We have macros in asm/asm.h to allow for use of the MOVN & MOVZ
instructions with compare-and-branch sequences providing compatibility
for ISA versions which don't include those instructions. However the
macros are unused, and appear to have always been unused. Delete the
dead code.
Signed-off-by: Paul Burton <paul.burton@mips.com>
Patchwork: https://patchwork.linux-mips.org/patch/20909/
Cc: linux-mips@linux-mips.org
Improve performance for the relevant systems and remove the DMA ordering
barrier from `readX_relaxed' and `writeX_relaxed' MMIO accessors, where
it is not needed according to our requirements[1]. For consistency make
the same arrangement with low-level port I/O accessors, but do not
actually provide any accessors making use of it.
References:
[1] "LINUX KERNEL MEMORY BARRIERS", Documentation/memory-barriers.txt,
Section "KERNEL I/O BARRIER EFFECTS"
Signed-off-by: Maciej W. Rozycki <macro@linux-mips.org>
Signed-off-by: Paul Burton <paul.burton@mips.com>
Patchwork: https://patchwork.linux-mips.org/patch/20865/
Cc: Ralf Baechle <ralf@linux-mips.org>
Architecturally the MIPS ISA does not specify ordering requirements for
uncached bus accesses such as MMIO operations normally use and therefore
explicit barriers have to be inserted between MMIO accesses where
unspecified ordering of operations would cause unpredictable results.
For example the R2020 write buffer implements write gathering and
combining[1] and as used with the DECstation models 2100 and 3100 for
MMIO accesses it bypasses the read buffer entirely, because conflicts
are resolved by the memory controller for DRAM accesses only[2] (NB the
R2020 and R3020 buffers are the same except for the maximum clock rate).
Consequently if a device has say a 16-bit control register at offset 0,
a 16-bit event mask register at offset 2 and a 16-bit reset register at
offset 4, and the initial value of the control register is 0x1111, then
in the absence of barriers a hypothetical code sequence like this:
u16 init_dev(u16 __iomem *dev);
u16 x;
write16(dev + 2, 0xffff);
write16(dev + 0, 0x2222);
x = read16(dev + 0);
write16(dev + 1, 0x3333);
write16(dev + 0, 0x4444);
return x;
}
will return 0x1111 and issue a single 32-bit write of 0x33334444 (in the
little-endian bus configuration) to offset 0 on the system bus.
This is because the read to set `x' from offset 0 bypasses the write of
0x2222 that is still in the write buffer pending the completion of the
write of 0xffff to the reset register. Then the write of 0x3333 to the
event mask register is merged with the preceding write to the control
register as they share the same word address, making it a 32-bit write
of 0x33332222 to offset 0. Finally the write of 0x4444 to the control
register is combined with the outstanding 32-bit write of 0x33332222 to
offset 0, because, again, it shares the same address.
This is an example from a legacy system, given here because it is well
documented and affects a machine we actually support. But likewise
modern MIPS systems may implement weak MMIO ordering, possibly even
without having it clearly documented except for being compliant with the
architecture specification with respect to the currently defined SYNC
instruction variants[3].
Considering the above and that we are required to implement MMIO
accessors such that individual accesses made with them are strongly
ordered with respect to each other[4], add the necessary barriers to our
`inX', `outX', `readX' and `writeX' handlers, as well the associated
special use variants. It's up to platforms then to possibly define the
respective barriers so as to expand to nil if no ordering enforcement is
actually needed for a given system; SYNC is supposed to be as cheap as
a NOP on strongly ordered MIPS implementations though.
Retain the option to generate weakly-ordered accessors, so that the
arrangement for `war_io_reorder_wmb' is not lost in case we need it for
fully raw accessors in the future. The reason for this is that it is
unclear from commit 1e820da3c9 ("MIPS: Loongson-3: Introduce
CONFIG_LOONGSON3_ENHANCEMENT") and especially commit 8faca49a67
("MIPS: Modify core io.h macros to account for the Octeon Errata
Core-301.") why they are needed there under the previous assumption that
these accessors can be weakly ordered.
References:
[1] "LR3020 Write Buffer", LSI Logic Corporation, September 1988,
Section "Byte Gathering", pp. 6-7
[2] "DECstation 3100 Desktop Workstation Functional Specification",
Digital Equipment Corporation, Revision 1.3, August 28, 1990,
Section 6.1 "Processor", p. 4
[3] "MIPS Architecture For Programmers, Volume II-A: The MIPS32
Instruction Set Manual", Imagination Technologies LTD, Document
Number: MD00086, Revision 6.06, December 15, 2016, Table 5.5
"Encodings of the Bits[10:6] of the SYNC instruction; the SType
Field", p. 409
[4] "LINUX KERNEL MEMORY BARRIERS", Documentation/memory-barriers.txt,
Section "KERNEL I/O BARRIER EFFECTS"
Signed-off-by: Maciej W. Rozycki <macro@linux-mips.org>
References: 8faca49a67 ("MIPS: Modify core io.h macros to account for the Octeon Errata Core-301.")
References: 1e820da3c9 ("MIPS: Loongson-3: Introduce CONFIG_LOONGSON3_ENHANCEMENT")
Signed-off-by: Paul Burton <paul.burton@mips.com>
Patchwork: https://patchwork.linux-mips.org/patch/20864/
Cc: Ralf Baechle <ralf@linux-mips.org>
Redefine `mmiowb' in terms of `iobarrier_w' so that it works correctly
for MIPS I platforms, which have no SYNC machine instruction and use a
call to `wbflush' instead.
