x86/build: Warn on orphan section placement

We don't want to depend on the linker's orphan section placement
heuristics as these can vary between linkers, and may change between
versions. All sections need to be explicitly handled in the linker script.

Now that all sections are explicitly handled, enable orphan section
warnings.

Signed-off-by: Kees Cook <keescook@chromium.org>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Reviewed-by: Nick Desaulniers <ndesaulniers@google.com>
Link: https://lore.kernel.org/r/20200902025347.2504702-5-keescook@chromium.org

arm/boot: Warn on orphan section placement

We don't want to depend on the linker's orphan section placement
heuristics as these can vary between linkers, and may change between
versions. All sections need to be explicitly handled in the linker script.

With all sections now handled, enable orphan section warning.

Signed-off-by: Kees Cook <keescook@chromium.org>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Reviewed-by: Nick Desaulniers <ndesaulniers@google.com>
Link: https://lore.kernel.org/r/20200902025347.2504702-4-keescook@chromium.org

arm/build: Warn on orphan section placement

We don't want to depend on the linker's orphan section placement
heuristics as these can vary between linkers, and may change between
versions. All sections need to be explicitly handled in the linker
script.

Specifically, this would have made a recently fixed bug very obvious:

ld: warning: orphan section `.fixup' from `arch/arm/lib/copy_from_user.o' being placed in section `.fixup'

With all sections handled, enable orphan section warning.

Signed-off-by: Kees Cook <keescook@chromium.org>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Reviewed-by: Nick Desaulniers <ndesaulniers@google.com>
Link: https://lore.kernel.org/r/20200902025347.2504702-3-keescook@chromium.org

arm64/build: Warn on orphan section placement

We don't want to depend on the linker's orphan section placement
heuristics as these can vary between linkers, and may change between
versions. All sections need to be explicitly handled in the linker
script.

With all sections now handled, enable orphan section warnings.

Signed-off-by: Kees Cook <keescook@chromium.org>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Acked-by: Will Deacon <will@kernel.org>
Reviewed-by: Nick Desaulniers <ndesaulniers@google.com>
Link: https://lore.kernel.org/r/20200902025347.2504702-2-keescook@chromium.org

x86/boot/compressed: Add missing debugging sections to output

Include the missing DWARF and STABS sections in the compressed image,
when they are present.

Signed-off-by: Kees Cook <keescook@chromium.org>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Link: https://lore.kernel.org/r/20200821194310.3089815-29-keescook@chromium.org

x86/boot/compressed: Remove, discard, or assert for unwanted sections

In preparation for warning on orphan sections, stop the linker from
generating the .eh_frame* sections, discard unwanted non-zero-sized
generated sections, and enforce other expected-to-be-zero-sized sections
(since discarding them might hide problems with them suddenly gaining
unexpected entries).

Signed-off-by: Kees Cook <keescook@chromium.org>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Link: https://lore.kernel.org/r/20200821194310.3089815-28-keescook@chromium.org

x86/boot/compressed: Reorganize zero-size section asserts

For readability, move the zero-sized sections to the end after DISCARDS.

Signed-off-by: Kees Cook <keescook@chromium.org>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Link: https://lore.kernel.org/r/20200821194310.3089815-27-keescook@chromium.org

x86/build: Add asserts for unwanted sections

In preparation for warning on orphan sections, enforce other
expected-to-be-zero-sized sections (since discarding them might hide
problems with them suddenly gaining unexpected entries).

Signed-off-by: Kees Cook <keescook@chromium.org>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Link: https://lore.kernel.org/r/20200821194310.3089815-25-keescook@chromium.org

x86/build: Enforce an empty .got.plt section

The .got.plt section should always be zero (or filled only with the
linker-generated lazy dispatch entry). Enforce this with an assert and
mark the section as INFO. This is more sensitive than just blindly
discarding the section.

Signed-off-by: Kees Cook <keescook@chromium.org>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Link: https://lore.kernel.org/r/20200821194310.3089815-24-keescook@chromium.org

x86/asm: Avoid generating unused kprobe sections

When !CONFIG_KPROBES, do not generate kprobe sections. This makes
sure there are no unexpected sections encountered by the linker scripts.

Signed-off-by: Kees Cook <keescook@chromium.org>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Link: https://lore.kernel.org/r/20200821194310.3089815-23-keescook@chromium.org

arm/boot: Handle all sections explicitly

In preparation for warning on orphan sections, use common macros for
debug sections, discards, and text stubs. Add discards for unwanted .note,
and .rel sections.

Signed-off-by: Kees Cook <keescook@chromium.org>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Cc: Russell King <linux@armlinux.org.uk>
Link: https://lore.kernel.org/r/20200821194310.3089815-21-keescook@chromium.org

arm/build: Assert for unwanted sections

In preparation for warning on orphan sections, enforce
expected-to-be-zero-sized sections (since discarding them might hide
problems with them suddenly gaining unexpected entries).

Signed-off-by: Kees Cook <keescook@chromium.org>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Cc: Russell King <linux@armlinux.org.uk>
Link: https://lore.kernel.org/r/20200821194310.3089815-19-keescook@chromium.org

arm/build: Add missing sections

Add missing text stub sections .vfp11_veneer and .v4_bx, as well as
missing DWARF sections, when present in the build.

Signed-off-by: Kees Cook <keescook@chromium.org>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Reviewed-by: Nick Desaulniers <ndesaulniers@google.com>
Cc: Russell King <linux@armlinux.org.uk>
Link: https://lore.kernel.org/r/20200821194310.3089815-18-keescook@chromium.org

arm/build: Explicitly keep .ARM.attributes sections

In preparation for adding --orphan-handling=warn, explicitly keep the
.ARM.attributes section (at address 0[1]) by expanding the existing
ELF_DETAILS macro into ARM_DETAILS.

[1] https://reviews.llvm.org/D85867

Suggested-by: Nick Desaulniers <ndesaulniers@google.com>
Signed-off-by: Kees Cook <keescook@chromium.org>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Reviewed-by: Fangrui Song <maskray@google.com>
Cc: Russell King <linux@armlinux.org.uk>
Link: https://lore.kernel.org/lkml/CAKwvOdk-racgq5pxsoGS6Vtifbtrk5fmkmnoLxrQMaOvV0nPWw@mail.gmail.com/
Link: https://lore.kernel.org/r/20200821194310.3089815-17-keescook@chromium.org

arm/build: Refactor linker script headers

In preparation for adding --orphan-handling=warn, refactor the linker
script header includes, and extract common macros.

Signed-off-by: Kees Cook <keescook@chromium.org>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Cc: Russell King <linux@armlinux.org.uk>
Link: https://lore.kernel.org/r/20200821194310.3089815-16-keescook@chromium.org

arm64/build: Assert for unwanted sections

In preparation for warning on orphan sections, discard
unwanted non-zero-sized generated sections, and enforce other
expected-to-be-zero-sized sections (since discarding them might hide
problems with them suddenly gaining unexpected entries).

Suggested-by: Ard Biesheuvel <ardb@kernel.org>
Signed-off-by: Kees Cook <keescook@chromium.org>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Cc: Catalin Marinas <catalin.marinas@arm.com>
Cc: Will Deacon <will@kernel.org>
Link: https://lore.kernel.org/r/20200821194310.3089815-14-keescook@chromium.org

arm64/build: Add missing DWARF sections

Explicitly include DWARF sections when they're present in the build.

Signed-off-by: Kees Cook <keescook@chromium.org>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Cc: Catalin Marinas <catalin.marinas@arm.com>
Cc: Will Deacon <will@kernel.org>
Link: https://lore.kernel.org/r/20200821194310.3089815-13-keescook@chromium.org

arm64/build: Use common DISCARDS in linker script

Use the common DISCARDS rule for the linker script in an effort to
regularize the linker script to prepare for warning on orphaned
sections. Additionally clean up left-over no-op macros.

Signed-off-by: Kees Cook <keescook@chromium.org>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Acked-by: Will Deacon <will@kernel.org>
Link: https://lore.kernel.org/r/20200821194310.3089815-12-keescook@chromium.org

arm64/build: Remove .eh_frame* sections due to unwind tables

Avoid .eh_frame* section generation by making sure both CFLAGS and AFLAGS
contain -fno-asychronous-unwind-tables and -fno-unwind-tables.

With all sources of .eh_frame now removed from the build, drop this
DISCARD so we can be alerted in the future if it returns unexpectedly
once orphan section warnings have been enabled.

Suggested-by: Ard Biesheuvel <ardb@kernel.org>
Signed-off-by: Kees Cook <keescook@chromium.org>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Acked-by: Will Deacon <will@kernel.org>
Link: https://lore.kernel.org/r/20200821194310.3089815-11-keescook@chromium.org

arm64/kernel: Remove needless Call Frame Information annotations

Remove last instance of an .eh_frame section by removing the needless Call
Frame Information annotations which were likely leftovers from 32-bit ARM.

Suggested-by: Ard Biesheuvel <ardb@kernel.org>
Signed-off-by: Kees Cook <keescook@chromium.org>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Acked-by: Will Deacon <will@kernel.org>
Link: https://lore.kernel.org/r/20200821194310.3089815-10-keescook@chromium.org

arm64/mm: Remove needless section quotes

Fix a case of needless quotes in __section(), which Clang doesn't like.

Signed-off-by: Kees Cook <keescook@chromium.org>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Reviewed-by: Nick Desaulniers <ndesaulniers@google.com>
Acked-by: Will Deacon <will@kernel.org>
Link: https://lore.kernel.org/r/20200821194310.3089815-9-keescook@chromium.org

efi/libstub: Disable -mbranch-protection

In preparation for adding --orphan-handling=warn to more architectures,
disable -mbranch-protection, as EFI does not yet support it[1].  This was
noticed due to it producing unwanted .note.gnu.property sections (prefixed
with .init due to the objcopy build step).

However, we must also work around a bug in Clang where the section is
still emitted for code-less object files[2], so also remove the section
during the objcopy.

[1] https://lore.kernel.org/lkml/CAMj1kXHck12juGi=E=P4hWP_8vQhQ+-x3vBMc3TGeRWdQ-XkxQ@mail.gmail.com
[2] https://bugs.llvm.org/show_bug.cgi?id=46480

Signed-off-by: Kees Cook <keescook@chromium.org>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Reviewed-by: Nick Desaulniers <ndesaulniers@google.com>
Acked-by: Ard Biesheuvel <ardb@kernel.org>
Link: https://lore.kernel.org/r/20200821194310.3089815-8-keescook@chromium.org

vmlinux.lds.h: Add PGO and AutoFDO input sections

Basically, consider .text.{hot|unlikely|unknown}.* part of .text, too.

When compiling with profiling information (collected via PGO
instrumentations or AutoFDO sampling), Clang will separate code into
.text.hot, .text.unlikely, or .text.unknown sections based on profiling
information. After D79600 (clang-11), these sections will have a
trailing `.` suffix, ie.  .text.hot., .text.unlikely., .text.unknown..

When using -ffunction-sections together with profiling infomation,
either explicitly (FGKASLR) or implicitly (LTO), code may be placed in
sections following the convention:
.text.hot.<foo>, .text.unlikely.<bar>, .text.unknown.<baz>
where <foo>, <bar>, and <baz> are functions.  (This produces one section
per function; we generally try to merge these all back via linker script
so that we don't have 50k sections).

For the above cases, we need to teach our linker scripts that such
sections might exist and that we'd explicitly like them grouped
together, otherwise we can wind up with code outside of the
_stext/_etext boundaries that might not be mapped properly for some
architectures, resulting in boot failures.

If the linker script is not told about possible input sections, then
where the section is placed as output is a heuristic-laiden mess that's
non-portable between linkers (ie. BFD and LLD), and has resulted in many
hard to debug bugs.  Kees Cook is working on cleaning this up by adding
--orphan-handling=warn linker flag used in ARCH=powerpc to additional
architectures. In the case of linker scripts, borrowing from the Zen of
Python: explicit is better than implicit.

Also, ld.bfd's internal linker script considers .text.hot AND
.text.hot.* to be part of .text, as well as .text.unlikely and
.text.unlikely.*. I didn't see support for .text.unknown.*, and didn't
see Clang producing such code in our kernel builds, but I see code in
LLVM that can produce such section names if profiling information is
missing. That may point to a larger issue with generating or collecting
profiles, but I would much rather be safe and explicit than have to
debug yet another issue related to orphan section placement.

