linux-stable.git/arch/x86/lib, branch linux-5.4.y

x86/insn: Fix PUSH instruction in x86 instruction decoder opcode map

2024-06-16T11:28:39+00:00

[ Upstream commit 59162e0c11d7257cde15f907d19fefe26da66692 ]

The x86 instruction decoder is used not only for decoding kernel
instructions. It is also used by perf uprobes (user space probes) and by
perf tools Intel Processor Trace decoding. Consequently, it needs to
support instructions executed by user space also.

Opcode 0x68 PUSH instruction is currently defined as 64-bit operand size
only i.e. (d64). That was based on Intel SDM Opcode Map. However that is
contradicted by the Instruction Set Reference section for PUSH in the
same manual.

Remove 64-bit operand size only annotation from opcode 0x68 PUSH
instruction.

Example:

  $ cat pushw.s
  .global  _start
  .text
  _start:
          pushw   $0x1234
          mov     $0x1,%eax   # system call number (sys_exit)
          int     $0x80
  $ as -o pushw.o pushw.s
  $ ld -s -o pushw pushw.o
  $ objdump -d pushw | tail -4
  0000000000401000 <.text>:
    401000:       66 68 34 12             pushw  $0x1234
    401004:       b8 01 00 00 00          mov    $0x1,%eax
    401009:       cd 80                   int    $0x80
  $ perf record -e intel_pt//u ./pushw
  [ perf record: Woken up 1 times to write data ]
  [ perf record: Captured and wrote 0.014 MB perf.data ]

 Before:

  $ perf script --insn-trace=disasm
  Warning:
  1 instruction trace errors
           pushw   10349 [000] 10586.869237014:            401000 [unknown] (/home/ahunter/git/misc/rtit-tests/pushw)           pushw $0x1234
           pushw   10349 [000] 10586.869237014:            401006 [unknown] (/home/ahunter/git/misc/rtit-tests/pushw)           addb %al, (%rax)
           pushw   10349 [000] 10586.869237014:            401008 [unknown] (/home/ahunter/git/misc/rtit-tests/pushw)           addb %cl, %ch
           pushw   10349 [000] 10586.869237014:            40100a [unknown] (/home/ahunter/git/misc/rtit-tests/pushw)           addb $0x2e, (%rax)
   instruction trace error type 1 time 10586.869237224 cpu 0 pid 10349 tid 10349 ip 0x40100d code 6: Trace doesn't match instruction

 After:

  $ perf script --insn-trace=disasm
             pushw   10349 [000] 10586.869237014:            401000 [unknown] (./pushw)           pushw $0x1234
             pushw   10349 [000] 10586.869237014:            401004 [unknown] (./pushw)           movl $1, %eax

Fixes: eb13296cfaf6 ("x86: Instruction decoder API")
Signed-off-by: Adrian Hunter 
Signed-off-by: Ingo Molnar 
Link: https://lore.kernel.org/r/20240502105853.5338-3-adrian.hunter@intel.com
Signed-off-by: Sasha Levin

x86/lib: Fix overflow when counting digits

2024-01-25T22:34:21+00:00

[ Upstream commit a24d61c609813963aacc9f6ec8343f4fcaac7243 ]

tl;dr: The num_digits() function has a theoretical overflow issue.
But it doesn't affect any actual in-tree users.  Fix it by using
a larger type for one of the local variables.

Long version:

There is an overflow in variable m in function num_digits when val
is >= 1410065408 which leads to the digit calculation loop to
iterate more times than required. This results in either more
digits being counted or in some cases (for example where val is
1932683193) the value of m eventually overflows to zero and the
while loop spins forever).

Currently the function num_digits is currently only being used for
small values of val in the SMP boot stage for digit counting on the
number of cpus and NUMA nodes, so the overflow is never encountered.
However it is useful to fix the overflow issue in case the function
is used for other purposes in the future. (The issue was discovered
while investigating the digit counting performance in various
kernel helper functions rather than any real-world use-case).

The simplest fix is to make m a long long, the overhead in
multiplication speed for a long long is very minor for small values
of val less than 10000 on modern processors. The alternative
fix is to replace the multiplication with a constant division
by 10 loop (this compiles down to an multiplication and shift)
without needing to make m a long long, but this is slightly slower
than the fix in this commit when measured on a range of x86
processors).

[ dhansen: subject and changelog tweaks ]

Fixes: 646e29a1789a ("x86: Improve the printout of the SMP bootup CPU table")
Signed-off-by: Colin Ian King 
Signed-off-by: Dave Hansen 
Link: https://lore.kernel.org/all/20231102174901.2590325-1-colin.i.king%40gmail.com
Signed-off-by: Sasha Levin

x86/asm: Fix an assembler warning with current binutils

2023-02-06T06:52:48+00:00

commit 55d235361fccef573990dfa5724ab453866e7816 upstream.

