I have run into the following build error when running ninja to build seL4test (following instructions at: https://docs.sel4.systems/GettingStarted#setting-up-your-machine)%C2%A0 Here is what I get for output. Any suggestions on a work-around? (Do I have the wrong version of a python tool or library?)
--
[65/263] Generating circular_includes_valid
FAILED: kernel/circular_includes_valid /home/dev/seL4test/build-x86/kernel/circular_includes_valid
cd /home/dev/seL4test/build-x86/kernel && /home/dev/seL4test/kernel/tools/circular_includes.py --ignore kernel_all_copy.c < kernel_all.i && touch circular_includes_valid
Traceback (most recent call last):
File "/home/dev/seL4test/kernel/tools/circular_includes.py", line 74, in <module>
sys.exit(main(args))
File "/home/dev/seL4test/kernel/tools/circular_includes.py", line 61, in main
while file_stack[-1] != header:
IndexError: list index out of range
[66/263] Invoking muslc build system
/home/dev/seL4test/projects/musllibc/src/prng/seed48.c:6:40: warning: argument 1 of type ‘short unsigned int *’ declared as a pointer [-Warray-parameter=]
6 | unsigned short *seed48(unsigned short *s)
| ~~~~~~~~~~~~~~~~^
In file included from /home/dev/seL4test/projects/musllibc/src/prng/seed48.c:1:
/home/dev/seL4test/projects/musllibc/include/stdlib.h:133:25: note: previously declared as an array ‘short unsigned int[3]’
133 | unsigned short *seed48 (unsigned short [3]);
| ^~~~~~~~~~~~~~~~~~
/home/dev/seL4test/projects/musllibc/src/unistd/getcwd.c: In function ‘getcwd’:
cc1: warning: function may return address of local variable [-Wreturn-local-addr]
/home/dev/seL4test/projects/musllibc/src/unistd/getcwd.c:9:14: note: declared here
9 | char tmp[PATH_MAX];
| ^~~
ninja: build stopped: subcommand failed.
"Ben" == Ben McCart ben@mccart.us writes:
Ben> I have run into the following build error when running ninja to Ben> build seL4test (following instructions Ben> at: https://docs.sel4.systems/GettingStarted#setting-up-your-machine)%C2%A0 Ben> Here is what I get for output. Any suggestions on a Ben> work-around? (Do I have the wrong version of a python tool or Ben> library?) --
On a clean install of Ubuntu Jammy everything works, using gcc 11.2.0
My suspicion is that you're not using this or a more recent GCC version. In the past (see https://github.com/seL4/seL4/pull/535 ) we've seen this error because the regex in the circular_includes.py script didn't match the format for file inclusion produced by the compiler.
Please check your gcc version. -- Dr Peter Chubb https://trustworthy.systems/ Trustworthy Systems Group CSE, UNSW Core hours: Mon 8am-3pm; Wed: 8am-5pm; Fri 8am-12pm.
---- On Tue, 07 May 2024 17:15:04 -0600 Peter Chubb via Devel devel@sel4.systems wrote ---
"Ben" == Ben McCart mailto:ben@mccart.us writes:
Ben> I have run into the following build error when running ninja to Ben> build seL4test (following instructions Ben> at: https://docs.sel4.systems/GettingStarted#setting-up-your-machine)%C2%A0 Ben> Here is what I get for output. Any suggestions on a Ben> work-around? (Do I have the wrong version of a python tool or Ben> library?) --
On a clean install of Ubuntu Jammy everything works, using gcc 11.2.0
My suspicion is that you're not using this or a more recent GCC version. In the past (see https://github.com/seL4/seL4/pull/535 ) we've seen this error because the regex in the circular_includes.py script didn't match the format for file inclusion produced by the compiler.
Please check your gcc version.
---
My GCC version is 11.4.0, I gave up the manual setup approach and just pulled the trustworthysystems/sel4 image from docker-hub and am using that to build and run the seL4test without any trouble. I'll just use the docker containers for my investigation for now.