Reported-by: Jian Cai <jiancai@google.com>
Suggested-by: Fāng-ruì Sòng <maskray@google.com>
Signed-off-by: Nick Desaulniers <ndesaulniers@google.com>
Signed-off-by: Kees Cook <keescook@chromium.org>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Tested-by: Luis Lozano <llozano@google.com>
Tested-by: Manoj Gupta <manojgupta@google.com>
Acked-by: Kees Cook <keescook@chromium.org>
Cc: linux-arch@vger.kernel.org
Cc: stable@vger.kernel.org
Link: https://sourceware.org/git/?p=binutils-gdb.git;a=commitdiff;h=add44f8d5c5c05e08b11e033127a744d61c26aee
Link: https://sourceware.org/git/?p=binutils-gdb.git;a=commitdiff;h=1de778ed23ce7492c523d5850c6c6dbb34152655
Link: https://reviews.llvm.org/D79600
Link: https://bugs.chromium.org/p/chromium/issues/detail?id=1084760
Link: https://lore.kernel.org/r/20200821194310.3089815-7-keescook@chromium.org
Debugged-by: Luis Lozano <llozano@google.com>

vmlinux.lds.h: Add .symtab, .strtab, and .shstrtab to ELF_DETAILS

When linking vmlinux with LLD, the synthetic sections .symtab, .strtab,
and .shstrtab are listed as orphaned. Add them to the ELF_DETAILS section
so there will be no warnings when --orphan-handling=warn is used more
widely. (They are added above comment as it is the more common
order[1].)

ld.lld: warning: <internal>:(.symtab) is being placed in '.symtab'
ld.lld: warning: <internal>:(.shstrtab) is being placed in '.shstrtab'
ld.lld: warning: <internal>:(.strtab) is being placed in '.strtab'

[1] https://lore.kernel.org/lkml/20200622224928.o2a7jkq33guxfci4@google.com/

Reported-by: Fangrui Song <maskray@google.com>
Signed-off-by: Kees Cook <keescook@chromium.org>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Cc: linux-arch@vger.kernel.org
Link: https://lore.kernel.org/r/20200821194310.3089815-6-keescook@chromium.org

vmlinux.lds.h: Split ELF_DETAILS from STABS_DEBUG

The .comment section doesn't belong in STABS_DEBUG. Split it out into a
new macro named ELF_DETAILS. This will gain other non-debug sections
that need to be accounted for when linking with --orphan-handling=warn.

Signed-off-by: Kees Cook <keescook@chromium.org>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Cc: linux-arch@vger.kernel.org
Link: https://lore.kernel.org/r/20200821194310.3089815-5-keescook@chromium.org

vmlinux.lds.h: Avoid KASAN and KCSAN's unwanted sections

KASAN (-fsanitize=kernel-address) and KCSAN (-fsanitize=thread)
produce unwanted[1] .eh_frame and .init_array.* sections. Add them to
COMMON_DISCARDS, except with CONFIG_CONSTRUCTORS, which wants to keep
.init_array.* sections.

[1] https://bugs.llvm.org/show_bug.cgi?id=46478

Signed-off-by: Kees Cook <keescook@chromium.org>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Tested-by: Marco Elver <elver@google.com>
Cc: linux-arch@vger.kernel.org
Link: https://lore.kernel.org/r/20200821194310.3089815-4-keescook@chromium.org

vmlinux.lds.h: Add .gnu.version* to COMMON_DISCARDS

For vmlinux linking, no architecture uses the .gnu.version* sections,
so remove it via the COMMON_DISCARDS macro in preparation for adding
--orphan-handling=warn more widely. This is a work-around for what
appears to be a bug[1] in ld.bfd which warns for this synthetic section
even when none is found in input objects, and even when no section is
emitted for an output object[2].

[1] https://sourceware.org/bugzilla/show_bug.cgi?id=26153
[2] https://lore.kernel.org/lkml/202006221524.CEB86E036B@keescook/

Signed-off-by: Kees Cook <keescook@chromium.org>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Reviewed-by: Fangrui Song <maskray@google.com>
Cc: linux-arch@vger.kernel.org
Link: https://lore.kernel.org/r/20200821194310.3089815-3-keescook@chromium.org

vmlinux.lds.h: Create COMMON_DISCARDS

Collect the common DISCARD sections for architectures that need more
specialized discard control than what the standard DISCARDS section
provides.

Signed-off-by: Kees Cook <keescook@chromium.org>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Cc: linux-arch@vger.kernel.org
Link: https://lore.kernel.org/r/20200821194310.3089815-2-keescook@chromium.org

x86/boot: Check that there are no run-time relocations

Add a linker script check that there are no run-time relocations, and
remove the old one that tries to check via looking for specially-named
sections in the object files.

Drop the tests for -fPIE compiler option and -pie linker option, as they
are available in all supported gcc and binutils versions (as well as
clang and lld).

Signed-off-by: Arvind Sankar <nivedita@alum.mit.edu>
Signed-off-by: Kees Cook <keescook@chromium.org>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Tested-by: Nick Desaulniers <ndesaulniers@google.com>
Tested-by: Sedat Dilek <sedat.dilek@gmail.com>
Reviewed-by: Kees Cook <keescook@chromium.org>
Reviewed-by: Ard Biesheuvel <ardb@kernel.org>
Reviewed-by: Fangrui Song <maskray@google.com>
Reviewed-by: Sedat Dilek <sedat.dilek@gmail.com>
Link: https://lore.kernel.org/r/20200731230820.1742553-8-keescook@chromium.org

x86/boot: Remove run-time relocations from head_{32,64}.S

The BFD linker generates run-time relocations for z_input_len and
z_output_len, even though they are absolute symbols.

This is fixed for binutils-2.35 [1]. Work around this for earlier
versions by defining two variables input_len and output_len in addition
to the symbols, and use them via position-independent references.

This eliminates the last two run-time relocations in the head code and
allows us to drop the -z noreloc-overflow flag to the linker.

Move the -pie and --no-dynamic-linker LDFLAGS to LDFLAGS_vmlinux instead
of KBUILD_LDFLAGS. There shouldn't be anything else getting linked, but
this is the more logical location for these flags, and modversions might
call the linker if an EXPORT_SYMBOL is left over accidentally in one of
the decompressors.

[1] https://sourceware.org/bugzilla/show_bug.cgi?id=25754

Signed-off-by: Arvind Sankar <nivedita@alum.mit.edu>
Signed-off-by: Kees Cook <keescook@chromium.org>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Tested-by: Nick Desaulniers <ndesaulniers@google.com>
Tested-by: Sedat Dilek <sedat.dilek@gmail.com>
Reviewed-by: Kees Cook <keescook@chromium.org>
Reviewed-by: Ard Biesheuvel <ardb@kernel.org>
Reviewed-by: Fangrui Song <maskray@google.com>
Link: https://lore.kernel.org/r/20200731230820.1742553-7-keescook@chromium.org

x86/boot: Remove run-time relocations from .head.text code

The assembly code in head_{32,64}.S, while meant to be
position-independent, generates run-time relocations because it uses
instructions such as:

leal gdt(%edx), %eax

which make the assembler and linker think that the code is using %edx as
an index into gdt, and hence gdt needs to be relocated to its run-time
address.

On 32-bit, with lld Dmitry Golovin reports that this results in a
link-time error with default options (i.e. unless -z notext is
explicitly passed):

  LD      arch/x86/boot/compressed/vmlinux
  ld.lld: error: can't create dynamic relocation R_386_32 against local
  symbol in readonly segment; recompile object files with -fPIC or pass
  '-Wl,-z,notext' to allow text relocations in the output

With the BFD linker, this generates a warning during the build, if
--warn-shared-textrel is enabled, which at least Gentoo enables by
default:

  LD      arch/x86/boot/compressed/vmlinux
  ld: arch/x86/boot/compressed/head_32.o: warning: relocation in read-only section `.head.text'
  ld: warning: creating a DT_TEXTREL in object

On 64-bit, it is not possible to link the kernel as -pie with lld, and
it is only possible with a BFD linker that supports -z noreloc-overflow,
i.e. versions >2.26. This is because these instructions cannot really be
relocated: the displacement field is only 32-bits wide, and thus cannot
be relocated for a 64-bit load address. The -z noreloc-overflow option
simply overrides the linker error, and results in R_X86_64_RELATIVE
relocations that apply a 64-bit relocation to a 32-bit field anyway.
This happens to work because nothing will process these run-time
relocations.

Start fixing this by removing relocations from .head.text:

- On 32-bit, use a base register that holds the address of the GOT and
  reference symbol addresses using @GOTOFF, i.e.
leal gdt@GOTOFF(%edx), %eax

- On 64-bit, most of the code can (and already does) use %rip-relative
  addressing, however the .code32 bits can't, and the 64-bit code also
  needs to reference symbol addresses as they will be after moving the
  compressed kernel to the end of the decompression buffer.
  For these cases, reference the symbols as an offset to startup_32 to
  avoid creating relocations, i.e.:

leal (gdt-startup_32)(%bp), %eax

  This only works in .head.text as the subtraction cannot be represented
  as a PC-relative relocation unless startup_32 is in the same section
  as the code. Move efi32_pe_entry into .head.text so that it can use
  the same method to avoid relocations.

Reported-by: Dmitry Golovin <dima@golovin.in>
Signed-off-by: Arvind Sankar <nivedita@alum.mit.edu>
Signed-off-by: Kees Cook <keescook@chromium.org>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Tested-by: Nick Desaulniers <ndesaulniers@google.com>
Tested-by: Sedat Dilek <sedat.dilek@gmail.com>
Reviewed-by: Kees Cook <keescook@chromium.org>
Reviewed-by: Ard Biesheuvel <ardb@kernel.org>
Reviewed-by: Fangrui Song <maskray@google.com>
Link: https://lore.kernel.org/r/20200731230820.1742553-6-keescook@chromium.org

x86/boot: Add .text.* to setup.ld

GCC puts the main function into .text.startup when compiled with -Os (or
-O2). This results in arch/x86/boot/main.c having a .text.startup
section which is currently not included explicitly in the linker script
setup.ld in the same directory.

The BFD linker places this orphan section immediately after .text, so
this still works. However, LLD git, since [1], is choosing to place it
immediately after the .bstext section instead (this is the first code
section). This plays havoc with the section layout that setup.elf
requires to create the setup header, for eg on 64-bit:

    LD      arch/x86/boot/setup.elf
  ld.lld: error: section .text.startup file range overlaps with .header
  >>> .text.startup range is [0x200040, 0x2001FE]
  >>> .header range is [0x2001EF, 0x20026B]

  ld.lld: error: section .header file range overlaps with .bsdata
  >>> .header range is [0x2001EF, 0x20026B]
  >>> .bsdata range is [0x2001FF, 0x200398]

  ld.lld: error: section .bsdata file range overlaps with .entrytext
  >>> .bsdata range is [0x2001FF, 0x200398]
  >>> .entrytext range is [0x20026C, 0x2002D3]

  ld.lld: error: section .text.startup virtual address range overlaps
  with .header
  >>> .text.startup range is [0x40, 0x1FE]
  >>> .header range is [0x1EF, 0x26B]

  ld.lld: error: section .header virtual address range overlaps with
  .bsdata
  >>> .header range is [0x1EF, 0x26B]
  >>> .bsdata range is [0x1FF, 0x398]

  ld.lld: error: section .bsdata virtual address range overlaps with
  .entrytext
  >>> .bsdata range is [0x1FF, 0x398]
  >>> .entrytext range is [0x26C, 0x2D3]

  ld.lld: error: section .text.startup load address range overlaps with
  .header
  >>> .text.startup range is [0x40, 0x1FE]
  >>> .header range is [0x1EF, 0x26B]

  ld.lld: error: section .header load address range overlaps with
  .bsdata
  >>> .header range is [0x1EF, 0x26B]
  >>> .bsdata range is [0x1FF, 0x398]

  ld.lld: error: section .bsdata load address range overlaps with
  .entrytext
  >>> .bsdata range is [0x1FF, 0x398]
  >>> .entrytext range is [0x26C, 0x2D3]

Add .text.* to the .text output section to fix this, and also prevent
any future surprises if the compiler decides to create other such
sections.

[1] https://reviews.llvm.org/D75225

Signed-off-by: Arvind Sankar <nivedita@alum.mit.edu>
Signed-off-by: Kees Cook <keescook@chromium.org>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Tested-by: Nick Desaulniers <ndesaulniers@google.com>
Tested-by: Sedat Dilek <sedat.dilek@gmail.com>
Reviewed-by: Kees Cook <keescook@chromium.org>
Reviewed-by: Ard Biesheuvel <ardb@kernel.org>
Reviewed-by: Fangrui Song <maskray@google.com>
Link: https://lore.kernel.org/r/20200731230820.1742553-5-keescook@chromium.org

x86/boot/compressed: Get rid of GOT fixup code

In a previous patch, we have eliminated GOT entries from the decompressor
binary and added an assertion that the .got section is empty. This means
that the GOT fixup routines that exist in both the 32-bit and 64-bit
startup routines have become dead code, and can be removed.

