From: Ben Hutchings <ben@decadent.org.uk>
To: linux-kernel@vger.kernel.org, stable@vger.kernel.org
Cc: akpm@linux-foundation.org,
"Jinpu Wang" <jinpu.wang@profitbricks.com>,
kernel-hardening@lists.openwall.com,
"Jiri Slaby" <jslaby@suse.cz>,
"Andy Lutomirski" <luto@kernel.org>,
"Linus Torvalds" <torvalds@linux-foundation.org>,
"Jan Beulich" <JBeulich@suse.com>,
alan@linux.intel.com, "Thomas Gleixner" <tglx@linutronix.de>,
"Dan Williams" <dan.j.williams@intel.com>,
gregkh@linuxfoundation.org, linux-arch@vger.kernel.org
Subject: [PATCH 3.16 56/76] x86/syscall: Sanitize syscall table de-references under speculation
Date: Mon, 12 Mar 2018 03:06:12 +0000 [thread overview]
Message-ID: <lsq.1520823972.93871245@decadent.org.uk> (raw)
In-Reply-To: <lsq.1520823971.5976735@decadent.org.uk>
3.16.56-rc1 review patch. If anyone has any objections, please let me know.
------------------
From: Ben Hutchings <ben@decadent.org.uk>
commit 2fbd7af5af8665d18bcefae3e9700be07e22b681 upstream.
The upstream version of this, touching C code, was written by Dan Williams,
with the following description:
> The syscall table base is a user controlled function pointer in kernel
> space. Use array_index_nospec() to prevent any out of bounds speculation.
>
> While retpoline prevents speculating into a userspace directed target it
> does not stop the pointer de-reference, the concern is leaking memory
> relative to the syscall table base, by observing instruction cache
> behavior.
The x86_64 assembly version for 4.4 was written by Jiri Slaby, with
the following description:
> In 4.4.118, we have commit c8961332d6da (x86/syscall: Sanitize syscall
> table de-references under speculation), which is a backport of upstream
> commit 2fbd7af5af86. But it fixed only the C part of the upstream patch
> -- the IA32 sysentry. So it ommitted completely the assembly part -- the
> 64bit sysentry.
>
> Fix that in this patch by explicit array_index_mask_nospec written in
> assembly. The same was used in lib/getuser.S.
>
> However, to have "sbb" working properly, we have to switch from "cmp"
> against (NR_syscalls-1) to (NR_syscalls), otherwise the last syscall
> number would be "and"ed by 0. It is because the original "ja" relies on
> "CF" or "ZF", but we rely only on "CF" in "sbb". That means: switch to
> "jae" conditional jump too.
>
> Final note: use rcx for mask as this is exactly what is overwritten by
> the 4th syscall argument (r10) right after.
In 3.16 the x86_32 syscall table lookup is also written in assembly.
So I've taken Jiri's version and added similar masking in entry_32.S,
using edx as the temporary. edx is clobbered by SAVE_REGS and seems
to be free at this point.
Cc: Dan Williams <dan.j.williams@intel.com>
Cc: Jiri Slaby <jslaby@suse.cz>
Cc: Jan Beulich <JBeulich@suse.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: linux-arch@vger.kernel.org
Cc: kernel-hardening@lists.openwall.com
Cc: gregkh@linuxfoundation.org
Cc: Andy Lutomirski <luto@kernel.org>
Cc: alan@linux.intel.com
Cc: Jinpu Wang <jinpu.wang@profitbricks.com>
Signed-off-by: Ben Hutchings <ben@decadent.org.uk>
---
--- a/arch/x86/kernel/entry_32.S
+++ b/arch/x86/kernel/entry_32.S
@@ -426,6 +426,8 @@ sysenter_past_esp:
sysenter_do_call:
cmpl $(NR_syscalls), %eax
jae sysenter_badsys
+ sbb %edx, %edx /* array_index_mask_nospec() */
+ and %edx, %eax
call *sys_call_table(,%eax,4)