Fix a warning: "found `movsd'; assuming `movsl' was meant"

Signed-off-by: Mikulas Patocka 
Signed-off-by: Ingo Molnar 
Cc: linux-kernel@vger.kernel.org
Signed-off-by: Greg Kroah-Hartman

x86/delay: Fix the wrong asm constraint in delay_loop()

2022-06-14T16:11:31+00:00

[ Upstream commit b86eb74098a92afd789da02699b4b0dd3f73b889 ]

The asm constraint does not reflect the fact that the asm statement can
modify the value of the local variable loops. Which it does.

Specifying the wrong constraint may lead to undefined behavior, it may
clobber random stuff (e.g. local variable, important temporary value in
regs, etc.). This is especially dangerous when the compiler decides to
inline the function and since it doesn't know that the value gets
modified, it might decide to use it from a register directly without
reloading it.

Change the constraint to "+a" to denote that the first argument is an
input and an output argument.

  [ bp: Fix typo, massage commit message. ]

Fixes: e01b70ef3eb3 ("x86: fix bug in arch/i386/lib/delay.c file, delay_loop function")
Signed-off-by: Ammar Faizi 
Signed-off-by: Borislav Petkov 
Link: https://lore.kernel.org/r/20220329104705.65256-2-ammarfaizi2@gnuweeb.org
Signed-off-by: Sasha Levin

x86: xen: insn: Decode Xen and KVM emulate-prefix signature

2022-05-15T17:54:47+00:00

commit 4d65adfcd1196818659d3bd9b42dccab291e1751 upstream.

Decode Xen and KVM's emulate-prefix signature by x86 insn decoder.
It is called "prefix" but actually not x86 instruction prefix, so
this adds insn.emulate_prefix_size field instead of reusing
insn.prefixes.

If x86 decoder finds a special sequence of instructions of
XEN_EMULATE_PREFIX and 'ud2a; .ascii "kvm"', it just counts the
length, set insn.emulate_prefix_size and fold it with the next
instruction. In other words, the signature and the next instruction
is treated as a single instruction.

Signed-off-by: Masami Hiramatsu 
Signed-off-by: Peter Zijlstra (Intel) 
Acked-by: Josh Poimboeuf 
Cc: Juergen Gross 
Cc: x86@kernel.org
Cc: Boris Ostrovsky 
Cc: Ingo Molnar 
Cc: Stefano Stabellini 
Cc: Andrew Cooper 
Cc: Borislav Petkov 
Cc: xen-devel@lists.xenproject.org
Cc: Randy Dunlap 
Link: https://lkml.kernel.org/r/156777564986.25081.4964537658500952557.stgit@devnote2
[mheyne: resolved contextual conflict in tools/objtools/sync-check.sh]
Signed-off-by: Maximilian Heyne 
Signed-off-by: Greg Kroah-Hartman

x86: __memcpy_flushcache: fix wrong alignment if size > 2^32

2022-05-09T07:03:26+00:00

[ Upstream commit a6823e4e360fe975bd3da4ab156df7c74c8b07f3 ]

The first "if" condition in __memcpy_flushcache is supposed to align the
"dest" variable to 8 bytes and copy data up to this alignment.  However,
this condition may misbehave if "size" is greater than 4GiB.

The statement min_t(unsigned, size, ALIGN(dest, 8) - dest); casts both
arguments to unsigned int and selects the smaller one.  However, the
cast truncates high bits in "size" and it results in misbehavior.

For example:

	suppose that size == 0x100000001, dest == 0x200000002
	min_t(unsigned, size, ALIGN(dest, 8) - dest) == min_t(0x1, 0xe) == 0x1;
	...
	dest += 0x1;

so we copy just one byte "and" dest remains unaligned.

This patch fixes the bug by replacing unsigned with size_t.

Signed-off-by: Mikulas Patocka 
Signed-off-by: Linus Torvalds 
Signed-off-by: Sasha Levin

x86/msr: Fix wr/rdmsr_safe_regs_on_cpu() prototypes

2021-05-22T09:38:27+00:00

commit 396a66aa1172ef2b78c21651f59b40b87b2e5e1e upstream.

gcc-11 warns about mismatched prototypes here:

  arch/x86/lib/msr-smp.c:255:51: error: argument 2 of type ‘u32 *’ {aka ‘unsigned int *’} declared as a pointer [-Werror=array-parameter=]
    255 | int rdmsr_safe_regs_on_cpu(unsigned int cpu, u32 *regs)
        |                                              ~~~~~^~~~
  arch/x86/include/asm/msr.h:347:50: note: previously declared as an array ‘u32[8]’ {aka ‘unsigned int[8]’}

GCC is right here - fix up the types.