While at it, drop the KEEP() from the linker script, as it has no effect
on the contents of output sections that are created by the linker itself.

Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
Signed-off-by: Arvind Sankar <nivedita@alum.mit.edu>
Signed-off-by: Kees Cook <keescook@chromium.org>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Tested-by: Nick Desaulniers <ndesaulniers@google.com>
Tested-by: Sedat Dilek <sedat.dilek@gmail.com>
Reviewed-by: Kees Cook <keescook@chromium.org>
Acked-by: Arvind Sankar <nivedita@alum.mit.edu>
Link: https://lore.kernel.org/r/20200731230820.1742553-4-keescook@chromium.org

x86/boot/compressed: Force hidden visibility for all symbol references

Eliminate all GOT entries in the decompressor binary, by forcing hidden
visibility for all symbol references, which informs the compiler that
such references will be resolved at link time without the need for
allocating GOT entries.

To ensure that no GOT entries will creep back in, add an assertion to
the decompressor linker script that will fire if the .got section has
a non-zero size.

[Arvind: move hidden.h to include/linux instead of making a copy]

Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
Signed-off-by: Arvind Sankar <nivedita@alum.mit.edu>
Signed-off-by: Kees Cook <keescook@chromium.org>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Tested-by: Nick Desaulniers <ndesaulniers@google.com>
Tested-by: Sedat Dilek <sedat.dilek@gmail.com>
Reviewed-by: Kees Cook <keescook@chromium.org>
Acked-by: Arvind Sankar <nivedita@alum.mit.edu>
Link: https://lore.kernel.org/r/20200731230820.1742553-3-keescook@chromium.org

x86/boot/compressed: Move .got.plt entries out of the .got section

The .got.plt section contains the part of the GOT which is used by PLT
entries, and which gets updated lazily by the dynamic loader when
function calls are dispatched through those PLT entries.

On fully linked binaries such as the kernel proper or the decompressor,
this never happens, and so in practice, the .got.plt section consists
only of the first 3 magic entries that are meant to point at the _DYNAMIC
section and at the fixup routine in the loader. However, since we don't
use a dynamic loader, those entries are never populated or used.

This means that treating those entries like ordinary GOT entries, and
updating their values based on the actual placement of the executable in
memory is completely pointless, and we can just ignore the .got.plt
section entirely, provided that it has no additional entries beyond
the first 3 ones.

So add an assertion in the linker script to ensure that this assumption
holds, and move the contents out of the [_got, _egot) memory range that
is modified by the GOT fixup routines.

While at it, drop the KEEP(), since it has no effect on the contents
of output sections that are created by the linker itself.

Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
Signed-off-by: Arvind Sankar <nivedita@alum.mit.edu>
Signed-off-by: Kees Cook <keescook@chromium.org>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Tested-by: Sedat Dilek <sedat.dilek@gmail.com>
Tested-by: Nick Desaulniers <ndesaulniers@google.com>
Reviewed-by: Kees Cook <keescook@chromium.org>
Acked-by: Arvind Sankar <nivedita@alum.mit.edu>
Link: https://lore.kernel.org/r/20200731230820.1742553-2-keescook@chromium.org

Merge tag 'rtc-5.9' of git://git./linux/kernel/git/abelloni/linux

Pull RTC updates from Alexandre Belloni:
 "Not much this cycle - mostly non urgent driver fixes:

   - ds1374: use watchdog core

   - pcf2127: add alarm and pcf2129 support"

* tag 'rtc-5.9' of git://git.kernel.org/pub/scm/linux/kernel/git/abelloni/linux:
  rtc: pcf2127: fix alarm handling
  rtc: pcf2127: add alarm support
  rtc: pcf2127: add pca2129 device id
  rtc: max77686: Fix wake-ups for max77620
  rtc: ds1307: provide an indication that the watchdog has fired
  rtc: ds1374: remove unused define
  rtc: ds1374: fix RTC_DRV_DS1374_WDT dependencies
  rtc: cleanup obsolete comment about struct rtc_class_ops
  rtc: pl031: fix set_alarm by adding back call to alarm_irq_enable
  rtc: ds1374: wdt: Use watchdog core for watchdog part
  rtc: Replace HTTP links with HTTPS ones
  rtc: goldfish: Enable interrupt in set_alarm() when necessary
  rtc: max77686: Do not allow interrupt to fire before system resume
  rtc: imxdi: fix trivial typos
  rtc: cpcap: fix range

Merge tag 'ceph-for-5.9-rc1' of git://github.com/ceph/ceph-client

Pull ceph updates from Ilya Dryomov:
 "Xiubo has completed his work on filesystem client metrics, they are
  sent to all available MDSes once per second now.

  Other than that, we have a lot of fixes and cleanups all around the
  filesystem, including a tweak to cut down on MDS request resends in
  multi-MDS setups from Yanhu and fixups for SELinux symlink labeling
  and MClientSession message decoding from Jeff"

* tag 'ceph-for-5.9-rc1' of git://github.com/ceph/ceph-client: (22 commits)
  ceph: handle zero-length feature mask in session messages
  ceph: use frag's MDS in either mode
  ceph: move sb->wb_pagevec_pool to be a global mempool
  ceph: set sec_context xattr on symlink creation
  ceph: remove redundant initialization of variable mds
  ceph: fix use-after-free for fsc->mdsc
  ceph: remove unused variables in ceph_mdsmap_decode()
  ceph: delete repeated words in fs/ceph/
  ceph: send client provided metric flags in client metadata
  ceph: periodically send perf metrics to MDSes
  ceph: check the sesion state and return false in case it is closed
  libceph: replace HTTP links with HTTPS ones
  ceph: remove unnecessary cast in kfree()
  libceph: just have osd_req_op_init() return a pointer
  ceph: do not access the kiocb after aio requests
  ceph: clean up and optimize ceph_check_delayed_caps()
  ceph: fix potential mdsc use-after-free crash
  ceph: switch to WARN_ON_ONCE in encode_supported_features()
  ceph: add global total_caps to count the mdsc's total caps number
  ceph: add check_session_state() helper and make it global
  ...

Merge branch 'parisc-5.9-2' of git://git./linux/kernel/git/deller/parisc-linux

Pull more parisc updates from Helge Deller:

 - Oscar Carter contributed a patch which fixes parisc's usage of
   dereference_function_descriptor() and thus will allow using the
   -Wcast-function-type compiler option in the top-level Makefile

 - Sven Schnelle fixed a bug in the SBA code to prevent crashes during
   kexec

 - John David Anglin provided implementations for __smp_store_release()
   and __smp_load_acquire barriers() which avoids using the sync
   assembler instruction and thus speeds up barrier paths

 - Some whitespace cleanups in parisc's atomic.h header file

* 'parisc-5.9-2' of git://git.kernel.org/pub/scm/linux/kernel/git/deller/parisc-linux:
  parisc: Implement __smp_store_release and __smp_load_acquire barriers
  parisc: mask out enable and reserved bits from sba imask
  parisc: Whitespace cleanups in atomic.h
  parisc/kernel/ftrace: Remove function callback casts
  sections.h: dereference_function_descriptor() returns void pointer

Merge tag 'for-linus' of git://git./virt/kvm/kvm

Pull more KVM updates from Paolo Bonzini:
 "PPC:
   - Improvements and bugfixes for secure VM support, giving reduced
     startup time and memory hotplug support.

   - Locking fixes in nested KVM code

   - Increase number of guests supported by HV KVM to 4094

   - Preliminary POWER10 support

  ARM:
   - Split the VHE and nVHE hypervisor code bases, build the EL2 code
     separately, allowing for the VHE code to now be built with
     instrumentation

   - Level-based TLB invalidation support

   - Restructure of the vcpu register storage to accomodate the NV code

   - Pointer Authentication available for guests on nVHE hosts

   - Simplification of the system register table parsing

   - MMU cleanups and fixes

   - A number of post-32bit cleanups and other fixes

  MIPS:
   - compilation fixes

  x86:
   - bugfixes

   - support for the SERIALIZE instruction"

* tag 'for-linus' of git://git.kernel.org/pub/scm/virt/kvm/kvm: (70 commits)
  KVM: MIPS/VZ: Fix build error caused by 'kvm_run' cleanup
  x86/kvm/hyper-v: Synic default SCONTROL MSR needs to be enabled
  MIPS: KVM: Convert a fallthrough comment to fallthrough
  MIPS: VZ: Only include loongson_regs.h for CPU_LOONGSON64
  x86: Expose SERIALIZE for supported cpuid
  KVM: x86: Don't attempt to load PDPTRs when 64-bit mode is enabled
  KVM: arm64: Move S1PTW S2 fault logic out of io_mem_abort()
  KVM: arm64: Don't skip cache maintenance for read-only memslots
  KVM: arm64: Handle data and instruction external aborts the same way
  KVM: arm64: Rename kvm_vcpu_dabt_isextabt()
  KVM: arm: Add trace name for ARM_NISV
  KVM: arm64: Ensure that all nVHE hyp code is in .hyp.text
  KVM: arm64: Substitute RANDOMIZE_BASE for HARDEN_EL2_VECTORS
  KVM: arm64: Make nVHE ASLR conditional on RANDOMIZE_BASE
  KVM: PPC: Book3S HV: Rework secure mem slot dropping
  KVM: PPC: Book3S HV: Move kvmppc_svm_page_out up
  KVM: PPC: Book3S HV: Migrate hot plugged memory
  KVM: PPC: Book3S HV: In H_SVM_INIT_DONE, migrate remaining normal-GFNs to secure-GFNs
  KVM: PPC: Book3S HV: Track the state GFNs associated with secure VMs
  KVM: PPC: Book3S HV: Disable page merging in H_SVM_INIT_START
  ...

Merge tag 'clk-for-linus' of git://git./linux/kernel/git/clk/linux

Pull more clk updates from Stephen Boyd:
 "Here's some more updates that missed the last pull request because I
  happened to tag the tree at an earlier point in the history of
  clk-next. I must have fat fingered it and checked out an older version
  of clk-next on this second computer I'm using.

  This time it actually includes more code for Qualcomm SoCs, the AT91
  major updates, and some Rockchip SoC clk driver updates as well. I've
  corrected this flow so this shouldn't happen again"

* tag 'clk-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/clk/linux: (83 commits)
  clk: bcm2835: Do not use prediv with bcm2711's PLLs
  clk: drop unused function __clk_get_flags
  clk: hsdk: Fix bad dependency on IOMEM
  dt-bindings: clock: Fix YAML schemas for LPASS clocks on SC7180
  clk: mmp: avoid missing prototype warning
  clk: sparx5: Add Sparx5 SoC DPLL clock driver
  dt-bindings: clock: sparx5: Add bindings include file
  clk: qoriq: add LS1021A core pll mux options
  clk: clk-atlas6: fix return value check in atlas6_clk_init()
  clk: tegra: pll: Improve PLLM enable-state detection
  clk: X1000: Add support for calculat REFCLK of USB PHY.
  clk: JZ4780: Reformat the code to align it.
  clk: JZ4780: Add functions for enable and disable USB PHY.
  clk: Ingenic: Add RTC related clocks for Ingenic SoCs.
  dt-bindings: clock: Add tabs to align code.
  dt-bindings: clock: Add RTC related clocks for Ingenic SoCs.
  clk: davinci: Use fallthrough pseudo-keyword
  clk: imx: Use fallthrough pseudo-keyword
  clk: qcom: gcc-sdm660: Fix up gcc_mss_mnoc_bimc_axi_clk
  clk: qcom: gcc-sdm660: Add missing modem reset
  ...

Merge tag 'linux-watchdog-5.9-rc1' of git://linux-watchdog.org/linux-watchdog

Pull watchdog updates from Wim Van Sebroeck:

 - f71808e_wdt imporvements

 - dw_wdt improvements

 - mlx-wdt: support new watchdog type with longer timeout period

 - fallthrough pseudo-keyword replacements

 - overall small fixes and improvements

* tag 'linux-watchdog-5.9-rc1' of git://www.linux-watchdog.org/linux-watchdog: (35 commits)
  watchdog: rti-wdt: balance pm runtime enable calls
  watchdog: rti-wdt: attach to running watchdog during probe
  watchdog: add support for adjusting last known HW keepalive time
  watchdog: use __watchdog_ping in startup
  watchdog: softdog: Add options 'soft_reboot_cmd' and 'soft_active_on_boot'
  watchdog: pcwd_usb: remove needless check before usb_free_coherent()
  watchdog: Replace HTTP links with HTTPS ones
  dt-bindings: watchdog: renesas,wdt: Document r8a774e1 support
  watchdog: initialize device before misc_register
  watchdog: booke_wdt: Add common nowayout parameter driver
  watchdog: scx200_wdt: Use fallthrough pseudo-keyword
  watchdog: Use fallthrough pseudo-keyword
  watchdog: f71808e_wdt: do stricter parameter validation
  watchdog: f71808e_wdt: clear watchdog timeout occurred flag
  watchdog: f71808e_wdt: remove use of wrong watchdog_info option
  watchdog: f71808e_wdt: indicate WDIOF_CARDRESET support in watchdog_info.options
  docs: watchdog: codify ident.options as superset of possible status flags
  dt-bindings: watchdog: Add compatible for QCS404, SC7180, SDM845, SM8150
  dt-bindings: watchdog: Convert QCOM watchdog timer bindings to YAML
  watchdog: dw_wdt: Add DebugFS files
  ...

