jaeger changed the topic of #crux-devel to: CRUX (https://crux.nu/) development channel | Logs: https://libera.irclog.whitequark.org/crux-devel/
chrcav has quit [Ping timeout: 244 seconds]
chrcav has joined #crux-devel
darfo has quit [Quit: Leaving]
<jaeger> ok, sandbox pages have been moved
groovy2shoes has quit [Remote host closed the connection]
groovy2shoes has joined #crux-devel
SiFuh has quit [Read error: Connection reset by peer]
SiFuh has joined #crux-devel
groovy3shoes has joined #crux-devel
groovy2shoes has quit [*.net *.split]
stenur has quit [*.net *.split]
pitillo_ has quit [*.net *.split]
pitillo_ has joined #crux-devel
stenur has joined #crux-devel
SiFuh has quit [Ping timeout: 240 seconds]
SiFuh has joined #crux-devel
pitillo_ has quit [Ping timeout: 244 seconds]
groovy3shoes has quit [Read error: Connection reset by peer]
groovy4shoes has joined #crux-devel
jue has joined #crux-devel
darfo has joined #crux-devel
<jaeger> Been a busy weekend but I got the docs moved over last night and I'm starting a bootstrap today
ivandi has quit [Quit: WeeChat 3.6]
<SiFuh> Cool jaeger
groovy4shoes has quit [Remote host closed the connection]
groovy4shoes has joined #crux-devel
<jaeger> hrm... maybe...
<jaeger> The kernel build is now failing for some reason with a double free error
<SiFuh> Oops
<SiFuh> Wasn't there a GCC or Glibc update recently?
<jaeger> end of august, yeah
<SiFuh> Anyone running them they should try to compile the 5.15.55 kernel
<SiFuh> jaeger: you might have to change the kernel to 5.15.65
<jaeger> worth a look
<jaeger> hrmm, same issue with 5.15.65
<jaeger> GEN /usr/src/iso/kernel/linux-5.15.65/tools/objtool/arch/x86/lib/inat-tables.c
<jaeger> free(): double free detected in tcache 2
<jaeger> for reference
<jaeger> make[5]: *** [arch/x86/Build:9: /usr/src/iso/kernel/linux-5.15.65/tools/objtool/arch/x86/lib/inat-tables.c] Aborted
<jaeger> There was a similar bug report from binutils 2.34 but that was pretty old
<SiFuh> Some of the errors are 5 days noted.
<SiFuh> So it might not be a good thing to push 3.7 with GCC 12.2.0
<SiFuh> Or, hold back on the 3.7 release until it is fixed. But I am pretty sure that the kernel issues are to do with gc-12
<SiFuh> gcc-12*
<jaeger> That gcc update happened quite a bit before this breakage, though, which is weird
<jaeger> I wonder if my build system has broken somehow
<SiFuh> Hmm
<SiFuh> My iso is using gcc-12.1.0
<SiFuh> Wasn't the update 12.2.0 ?
<jaeger> yes, aug 21
<SiFuh> You built rc4 before then?
<jaeger> I've built probably 2 or 3 local ISOs between then and now... even just a couple days ago
<SiFuh> Maybe you don't break something
<SiFuh> did*
<jaeger> Gonna do some toolchain rebuilding on the build system and see if that helps
<SiFuh> Cool, by the way now that you have 5.15.65 you going to stick with it or go back to 55?
<jaeger> I don't have a strong preference, could stick with it
<SiFuh> If you do decide, let me know, so I can look if there are minor differences for the modular kernel config.
<SiFuh> There might be some fancy new changes like a food replicator or a holodeck support built in
jue has quit [Ping timeout: 264 seconds]
<farkuhar> I'm here to report a successful build of Linux 5.15.55 using gcc-12.2.0 (but my config does not enable as many modules as the ISO).
<SiFuh> Kind of beside the point, but the success is a success
<SiFuh> Did you compile the inat-tables module?
<farkuhar> No, that was one of the modules I skipped.
<farkuhar> I'll run make menuconfig and enable it, then try the build again.
<SiFuh> Hahaha I was just imaging this "I will test drive your car, by changing the oil in my motorcycle"
<farkuhar> Just to make sure I'm in the right menu, is it CONFIG_IP_NF_NAT that enables the problematic module?
<farkuhar> apparently not. I just ran 'make modules' and it compiled iptable_nat.ko (again without any error, though).
<SiFuh> I think so, let me check
groovy4shoes has quit [Remote host closed the connection]
<farkuhar> Actually I think inat-tables.c just provides library code, it doesn't correspond to a single kernel module. I probably did compile it without error, since all the netfilter *.ko modules did get created successfully.
<SiFuh> It is included in a lot
<SiFuh> So yeah, best bet is to try compile with the iso config and see how you go
<farkuhar> Here are all the netfilter modules I just compiled: http://sprunge.us/zuO5of
<SiFuh> I didn't see any netfiler associated with it
<farkuhar> Okay, I'll grab the iso config from gitweb and try it again that way.
<SiFuh> I think it is the safest way to test to get a foot hold on what is going on.
<farkuhar> Successfully built Linux 5.15.55 using the iso config and gcc 12.2.0.
<SiFuh> Awesome!