!xmLtiCaAJxfhURjrXl:matrix.org

NixOS RISC-V

196 Members
NixOS on RISC-V https://wiki.nixos.org/wiki/RISC-V https://pad.lassul.us/NixOS-riscv64-linux https://github.com/orgs/NixOS/teams/risc-v58 Servers

Load older messages


SenderMessageTime
12 Nov 2025
@joerg:thalheim.ioMic92* Is riscv 4k page aligned? Usually the first page should not be mapped in userspace 06:06:36
@joerg:thalheim.ioMic92Not sure if this is your thing, but llms are pretty good at having read all of the cpu architecture specs and have found in the past for me pretty quickly oddities in readelf. 06:09:20
@rosscomputerguy:matrix.orgTristan RossI've been rubber ducking ChatGPT since Sunday with this problem06:09:44
@rosscomputerguy:matrix.orgTristan RossIt believe AppArmor is the problem, despite me disabling it thinking that was the problem06:10:10
@joerg:thalheim.ioMic92Feels unlikely. 06:10:53
@joerg:thalheim.ioMic92You would see the audit log in dmesg06:11:16
@rosscomputerguy:matrix.orgTristan RossYeah, which I don't but ChatGPT is just confident it's AppArmor lol06:11:33
@joerg:thalheim.ioMic92It's definitely the mmap syscall requesting that returns eperm 06:12:33
@joerg:thalheim.ioMic92You could even ask it to write an mmap call on these parameters to reproduce it06:13:30
@joerg:thalheim.ioMic92I just don't know which load segment is off. My brain cannot parse these unaligned columns from the phone06:14:34
@rosscomputerguy:matrix.orgTristan Ross
$ zcat /proc/config.gz | grep CONFIG_STRICT_KERNEL_RWX
CONFIG_STRICT_KERNEL_RWX=y

Could this be a problem?

06:14:54
@rosscomputerguy:matrix.orgTristan RossChatGPT says it could be lol06:15:06
@joerg:thalheim.ioMic92No we don't map as executable here if i interpret correctly06:16:06
@joerg:thalheim.ioMic92 mmap(0xf000, 4096, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_FIXED|MAP_DENYWRITE, 3, 0) = -1 EPERM (Operation not permitted) 06:16:19
@joerg:thalheim.ioMic92Only read/write06:16:32
@joerg:thalheim.ioMic92Figure out what 3 is here06:16:59
@joerg:thalheim.ioMic92Run strace with -yy06:17:15
@dramforever:matrix.orgdramforever0xf000 is sus06:17:29
@joerg:thalheim.ioMic92This gives you file descriptors06:17:33
@dramforever:matrix.orgdramforeveris it below the minimum address limit06:17:38
@joerg:thalheim.ioMic92
In reply to @dramforever:matrix.org
0xf000 is sus
I agree
06:17:47
@joerg:thalheim.ioMic92What's allowed on RISC-V? 06:18:05
@rosscomputerguy:matrix.orgTristan Ross I set sudo sysctl -w vm.mmap_min_addr=4096 and GCC works 06:18:15
@dramforever:matrix.orgdramforeverthe other thing is it may be relevant to us enabling PIE recently06:18:41
@dramforever:matrix.orgdramforeverish06:18:45
@dramforever:matrix.orgdramforeverrecent-ish06:18:49
@dramforever:matrix.orgdramforeveri'm starting to think maybe you shouldn't be on 6.606:19:14
@rosscomputerguy:matrix.orgTristan RossThe bootstrap-tools GCC is from last year, I thought PIE was within the last couple months06:19:14
@dramforever:matrix.orgdramforever although it really also shouldn't be breaking 06:19:35
@joerg:thalheim.ioMic92Check if the binary is supposes to be relocatable with checkseck06:19:44

Show newer messages


Back to Room ListRoom Version: 10