user/sven/linux.git/scripts, branch v4.15.5

kasan: rework Kconfig settings

2018-02-16T19:06:53Z

commit e7c52b84fb18f08ce49b6067ae6285aca79084a8 upstream. We get a lot of very large stack frames using gcc-7.0.1 with the default -fsanitize-address-use-after-scope --param asan-stack=1 options, which can easily cause an overflow of the kernel stack, e.g. drivers/gpu/drm/i915/gvt/handlers.c:2434:1: warning: the frame size of 46176 bytes is larger than 3072 bytes drivers/net/wireless/ralink/rt2x00/rt2800lib.c:5650:1: warning: the frame size of 23632 bytes is larger than 3072 bytes lib/atomic64_test.c:250:1: warning: the frame size of 11200 bytes is larger than 3072 bytes drivers/gpu/drm/i915/gvt/handlers.c:2621:1: warning: the frame size of 9208 bytes is larger than 3072 bytes drivers/media/dvb-frontends/stv090x.c:3431:1: warning: the frame size of 6816 bytes is larger than 3072 bytes fs/fscache/stats.c:287:1: warning: the frame size of 6536 bytes is larger than 3072 bytes To reduce this risk, -fsanitize-address-use-after-scope is now split out into a separate CONFIG_KASAN_EXTRA Kconfig option, leading to stack frames that are smaller than 2 kilobytes most of the time on x86_64. An earlier version of this patch also prevented combining KASAN_EXTRA with KASAN_INLINE, but that is no longer necessary with gcc-7.0.1. All patches to get the frame size below 2048 bytes with CONFIG_KASAN=y and CONFIG_KASAN_EXTRA=n have been merged by maintainers now, so we can bring back that default now. KASAN_EXTRA=y still causes lots of warnings but now defaults to !COMPILE_TEST to disable it in allmodconfig, and it remains disabled in all other defconfigs since it is a new option. I arbitrarily raise the warning limit for KASAN_EXTRA to 3072 to reduce the noise, but an allmodconfig kernel still has around 50 warnings on gcc-7. I experimented a bit more with smaller stack frames and have another follow-up series that reduces the warning limit for 64-bit architectures to 1280 bytes (without CONFIG_KASAN). With earlier versions of this patch series, I also had patches to address the warnings we get with KASAN and/or KASAN_EXTRA, using a "noinline_if_stackbloat" annotation. That annotation now got replaced with a gcc-8 bugfix (see https://gcc.gnu.org/bugzilla/show_bug.cgi?id=81715) and a workaround for older compilers, which means that KASAN_EXTRA is now just as bad as before and will lead to an instant stack overflow in a few extreme cases. This reverts parts of commit 3f181b4d8652 ("lib/Kconfig.debug: disable -Wframe-larger-than warnings with KASAN=y"). Two patches in linux-next should be merged first to avoid introducing warnings in an allmodconfig build: 3cd890dbe2a4 ("media: dvb-frontends: fix i2c access helpers for KASAN") 16c3ada89cff ("media: r820t: fix r820t_write_reg for KASAN") Do we really need to backport this? I think we do: without this patch, enabling KASAN will lead to unavoidable kernel stack overflow in certain device drivers when built with gcc-7 or higher on linux-4.10+ or any version that contains a backport of commit c5caf21ab0cf8. Most people are probably still on older compilers, but it will get worse over time as they upgrade their distros. The warnings we get on kernels older than this should all be for code that uses dangerously large stack frames, though most of them do not cause an actual stack overflow by themselves.The asan-stack option was added in linux-4.0, and commit 3f181b4d8652 ("lib/Kconfig.debug: disable -Wframe-larger-than warnings with KASAN=y") effectively turned off the warning for allmodconfig kernels, so I would like to see this fix backported to any kernels later than 4.0. I have done dozens of fixes for individual functions with stack frames larger than 2048 bytes with asan-stack, and I plan to make sure that all those fixes make it into the stable kernels as well (most are already there). Part of the complication here is that asan-stack (from 4.0) was originally assumed to always require much larger stacks, but that turned out to be a combination of multiple gcc bugs that we have now worked around and fixed, but sanitize-address-use-after-scope (from v4.10) has a much higher inherent stack usage and also suffers from at least three other problems that we have analyzed but not yet fixed upstream, each of them makes the stack usage more severe than it should be. Link: http://lkml.kernel.org/r/20171221134744.2295529-1-arnd@arndb.de Signed-off-by: Arnd Bergmann Acked-by: Andrey Ryabinin Cc: Mauro Carvalho Chehab Cc: Andrey Ryabinin Cc: Alexander Potapenko Cc: Dmitry Vyukov Cc: Andrey Konovalov Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds Signed-off-by: Greg Kroah-Hartman

