993145 - Skip sandbox initialization on non-seccomp-capable Linux systems

Status: RESOLVED FIXED

(Core :: Security, defect)

Description

[Jed Davis [:jld] ⟨⏰|UTC-8⟩ ⟦he/him⟧]

As a workaround for bug 985775, we could check for seccomp support and, if not present, completely skip sandbox initialization and thus avoid the use of realtime signals that confuses old versions of gdbserver.  This would allow us to avoid changing manifests (and potentially introducing breakage) for a large number of pre-KitKat B2G devices.

Method 1. prctl(PR_GET_SECCOMP) will fail with EINVAL if no seccomp support is present.  This seems to be the case on Android kernels that haven't specifically been patched for seccomp-bpf.

Method 2. prctl(PR_SET_SECCOMP, SECCOMP_MODE_FILTER, nullptr) should work to test more specifically for seccomp-bpf — it will fail with EFAULT if seccomp-bpf is present and EINVAL otherwise.  But I don't know that this trick would gain us anything in practice.

Comment 1

[Jed Davis [:jld] ⟨⏰|UTC-8⟩ ⟦he/him⟧]

Attachment: bug933145-omit-needless-sandbox-hg0.diff

Comment 2

[Guillaume Destuynder [:kang] [this account is no longer active]]

given that this is supposed to be used for desktop too, i would use method 2.

Comment 3

[Guillaume Destuynder [:kang] [this account is no longer active]]

Comment on attachment 8403034 [details] [diff] [review]
bug933145-omit-needless-sandbox-hg0.diff

Review of attachment 8403034 [details] [diff] [review]:
-----------------------------------------------------------------

Other than the method this looks fine

Comment 4

[Jed Davis [:jld] ⟨⏰|UTC-8⟩ ⟦he/him⟧]

(In reply to Guillaume Destuynder [:kang] (use NEEDINFO!) from comment #2)
> given that this is supposed to be used for desktop too, i would use method 2.

On desktop we'll hopefully have a more up-to-date toolchain.

Comment 5

[Jed Davis [:jld] ⟨⏰|UTC-8⟩ ⟦he/him⟧]

Attachment: bug993145-omit-needless-sandbox-hg1.diff

Fixed bug number in commit message; carrying over r+.

Comment 6

[Carsten Book [:Tomcat]]

https://hg.mozilla.org/integration/mozilla-inbound/rev/d0137753c9d6

Comment 7

[Ryan VanderMeulen [:RyanVM]]

https://hg.mozilla.org/mozilla-central/rev/d0137753c9d6