In general, I like the concept. I haven't tested it yet, but it looks promising.
- Jesse On 2021-12-16 7:15 a.m., Alexander Vickberg wrote: > Hi Jesse, > > It seems there's a missing '#include <sys/sysmacros.h>' in my patch. > However, > before I send a v2, please let me know what you think about it in general. > > Best regards > Alexander Vickberg > > Den tors 16 dec. 2021 kl 00:06 skrev Jesse Smith > <jsm...@resonatingmedia.com <mailto:jsm...@resonatingmedia.com>>: > > Alexander, > > Thank you for sending over the patches. I'll give them a test run > in the > coming days. > > Jesse > > > On 2021-12-15 12:43 p.m., Alexander Vickberg wrote: > > Hello list, > > > > Bootlogd is eating the log data and doesn't forward it to > console. The issue > > arised with commit 986bee6. The log is correctly written to the > file. My system > > does not specify console as boot argument but instead uses > 'stdout-path' in the > > device-tree, e.g: > > > > chosen { > > bootargs = "earlycon uio_pdrv_genirq.of_id=generic-uio > ubi.mtd=ubi"; > > stdout-path = "serial0:115200n8"; > > }; > > > > With this configuration bootlogd fails to detect current > console. I have made > > a patch to fix it by bringing back removed code which checks > either which tty > > is 'stdin' or using TIOCGDEV ioctl on 'stdin'. Both worked on my > system (Linux > > 5.10). Patches are proposals and have not been tested on any > other system than > > Linux. > > > > Best regards > > Alexander Vickberg > > > > Alexander Vickberg (2): > > bootlogd: Use consistent indentation > > Fix bootlogd not printing to console without console= in cmdline > > > > src/bootlogd.c | 212 > +++++++++++++++++++++++++++---------------------- > > 1 file changed, 119 insertions(+), 93 deletions(-) > > > >
