diff options
Diffstat (limited to 'results/classifier/118/none/1832916')
| -rw-r--r-- | results/classifier/118/none/1832916 | 49 |
1 files changed, 49 insertions, 0 deletions
diff --git a/results/classifier/118/none/1832916 b/results/classifier/118/none/1832916 new file mode 100644 index 000000000..819564270 --- /dev/null +++ b/results/classifier/118/none/1832916 @@ -0,0 +1,49 @@ +x86: 0.583 +architecture: 0.554 +device: 0.469 +graphic: 0.426 +mistranslation: 0.352 +network: 0.329 +socket: 0.283 +kernel: 0.251 +semantic: 0.239 +register: 0.214 +vnc: 0.205 +ppc: 0.197 +files: 0.185 +boot: 0.184 +permissions: 0.175 +PID: 0.148 +TCG: 0.128 +risc-v: 0.125 +VMM: 0.113 +virtual: 0.111 +peripherals: 0.110 +i386: 0.084 +performance: 0.083 +arm: 0.078 +debug: 0.069 +user-level: 0.066 +hypervisor: 0.047 +KVM: 0.034 +assembly: 0.008 + +linux-user does not check PROT_EXEC + +At no point do we actually verify that a page is PROT_EXEC before translating. All we end up verifying is that the page is readable. Not the same thing, obviously. + +The following test case should work for any architecture, though I've only validated it for x86_64 and aarch64. + + + +It turns out we can't fix this without also fixing +our implementation of signal trampolines. + + +This is an automated cleanup. This bug report has been moved to QEMU's +new bug tracker on gitlab.com and thus gets marked as 'expired' now. +Please continue with the discussion here: + + https://gitlab.com/qemu-project/qemu/-/issues/122 + + |