Merge tag 'vfio-v5.9-rc1' of git://github.com/awilliam/linux-vfio

Pull VFIO updates from Alex Williamson:

 - Inclusive naming updates (Alex Williamson)

 - Intel X550 INTx quirk (Alex Williamson)

 - Error path resched between unmaps (Xiang Zheng)

 - SPAPR IOMMU pin_user_pages() conversion (John Hubbard)

 - Trivial mutex simplification (Alex Williamson)

 - QAT device denylist (Giovanni Cabiddu)

 - type1 IOMMU ioctl refactor (Liu Yi L)

* tag 'vfio-v5.9-rc1' of git://github.com/awilliam/linux-vfio:
  vfio/type1: Refactor vfio_iommu_type1_ioctl()
  vfio/pci: Add QAT devices to denylist
  vfio/pci: Add device denylist
  PCI: Add Intel QuickAssist device IDs
  vfio/pci: Hold igate across releasing eventfd contexts
  vfio/spapr_tce: convert get_user_pages() --> pin_user_pages()
  vfio/type1: Add conditional rescheduling after iommu map failed
  vfio/pci: Add Intel X550 to hidden INTx devices
  vfio: Cleanup allowed driver naming

Merge tag 'drm-next-2020-08-12' of git://anongit.freedesktop.org/drm/drm

Pull drm fixes from Dave Airlie:
 "This has a few vmwgfx regression fixes we hit from the merge window
  (one in TTM), it also has a bunch of amdgpu fixes along with a
  scattering everywhere else.

  core:
   - Fix drm_dp_mst_port refcount leaks in drm_dp_mst_allocate_vcpi
   - Remove null check for kfree in drm_dev_release.
   - Fix DRM_FORMAT_MOD_AMLOGIC_FBC definition.
   - re-added docs for drm_gem_flink_ioctl()
   - add orientation quirk for ASUS T103HAF

  ttm:
   - ttm: fix page-offset calculation within TTM
   - revert patch causing vmwgfx regressions

  fbcon:
   - Fix a fbcon OOB read in fbdev, found by syzbot.

  vga:
   - Mark vga_tryget static as it's not used elsewhere.

  amdgpu:
   - Re-add spelling typo fix
   - Sienna Cichlid fixes
   - Navy Flounder fixes
   - DC fixes
   - SMU i2c fix
   - Power fixes

  vmwgfx:
   - regression fixes for modesetting crashes
   - misc fixes

  xlnx:
   - Small fixes to xlnx.

  omap:
   - Fix mode initialization in omap_connector_mode_valid().
   - force runtime PM suspend on system suspend

  tidss:
   - fix modeset init for DPI panels"

* tag 'drm-next-2020-08-12' of git://anongit.freedesktop.org/drm/drm: (70 commits)
  drm/ttm: revert "drm/ttm: make TT creation purely optional v3"
  drm/vmwgfx: fix spelling mistake "Cant" -> "Can't"
  drm/vmwgfx: fix spelling mistake "Cound" -> "Could"
  drm/vmwgfx/ldu: Use drm_mode_config_reset
  drm/vmwgfx/sou: Use drm_mode_config_reset
  drm/vmwgfx/stdu: Use drm_mode_config_reset
  drm/vmwgfx: Fix two list_for_each loop exit tests
  drm/vmwgfx: Use correct vmw_legacy_display_unit pointer
  drm/vmwgfx: Use struct_size() helper
  drm/amdgpu: Fix bug where DPM is not enabled after hibernate and resume
  drm/amd/powerplay: put VCN/JPEG into PG ungate state before dpm table setup(V3)
  drm/amd/powerplay: update swSMU VCN/JPEG PG logics
  drm/amdgpu: use mode1 reset by default for sienna_cichlid
  drm/amdgpu/smu: rework i2c adpater registration
  drm/amd/display: Display goes blank after inst
  drm/amd/display: Change null plane state swizzle mode to 4kb_s
  drm/amd/display: Use helper function to check for HDMI signal
  drm/amd/display: AMD OUI (DPCD 0x00300) skipped on some sink
  drm/amd/display: Fix logger context
  drm/amd/display: populate new dml variable
  ...

Merge branch 'akpm' (patches from Andrew)

Merge more updates from Andrew Morton:

 - most of the rest of MM (memcg, hugetlb, vmscan, proc, compaction,
   mempolicy, oom-kill, hugetlbfs, migration, thp, cma, util,
   memory-hotplug, cleanups, uaccess, migration, gup, pagemap),

 - various other subsystems (alpha, misc, sparse, bitmap, lib, bitops,
   checkpatch, autofs, minix, nilfs, ufs, fat, signals, kmod, coredump,
   exec, kdump, rapidio, panic, kcov, kgdb, ipc).

* emailed patches from Andrew Morton <akpm@linux-foundation.org>: (164 commits)
  mm/gup: remove task_struct pointer for all gup code
  mm: clean up the last pieces of page fault accountings
  mm/xtensa: use general page fault accounting
  mm/x86: use general page fault accounting
  mm/sparc64: use general page fault accounting
  mm/sparc32: use general page fault accounting
  mm/sh: use general page fault accounting
  mm/s390: use general page fault accounting
  mm/riscv: use general page fault accounting
  mm/powerpc: use general page fault accounting
  mm/parisc: use general page fault accounting
  mm/openrisc: use general page fault accounting
  mm/nios2: use general page fault accounting
  mm/nds32: use general page fault accounting
  mm/mips: use general page fault accounting
  mm/microblaze: use general page fault accounting
  mm/m68k: use general page fault accounting
  mm/ia64: use general page fault accounting
  mm/hexagon: use general page fault accounting
  mm/csky: use general page fault accounting
  ...

mm/gup: remove task_struct pointer for all gup code

After the cleanup of page fault accounting, gup does not need to pass
task_struct around any more.  Remove that parameter in the whole gup
stack.

Signed-off-by: Peter Xu <peterx@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Reviewed-by: John Hubbard <jhubbard@nvidia.com>
Link: http://lkml.kernel.org/r/20200707225021.200906-26-peterx@redhat.com
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

mm: clean up the last pieces of page fault accountings

Here're the last pieces of page fault accounting that were still done
outside handle_mm_fault() where we still have regs==NULL when calling
handle_mm_fault():

arch/powerpc/mm/copro_fault.c:   copro_handle_mm_fault
arch/sparc/mm/fault_32.c:        force_user_fault
arch/um/kernel/trap.c:           handle_page_fault
mm/gup.c:                        faultin_page
                                 fixup_user_fault
mm/hmm.c:                        hmm_vma_fault
mm/ksm.c:                        break_ksm

Some of them has the issue of duplicated accounting for page fault
retries.  Some of them didn't do the accounting at all.

This patch cleans all these up by letting handle_mm_fault() to do per-task
page fault accounting even if regs==NULL (though we'll still skip the perf
event accountings).  With that, we can safely remove all the outliers now.

There's another functional change in that now we account the page faults
to the caller of gup, rather than the task_struct that passed into the gup
code.  More information of this can be found at [1].

After this patch, below things should never be touched again outside
handle_mm_fault():

  - task_struct.[maj|min]_flt
  - PERF_COUNT_SW_PAGE_FAULTS_[MAJ|MIN]

[1] https://lore.kernel.org/lkml/CAHk-=wj_V2Tps2QrMn20_W0OJF9xqNh52XSGA42s-ZJ8Y+GyKw@mail.gmail.com/

Signed-off-by: Peter Xu <peterx@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Cc: Albert Ou <aou@eecs.berkeley.edu>
Cc: Alexander Gordeev <agordeev@linux.ibm.com>
Cc: Andy Lutomirski <luto@kernel.org>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: Borislav Petkov <bp@alien8.de>
Cc: Brian Cain <bcain@codeaurora.org>
Cc: Catalin Marinas <catalin.marinas@arm.com>
Cc: Christian Borntraeger <borntraeger@de.ibm.com>
Cc: Chris Zankel <chris@zankel.net>
Cc: Dave Hansen <dave.hansen@linux.intel.com>
Cc: David S. Miller <davem@davemloft.net>
Cc: Geert Uytterhoeven <geert@linux-m68k.org>
Cc: Gerald Schaefer <gerald.schaefer@de.ibm.com>
Cc: Greentime Hu <green.hu@gmail.com>
Cc: Guo Ren <guoren@kernel.org>
Cc: Heiko Carstens <heiko.carstens@de.ibm.com>
Cc: Helge Deller <deller@gmx.de>
Cc: H. Peter Anvin <hpa@zytor.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Ivan Kokshaysky <ink@jurassic.park.msu.ru>
Cc: James E.J. Bottomley <James.Bottomley@HansenPartnership.com>
Cc: John Hubbard <jhubbard@nvidia.com>
Cc: Jonas Bonn <jonas@southpole.se>
Cc: Ley Foon Tan <ley.foon.tan@intel.com>
Cc: "Luck, Tony" <tony.luck@intel.com>
Cc: Matt Turner <mattst88@gmail.com>
Cc: Max Filippov <jcmvbkbc@gmail.com>
Cc: Michael Ellerman <mpe@ellerman.id.au>
Cc: Michal Simek <monstr@monstr.eu>
Cc: Nick Hu <nickhu@andestech.com>
Cc: Palmer Dabbelt <palmer@dabbelt.com>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Paul Walmsley <paul.walmsley@sifive.com>
Cc: Pekka Enberg <penberg@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Richard Henderson <rth@twiddle.net>
Cc: Rich Felker <dalias@libc.org>
Cc: Russell King <linux@armlinux.org.uk>
Cc: Stafford Horne <shorne@gmail.com>
Cc: Stefan Kristiansson <stefan.kristiansson@saunalahti.fi>
Cc: Thomas Bogendoerfer <tsbogend@alpha.franken.de>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Vasily Gorbik <gor@linux.ibm.com>
Cc: Vincent Chen <deanbo422@gmail.com>
Cc: Vineet Gupta <vgupta@synopsys.com>
Cc: Will Deacon <will@kernel.org>
Cc: Yoshinori Sato <ysato@users.sourceforge.jp>
Link: http://lkml.kernel.org/r/20200707225021.200906-25-peterx@redhat.com
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

mm/xtensa: use general page fault accounting

Use the general page fault accounting by passing regs into
handle_mm_fault().  It naturally solve the issue of multiple page fault
accounting when page fault retry happened.

Remove the PERF_COUNT_SW_PAGE_FAULTS_[MAJ|MIN] perf events because it's
now also done in handle_mm_fault().

Move the PERF_COUNT_SW_PAGE_FAULTS event higher before taking mmap_sem for
the fault, then it'll match with the rest of the archs.

Signed-off-by: Peter Xu <peterx@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Acked-by: Max Filippov <jcmvbkbc@gmail.com>
Cc: Chris Zankel <chris@zankel.net>
Link: http://lkml.kernel.org/r/20200707225021.200906-24-peterx@redhat.com
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

mm/x86: use general page fault accounting

Use the general page fault accounting by passing regs into
handle_mm_fault().

Signed-off-by: Peter Xu <peterx@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Cc: Dave Hansen <dave.hansen@linux.intel.com>
Cc: Andy Lutomirski <luto@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Borislav Petkov <bp@alien8.de>
Cc: H. Peter Anvin <hpa@zytor.com>
Link: http://lkml.kernel.org/r/20200707225021.200906-23-peterx@redhat.com
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

mm/sparc64: use general page fault accounting

Use the general page fault accounting by passing regs into
handle_mm_fault().  It naturally solve the issue of multiple page fault
accounting when page fault retry happened.

Signed-off-by: Peter Xu <peterx@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Acked-by: David S. Miller <davem@davemloft.net>
Link: http://lkml.kernel.org/r/20200707225021.200906-22-peterx@redhat.com
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

mm/sparc32: use general page fault accounting

Use the general page fault accounting by passing regs into
handle_mm_fault().  It naturally solve the issue of multiple page fault
accounting when page fault retry happened.

