linux.git/arch/x86/entry/common.c, branch v6.2

Merge tag 'x86-entry-2021-06-29' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip

2021-06-29T19:44:51+00:00

Pull x86 entry code related updates from Thomas Gleixner:

 - Consolidate the macros for .byte ... opcode sequences

 - Deduplicate register offset defines in include files

 - Simplify the ia32,x32 compat handling of the related syscall tables
   to get rid of #ifdeffery.

 - Clear all EFLAGS which are not required for syscall handling

 - Consolidate the syscall tables and switch the generation over to the
   generic shell script and remove the CFLAGS tweaks which are not
   longer required.

 - Use 'int' type for system call numbers to match the generic code.

 - Add more selftests for syscalls

* tag 'x86-entry-2021-06-29' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  x86/syscalls: Don't adjust CFLAGS for syscall tables
  x86/syscalls: Remove -Wno-override-init for syscall tables
  x86/uml/syscalls: Remove array index from syscall initializers
  x86/syscalls: Clear 'offset' and 'prefix' in case they are set in env
  x86/entry: Use int everywhere for system call numbers
  x86/entry: Treat out of range and gap system calls the same
  x86/entry/64: Sign-extend system calls on entry to int
  selftests/x86/syscall: Add tests under ptrace to syscall_numbering_64
  selftests/x86/syscall: Simplify message reporting in syscall_numbering
  selftests/x86/syscall: Update and extend syscall_numbering_64
  x86/syscalls: Switch to generic syscallhdr.sh
  x86/syscalls: Use __NR_syscalls instead of __NR_syscall_max
  x86/unistd: Define X32_NR_syscalls only for 64-bit kernel
  x86/syscalls: Stop filling syscall arrays with *_sys_ni_syscall
  x86/syscalls: Switch to generic syscalltbl.sh
  x86/entry/x32: Rename __x32_compat_sys_* to __x64_compat_sys_*

Merge tag 'x86-asm-2021-06-28' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip

2021-06-28T19:57:11+00:00

Pull x86 asm updates from Ingo Molnar:

 - Micro-optimize and standardize the do_syscall_64() calling convention

 - Make syscall entry flags clearing more conservative

 - Clean up syscall table handling

 - Clean up & standardize assembly macros, in preparation of FRED

 - Misc cleanups and fixes

* tag 'x86-asm-2021-06-28' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  x86/asm: Make  valid on cross-builds as well
  x86/regs: Syscall_get_nr() returns -1 for a non-system call
  x86/entry: Split PUSH_AND_CLEAR_REGS into two submacros
  x86/syscall: Maximize MSR_SYSCALL_MASK
  x86/syscall: Unconditionally prototype {ia32,x32}_sys_call_table[]
  x86/entry: Reverse arguments to do_syscall_64()
  x86/entry: Unify definitions from  and 
  x86/asm: Use _ASM_BYTES() in 
  x86/asm: Add _ASM_BYTES() macro for a .byte ... opcode sequence
  x86/asm: Have the __ASM_FORM macros handle commas in arguments

x86/xen: Fix noinstr fail in xen_pv_evtchn_do_upcall()

2021-06-22T11:56:42+00:00

Fix:

  vmlinux.o: warning: objtool: xen_pv_evtchn_do_upcall()+0x23: call to irq_enter_rcu() leaves .noinstr.text section

Fixes: 359f01d1816f ("x86/entry: Use run_sysvec_on_irqstack_cond() for XEN upcall")
Signed-off-by: Peter Zijlstra (Intel) 
Signed-off-by: Ingo Molnar 
Link: https://lore.kernel.org/r/20210621120120.532960208@infradead.org

x86/entry: Fix noinstr fail in __do_fast_syscall_32()

2021-06-22T11:56:42+00:00

Fix:

  vmlinux.o: warning: objtool: __do_fast_syscall_32()+0xf5: call to trace_hardirqs_off() leaves .noinstr.text section

Fixes: 5d5675df792f ("x86/entry: Fix entry/exit mismatch on failed fast 32-bit syscalls")
Signed-off-by: Peter Zijlstra (Intel) 
Signed-off-by: Ingo Molnar 
Link: https://lore.kernel.org/r/20210621120120.467898710@infradead.org

x86/entry: Use int everywhere for system call numbers

2021-05-25T08:07:00+00:00

System call numbers are defined as int, so use int everywhere for system
call numbers. This is strictly a cleanup; it should not change anything
user visible; all ABI changes have been done in the preceeding patches.

