| 12 Nov 2025 |
 Mic92 | I think there might be something your gcc tries to map to an fixed address space adress that the kernel even before staring ld.so outrightly refuses | 05:55:56 |
| Mic92 | * I think there might be something your gcc tries to map to an fixed address space adress that the kernel even before starting ld.so outrightly refuses | 05:56:15 |
 Tristan Ross | roma@mu-gundam:~/nixpkgs$ strace -E LD_DEBUG=all /nix/store/kg61gdq4svvph0c4pcsf867r8nkbiqs7-bootstrap-tools/lib/ld-linux-riscv64-lp64d.so.1 /nix/store/kg61gdq4svvph0c4pcsf867rc
execve("/nix/store/kg61gdq4svvph0c4pcsf867r8nkbiqs7-bootstrap-tools/lib/ld-linux-riscv64-lp64d.so.1", ["/nix/store/kg61gdq4svvph0c4pcsf8"..., "/nix/store/kg61gdq4svvph0c4pcsf8"0
brk(NULL) = 0x7fff84224000
openat(AT_FDCWD, "/nix/store/kg61gdq4svvph0c4pcsf867r8nkbiqs7-bootstrap-tools/bin/gcc", O_RDONLY|O_CLOEXEC) = 3
read(3, "\177ELF\2\1\1\3\0\0\0\0\0\0\0\0\2\0\363\0\1\0\0\0\230\312\2\0\0\0\0\0"..., 832) = 832
getpid() = 6665
writev(2, [{iov_base=" 6665:\t", iov_len=12}, {iov_base="file=", iov_len=5}, {iov_base="/nix/store/kg61gdq4svvph0c4pcsf8"..., iov_len=67}, {iov_base=" [", iov_len=2}, {iovp
) = 111
mmap(0xf000, 4096, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_FIXED|MAP_DENYWRITE, 3, 0) = -1 EPERM (Operation not permitted)
close(3) = 0
writev(2, [{iov_base="/nix/store/kg61gdq4svvph0c4pcsf8"..., iov_len=67}, {iov_base=": ", iov_len=2}, {iov_base="error while loading shared libra"..., iov_len=36}, {iov_base=": t
) = 217
exit_group(127) = ?
+++ exited with 127 +++
| 05:56:34 |
| Tristan Ross | Best I can get out of the UART atm | 05:56:46 |
| Mic92 |
mmap(0xf000, 4096, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_FIXED|MAP_DENYWRITE, 3, 0) = -1 EPERM (Operation not permitted) | 05:57:53 |
| Mic92 | Now do readelf and look for suspicious addresses. | 05:58:18 |
| Mic92 | 0xf000 | 05:58:33 |
| Mic92 | But might not exactly the same | 05:58:49 |
| Mic92 | Potentially it applied an offset already | 05:59:09 |
| Tristan Ross | Program Headers:
Type Offset VirtAddr PhysAddr
FileSiz MemSiz Flags Align
PHDR 0x0000000000000040 0x000000000000f040 0x000000000000f040
0x00000000000002a0 0x00000000000002a0 R 0x8
RISCV_ATTRIBUT 0x0000000000219e92 0x0000000000000000 0x0000000000000000
0x0000000000000057 0x0000000000000000 R 0x1
GNU_STACK 0x0000000000001000 0x0000000000000000 0x0000000000000000
0x0000000000000000 0x0000000000000000 RW 0x10
LOAD 0x0000000000000000 0x000000000000f000 0x000000000000f000
0x0000000000001000 0x0000000000001000 RW 0x1000
LOAD 0x0000000000001000 0x0000000000010000 0x0000000000010000
0x000000000020ded0 0x000000000020ded0 R E 0x1000
INTERP 0x00000000000142d0 0x00000000000232d0 0x00000000000232d0
0x000000000000005c 0x000000000000005c R 0x1
[Requesting program interpreter: /nix/store/kg61gdq4svvph0c4pcsf867r8nkbiqs7-bootstrap-tools/lib/ld-linux-riscv64-lp64d.so.1]
NOTE 0x0000000000014330 0x0000000000023330 0x0000000000023330
0x0000000000000020 0x0000000000000020 R 0x4
GNU_EH_FRAME 0x00000000001c41d8 0x00000000001d31d8 0x00000000001d31d8
0x000000000000967c 0x000000000000967c R 0x4
LOAD 0x000000000020ef58 0x000000000021ef58 0x000000000021ef58
0x000000000000af28 0x00000000000103b8 RW 0x1000
TLS 0x000000000020ef58 0x000000000021ef58 0x000000000021ef58
0x0000000000000000 0x0000000000000010 R 0x8
GNU_RELRO 0x000000000020ef58 0x000000000021ef58 0x000000000021ef58
0x00000000000080a8 0x00000000000080a8 R 0x1
DYNAMIC 0x0000000000216dd0 0x0000000000226dd0 0x0000000000226dd0
0x0000000000000230 0x0000000000000230 RW 0x8
| 05:59:14 |
| Mic92 | VirtAddr looks suspicious | 06:00:09 |
| Tristan Ross | Of the PHDR? | 06:01:03 |
| Tristan Ross | ls works so presumably, it should look like this:
Program Headers:
Type Offset VirtAddr PhysAddr
FileSiz MemSiz Flags Align
PHDR 0x0000000000000040 0x0000000000010040 0x0000000000010040
0x0000000000000230 0x0000000000000230 R 0x8
RISCV_ATTRIBUT 0x000000000012b89a 0x0000000000000000 0x0000000000000000
0x0000000000000057 0x0000000000000000 R 0x1
GNU_STACK 0x0000000000000000 0x0000000000000000 0x0000000000000000
0x0000000000000000 0x0000000000000000 RW 0x10