Signed-off-by: Peter Xu <peterx@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Acked-by: David S. Miller <davem@davemloft.net>
Link: http://lkml.kernel.org/r/20200707225021.200906-21-peterx@redhat.com
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

mm/sh: use general page fault accounting

Use the general page fault accounting by passing regs into
handle_mm_fault().  It naturally solve the issue of multiple page fault
accounting when page fault retry happened.

Signed-off-by: Peter Xu <peterx@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Cc: Yoshinori Sato <ysato@users.sourceforge.jp>
Cc: Rich Felker <dalias@libc.org>
Link: http://lkml.kernel.org/r/20200707225021.200906-20-peterx@redhat.com
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

mm/s390: use general page fault accounting

Use the general page fault accounting by passing regs into
handle_mm_fault().  It naturally solve the issue of multiple page fault
accounting when page fault retry happened.

Signed-off-by: Peter Xu <peterx@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Reviewed-by: Gerald Schaefer <gerald.schaefer@de.ibm.com>
Acked-by: Gerald Schaefer <gerald.schaefer@de.ibm.com>
Cc: Alexander Gordeev <agordeev@linux.ibm.com>
Cc: Heiko Carstens <heiko.carstens@de.ibm.com>
Cc: Vasily Gorbik <gor@linux.ibm.com>
Cc: Christian Borntraeger <borntraeger@de.ibm.com>
Link: http://lkml.kernel.org/r/20200707225021.200906-19-peterx@redhat.com
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

mm/riscv: use general page fault accounting

Use the general page fault accounting by passing regs into
handle_mm_fault().  It naturally solve the issue of multiple page fault
accounting when page fault retry happened.

Signed-off-by: Peter Xu <peterx@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Reviewed-by: Pekka Enberg <penberg@kernel.org>
Acked-by: Palmer Dabbelt <palmerdabbelt@google.com>
Cc: Paul Walmsley <paul.walmsley@sifive.com>
Cc: Albert Ou <aou@eecs.berkeley.edu>
Link: http://lkml.kernel.org/r/20200707225021.200906-18-peterx@redhat.com
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

mm/powerpc: use general page fault accounting

Use the general page fault accounting by passing regs into
handle_mm_fault().

Signed-off-by: Peter Xu <peterx@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Acked-by: Michael Ellerman <mpe@ellerman.id.au>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: Paul Mackerras <paulus@samba.org>
Link: http://lkml.kernel.org/r/20200707225021.200906-17-peterx@redhat.com
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

mm/parisc: use general page fault accounting

Use the general page fault accounting by passing regs into
handle_mm_fault().  It naturally solve the issue of multiple page fault
accounting when page fault retry happened.

Add the missing PERF_COUNT_SW_PAGE_FAULTS perf events too.  Note, the
other two perf events (PERF_COUNT_SW_PAGE_FAULTS_[MAJ|MIN]) were done in
handle_mm_fault().

Signed-off-by: Peter Xu <peterx@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Cc: James E.J. Bottomley <James.Bottomley@HansenPartnership.com>
Cc: Helge Deller <deller@gmx.de>
Link: http://lkml.kernel.org/r/20200707225021.200906-16-peterx@redhat.com
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

mm/openrisc: use general page fault accounting

Use the general page fault accounting by passing regs into
handle_mm_fault().  It naturally solve the issue of multiple page fault
accounting when page fault retry happened.

Add the missing PERF_COUNT_SW_PAGE_FAULTS perf events too.  Note, the
other two perf events (PERF_COUNT_SW_PAGE_FAULTS_[MAJ|MIN]) were done in
handle_mm_fault().

Signed-off-by: Peter Xu <peterx@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Acked-by: Stafford Horne <shorne@gmail.com>
Cc: Jonas Bonn <jonas@southpole.se>
Cc: Stefan Kristiansson <stefan.kristiansson@saunalahti.fi>
Link: http://lkml.kernel.org/r/20200707225021.200906-15-peterx@redhat.com
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

mm/nios2: use general page fault accounting

Use the general page fault accounting by passing regs into
handle_mm_fault().  It naturally solve the issue of multiple page fault
accounting when page fault retry happened.

Add the missing PERF_COUNT_SW_PAGE_FAULTS perf events too.  Note, the
other two perf events (PERF_COUNT_SW_PAGE_FAULTS_[MAJ|MIN]) were done in
handle_mm_fault().

Signed-off-by: Peter Xu <peterx@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Cc: Ley Foon Tan <ley.foon.tan@intel.com>
Link: http://lkml.kernel.org/r/20200707225021.200906-14-peterx@redhat.com
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

mm/nds32: use general page fault accounting

Use the general page fault accounting by passing regs into
handle_mm_fault().  It naturally solve the issue of multiple page fault
accounting when page fault retry happened.

Fix PERF_COUNT_SW_PAGE_FAULTS perf event manually for page fault retries,
by moving it before taking mmap_sem.

Signed-off-by: Peter Xu <peterx@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Acked-by: Greentime Hu <green.hu@gmail.com>
Cc: Nick Hu <nickhu@andestech.com>
Cc: Vincent Chen <deanbo422@gmail.com>
Link: http://lkml.kernel.org/r/20200707225021.200906-13-peterx@redhat.com
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

mm/mips: use general page fault accounting

Use the general page fault accounting by passing regs into
handle_mm_fault().  It naturally solve the issue of multiple page fault
accounting when page fault retry happened.

Fix PERF_COUNT_SW_PAGE_FAULTS perf event manually for page fault retries,
by moving it before taking mmap_sem.

Signed-off-by: Peter Xu <peterx@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Acked-by: Thomas Bogendoerfer <tsbogend@alpha.franken.de>
Link: http://lkml.kernel.org/r/20200707225021.200906-12-peterx@redhat.com
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

mm/microblaze: use general page fault accounting

Use the general page fault accounting by passing regs into
handle_mm_fault().  It naturally solve the issue of multiple page fault
accounting when page fault retry happened.

Add the missing PERF_COUNT_SW_PAGE_FAULTS perf events too.  Note, the
other two perf events (PERF_COUNT_SW_PAGE_FAULTS_[MAJ|MIN]) were done in
handle_mm_fault().

Signed-off-by: Peter Xu <peterx@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Cc: Michal Simek <monstr@monstr.eu>
Link: http://lkml.kernel.org/r/20200707225021.200906-11-peterx@redhat.com
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

mm/m68k: use general page fault accounting

Use the general page fault accounting by passing regs into
handle_mm_fault().  It naturally solve the issue of multiple page fault
accounting when page fault retry happened.

Add the missing PERF_COUNT_SW_PAGE_FAULTS perf events too.  Note, the
other two perf events (PERF_COUNT_SW_PAGE_FAULTS_[MAJ|MIN]) were done in
handle_mm_fault().

Signed-off-by: Peter Xu <peterx@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Cc: Geert Uytterhoeven <geert@linux-m68k.org>
Link: http://lkml.kernel.org/r/20200707225021.200906-10-peterx@redhat.com
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

mm/ia64: use general page fault accounting

Use the general page fault accounting by passing regs into
handle_mm_fault().  It naturally solve the issue of multiple page fault
accounting when page fault retry happened.

Add the missing PERF_COUNT_SW_PAGE_FAULTS perf events too.  Note, the
other two perf events (PERF_COUNT_SW_PAGE_FAULTS_[MAJ|MIN]) were done in
handle_mm_fault().

Signed-off-by: Peter Xu <peterx@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Cc: "Luck, Tony" <tony.luck@intel.com>
Link: http://lkml.kernel.org/r/20200707225021.200906-9-peterx@redhat.com
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

mm/hexagon: use general page fault accounting

Use the general page fault accounting by passing regs into
handle_mm_fault().  It naturally solve the issue of multiple page fault
accounting when page fault retry happened.

Add the missing PERF_COUNT_SW_PAGE_FAULTS perf events too.  Note, the
other two perf events (PERF_COUNT_SW_PAGE_FAULTS_[MAJ|MIN]) were done in
handle_mm_fault().

Signed-off-by: Peter Xu <peterx@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Acked-by: Brian Cain <bcain@codeaurora.org>
Link: http://lkml.kernel.org/r/20200707225021.200906-8-peterx@redhat.com
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

mm/csky: use general page fault accounting

Use the general page fault accounting by passing regs into
handle_mm_fault().  It naturally solve the issue of multiple page fault
accounting when page fault retry happened.

Signed-off-by: Peter Xu <peterx@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Acked-by: Guo Ren <guoren@kernel.org>
Link: http://lkml.kernel.org/r/20200707225021.200906-7-peterx@redhat.com
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

mm/arm64: use general page fault accounting

Use the general page fault accounting by passing regs into
handle_mm_fault().  It naturally solve the issue of multiple page fault
accounting when page fault retry happened.  To do this, we pass pt_regs
pointer into __do_page_fault().

Signed-off-by: Peter Xu <peterx@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Acked-by: Will Deacon <will@kernel.org>
Cc: Catalin Marinas <catalin.marinas@arm.com>
Link: http://lkml.kernel.org/r/20200707225021.200906-6-peterx@redhat.com
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

mm/arm: use general page fault accounting

Use the general page fault accounting by passing regs into
handle_mm_fault().  It naturally solve the issue of multiple page fault
accounting when page fault retry happened.  To do this, we need to pass
the pt_regs pointer into __do_page_fault().

Fix PERF_COUNT_SW_PAGE_FAULTS perf event manually for page fault retries,
by moving it before taking mmap_sem.

Signed-off-by: Peter Xu <peterx@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Cc: Russell King <linux@armlinux.org.uk>
Cc: Will Deacon <will@kernel.org>
Link: http://lkml.kernel.org/r/20200707225021.200906-5-peterx@redhat.com
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

mm/arc: use general page fault accounting

Use the general page fault accounting by passing regs into
handle_mm_fault().  It naturally solve the issue of multiple page fault
accounting when page fault retry happened.

Fix PERF_COUNT_SW_PAGE_FAULTS perf event manually for page fault retries,
by moving it before taking mmap_sem.

Signed-off-by: Peter Xu <peterx@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Cc: Vineet Gupta <vgupta@synopsys.com>
Link: http://lkml.kernel.org/r/20200707225021.200906-4-peterx@redhat.com
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

mm/alpha: use general page fault accounting

Use the general page fault accounting by passing regs into
handle_mm_fault().

Add the missing PERF_COUNT_SW_PAGE_FAULTS perf events too.  Note, the
other two perf events (PERF_COUNT_SW_PAGE_FAULTS_[MAJ|MIN]) were done in
handle_mm_fault().

Signed-off-by: Peter Xu <peterx@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Cc: Richard Henderson <rth@twiddle.net>
Cc: Ivan Kokshaysky <ink@jurassic.park.msu.ru>
Cc: Matt Turner <mattst88@gmail.com>
Link: http://lkml.kernel.org/r/20200707225021.200906-3-peterx@redhat.com
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

mm: do page fault accounting in handle_mm_fault

Patch series "mm: Page fault accounting cleanups", v5.

This is v5 of the pf accounting cleanup series.  It originates from Gerald
Schaefer's report on an issue a week ago regarding to incorrect page fault
accountings for retried page fault after commit 4064b9827063 ("mm: allow
VM_FAULT_RETRY for multiple times"):

  https://lore.kernel.org/lkml/20200610174811.44b94525@thinkpad/

What this series did:

  - Correct page fault accounting: we do accounting for a page fault
    (no matter whether it's from #PF handling, or gup, or anything else)
    only with the one that completed the fault.  For example, page fault
    retries should not be counted in page fault counters.  Same to the
    perf events.

  - Unify definition of PERF_COUNT_SW_PAGE_FAULTS: currently this perf
    event is used in an adhoc way across different archs.

    Case (1): for many archs it's done at the entry of a page fault
    handler, so that it will also cover e.g.  errornous faults.

    Case (2): for some other archs, it is only accounted when the page
    fault is resolved successfully.

    Case (3): there're still quite some archs that have not enabled
    this perf event.

    Since this series will touch merely all the archs, we unify this
    perf event to always follow case (1), which is the one that makes most
    sense.  And since we moved the accounting into handle_mm_fault, the
    other two MAJ/MIN perf events are well taken care of naturally.

  - Unify definition of "major faults": the definition of "major
    fault" is slightly changed when used in accounting (not
    VM_FAULT_MAJOR).  More information in patch 1.

  - Always account the page fault onto the one that triggered the page
    fault.  This does not matter much for #PF handlings, but mostly for
    gup.  More information on this in patch 25.