sysenter_after_call:
movl %eax,PT_EAX(%esp)
@@ -503,6 +505,8 @@ ENTRY(system_call)
cmpl $(NR_syscalls), %eax
jae syscall_badsys
syscall_call:
+ sbb %edx, %edx /* array_index_mask_nospec() */
+ and %edx, %eax
call *sys_call_table(,%eax,4)
syscall_after_call:
movl %eax,PT_EAX(%esp) # store the return value
--- a/arch/x86/kernel/entry_64.S
+++ b/arch/x86/kernel/entry_64.S
@@ -445,12 +445,14 @@ GLOBAL(system_call_after_swapgs)
jnz tracesys
system_call_fastpath:
#if __SYSCALL_MASK == ~0
- cmpq $__NR_syscall_max,%rax
+ cmpq $NR_syscalls, %rax
#else
andl $__SYSCALL_MASK,%eax
- cmpl $__NR_syscall_max,%eax
+ cmpl $NR_syscalls, %eax
#endif
- ja badsys
+ jae badsys
+ sbb %rcx, %rcx /* array_index_mask_nospec() */
+ and %rcx, %rax
movq %r10,%rcx
#ifdef CONFIG_RETPOLINE
movq sys_call_table(, %rax, 8), %rax
@@ -577,12 +579,14 @@ tracesys:
LOAD_ARGS ARGOFFSET, 1
RESTORE_REST
#if __SYSCALL_MASK == ~0
- cmpq $__NR_syscall_max,%rax
+ cmpq $NR_syscalls, %rax
#else
andl $__SYSCALL_MASK,%eax
- cmpl $__NR_syscall_max,%eax
+ cmpl $NR_syscalls, %eax
#endif
- ja int_ret_from_sys_call /* RAX(%rsp) set to -ENOSYS above */
+ jae int_ret_from_sys_call /* RAX(%rsp) set to -ENOSYS above */
+ sbb %rcx, %rcx /* array_index_mask_nospec() */
+ and %rcx, %rax
movq %r10,%rcx /* fixup for C */
#ifdef CONFIG_RETPOLINE
movq sys_call_table(, %rax, 8), %rax
next prev parent reply other threads:[~2018-03-12 3:06 UTC|newest]
Thread overview: 81+ messages / expand[flat|nested] mbox.gz Atom feed top
2018-03-12 3:06 [PATCH 3.16 00/76] 3.16.56-rc1 review Ben Hutchings
2018-03-12 3:06 ` [PATCH 3.16 08/76] x86/cpufeatures: Add X86_BUG_SPECTRE_V[12] Ben Hutchings
2018-03-12 3:06 ` [PATCH 3.16 24/76] x86/spectre: Add boot time option to select Spectre v2 mitigation Ben Hutchings
2018-03-12 3:06 ` [PATCH 3.16 38/76] x86/pti: Document fix wrong index Ben Hutchings
2018-03-12 3:06 ` [PATCH 3.16 55/76] x86/get_user: Use pointer masking to limit speculation Ben Hutchings
2018-03-12 3:06 ` [PATCH 3.16 36/76] kprobes/x86: Blacklist indirect thunk functions for kprobes Ben Hutchings
2018-03-12 3:06 ` [PATCH 3.16 09/76] x86/cpu: Merge bugs.c and bugs_64.c Ben Hutchings
2018-03-12 3:06 ` [PATCH 3.16 28/76] x86/retpoline/hyperv: Convert assembler indirect jumps Ben Hutchings
2018-03-12 3:06 ` [PATCH 3.16 58/76] nl80211: Sanitize array index in parse_txq_params Ben Hutchings
2018-03-12 3:06 ` [PATCH 3.16 52/76] array_index_nospec: Sanitize speculative array de-references Ben Hutchings
2018-03-12 3:06 ` [PATCH 3.16 02/76] x86/Documentation: Add PTI description Ben Hutchings
2018-03-12 3:06 ` [PATCH 3.16 47/76] x86/nospec: Fix header guards names Ben Hutchings
2018-03-12 3:06 ` [PATCH 3.16 11/76] x86/cpu: Implement CPU vulnerabilites sysfs functions Ben Hutchings
2018-03-12 3:06 ` [PATCH 3.16 53/76] x86: Implement array_index_mask_nospec Ben Hutchings
2018-03-12 3:06 ` [PATCH 3.16 74/76] x86/usercopy: Replace open coded stac/clac with __uaccess_{begin, end} Ben Hutchings
2018-03-12 3:06 ` [PATCH 3.16 35/76] retpoline: Introduce start/end markers of indirect thunk Ben Hutchings
2018-03-12 3:06 ` [PATCH 3.16 62/76] x86/kvm: Update spectre-v1 mitigation Ben Hutchings
2018-03-12 3:06 ` [PATCH 3.16 61/76] x86/paravirt: Remove 'noreplace-paravirt' cmdline option Ben Hutchings
2018-03-12 3:06 ` [PATCH 3.16 27/76] x86/retpoline/ftrace: Convert ftrace assembler indirect jumps Ben Hutchings