[ tglx: Replaced the unsigned long cast ]

Signed-off-by: H. Peter Anvin (Intel) 
Signed-off-by: Thomas Gleixner 
Link: https://lore.kernel.org/r/20210518191303.4135296-7-hpa@zytor.com

x86/entry: Treat out of range and gap system calls the same

2021-05-20T13:19:49+00:00

The current 64-bit system call entry code treats out-of-range system
calls differently than system calls that map to a hole in the system
call table.

This is visible to the user if system calls are intercepted via ptrace or
seccomp and the return value (regs->ax) is modified: in the former case,
the return value is preserved, and in the latter case, sys_ni_syscall() is
called and the return value is forced to -ENOSYS.

The API spec in  is very clear that only
(int)-1 is the non-system-call sentinel value, so make the system call
behavior consistent by calling sys_ni_syscall() for all invalid system
call numbers except for -1.

Although currently sys_ni_syscall() simply returns -ENOSYS, calling it
explicitly is friendly for tracing and future possible extensions, and
as this is an error path there is no reason to optimize it.

Signed-off-by: H. Peter Anvin (Intel) 
Signed-off-by: Thomas Gleixner 
Link: https://lore.kernel.org/r/20210518191303.4135296-6-hpa@zytor.com

x86/entry: Reverse arguments to do_syscall_64()

2021-05-12T08:49:14+00:00

Reverse the order of arguments to do_syscall_64() so that the first
argument is the pt_regs pointer. This is not only consistent with
*all* other entry points from assembly, but it actually makes the
compiled code slightly better.

Signed-off-by: H. Peter Anvin (Intel) 
Signed-off-by: Ingo Molnar 
Link: https://lore.kernel.org/r/20210510185316.3307264-3-hpa@zytor.com

x86/entry: Enable random_kstack_offset support

2021-04-08T12:05:20+00:00

Allow for a randomized stack offset on a per-syscall basis, with roughly
5-6 bits of entropy, depending on compiler and word size. Since the
method of offsetting uses macros, this cannot live in the common entry
code (the stack offset needs to be retained for the life of the syscall,
which means it needs to happen at the actual entry point).

Signed-off-by: Kees Cook 
Signed-off-by: Thomas Gleixner 
Reviewed-by: Thomas Gleixner 
Link: https://lore.kernel.org/r/20210401232347.2791257-5-keescook@chromium.org

x86/entry: Fix entry/exit mismatch on failed fast 32-bit syscalls

2021-03-06T12:10:06+00:00

On a 32-bit fast syscall that fails to read its arguments from user
memory, the kernel currently does syscall exit work but not
syscall entry work.  This confuses audit and ptrace.  For example:

    $ ./tools/testing/selftests/x86/syscall_arg_fault_32
    ...
    strace: pid 264258: entering, ptrace_syscall_info.op == 2
    ...

This is a minimal fix intended for ease of backporting.  A more
complete cleanup is coming.

Fixes: 0b085e68f407 ("x86/entry: Consolidate 32/64 bit syscall entry")
Signed-off-by: Andy Lutomirski 
Signed-off-by: Thomas Gleixner 
Signed-off-by: Borislav Petkov 
Cc: stable@vger.kernel.org
Link: https://lore.kernel.org/r/8c82296ddf803b91f8d1e5eac89e5803ba54ab0e.1614884673.git.luto@kernel.org

Merge branch 'x86/paravirt' into x86/entry

2021-02-12T12:36:43+00:00

Merge in the recent paravirt changes to resolve conflicts caused
by objtool annotations.

Conflicts:
	arch/x86/xen/xen-asm.S

Signed-off-by: Ingo Molnar