Patchset layout:

Patch 1:     Introduced the accounting in handle_mm_fault(), not enabled.
Patch 2-23:  Enable the new accounting for arch #PF handlers one by one.
Patch 24:    Enable the new accounting for the rest outliers (gup, iommu, etc.)
Patch 25:    Cleanup GUP task_struct pointer since it's not needed any more

This patch (of 25):

This is a preparation patch to move page fault accountings into the
general code in handle_mm_fault().  This includes both the per task
flt_maj/flt_min counters, and the major/minor page fault perf events.  To
do this, the pt_regs pointer is passed into handle_mm_fault().

PERF_COUNT_SW_PAGE_FAULTS should still be kept in per-arch page fault
handlers.

So far, all the pt_regs pointer that passed into handle_mm_fault() is
NULL, which means this patch should have no intented functional change.

Suggested-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Peter Xu <peterx@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Cc: Albert Ou <aou@eecs.berkeley.edu>
Cc: Alexander Gordeev <agordeev@linux.ibm.com>
Cc: Andy Lutomirski <luto@kernel.org>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: Borislav Petkov <bp@alien8.de>
Cc: Brian Cain <bcain@codeaurora.org>
Cc: Catalin Marinas <catalin.marinas@arm.com>
Cc: Christian Borntraeger <borntraeger@de.ibm.com>
Cc: Chris Zankel <chris@zankel.net>
Cc: Dave Hansen <dave.hansen@linux.intel.com>
Cc: David S. Miller <davem@davemloft.net>
Cc: Geert Uytterhoeven <geert@linux-m68k.org>
Cc: Gerald Schaefer <gerald.schaefer@de.ibm.com>
Cc: Greentime Hu <green.hu@gmail.com>
Cc: Guo Ren <guoren@kernel.org>
Cc: Heiko Carstens <heiko.carstens@de.ibm.com>
Cc: Helge Deller <deller@gmx.de>
Cc: H. Peter Anvin <hpa@zytor.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Ivan Kokshaysky <ink@jurassic.park.msu.ru>
Cc: James E.J. Bottomley <James.Bottomley@HansenPartnership.com>
Cc: John Hubbard <jhubbard@nvidia.com>
Cc: Jonas Bonn <jonas@southpole.se>
Cc: Ley Foon Tan <ley.foon.tan@intel.com>
Cc: "Luck, Tony" <tony.luck@intel.com>
Cc: Matt Turner <mattst88@gmail.com>
Cc: Max Filippov <jcmvbkbc@gmail.com>
Cc: Michael Ellerman <mpe@ellerman.id.au>
Cc: Michal Simek <monstr@monstr.eu>
Cc: Nick Hu <nickhu@andestech.com>
Cc: Palmer Dabbelt <palmer@dabbelt.com>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Paul Walmsley <paul.walmsley@sifive.com>
Cc: Pekka Enberg <penberg@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Richard Henderson <rth@twiddle.net>
Cc: Rich Felker <dalias@libc.org>
Cc: Russell King <linux@armlinux.org.uk>
Cc: Stafford Horne <shorne@gmail.com>
Cc: Stefan Kristiansson <stefan.kristiansson@saunalahti.fi>
Cc: Thomas Bogendoerfer <tsbogend@alpha.franken.de>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Vasily Gorbik <gor@linux.ibm.com>
Cc: Vincent Chen <deanbo422@gmail.com>
Cc: Vineet Gupta <vgupta@synopsys.com>
Cc: Will Deacon <will@kernel.org>
Cc: Yoshinori Sato <ysato@users.sourceforge.jp>
Link: http://lkml.kernel.org/r/20200707225021.200906-1-peterx@redhat.com
Link: http://lkml.kernel.org/r/20200707225021.200906-2-peterx@redhat.com
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

mm/gup: use a standard migration target allocation callback

There is a well-defined migration target allocation callback. Use it.

Signed-off-by: Joonsoo Kim <iamjoonsoo.kim@lge.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Acked-by: Vlastimil Babka <vbabka@suse.cz>
Acked-by: Michal Hocko <mhocko@suse.com>
Cc: "Aneesh Kumar K . V" <aneesh.kumar@linux.ibm.com>
Cc: Christoph Hellwig <hch@infradead.org>
Cc: Mike Kravetz <mike.kravetz@oracle.com>
Cc: Naoya Horiguchi <n-horiguchi@ah.jp.nec.com>
Cc: Roman Gushchin <guro@fb.com>
Link: http://lkml.kernel.org/r/1596180906-8442-3-git-send-email-iamjoonsoo.kim@lge.com
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

mm/hugetlb: make hugetlb migration callback CMA aware

new_non_cma_page() in gup.c requires to allocate the new page that is not
on the CMA area.  new_non_cma_page() implements it by using allocation
scope APIs.

However, there is a work-around for hugetlb.  Normal hugetlb page
allocation API for migration is alloc_huge_page_nodemask().  It consists
of two steps.  First is dequeing from the pool.  Second is, if there is no
available page on the queue, allocating by using the page allocator.

new_non_cma_page() can't use this API since first step (deque) isn't aware
of scope API to exclude CMA area.  So, new_non_cma_page() exports hugetlb
internal function for the second step, alloc_migrate_huge_page(), to
global scope and uses it directly.  This is suboptimal since hugetlb pages
on the queue cannot be utilized.

This patch tries to fix this situation by making the deque function on
hugetlb CMA aware.  In the deque function, CMA memory is skipped if
PF_MEMALLOC_NOCMA flag is found.

Signed-off-by: Joonsoo Kim <iamjoonsoo.kim@lge.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Acked-by: Mike Kravetz <mike.kravetz@oracle.com>
Acked-by: Vlastimil Babka <vbabka@suse.cz>
Acked-by: Michal Hocko <mhocko@suse.com>
Cc: "Aneesh Kumar K . V" <aneesh.kumar@linux.ibm.com>
Cc: Christoph Hellwig <hch@infradead.org>
Cc: Naoya Horiguchi <n-horiguchi@ah.jp.nec.com>
Cc: Roman Gushchin <guro@fb.com>
Link: http://lkml.kernel.org/r/1596180906-8442-2-git-send-email-iamjoonsoo.kim@lge.com
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

mm/gup: restrict CMA region by using allocation scope API

We have well defined scope API to exclude CMA region.  Use it rather than
manipulating gfp_mask manually.  With this change, we can now restore
__GFP_MOVABLE for gfp_mask like as usual migration target allocation.  It
would result in that the ZONE_MOVABLE is also searched by page allocator.
For hugetlb, gfp_mask is redefined since it has a regular allocation mask
filter for migration target.  __GPF_NOWARN is added to hugetlb gfp_mask
filter since a new user for gfp_mask filter, gup, want to be silent when
allocation fails.

Note that this can be considered as a fix for the commit 9a4e9f3b2d73
("mm: update get_user_pages_longterm to migrate pages allocated from CMA
region").  However, "Fixes" tag isn't added here since it is just
suboptimal but it doesn't cause any problem.

Suggested-by: Michal Hocko <mhocko@suse.com>
Signed-off-by: Joonsoo Kim <iamjoonsoo.kim@lge.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Acked-by: Vlastimil Babka <vbabka@suse.cz>
Cc: Christoph Hellwig <hch@infradead.org>
Cc: Roman Gushchin <guro@fb.com>
Cc: Mike Kravetz <mike.kravetz@oracle.com>
Cc: Naoya Horiguchi <n-horiguchi@ah.jp.nec.com>
Cc: "Aneesh Kumar K . V" <aneesh.kumar@linux.ibm.com>
Link: http://lkml.kernel.org/r/1596180906-8442-1-git-send-email-iamjoonsoo.kim@lge.com
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

mm/page_alloc: remove a wrapper for alloc_migration_target()

There is a well-defined standard migration target callback.  Use it
directly.

Signed-off-by: Joonsoo Kim <iamjoonsoo.kim@lge.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Acked-by: Michal Hocko <mhocko@suse.com>
Acked-by: Vlastimil Babka <vbabka@suse.cz>
Cc: Christoph Hellwig <hch@infradead.org>
Cc: Mike Kravetz <mike.kravetz@oracle.com>
Cc: Naoya Horiguchi <n-horiguchi@ah.jp.nec.com>
Cc: Roman Gushchin <guro@fb.com>
Link: http://lkml.kernel.org/r/1594622517-20681-8-git-send-email-iamjoonsoo.kim@lge.com
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

mm/mempolicy: use a standard migration target allocation callback

There is a well-defined migration target allocation callback.  Use it.

Signed-off-by: Joonsoo Kim <iamjoonsoo.kim@lge.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Acked-by: Michal Hocko <mhocko@suse.com>
Acked-by: Vlastimil Babka <vbabka@suse.cz>
Cc: Christoph Hellwig <hch@infradead.org>
Cc: Mike Kravetz <mike.kravetz@oracle.com>
Cc: Naoya Horiguchi <n-horiguchi@ah.jp.nec.com>
Cc: Roman Gushchin <guro@fb.com>
Link: http://lkml.kernel.org/r/1594622517-20681-7-git-send-email-iamjoonsoo.kim@lge.com
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

mm/migrate: introduce a standard migration target allocation function

There are some similar functions for migration target allocation.  Since
there is no fundamental difference, it's better to keep just one rather
than keeping all variants.  This patch implements base migration target
allocation function.  In the following patches, variants will be converted
to use this function.

Changes should be mechanical, but, unfortunately, there are some
differences.  First, some callers' nodemask is assgined to NULL since NULL
nodemask will be considered as all available nodes, that is,
&node_states[N_MEMORY].  Second, for hugetlb page allocation, gfp_mask is
redefined as regular hugetlb allocation gfp_mask plus __GFP_THISNODE if
user provided gfp_mask has it.  This is because future caller of this
function requires to set this node constaint.  Lastly, if provided nodeid
is NUMA_NO_NODE, nodeid is set up to the node where migration source
lives.  It helps to remove simple wrappers for setting up the nodeid.

Note that PageHighmem() call in previous function is changed to open-code
"is_highmem_idx()" since it provides more readability.

[akpm@linux-foundation.org: tweak patch title, per Vlastimil]
[akpm@linux-foundation.org: fix typo in comment]

Signed-off-by: Joonsoo Kim <iamjoonsoo.kim@lge.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Acked-by: Vlastimil Babka <vbabka@suse.cz>
Acked-by: Michal Hocko <mhocko@suse.com>
Cc: Christoph Hellwig <hch@infradead.org>
Cc: Mike Kravetz <mike.kravetz@oracle.com>
Cc: Naoya Horiguchi <n-horiguchi@ah.jp.nec.com>
Cc: Roman Gushchin <guro@fb.com>
Link: http://lkml.kernel.org/r/1594622517-20681-6-git-send-email-iamjoonsoo.kim@lge.com
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

mm/migrate: clear __GFP_RECLAIM to make the migration callback consistent with regular THP allocations

new_page_nodemask is a migration callback and it tries to use a common gfp
flags for the target page allocation whether it is a base page or a THP.
The later only adds GFP_TRANSHUGE to the given mask.  This results in the
allocation being slightly more aggressive than necessary because the
resulting gfp mask will contain also __GFP_RECLAIM_KSWAPD.  THP
allocations usually exclude this flag to reduce over eager background
reclaim during a high THP allocation load which has been seen during large
mmaps initialization.  There is no indication that this is a problem for
migration as well but theoretically the same might happen when migrating
large mappings to a different node.  Make the migration callback
consistent with regular THP allocations.

[akpm@linux-foundation.org: fix comment typo, per Vlastimil]

Signed-off-by: Joonsoo Kim <iamjoonsoo.kim@lge.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Acked-by: Michal Hocko <mhocko@suse.com>
Acked-by: Vlastimil Babka <vbabka@suse.cz>
Cc: Christoph Hellwig <hch@infradead.org>
Cc: Mike Kravetz <mike.kravetz@oracle.com>
Cc: Naoya Horiguchi <n-horiguchi@ah.jp.nec.com>
Cc: Roman Gushchin <guro@fb.com>
Link: http://lkml.kernel.org/r/1594622517-20681-5-git-send-email-iamjoonsoo.kim@lge.com
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

mm/hugetlb: unify migration callbacks

There is no difference between two migration callback functions,
alloc_huge_page_node() and alloc_huge_page_nodemask(), except
__GFP_THISNODE handling.  It's redundant to have two almost similar
functions in order to handle this flag.  So, this patch tries to remove
one by introducing a new argument, gfp_mask, to
alloc_huge_page_nodemask().

After introducing gfp_mask argument, it's caller's job to provide correct
gfp_mask.  So, every callsites for alloc_huge_page_nodemask() are changed
to provide gfp_mask.

Note that it's safe to remove a node id check in alloc_huge_page_node()
since there is no caller passing NUMA_NO_NODE as a node id.