2018-03-12 3:06 ` [PATCH 3.16 33/76] x86/retpoline: Remove compile time warning Ben Hutchings
2018-03-12 3:06 ` [PATCH 3.16 18/76] x86/cpu/AMD: Use LFENCE_RDTSC in preference to MFENCE_RDTSC Ben Hutchings
2018-03-12 3:06 ` [PATCH 3.16 44/76] KVM: x86: Make indirect calls in emulator speculation safe Ben Hutchings
2018-03-12 3:06 ` [PATCH 3.16 45/76] KVM: VMX: Make indirect call " Ben Hutchings
2018-03-12 3:06 ` [PATCH 3.16 69/76] nospec: Kill array_index_nospec_mask_check() Ben Hutchings
2018-03-12 3:06 ` [PATCH 3.16 31/76] x86/retpoline/irq32: Convert assembler indirect jumps Ben Hutchings
2018-03-12 3:06 ` [PATCH 3.16 39/76] x86/retpoline: Optimize inline assembler for vmexit_fill_RSB Ben Hutchings
2018-03-12 3:06 ` [PATCH 3.16 29/76] x86/retpoline/xen: Convert Xen hypercall indirect jumps Ben Hutchings
2018-03-12 3:06 ` [PATCH 3.16 13/76] x86/alternatives: Guard NOPs optimization Ben Hutchings
2018-03-12 3:06 ` [PATCH 3.16 46/76] module/retpoline: Warn about missing retpoline in module Ben Hutchings
2018-03-12 3:06 ` [PATCH 3.16 75/76] x86/uaccess: Use __uaccess_begin_nospec() and uaccess_try_nospec Ben Hutchings
2018-03-12 3:06 ` [PATCH 3.16 03/76] x86/cpu: Factor out application of forced CPU caps Ben Hutchings
2018-03-12 3:06 ` [PATCH 3.16 37/76] kprobes/x86: Disable optimizing on the function jumps to indirect thunk Ben Hutchings
2018-03-12 3:06 ` [PATCH 3.16 71/76] x86: reorganize SMAP handling in user space accesses Ben Hutchings
2018-03-12 3:06 ` [PATCH 3.16 19/76] x86/asm: Make asm/alternative.h safe from assembly Ben Hutchings
2018-03-12 3:06 ` [PATCH 3.16 63/76] x86/retpoline: Avoid retpolines for built-in __init functions Ben Hutchings
2018-03-12 3:06 ` [PATCH 3.16 10/76] sysfs/cpu: Add vulnerability folder Ben Hutchings
2018-03-12 3:06 ` [PATCH 3.16 57/76] vfs, fdtable: Prevent bounds-check bypass via speculative execution Ben Hutchings
2018-03-12 3:06 ` [PATCH 3.16 05/76] x86/cpufeatures: Add X86_BUG_CPU_INSECURE Ben Hutchings
2018-03-12 3:06 ` [PATCH 3.16 66/76] x86/cpufeatures: Clean up Spectre v2 related CPUID flags Ben Hutchings
2018-03-12 3:06 ` [PATCH 3.16 15/76] x86/alternatives: Make optimize_nops() interrupt safe and synced Ben Hutchings
2018-03-12 3:06 ` [PATCH 3.16 48/76] x86/bugs: Drop one "mitigation" from dmesg Ben Hutchings
2018-03-12 3:06 ` [PATCH 3.16 54/76] x86: Introduce barrier_nospec Ben Hutchings
2018-03-12 3:06 ` [PATCH 3.16 25/76] x86/retpoline/crypto: Convert crypto assembler indirect jumps Ben Hutchings
2018-03-12 3:06 ` [PATCH 3.16 26/76] x86/retpoline/entry: Convert entry " Ben Hutchings
2018-03-12 3:06 ` [PATCH 3.16 40/76] x86/cpu/intel: Introduce macros for Intel family numbers Ben Hutchings
2018-03-12 3:06 ` [PATCH 3.16 64/76] x86/spectre: Simplify spectre_v2 command line parsing Ben Hutchings
2018-03-12 3:06 ` [PATCH 3.16 22/76] x86/asm: Use register variable to get stack pointer value Ben Hutchings
2018-03-12 3:06 ` [PATCH 3.16 17/76] x86/cpu/AMD: Make LFENCE a serializing instruction Ben Hutchings
2018-03-12 3:06 ` [PATCH 3.16 70/76] nospec: Include <asm/barrier.h> dependency Ben Hutchings
2018-03-12 3:06 ` [PATCH 3.16 23/76] x86/retpoline: Add initial retpoline support Ben Hutchings
2018-03-12 3:06 ` [PATCH 3.16 59/76] x86/spectre: Report get_user mitigation for spectre_v1 Ben Hutchings
2018-03-12 3:06 ` [PATCH 3.16 16/76] x86/alternatives: Fix optimize_nops() checking Ben Hutchings
