qemu: Add QemuConfig to set qemu args via a struct #2339
Conversation
Marcondiro
force-pushed
the
main
branch
7 times, most recently
from
3903120 to
44536d8
Compare
Marcondiro
force-pushed
the
main
branch
19 times, most recently
from
8ca1cd3 to
fd66199
Compare
Marcondiro
changed the title
|
Also, I just thought about it but we could have a look at the debug flags for systemmode ( |
Marcondiro
force-pushed
the
main
branch
2 times, most recently
from
5b9bc5f to
5b10934
Compare
Marcondiro
force-pushed
the
main
branch
3 times, most recently
from
ad626a4 to
ea428d1
Compare
| .no_graphic(true) | ||
| .snapshot(true) | ||
| .drives([qemu_config::Drive::builder() | ||
| .interface(qemu_config::DriveInterface::none) |
There was a problem hiding this comment.
Stupid questions without understanding the details:
Do you actually need to specify none here? And, if so, could Option::None make sense?
There was a problem hiding this comment.
Do you actually need to specify none here?
I'd say yes because, in Qemu, none is not the default interface
And, if so, could Option::None make sense?
mm I don't think so
| #[allow(non_camel_case_types)] | ||
| #[derive(Debug, strum_macros::Display, Clone)] | ||
| #[strum(prefix = "-monitor ")] | ||
| pub enum Monitor { |
There was a problem hiding this comment.
Monitor and Serial have exactly the same values, It'd make sense to combine them to OutputKind or similar.
Also, what's the difference of none and null?
Also, shouldn't enums be CamelCase?
There was a problem hiding this comment.
Monitor and Serial have exactly the same values, It'd make sense to combine them to
OutputKindor similar.
It's not clear to me how you would deduplicate it, can you elaborate more?
Also, what's the difference of none and null?
They are 2 different settings in qemu, basically none means no device at all while null is discard its output. https://www.qemu.org/docs/master/system/invocation.html
There was a problem hiding this comment.
Yes with the prefix hack I guess you can not
There was a problem hiding this comment.
Ah you could have a common enum and then implement From for the individual enums but it might be overcomplicated without clear benefit
Marcondiro
force-pushed
the
main
branch
3 times, most recently
from
8ccc39f to
6398310
Compare
|
@rmalmain Finished `dev` profile [optimized + debuginfo] target(s) in 4m 06s
[cargo-make] INFO - Running Task: test
+ cd /__w/LibAFL/LibAFL/fuzzers/qemu/qemu_launcher
Profile: dev
+ echo Profile: dev
+ cd injection_test
+ make
gcc -g -o static sqltest.c -l sqlite3 -lm -static -lpthread -ldl
/usr/bin/ld: /usr/lib/gcc/x86_64-linux-gnu/11/../../../x86_64-linux-gnu/libsqlite3.a(os_unix.o): in function `unixDlOpen':
(.text+0x8ad): warning: Using 'dlopen' in statically linked applications requires at runtime the shared libraries from the glibc version used for linking
gcc -g -o sqltest sqltest.c -l sqlite3 -lm -lpthread
+ mkdir in
+ echo aaaaaaaaaa
+ find /__w/LibAFL/LibAFL/target/x86_64 -name qemu_launcher
+ timeout 10s /__w/LibAFL/LibAFL/target/x86_64/debug/qemu_launcher -o out -i in -j ../injections.toml -v -- ./static
+ true
+ grep -Ei found.*injection fuzz.log
+ [ -z ]
+ echo Fuzzer does not generate any testcases or any crashes
Fuzzer does not generate any testcases or any crashes
+ echo Logs:
Logs:
+ cat fuzz.log
+ exit 1
Error while executing command, exit code: 1
Error: Process completed with exit code 1. |
|
looks good to me by now, should we merge once the fix pr is merged? |
No description provided.