IRC channel logs
2026-01-15.log
back to list of logs
<sobkas>Well I got no more room in ffffffffd7ee8290 ((...bc/elf/ld-x86-64.so.1(2596)) <damo22>because it could have interrupted from kernel <damo22>it would be easier if they didnt touch percpu area <damo22>OK the problem is this: SWAPGS_ENTRY is correct because we always know at a line of code in the kernel that we are ENTERING the kernel. SWAPGS_EXIT is currently undefined because how do we know if we are exiting to USER or KERNEL again? <damo22>all we know is we are going to call iretq <damo22>we can get an interrupt from a kernel thread, or a user thread <azert>damo22: do you remember I posted the Linux swapgs original code and it was saving the information on ebx <azert>you need to save the information on SWAPGS_ENTRY to use it at exit <azert>you could also store it on the stack <damo22>but i think my stack layout is now different and it might not be working correctly <damo22>trap_from_kernel can interrupt while taking an interrupt? <aculnaig>Hi, i would like to introduce HTTPS support in the httpfs translator through GnuTLS. Who is the person to ask advice for ? <aculnaig>I mean, if I have question arount HTTP + TLS <azert>aculnaig: I think Amos on the mailing list told you he knows a lot <azert>aculnaig: are you aware of libneon? <aculnaig>azert: i have heard of it but never use it. it is a good library ? <sobkas>Uh, so where core is stored after crash? <jack-sparrow>If you feel like joining the Gentoo/Hurd adventure, I invite you to join #gentoo-hurd on irc.libera.chat <youpi>sobkas: core files are stored in the cwd directory of the program iirc <youpi>so most often the program you started it from <sobkas>Welp, trying to build libc puts system into unusable state, programs just hang <sobkas>I just ran shutdown -h now and it just sits there... <sobkas>So good night, with a forced shutdown