suggesting that I go and read the linux kernel source, in order to find out what kernel version supports sendmsg controllen? |
Beta Was this translation helpful? Give feedback.
-
|
Also, are you suggesting that I post to the linux kernel mailing list - that doesn't seem right. Is there a discussion group for android/linux kernel? |
Beta Was this translation helpful? Give feedback.
-
if it's not documented in the man page, reading the source is the only alternative i know of. (if you find anything interesting but don't want to go through the bureaucracy of trying to get the man page improved, let me know and i can try to upstream it for you.) |
Beta Was this translation helpful? Give feedback.
-
|
While I was quite incredulous, I looked up the source for linux kernel for the revision on this dev phone, read the networking stack code, and walah, the problem revealed itself. I'm still not sure how to feel about this. |
Beta Was this translation helpful? Give feedback.
if it's not documented in the man page, reading the source is the only alternative i know of. (if you find anything interesting but don't want to go through the bureaucracy of trying to get the man page improved, let me know and i can try to upstream it for you.)