kasan: don't emit builtin calls when sanitization is off

2018-02-16T19:06:53Z

commit 0e410e158e5baa1300bdf678cea4f4e0cf9d8b94 upstream. With KASAN enabled the kernel has two different memset() functions, one with KASAN checks (memset) and one without (__memset). KASAN uses some macro tricks to use the proper version where required. For example memset() calls in mm/slub.c are without KASAN checks, since they operate on poisoned slab object metadata. The issue is that clang emits memset() calls even when there is no memset() in the source code. They get linked with improper memset() implementation and the kernel fails to boot due to a huge amount of KASAN reports during early boot stages. The solution is to add -fno-builtin flag for files with KASAN_SANITIZE := n marker. Link: http://lkml.kernel.org/r/8ffecfffe04088c52c42b92739c2bd8a0bcb3f5e.1516384594.git.andreyknvl@google.com Signed-off-by: Andrey Konovalov Acked-by: Nick Desaulniers Cc: Masahiro Yamada Cc: Michal Marek Cc: Andrey Ryabinin Cc: Alexander Potapenko Cc: Dmitry Vyukov Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds Signed-off-by: Greg Kroah-Hartman

module/retpoline: Warn about missing retpoline in module

2018-02-07T19:14:06Z

commit caf7501a1b4ec964190f31f9c3f163de252273b8 There's a risk that a kernel which has full retpoline mitigations becomes vulnerable when a module gets loaded that hasn't been compiled with the right compiler or the right option. To enable detection of that mismatch at module load time, add a module info string "retpoline" at build time when the module was compiled with retpoline support. This only covers compiled C source, but assembler source or prebuilt object files are not checked. If a retpoline enabled kernel detects a non retpoline protected module at load time, print a warning and report it in the sysfs vulnerability file. [ tglx: Massaged changelog ] Signed-off-by: Andi Kleen Signed-off-by: Thomas Gleixner Cc: David Woodhouse Cc: gregkh@linuxfoundation.org Cc: torvalds@linux-foundation.org Cc: jeyu@kernel.org Cc: arjan@linux.intel.com Link: https://lkml.kernel.org/r/20180125235028.31211-1-andi@firstfloor.org Signed-off-by: Greg Kroah-Hartman

scripts/gdb/linux/tasks.py: fix get_thread_info

2018-01-19T18:09:41Z

Since kernel 4.9, the thread_info has been moved into task_struct, no longer locates at the bottom of kernel stack. See commits c65eacbe290b ("sched/core: Allow putting thread_info into task_struct") and 15f4eae70d36 ("x86: Move thread_info into task_struct"). Before fix: (gdb) set $current = $lx_current() (gdb) p $lx_thread_info($current) $1 = {flags = 1470918301} (gdb) p $current.thread_info $2 = {flags = 2147483648} After fix: (gdb) p $lx_thread_info($current) $1 = {flags = 2147483648} (gdb) p $current.thread_info $2 = {flags = 2147483648} Link: http://lkml.kernel.org/r/20180118210159.17223-1-imxikangjie@gmail.com Fixes: 15f4eae70d36 ("x86: Move thread_info into task_struct") Signed-off-by: Xi Kangjie Acked-by: Jan Kiszka Acked-by: Kieran Bingham Cc: Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds

scripts/decodecode: fix decoding for AArch64 (arm64) instructions

2018-01-19T18:09:41Z

There are a couple of problems with the decodecode script and arm64: 1. AArch64 objdump refuses to disassemble .4byte directives as instructions, insisting that they are data values and displaying them as: a94153f3 .word 0xa94153f3 <-- trapping instruction This is resolved by using the .inst directive instead. 2. Disassembly of branch instructions attempts to provide the target as an offset from a symbol, e.g.: 0: 34000082 cbz w2, 10 <.text+0x10> however this falls foul of the grep -v, which matches lines containing ".text" and ends up removing all branch instructions from the dump. This patch resolves both issues by using the .inst directive for 4-byte quantities on arm64 and stripping the resulting binaries (as is done on arm already) to remove the mapping symbols. Link: http://lkml.kernel.org/r/1506596147-23630-1-git-send-email-will.deacon@arm.com Signed-off-by: Will Deacon Reviewed-by: Dave Martin Cc: Michal Marek Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds

Merge branch 'x86-pti-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip

2018-01-17T19:54:56Z

Pull x86 pti bits and fixes from Thomas Gleixner: "This last update contains: - An objtool fix to prevent a segfault with the gold linker by changing the invocation order. That's not just for gold, it's a general robustness improvement. - An improved error message for objtool which spares tearing hairs. - Make KASAN fail loudly if there is not enough memory instead of oopsing at some random place later - RSB fill on context switch to prevent RSB underflow and speculation through other units. - Make the retpoline/RSB functionality work reliably for both Intel and AMD - Add retpoline to the module version magic so mismatch can be detected - A small (non-fix) update for cpufeatures which prevents cpu feature clashing for the upcoming extra mitigation bits to ease backporting" * 'x86-pti-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip: module: Add retpoline tag to VERMAGIC x86/cpufeature: Move processor tracing out of scattered features objtool: Improve error message for bad file argument objtool: Fix seg fault with gold linker x86/retpoline: Add LFENCE to the retpoline/RSB filling RSB macros x86/retpoline: Fill RSB on context switch for affected CPUs x86/kasan: Panic if there is not enough memory to boot

objtool: Fix seg fault with gold linker

2018-01-16T00:27:27Z

Objtool segfaults when the gold linker is used with CONFIG_MODVERSIONS=y and CONFIG_UNWINDER_ORC=y. With CONFIG_MODVERSIONS=y, the .o file gets passed to the linker before being passed to objtool. The gold linker seems to strip unused ELF symbols by default, which confuses objtool and causes the seg fault when it's trying to generate ORC metadata. Objtool should really be running immediately after GCC anyway, without a linker call in between. Change the makefile ordering so that objtool is called before the linker. Reported-and-tested-by: Markus Signed-off-by: Josh Poimboeuf Cc: Linus Torvalds Cc: Peter Zijlstra Cc: Thomas Gleixner Fixes: ee9f8fce9964 ("x86/unwind: Add the ORC unwinder") Link: http://lkml.kernel.org/r/355f04da33581f4a3bf82e5b512973624a1e23a2.1516025651.git.jpoimboe@redhat.com Signed-off-by: Ingo Molnar

Merge tag 'kbuild-fixes-v4.15' of git://git.kernel.org/pub/scm/linux/kernel/git/masahiroy/linux-kbuild

2018-01-13T21:24:56Z

Pull Kbuild fixes from Masahiro Yamada: - fix cross-compilation for architectures that setup CROSS_COMPILE in their arch Makefile - fix Kconfig rational operators for bool / tristate - drop a gperf-generated file from .gitignore * tag 'kbuild-fixes-v4.15' of git://git.kernel.org/pub/scm/linux/kernel/git/masahiroy/linux-kbuild: genksyms: drop *.hash.c from .gitignore kconfig: fix relational operators for bool and tristate symbols kbuild: move cc-option and cc-disable-warning after incl. arch Makefile

genksyms: drop *.hash.c from .gitignore

2018-01-13T12:50:13Z

This is a left-over of commit bb3290d91695 ("Remove gperf usage from toolchain"). We do not generate a hash function any more. Signed-off-by: Masahiro Yamada

kconfig: fix relational operators for bool and tristate symbols

2018-01-05T17:31:23Z

Since commit 31847b67bec0 ("kconfig: allow use of relations other than (in)equality") it is possible to use relational operators in Kconfig statements. However, those operators give unexpected results when applied to bool/tristate values: (n < y) = y (correct) (m < y) = y (correct) (n < m) = n (wrong) This happens because relational operators process bool and tristate symbols as strings and m sorts before n. It makes little sense to do a lexicographical compare on bool and tristate values though. Documentation/kbuild/kconfig-language.txt states that expression can have a value of 'n', 'm' or 'y' (or 0, 1, 2 respectively for calculations). Let's make it so for relational comparisons with bool/tristate expressions as well and document them. If at least one symbol is an actual string then the lexicographical compare works just as before. Signed-off-by: Nicolas Pitre Acked-by: Randy Dunlap Tested-by: Randy Dunlap Signed-off-by: Masahiro Yamada