Signed-off-by: Joonsoo Kim <iamjoonsoo.kim@lge.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Reviewed-by: Mike Kravetz <mike.kravetz@oracle.com>
Reviewed-by: Vlastimil Babka <vbabka@suse.cz>
Acked-by: Michal Hocko <mhocko@suse.com>
Cc: Christoph Hellwig <hch@infradead.org>
Cc: Naoya Horiguchi <n-horiguchi@ah.jp.nec.com>
Cc: Roman Gushchin <guro@fb.com>
Link: http://lkml.kernel.org/r/1594622517-20681-4-git-send-email-iamjoonsoo.kim@lge.com
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

mm/migrate: move migration helper from .h to .c

It's not performance sensitive function.  Move it to .c.  This is a
preparation step for future change.

Signed-off-by: Joonsoo Kim <iamjoonsoo.kim@lge.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Reviewed-by: Vlastimil Babka <vbabka@suse.cz>
Acked-by: Mike Kravetz <mike.kravetz@oracle.com>
Acked-by: Michal Hocko <mhocko@suse.com>
Cc: Christoph Hellwig <hch@infradead.org>
Cc: Naoya Horiguchi <n-horiguchi@ah.jp.nec.com>
Cc: Roman Gushchin <guro@fb.com>
Link: http://lkml.kernel.org/r/1594622517-20681-3-git-send-email-iamjoonsoo.kim@lge.com
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

mm/page_isolation: prefer the node of the source page

Patch series "clean-up the migration target allocation functions", v5.

This patch (of 9):

For locality, it's better to migrate the page to the same node rather than
the node of the current caller's cpu.

Signed-off-by: Joonsoo Kim <iamjoonsoo.kim@lge.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Reviewed-by: Vlastimil Babka <vbabka@suse.cz>
Acked-by: Roman Gushchin <guro@fb.com>
Acked-by: Michal Hocko <mhocko@suse.com>
Cc: Christoph Hellwig <hch@infradead.org>
Cc: Mike Kravetz <mike.kravetz@oracle.com>
Cc: Naoya Horiguchi <n-horiguchi@ah.jp.nec.com>
Link: http://lkml.kernel.org/r/1594622517-20681-1-git-send-email-iamjoonsoo.kim@lge.com
Link: http://lkml.kernel.org/r/1594622517-20681-2-git-send-email-iamjoonsoo.kim@lge.com
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

ipc/shm.c: remove the superfluous break

Remove the superfuous break, as there is a 'return' before it.

Signed-off-by: Liao Pingfang <liao.pingfang@zte.com.cn>
Signed-off-by: Yi Wang <wang.yi59@zte.com.cn>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Reviewed-by: Andrew Morton <akpm@linux-foundation.org>
Link: http://lkml.kernel.org/r/1594724361-11525-1-git-send-email-wang.yi59@zte.com.cn
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

ipc: uninline functions

Two functions are only called via function pointers, don't bother
inlining them.

Signed-off-by: Alexey Dobriyan <adobriyan@gmail.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Cc: Manfred Spraul <manfred@colorfullife.com>
Cc: Davidlohr Bueso <dave@stgolabs.net>
Link: http://lkml.kernel.org/r/20200710200312.GA960353@localhost.localdomain
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

scripts/gdb: fix python 3.8 SyntaxWarning

Fixes the observed warnings:
scripts/gdb/linux/rbtree.py:20: SyntaxWarning: "is" with a literal. Did
you mean "=="?
  if node is 0:
scripts/gdb/linux/rbtree.py:36: SyntaxWarning: "is" with a literal. Did
you mean "=="?
  if node is 0:

It looks like this is a new warning added in Python 3.8. I've only seen
this once after adding the add-auto-load-safe-path rule to my ~/.gdbinit
for a new tree.

Fixes: commit 449ca0c95ea2 ("scripts/gdb: add rb tree iterating utilities")
Signed-off-by: Nick Desaulniers <ndesaulniers@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Reviewed-by: Stephen Boyd <swboyd@chromium.org>
Cc: Jan Kiszka <jan.kiszka@siemens.com>
Cc: Kieran Bingham <kbingham@kernel.org>
Cc: Aymeric Agon-Rambosson <aymeric.agon@yandex.com>
Link: http://lkml.kernel.org/r/20200805225015.2847624-1-ndesaulniers@google.com
Link: https://adamj.eu/tech/2020/01/21/why-does-python-3-8-syntaxwarning-for-is-literal/
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

kcov: make some symbols static

Fix sparse build warnings:

kernel/kcov.c:99:1: warning:
 symbol '__pcpu_scope_kcov_percpu_data' was not declared. Should it be static?
kernel/kcov.c:778:6: warning:
 symbol 'kcov_remote_softirq_start' was not declared. Should it be static?
kernel/kcov.c:795:6: warning:
 symbol 'kcov_remote_softirq_stop' was not declared. Should it be static?

Reported-by: Hulk Robot <hulkci@huawei.com>
Signed-off-by: Wei Yongjun <weiyongjun1@huawei.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Reviewed-by: Andrey Konovalov <andreyknvl@google.com>
Link: http://lkml.kernel.org/r/20200702115501.73077-1-weiyongjun1@huawei.com
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

kcov: unconditionally add -fno-stack-protector to compiler options

Unconditionally add -fno-stack-protector to KCOV's compiler options, as
all supported compilers support the option.  This saves a compiler
invocation to determine if the option is supported.

Because Clang does not support -fno-conserve-stack, and
-fno-stack-protector was wrapped in the same cc-option, we were missing
-fno-stack-protector with Clang. Unconditionally adding this option
fixes this for Clang.

Suggested-by: Nick Desaulniers <ndesaulniers@google.com>
Signed-off-by: Marco Elver <elver@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Reviewed-by: Nick Desaulniers <ndesaulniers@google.com>
Reviewed-by: Andrey Konovalov <andreyknvl@google.com>
Cc: Dmitry Vyukov <dvyukov@google.com>
Cc: Alexander Potapenko <glider@google.com>
Link: http://lkml.kernel.org/r/20200615184302.7591-1-elver@google.com
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

panic: make print_oops_end_marker() static

Since print_oops_end_marker() is not used externally, also remove it in
kernel.h at the same time.

Signed-off-by: Yue Hu <huyue2@yulong.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Cc: Kees Cook <keescook@chromium.org>
Link: http://lkml.kernel.org/r/20200724011516.12756-1-zbestahu@gmail.com
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

lib/Kconfig.debug: fix typo in the help text of CONFIG_PANIC_TIMEOUT

There exists duplicated "the" in the help text of CONFIG_PANIC_TIMEOUT,
Remove it.

Signed-off-by: Tiezhu Yang <yangtiezhu@loongson.cn>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Reviewed-by: Kees Cook <keescook@chromium.org>
Cc: Xuefeng Li <lixuefeng@loongson.cn>
Link: http://lkml.kernel.org/r/1591103358-32087-2-git-send-email-yangtiezhu@loongson.cn
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

kernel/panic.c: make oops_may_print() return bool

The return value of oops_may_print() is true or false, so change its type
to reflect that.

Signed-off-by: Tiezhu Yang <yangtiezhu@loongson.cn>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Reviewed-by: Kees Cook <keescook@chromium.org>
Cc: Xuefeng Li <lixuefeng@loongson.cn>
Link: http://lkml.kernel.org/r/1591103358-32087-1-git-send-email-yangtiezhu@loongson.cn
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

rapidio/rio_mport_cdev: use array_size() helper in copy_{from,to}_user()

Use array_size() helper instead of the open-coded version in
copy_{from,to}_user().  These sorts of multiplication factors need to be
wrapped in array_size().

This issue was found with the help of Coccinelle and, audited and fixed
manually.

Signed-off-by: Gustavo A. R. Silva <gustavoars@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Reviewed-by: Kees Cook <keescook@chromium.org>
Cc: Matt Porter <mporter@kernel.crashing.org>
Cc: Alexandre Bounine <alex.bou9@gmail.com>
Link: http://lkml.kernel.org/r/20200616183050.GA31840@embeddedor
Addresses-KSPP-ID: https://github.com/KSPP/linux/issues/83
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

drivers/rapidio/rio-scan.c: use struct_size() helper

Make use of the struct_size() helper instead of an open-coded version in
order to avoid any potential type mistakes.

Also, while there, use the preferred form for passing a size of a struct.
The alternative form where struct name is spelled out hurts readability
and introduces an opportunity for a bug when the pointer variable type is
changed but the corresponding sizeof that is passed as argument is not.

This issue was found with the help of Coccinelle and, audited and fixed
manually.

Addresses KSPP ID: https://github.com/KSPP/linux/issues/83

Signed-off-by: Gustavo A. R. Silva <gustavoars@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Cc: Matt Porter <mporter@kernel.crashing.org>
Cc: Alexandre Bounine <alex.bou9@gmail.com>
Link: http://lkml.kernel.org/r/20200619170445.GA22641@embeddedor
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

drivers/rapidio/devices/rio_mport_cdev.c: use struct_size() helper

Make use of the struct_size() helper instead of an open-coded version in
order to avoid any potential type mistakes.

This issue was found with the help of Coccinelle and, audited and fixed
manually.

Addresses KSPP ID: https://github.com/KSPP/linux/issues/83

Signed-off-by: Gustavo A. R. Silva <gustavoars@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Cc: Matt Porter <mporter@kernel.crashing.org>
Cc: Alexandre Bounine <alex.bou9@gmail.com>
Link: http://lkml.kernel.org/r/20200619170843.GA24923@embeddedor
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

kdump: append kernel build-id string to VMCOREINFO

Make kernel GNU build-id available in VMCOREINFO.  Having build-id in
VMCOREINFO facilitates presenting appropriate kernel namelist image with
debug information file to kernel crash dump analysis tools.  Currently
VMCOREINFO lacks uniquely identifiable key for crash analysis automation.

Regarding if this patch is necessary or matching of linux_banner and
OSRELEASE in VMCOREINFO employed by crash(8) meets the need -- IMO,
build-id approach more foolproof, in most instances it is a cryptographic
hash generated using internal code/ELF bits unlike kernel version string
upon which linux_banner is based that is external to the code.  I feel
each is intended for a different purpose.  Also OSRELEASE is not suitable
when two different kernel builds from same version with different features
enabled.

Currently for most linux (and non-linux) systems build-id can be extracted
using standard methods for file types such as user mode crash dumps,
shared libraries, loadable kernel modules etc., This is an exception for
linux kernel dump.  Having build-id in VMCOREINFO brings some uniformity
for automation tools.

Tyler said:

: I think this is a nice improvement over today's linux_banner approach for
: correlating vmlinux to a kernel dump.
:
: The elf notes parsing in this patch lines up with what is described in in
: the "Notes (Nhdr)" section of the elf(5) man page.
:
: BUILD_ID_MAX is sufficient to hold a sha1 build-id, which is the default
: build-id type today in GNU ld(2).  It is also sufficient to hold the
: "fast" build-id, which is the default build-id type today in LLVM lld(2).

Signed-off-by: Vijay Balakrishna <vijayb@linux.microsoft.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Reviewed-by: Tyler Hicks <tyhicks@linux.microsoft.com>
Acked-by: Baoquan He <bhe@redhat.com>
Cc: Dave Young <dyoung@redhat.com>
Cc: Vivek Goyal <vgoyal@redhat.com>
Link: http://lkml.kernel.org/r/1591849672-34104-1-git-send-email-vijayb@linux.microsoft.com
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

exec: move path_noexec() check earlier

The path_noexec() check, like the regular file check, was happening too
late, letting LSMs see impossible execve()s.  Check it earlier as well in
may_open() and collect the redundant fs/exec.c path_noexec() test under
the same robustness comment as the S_ISREG() check.

My notes on the call path, and related arguments, checks, etc:

do_open_execat()
    struct open_flags open_exec_flags = {
        .open_flag = O_LARGEFILE | O_RDONLY | __FMODE_EXEC,
        .acc_mode = MAY_EXEC,
        ...
    do_filp_open(dfd, filename, open_flags)
        path_openat(nameidata, open_flags, flags)
            file = alloc_empty_file(open_flags, current_cred());
            do_open(nameidata, file, open_flags)
                may_open(path, acc_mode, open_flag)
                    /* new location of MAY_EXEC vs path_noexec() test */
                    inode_permission(inode, MAY_OPEN | acc_mode)
                        security_inode_permission(inode, acc_mode)
                vfs_open(path, file)
                    do_dentry_open(file, path->dentry->d_inode, open)
                        security_file_open(f)
                        open()
    /* old location of path_noexec() test */

Signed-off-by: Kees Cook <keescook@chromium.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Cc: Alexander Viro <viro@zeniv.linux.org.uk>
Cc: Aleksa Sarai <cyphar@cyphar.com>
Cc: Christian Brauner <christian.brauner@ubuntu.com>
Cc: Dmitry Vyukov <dvyukov@google.com>
Cc: Eric Biggers <ebiggers3@gmail.com>
Cc: Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp>
Link: http://lkml.kernel.org/r/20200605160013.3954297-4-keescook@chromium.org
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

exec: move S_ISREG() check earlier

The execve(2)/uselib(2) syscalls have always rejected non-regular files.
Recently, it was noticed that a deadlock was introduced when trying to
execute pipes, as the S_ISREG() test was happening too late.  This was
fixed in commit 73601ea5b7b1 ("fs/open.c: allow opening only regular files
during execve()"), but it was added after inode_permission() had already
run, which meant LSMs could see bogus attempts to execute non-regular
files.

