[PATCH] base-files: update min_free_kbytes configuration

2024-01-02 Thread quic_shoudil
From: Zhuo Fu set proper min_free_kbytes for small RAM device to avoid OOM during device boot up. Signed-off-by: shoudil --- package/base-files/files/etc/init.d/sysctl | 10 ++ 1 file changed, 6 insertions(+), 4 deletions(-) diff --git a/package/base-files/files/etc/init.d/sysctl b/p

[PATCH] build: enable thumb compile flag for 32bit architecture.

2024-01-02 Thread quic_shoudil
From: shoudil Enable thumb flag to reduce package size, which help optimize the size of system image as well. Signed-off-by: shoudil --- include/hardening.mk | 4 1 file changed, 4 insertions(+) diff --git a/include/hardening.mk b/include/hardening.mk index 6acd862f5c..8c59643866 100644

[PATCH 1/1] kernel/x86: merge "generic" into "legacy"

2024-01-02 Thread Elliott Mitchell
There is a desire to reduce the time spent on x86 OpenWRT platforms. "generic" is most suitable for removal. The few systems which could use "generic" must now opt for "legacy" (or for a lucky few "64"). Signed-off-by: Elliott Mitchell --- target/linux/x86/Makefile | 2 +-

[PATCH 0/1] Nuking x86 "generic"

2024-01-02 Thread Elliott Mitchell
For anyone who is reading this mailing list and expecting to read everything, a fair amount also gets discussed on GitHub. Seems consensus on reducing x86 by one target is emerging. In fact someone proposed getting rid of x86-32 completely. (https://github.com/openwrt/openwrt/pull/13924#discussio

Best way to start unprivileged service

2024-01-02 Thread W. Michael Petullo
I am packaging a service which I would like to run as an unprivileged user. This server does not support opening its ports before dropping its privileges. I would like the service to listen on a privileged port (i.e., <1024). What is the best way to accomplish this on OpenWrt? I have been looking

[PATCH netifd] system-linux: fix race condition in netlink socket error handing

2024-01-02 Thread Matthias Schiffer
The error handling needed for the buffer growth logic relies on uloop_fd's error flag, which is set based on epoll events. Doing so without handling recvmsg's error codes is racy, as an error state may be set between receiving epoll events and the next recvmsg, but calling recvmsg clears the error