[ mingo: Twiddled the changelog. ]

Signed-off-by: Arnd Bergmann 
Signed-off-by: Ingo Molnar 
Link: https://lore.kernel.org/r/20210322164541.912261-1-arnd@kernel.org
Signed-off-by: Greg Kroah-Hartman

x86/mmx: Use KFPU_387 for MMX string operations

2021-01-27T10:47:49+00:00

commit 67de8dca50c027ca0fa3b62a488ee5035036a0da upstream.

The default kernel_fpu_begin() doesn't work on systems that support XMM but
haven't yet enabled CR4.OSFXSR.  This causes crashes when _mmx_memcpy() is
called too early because LDMXCSR generates #UD when the aforementioned bit
is clear.

Fix it by using kernel_fpu_begin_mask(KFPU_387) explicitly.

Fixes: 7ad816762f9b ("x86/fpu: Reset MXCSR to default in kernel_fpu_begin()")
Reported-by: Krzysztof Mazur 
Signed-off-by: Andy Lutomirski 
Signed-off-by: Borislav Petkov 
Tested-by: Krzysztof Piotr Olędzki 
Tested-by: Krzysztof Mazur 
Cc: 
Link: https://lkml.kernel.org/r/e7bf21855fe99e5f3baa27446e32623358f69e8d.1611205691.git.luto@kernel.org
Signed-off-by: Greg Kroah-Hartman

x86/lib: Change .weak to SYM_FUNC_START_WEAK for arch/x86/lib/mem*_64.S

2020-12-16T09:56:55+00:00

commit 4d6ffa27b8e5116c0abb318790fd01d4e12d75e6 upstream.

Commit

  393f203f5fd5 ("x86_64: kasan: add interceptors for memset/memmove/memcpy functions")

added .weak directives to arch/x86/lib/mem*_64.S instead of changing the
existing ENTRY macros to WEAK. This can lead to the assembly snippet

  .weak memcpy
  ...
  .globl memcpy

which will produce a STB_WEAK memcpy with GNU as but STB_GLOBAL memcpy
with LLVM's integrated assembler before LLVM 12. LLVM 12 (since
https://reviews.llvm.org/D90108) will error on such an overridden symbol
binding.

Commit

  ef1e03152cb0 ("x86/asm: Make some functions local")

changed ENTRY in arch/x86/lib/memcpy_64.S to SYM_FUNC_START_LOCAL, which
was ineffective due to the preceding .weak directive.

Use the appropriate SYM_FUNC_START_WEAK instead.

Fixes: 393f203f5fd5 ("x86_64: kasan: add interceptors for memset/memmove/memcpy functions")
Fixes: ef1e03152cb0 ("x86/asm: Make some functions local")
Reported-by: Sami Tolvanen 
Signed-off-by: Fangrui Song 
Signed-off-by: Borislav Petkov 
Reviewed-by: Nick Desaulniers 
Tested-by: Nathan Chancellor 
Tested-by: Nick Desaulniers 
Cc: 
Link: https://lkml.kernel.org/r/20201103012358.168682-1-maskray@google.com
[nd: backport due to missing commit e9b9d020c487 ("x86/asm: Annotate aliases")]
Signed-off-by: Nick Desaulniers 
Signed-off-by: Greg Kroah-Hartman

x86/insn-eval: Use new for_each_insn_prefix() macro to loop over prefixes bytes

2020-12-11T12:23:33+00:00

commit 12cb908a11b2544b5f53e9af856e6b6a90ed5533 upstream

Since insn.prefixes.nbytes can be bigger than the size of
insn.prefixes.bytes[] when a prefix is repeated, the proper check must
be

  insn.prefixes.bytes[i] != 0 and i < 4

instead of using insn.prefixes.nbytes. Use the new
for_each_insn_prefix() macro which does it correctly.

Debugged by Kees Cook .

 [ bp: Massage commit message. ]

Fixes: 32d0b95300db ("x86/insn-eval: Add utility functions to get segment selector")
Reported-by: syzbot+9b64b619f10f19d19a7c@syzkaller.appspotmail.com
Signed-off-by: Masami Hiramatsu 
Signed-off-by: Borislav Petkov 
Cc: stable@vger.kernel.org
Link: https://lkml.kernel.org/r/160697104969.3146288.16329307586428270032.stgit@devnote2
[sudip: adjust context]
Signed-off-by: Sudip Mukherjee 
Signed-off-by: Greg Kroah-Hartman