Description: uninitialized kernel data leak in userspace coredumps
References:
 https://www.openwall.com/lists/oss-security/2020/05/06/1
 https://github.com/ruscur/linux/commit/a95cdec9fa0c08e6eeb410d461c03af8fd1fef0a
 https://lore.kernel.org/lkml/20200419100848.63472-1-glider@google.com/
 https://github.com/google/kmsan/issues/76
 https://twitter.com/grsecurity/status/1252558055629299712
 https://bugzilla.redhat.com/show_bug.cgi?id=1831399
Notes:
 carnil> Proposed fix has been accepted in linux-next tree, as per
 carnil> https://bugzilla.redhat.com/show_bug.cgi?id=1831399#c15
 bwh> The leak appears to have been introduced for x86 in 4.8 by
 bwh> commit 91c3dba7dbc1 "x86/fpu/xstate: Fix PTRACE frames for
 bwh> XSAVES".  A comparison of register sections in core dump
 bwh> under different Debian kernels agrees with that, as does
 bwh> @grsecurity (further down the thread).  Some other
 bwh> architectures are affected though.
Bugs:
upstream: released (5.7) [1d605416fb7175e1adf094251466caa52093b413]
4.19-upstream-stable: released (4.19.126) [61ce1733b30fdcf45e31308bc7795b6dc7f2ffba]
4.9-upstream-stable: released (4.9.226) [d228bc4b19e0b1c35f3eb404acbf1d607c01e64c]
3.16-upstream-stable: released (3.16.85) [d03daec2e50aa2a0b6de2c3572af5e1d61f9d132]
sid: released (5.6.14-2) [bugfix/all/fs-binfmt_elf.c-allocate-initialized-memory-in-fill_.patch]
4.19-buster-security: released (4.19.118-2+deb10u1) [bugfix/all/fs-binfmt_elf.c-allocate-initialized-memory-in-fill_.patch]
4.9-stretch-security: released (4.9.210-1+deb9u1) [bugfix/all/fs-binfmt_elf.c-allocate-initialized-memory-in-fill_.patch]
3.16-jessie-security: ignored "Does not affect supported architectures"