2018-03-12 3:06 ` [PATCH 3.16 04/76] x86/cpufeatures: Make CPU bugs sticky Ben Hutchings
2018-03-12 3:06 ` [PATCH 3.16 43/76] x86/retpoline: Remove the esp/rsp thunk Ben Hutchings
2018-03-12 3:06 ` [PATCH 3.16 21/76] x86: Clean up current_stack_pointer Ben Hutchings
2018-03-12 3:06 ` [PATCH 3.16 20/76] kconfig.h: use __is_defined() to check if MODULE is defined Ben Hutchings
2018-03-12 3:06 ` [PATCH 3.16 67/76] x86/spectre: Fix an error message Ben Hutchings
2018-03-12 3:06 ` [PATCH 3.16 34/76] x86/retpoline: Add LFENCE to the retpoline/RSB filling RSB macros Ben Hutchings
2018-03-12 3:06 ` [PATCH 3.16 73/76] x86: Introduce __uaccess_begin_nospec() and uaccess_try_nospec Ben Hutchings
2018-03-12 3:06 ` [PATCH 3.16 01/76] kvm: vmx: Scrub hardware GPRs at VM-exit Ben Hutchings
2018-03-12 3:06 ` [PATCH 3.16 14/76] x86/alternatives: Fix ALTERNATIVE_2 padding generation properly Ben Hutchings
2018-03-12 3:06 ` [PATCH 3.16 41/76] x86/retpoline: Fill RSB on context switch for affected CPUs Ben Hutchings
2018-03-12 3:06 ` [PATCH 3.16 06/76] x86/cpu, x86/pti: Do not enable PTI on AMD processors Ben Hutchings
2018-03-12 3:06 ` [PATCH 3.16 68/76] nospec: Move array_index_nospec() parameter checking into separate macro Ben Hutchings
2018-03-12 3:06 ` [PATCH 3.16 12/76] sysfs/cpu: Fix typos in vulnerability documentation Ben Hutchings
2018-03-12 3:06 ` [PATCH 3.16 32/76] x86/retpoline: Fill return stack buffer on vmexit Ben Hutchings
2018-03-12 3:06 ` [PATCH 3.16 07/76] x86/pti: Rename BUG_CPU_INSECURE to BUG_CPU_MELTDOWN Ben Hutchings
2018-03-12 3:06 ` [PATCH 3.16 49/76] x86/cpu/bugs: Make retpoline module warning conditional Ben Hutchings
2018-03-12 3:06 ` [PATCH 3.16 51/76] Documentation: Document array_index_nospec Ben Hutchings
2018-03-12 3:06 ` [PATCH 3.16 42/76] x86/cpu: Change type of x86_cache_size variable to unsigned int Ben Hutchings
2018-03-12 3:06 ` [PATCH 3.16 30/76] x86/retpoline/checksum32: Convert assembler indirect jumps Ben Hutchings
2018-03-12 3:06 ` Ben Hutchings [this message]
2018-03-12 7:32 ` [PATCH 3.16 56/76] x86/syscall: Sanitize syscall table de-references under speculation Jiri Slaby
2018-03-19 0:59 ` Ben Hutchings
2018-03-12 3:06 ` [PATCH 3.16 65/76] x86/speculation: Fix typo IBRS_ATT, which should be IBRS_ALL Ben Hutchings
2018-03-12 3:06 ` [PATCH 3.16 60/76] x86/spectre: Fix spelling mistake: "vunerable"-> "vulnerable" Ben Hutchings
2018-03-12 3:06 ` [PATCH 3.16 76/76] x86: fix build warnign with 32-bit PAE Ben Hutchings
2018-03-12 3:06 ` [PATCH 3.16 72/76] x86: fix SMAP in 32-bit environments Ben Hutchings
2018-03-12 15:00 ` [PATCH 3.16 00/76] 3.16.56-rc1 review Guenter Roeck
2018-03-12 16:45 ` Guenter Roeck
2018-03-20 17:25 ` Ben Hutchings
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=lsq.1520823972.93871245@decadent.org.uk \
--to=ben@decadent.org.uk \
--cc=JBeulich@suse.com \
--cc=akpm@linux-foundation.org \
--cc=alan@linux.intel.com \
--cc=dan.j.williams@intel.com \
--cc=gregkh@linuxfoundation.org \
--cc=jinpu.wang@profitbricks.com \
--cc=jslaby@suse.cz \
--cc=kernel-hardening@lists.openwall.com \
--cc=linux-arch@vger.kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=luto@kernel.org \
--cc=stable@vger.kernel.org \
--cc=tglx@linutronix.de \
--cc=torvalds@linux-foundation.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox
all inboxes | Powered by JetHome®