LOAD 0x0000000000000000 0x0000000000010000 0x0000000000010000
0x000000000011c610 0x000000000011c610 R E 0x1000
INTERP 0x0000000000000270 0x0000000000010270 0x0000000000010270
0x000000000000005c 0x000000000000005c R 0x1
[Requesting program interpreter: /nix/store/kg61gdq4svvph0c4pcsf867r8nkbiqs7-bootstrap-tools/lib/ld-linux-riscv64-lp64d.so.1]
NOTE 0x00000000000002e4 0x00000000000102e4 0x00000000000102e4
0x0000000000000044 0x0000000000000044 R 0x4
GNU_EH_FRAME 0x00000000001015b0 0x00000000001115b0 0x00000000001115b0
0x00000000000030bc 0x00000000000030bc R 0x4
LOAD 0x000000000011cf68 0x000000000012df68 0x000000000012df68
0x000000000000e920 0x0000000000023ba0 RW 0x1000
GNU_RELRO 0x000000000011cf68 0x000000000012df68 0x000000000012df68
0x000000000000a098 0x000000000000a098 R 0x1
DYNAMIC 0x0000000000126dd0 0x0000000000137dd0 0x0000000000137dd0
0x0000000000000230 0x0000000000000230 RW 0x8
| 06:01:56 |
| Mic92 | But i am not super familiar with virtual addresses on RISC-V to be honest. So i don't which adresses are acceptable on that architecture. You should consult the RISC-V os spec | 06:02:33 |
| Tristan Ross | Oh... | 06:02:44 |
| Mic92 | Maybe for now compare it with other binaries | 06:03:16 |
| Mic92 | And it's also hard to read for me. You want to look at what load addresses are requested | 06:05:14 |
| Mic92 | Is riscv 4k page aligned. Usually the first page should not be mapped in userspace | 06:06:27 |
| Tristan Ross | When I updated bootstrap-tools on my clone, I got awk to work and it works with this:
Program Headers:
Type Offset VirtAddr PhysAddr
FileSiz MemSiz Flags Align
PHDR 0x0000000000000040 0x0000000000000040 0x0000000000000040
0x00000000000002a0 0x00000000000002a0 R 0x8
RISCV_ATTRIBUT 0x00000000000b1022 0x0000000000000000 0x0000000000000000
0x000000000000006a 0x0000000000000000 R 0x1
LOAD 0x0000000000000000 0x0000000000000000 0x0000000000000000
0x00000000000ab84c 0x00000000000ab84c R E 0x1000
GNU_STACK 0x0000000000000000 0x0000000000000000 0x0000000000000000
0x0000000000000000 0x0000000000000000 RW 0x10
GNU_EH_FRAME 0x0000000000098fc8 0x0000000000098fc8 0x0000000000098fc8
0x00000000000017a4 0x00000000000017a4 R 0x4
NOTE 0x00000000000ab82c 0x00000000000ab82c 0x00000000000ab82c
0x0000000000000020 0x0000000000000020 R 0x4
LOAD 0x00000000000ac848 0x00000000000ac848 0x00000000000ac848
0x00000000000047c8 0x000000000000c818 RW 0x1000
GNU_RELRO 0x00000000000ac848 0x00000000000ac848 0x00000000000ac848
0x00000000000037b8 0x00000000000037b8 R 0x1
DYNAMIC 0x00000000000af2d0 0x00000000000af2d0 0x00000000000af2d0
0x0000000000000230 0x0000000000000230 RW 0x8
INTERP 0x00000000000b2000 0x00000000000ba000 0x00000000000ba000
0x000000000000005c 0x000000000000005c R 0x1
[Requesting program interpreter: /nix/store/95y6xmfgcvgxjmba9xm6ccxxnf6i61pw-bootstrap-tools/lib/ld-linux-riscv64-lp64d.so.1]
LOAD 0x00000000000b2000 0x00000000000ba000 0x00000000000ba000
0x0000000000000060 0x0000000000000060 RW 0x1000
LOAD 0x00000000000b3000 0x00000000000bb000 0x00000000000bb000
0x00000000000019d0 0x00000000000019d0 RW 0x1000
So maybe the virt addr PHDR isn't as important
| 06:06:33 |
| Mic92 | * Is riscv 4k page aligned? Usually the first page should not be mapped in userspace | 06:06:36 |
| Mic92 | Not 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 |
| Tristan Ross | I've been rubber ducking ChatGPT since Sunday with this problem | 06:09:44 |
| Tristan Ross | It believe AppArmor is the problem, despite me disabling it thinking that was the problem | 06:10:10 |
| Mic92 | Feels unlikely. | 06:10:53 |
| Mic92 | You would see the audit log in dmesg | 06:11:16 |
| Tristan Ross | Yeah, which I don't but ChatGPT is just confident it's AppArmor lol | 06:11:33 |
| Mic92 | It's definitely the mmap syscall requesting that returns eperm | 06:12:33 |
| Mic92 | You could even ask it to write an mmap call on these parameters to reproduce it | 06:13:30 |
| Mic92 | I just don't know which load segment is off. My brain cannot parse these unaligned columns from the phone | 06:14:34 |
| Tristan Ross | $ zcat /proc/config.gz | grep CONFIG_STRICT_KERNEL_RWX
CONFIG_STRICT_KERNEL_RWX=y
Could this be a problem?
| 06:14:54 |
