[<prev] [next>] [thread-next>] [day] [month] [year] [list]
Message-ID: <20190719170343.GA13680@linux.intel.com>
Date: Fri, 19 Jul 2019 10:03:44 -0700
From: Sean Christopherson <sean.j.christopherson@...el.com>
To: Thomas Gleixner <tglx@...utronix.de>,
Andy Lutomirski <luto@...nel.org>,
Vincenzo Frascino <vincenzo.frascino@....com>
Cc: x86@...nel.org, linux-kernel@...r.kernel.org
Subject: [5.2 REGRESSION] Generic vDSO breaks seccomp-enabled userspace on
i386
The generic vDSO implementation, starting with commit
7ac870747988 ("x86/vdso: Switch to generic vDSO implementation")
breaks seccomp-enabled userspace on 32-bit x86 (i386) kernels. Prior to
the generic implementation, the x86 vDSO used identical code for both
x86_64 and i386 kernels, which worked because it did all calcuations using
structs with naturally sized variables, i.e. didn't use __kernel_timespec.
The generic vDSO does its internal calculations using __kernel_timespec,
which in turn requires the i386 fallback syscall to use the 64-bit
variation, __NR_clock_gettime64.
Using __NR_clock_gettime64 instead of __NR_clock_gettime breaks userspace
applications that use seccomp filtering to block syscalls, as applications
are completely unaware of the newly added of __NR_clock_gettime64, e.g.
sshd gets zapped on syscall(403) when attempting to ssh into the system.
I can fudge around the issue easily enough, but I have no idea how to fix
this properly without duplicating __cvdso_clock_gettime, do_hres, etc...,
especially now that the i386 vDSO exposes __vdso_clock_gettime64().
Powered by blists - more mailing lists