Move the test into the other inode type checks (which already look for
other pathological conditions[1]).  Since there is no need to use
FMODE_EXEC while we still have access to "acc_mode", also switch the test
to MAY_EXEC.

Also include a comment with the redundant S_ISREG() checks at the end of
execve(2)/uselib(2) to note that they are present to avoid any mistakes.

My notes on the call path, and related arguments, checks, etc:

do_open_execat()
    struct open_flags open_exec_flags = {
        .open_flag = O_LARGEFILE | O_RDONLY | __FMODE_EXEC,
        .acc_mode = MAY_EXEC,
        ...
    do_filp_open(dfd, filename, open_flags)
        path_openat(nameidata, open_flags, flags)
            file = alloc_empty_file(open_flags, current_cred());
            do_open(nameidata, file, open_flags)
                may_open(path, acc_mode, open_flag)
    /* new location of MAY_EXEC vs S_ISREG() test */
                    inode_permission(inode, MAY_OPEN | acc_mode)
                        security_inode_permission(inode, acc_mode)
                vfs_open(path, file)
                    do_dentry_open(file, path->dentry->d_inode, open)
                        /* old location of FMODE_EXEC vs S_ISREG() test */
                        security_file_open(f)
                        open()

[1] https://lore.kernel.org/lkml/202006041910.9EF0C602@keescook/

Signed-off-by: Kees Cook <keescook@chromium.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Cc: Aleksa Sarai <cyphar@cyphar.com>
Cc: Alexander Viro <viro@zeniv.linux.org.uk>
Cc: Christian Brauner <christian.brauner@ubuntu.com>
Cc: Dmitry Vyukov <dvyukov@google.com>
Cc: Eric Biggers <ebiggers3@gmail.com>
Cc: Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp>
Link: http://lkml.kernel.org/r/20200605160013.3954297-3-keescook@chromium.org
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

exec: change uselib(2) IS_SREG() failure to EACCES

Patch series "Relocate execve() sanity checks", v2.

While looking at the code paths for the proposed O_MAYEXEC flag, I saw
some things that looked like they should be fixed up.

  exec: Change uselib(2) IS_SREG() failure to EACCES
This just regularizes the return code on uselib(2).

  exec: Move S_ISREG() check earlier
This moves the S_ISREG() check even earlier than it was already.

  exec: Move path_noexec() check earlier
This adds the path_noexec() check to the same place as the
S_ISREG() check.

This patch (of 3):

Change uselib(2)' S_ISREG() error return to EACCES instead of EINVAL so
the behavior matches execve(2), and the seemingly documented value.  The
"not a regular file" failure mode of execve(2) is explicitly
documented[1], but it is not mentioned in uselib(2)[2] which does,
however, say that open(2) and mmap(2) errors may apply.  The documentation
for open(2) does not include a "not a regular file" error[3], but mmap(2)
does[4], and it is EACCES.

[1] http://man7.org/linux/man-pages/man2/execve.2.html#ERRORS
[2] http://man7.org/linux/man-pages/man2/uselib.2.html#ERRORS
[3] http://man7.org/linux/man-pages/man2/open.2.html#ERRORS
[4] http://man7.org/linux/man-pages/man2/mmap.2.html#ERRORS

Signed-off-by: Kees Cook <keescook@chromium.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Acked-by: Christian Brauner <christian.brauner@ubuntu.com>
Cc: Aleksa Sarai <cyphar@cyphar.com>
Cc: Alexander Viro <viro@zeniv.linux.org.uk>
Cc: Dmitry Vyukov <dvyukov@google.com>
Cc: Eric Biggers <ebiggers3@gmail.com>
Cc: Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp>
Link: http://lkml.kernel.org/r/20200605160013.3954297-1-keescook@chromium.org
Link: http://lkml.kernel.org/r/20200605160013.3954297-2-keescook@chromium.org
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

coredump: add %f for executable filename

The document reads "%e" should be "executable filename" while actually it
could be changed by things like pr_ctl PR_SET_NAME.  People who uses "%e"
in core_pattern get surprised when they find out they get thread name
instead of executable filename.

This is either a bug of document or a bug of code.  Since the behavior of
"%e" is there for long time, it could bring another surprise for users if
we "fix" the code.

So we just "fix" the document.  And more, for users who really need the
"executable filename" in core_pattern, we introduce a new "%f" for the
real executable filename.  We already have "%E" for executable path in
kernel, so just reuse most of its code for the new added "%f" format.

Signed-off-by: Lepton Wu <ytht.net@gmail.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Link: http://lkml.kernel.org/r/20200701031432.2978761-1-ytht.net@gmail.com
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

test_kmod: avoid potential double free in trigger_config_run_type()

Reset the member "test_fs" of the test configuration after a call of the
function "kfree_const" to a null pointer so that a double memory release
will not be performed.

Fixes: d9c6a72d6fa2 ("kmod: add test driver to stress test the module loader")
Signed-off-by: Tiezhu Yang <yangtiezhu@loongson.cn>
Signed-off-by: Luis Chamberlain <mcgrof@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Acked-by: Luis Chamberlain <mcgrof@kernel.org>
Cc: Alexei Starovoitov <ast@kernel.org>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Cc: Christian Brauner <christian.brauner@ubuntu.com>
Cc: Chuck Lever <chuck.lever@oracle.com>
Cc: David Howells <dhowells@redhat.com>
Cc: David S. Miller <davem@davemloft.net>
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Cc: Jakub Kicinski <kuba@kernel.org>
Cc: James Morris <jmorris@namei.org>
Cc: Jarkko Sakkinen <jarkko.sakkinen@linux.intel.com>
Cc: J. Bruce Fields <bfields@fieldses.org>
Cc: Jens Axboe <axboe@kernel.dk>
Cc: Josh Triplett <josh@joshtriplett.org>
Cc: Kees Cook <keescook@chromium.org>
Cc: Lars Ellenberg <lars.ellenberg@linbit.com>
Cc: Nikolay Aleksandrov <nikolay@cumulusnetworks.com>
Cc: Philipp Reisner <philipp.reisner@linbit.com>
Cc: Roopa Prabhu <roopa@cumulusnetworks.com>
Cc: "Serge E. Hallyn" <serge@hallyn.com>
Cc: Sergei Trofimovich <slyfox@gentoo.org>
Cc: Sergey Kvachonok <ravenexp@gmail.com>
Cc: Shuah Khan <shuah@kernel.org>
Cc: Tony Vroon <chainsaw@gentoo.org>
Cc: Christoph Hellwig <hch@infradead.org>
Link: http://lkml.kernel.org/r/20200610154923.27510-4-mcgrof@kernel.org
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

kmod: remove redundant "be an" in the comment

There exists redundant "be an" in the comment, remove it.

Signed-off-by: Tiezhu Yang <yangtiezhu@loongson.cn>
Signed-off-by: Luis Chamberlain <mcgrof@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Acked-by: Luis Chamberlain <mcgrof@kernel.org>
Cc: Alexei Starovoitov <ast@kernel.org>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Cc: Christian Brauner <christian.brauner@ubuntu.com>
Cc: Chuck Lever <chuck.lever@oracle.com>
Cc: David Howells <dhowells@redhat.com>
Cc: David S. Miller <davem@davemloft.net>
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Cc: Jakub Kicinski <kuba@kernel.org>
Cc: James Morris <jmorris@namei.org>
Cc: Jarkko Sakkinen <jarkko.sakkinen@linux.intel.com>
Cc: J. Bruce Fields <bfields@fieldses.org>
Cc: Jens Axboe <axboe@kernel.dk>
Cc: Josh Triplett <josh@joshtriplett.org>
Cc: Kees Cook <keescook@chromium.org>
Cc: Lars Ellenberg <lars.ellenberg@linbit.com>
Cc: Nikolay Aleksandrov <nikolay@cumulusnetworks.com>
Cc: Philipp Reisner <philipp.reisner@linbit.com>
Cc: Roopa Prabhu <roopa@cumulusnetworks.com>
Cc: "Serge E. Hallyn" <serge@hallyn.com>
Cc: Sergei Trofimovich <slyfox@gentoo.org>
Cc: Sergey Kvachonok <ravenexp@gmail.com>
Cc: Shuah Khan <shuah@kernel.org>
Cc: Tony Vroon <chainsaw@gentoo.org>
Cc: Christoph Hellwig <hch@infradead.org>
Link: http://lkml.kernel.org/r/20200610154923.27510-3-mcgrof@kernel.org
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

selftests: kmod: use variable NAME in kmod_test_0001()

Patch series "kmod/umh: a few fixes".

Tiezhu Yang had sent out a patch set with a slew of kmod selftest fixes,
and one patch which modified kmod to return 254 when a module was not
found.  This opened up pandora's box about why that was being used for and
low and behold its because when UMH_WAIT_PROC is used we call a
kernel_wait4() call but have never unwrapped the error code.  The commit
log for that fix details the rationale for the approach taken.  I'd
appreciate some review on that, in particular nfs folks as it seems a case
was never really hit before.

This patch (of 5):

Use the variable NAME instead of "\000" directly in kmod_test_0001().

Signed-off-by: Tiezhu Yang <yangtiezhu@loongson.cn>
Signed-off-by: Luis Chamberlain <mcgrof@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Acked-by: Luis Chamberlain <mcgrof@kernel.org>
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Cc: Philipp Reisner <philipp.reisner@linbit.com>
Cc: Lars Ellenberg <lars.ellenberg@linbit.com>
Cc: Jens Axboe <axboe@kernel.dk>
Cc: J. Bruce Fields <bfields@fieldses.org>
Cc: Chuck Lever <chuck.lever@oracle.com>
Cc: Roopa Prabhu <roopa@cumulusnetworks.com>
Cc: Nikolay Aleksandrov <nikolay@cumulusnetworks.com>
Cc: David S. Miller <davem@davemloft.net>
Cc: Jakub Kicinski <kuba@kernel.org>
Cc: David Howells <dhowells@redhat.com>
Cc: Jarkko Sakkinen <jarkko.sakkinen@linux.intel.com>
Cc: James Morris <jmorris@namei.org>
Cc: "Serge E. Hallyn" <serge@hallyn.com>
Cc: Christian Brauner <christian.brauner@ubuntu.com>
Cc: Sergei Trofimovich <slyfox@gentoo.org>
Cc: Alexei Starovoitov <ast@kernel.org>
Cc: Kees Cook <keescook@chromium.org>
Cc: Josh Triplett <josh@joshtriplett.org>
Cc: Sergey Kvachonok <ravenexp@gmail.com>
Cc: Tony Vroon <chainsaw@gentoo.org>
Cc: Shuah Khan <shuah@kernel.org>
Cc: Christoph Hellwig <hch@infradead.org>
Link: http://lkml.kernel.org/r/20200610154923.27510-1-mcgrof@kernel.org
Link: http://lkml.kernel.org/r/20200610154923.27510-2-mcgrof@kernel.org
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

fs/signalfd.c: fix inconsistent return codes for signalfd4

The kernel signalfd4() syscall returns different error codes when called
either in compat or native mode.  This behaviour makes correct emulation
in qemu and testing programs like LTP more complicated.

Fix the code to always return -in both modes- EFAULT for unaccessible user
memory, and EINVAL when called with an invalid signal mask.

Signed-off-by: Helge Deller <deller@gmx.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Cc: Alexander Viro <viro@zeniv.linux.org.uk>
Cc: Laurent Vivier <laurent@vivier.eu>
Link: http://lkml.kernel.org/r/20200530100707.GA10159@ls3530.fritz.box
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

fat: fix fat_ra_init() for data clusters == 0

If data clusters == 0, fat_ra_init() calls the ->ent_blocknr() for the
cluster beyond ->max_clusters.

This checks the limit before initialization to suppress the warning.

Reported-by: syzbot+756199124937b31a9b7e@syzkaller.appspotmail.com
Signed-off-by: OGAWA Hirofumi <hirofumi@mail.parknet.co.jp>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Link: http://lkml.kernel.org/r/87mu462sv4.fsf@mail.parknet.co.jp
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>