blob: 6eab99bbb901caef0ee5b8206ccec9632165dd8f (
plain) (
blame)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
|
graphic: 0.759
device: 0.747
performance: 0.531
architecture: 0.489
register: 0.460
semantic: 0.447
debug: 0.396
network: 0.372
user-level: 0.274
vnc: 0.256
PID: 0.225
socket: 0.213
files: 0.205
arm: 0.188
permissions: 0.163
mistranslation: 0.139
VMM: 0.137
boot: 0.132
risc-v: 0.115
ppc: 0.111
virtual: 0.105
TCG: 0.104
peripherals: 0.094
kernel: 0.089
hypervisor: 0.073
i386: 0.072
x86: 0.060
KVM: 0.037
assembly: 0.029
qemu-binfmt-conf.sh handles errors inconsistently
Description of problem:
We are installing qemu via multiarch/qemu-user-static docker image. https://github.com/multiarch/qemu-user-static
What we have noticed is that because qemu-binfmt-conf.sh does not use `set -e`, its behavior with regards to failures is inconsistent. In short, registering the same thing into binfmt twice is an error (you get EEXIST). However, the exit code of qemu-binfmt-conf.sh itself seems to depend only on whether the last interpreter succeeded, leading to confusing and inconsistent results.
Steps to reproduce:
1. Register only qemu-arm-static interpreter with binfmt.
2. Run qemu-binfmt-conf.sh. Observe that the exit code is zero, and logs show the duplicate interpreter was rejected.
3. Remove all qemu interpreters.
3. Register only qemu-loongarch64-static interpreter (currently last in qemu_target_list) with binfmt.
3. Run qemu-binfmt-conf.sh. Observe that the exit code is non-zero, and logs show the duplicate interpreter was rejected.
|