This doesn't change the semantics for CONFIG_CPU_CAVIUM_OCTEON, because
`iobarrier_w' expands to `wmb', which is ultimately the same as the
current arrangement. For MIPS I platforms this not only makes any code
that would happen to use `mmiowb' build and run, but it actually
enforces the ordering required as well, as `iobarrier_w' has it already
covered with the use of `wmb'.
Signed-off-by: Maciej W. Rozycki <macro@linux-mips.org>
Signed-off-by: Paul Burton <paul.burton@mips.com>
Patchwork: https://patchwork.linux-mips.org/patch/20863/
Cc: Ralf Baechle <ralf@linux-mips.org>
Define MMIO ordering barriers as separate operations so as to allow
making places where such a barrier is required distinct from places
where a memory or a DMA barrier is needed.
Architecturally MIPS does not specify ordering requirements for uncached
bus accesses such as MMIO operations normally use and therefore explicit
barriers have to be inserted between MMIO accesses where unspecified
ordering of operations would cause unpredictable results.
MIPS MMIO ordering barriers are implemented using the same underlying
mechanism that memory or a DMA barrier ordering barriers use, that is
either a suitable SYNC instruction or a platform-specific `wbflush'
call. However platforms may implement different ordering rules for
different kinds of bus activity, so having a separate API makes it
possible to remove unnecessary barriers and avoid a performance hit they
may cause due to unrelated bus activity by making their implementation
expand to nil while keeping the necessary ones.
Also having distinct barriers for each kind of use makes it easier for
the reader to understand what code has been intended to do.
Signed-off-by: Maciej W. Rozycki <macro@linux-mips.org>
Signed-off-by: Paul Burton <paul.burton@mips.com>
Patchwork: https://patchwork.linux-mips.org/patch/20862/
Cc: Ralf Baechle <ralf@linux-mips.org>
Rewrite to use the `reorder' assembly mode and remove manually scheduled
delay slots except where GAS cannot schedule a delay-slot instruction
due to a data dependency or a section switch (as is the case with the EX
macro). No change in machine code produced.
Signed-off-by: Maciej W. Rozycki <macro@linux-mips.org>
[paul.burton@mips.com:
Fix conflict with commit 932afdeec1 ("MIPS: Add Kconfig variable for
CPUs with unaligned load/store instructions")]
Signed-off-by: Paul Burton <paul.burton@mips.com>
Patchwork: https://patchwork.linux-mips.org/patch/20834/
Cc: Ralf Baechle <ralf@linux-mips.org>
Fix a commit 8a8158c85e ("MIPS: memset.S: EVA & fault support for
small_memset") regression and remove assembly warnings:
arch/mips/lib/memset.S: Assembler messages:
arch/mips/lib/memset.S:243: Warning: Macro instruction expanded into multiple instructions in a branch delay slot
triggering with the CPU_DADDI_WORKAROUNDS option set and this code:
PTR_SUBU a2, t1, a0
jr ra
PTR_ADDIU a2, 1
This is because with that option in place the DADDIU instruction, which
the PTR_ADDIU CPP macro expands to, becomes a GAS macro, which in turn
expands to an LI/DADDU (or actually ADDIU/DADDU) sequence:
13c: 01a4302f dsubu a2,t1,a0
140: 03e00008 jr ra
144: 24010001 li at,1
148: 00c1302d daddu a2,a2,at
...
Correct this by switching off the `noreorder' assembly mode and letting
GAS schedule this jump's delay slot, as there is nothing special about
it that would require manual scheduling. With this change in place
correct code is produced:
13c: 01a4302f dsubu a2,t1,a0
140: 24010001 li at,1
144: 03e00008 jr ra
148: 00c1302d daddu a2,a2,at
...
Signed-off-by: Maciej W. Rozycki <macro@linux-mips.org>
Signed-off-by: Paul Burton <paul.burton@mips.com>
Fixes: 8a8158c85e ("MIPS: memset.S: EVA & fault support for small_memset")
Patchwork: https://patchwork.linux-mips.org/patch/20833/
Cc: Ralf Baechle <ralf@linux-mips.org>
Cc: stable@vger.kernel.org # 4.17+
In much the same vein as commit ac41f9c462 ("MIPS: Remove a temporary
hack for debugging cache flushes in SMTC configuration") and commit
eb75ecb113 ("MIPS: MT: Remove unused MT single-threaded cache flush
code"), remove the long obsolete ndflush & niflush command line
arguments which provided a hack that should not be useful outside of
debug sessions performed long ago.
Signed-off-by: Paul Burton <paul.burton@mips.com>
Commit ac41f9c462 ("MIPS: Remove a temporary hack for debugging cache
flushes in SMTC configuration") removed an ugly hack that allowed cache
flushing to be performed single-threaded, something which should not be
necessary outside of debug sessions performed long ago.
Whilst the hack was removed from the cache flush code itself, the
mt_protdflush & mt_protiflush variables were left behind along with code
providing the protdflush & protiflush command line arguments. The
mt_cflush_lockdown() & mt_cflush_release() functions were also left
behind but are now entirely unused.
Remove all the unused code to complete the removal of the MT ASE
single-threaded cache flush hack.
Signed-off-by: Paul Burton <paul.burton@mips.com>
MIPSR6 CPUs do not support unaligned load/store instructions
(LWL, LWR, SWL, SWR and LDL, LDR, SDL, SDR for 64bit).
Currently the MIPS tree has some special cases to avoid these
instructions, and the code is testing for !CONFIG_CPU_MIPSR6.
This patch declares a new Kconfig variable:
CONFIG_CPU_HAS_LOAD_STORE_LR.
This variable indicates that the CPU supports these instructions.
Then, the patch does the following:
- Carefully selects this option on all CPUs except MIPSR6.
- Switches all the special cases to test for the new variable,
and inverts the logic:
'#ifndef CONFIG_CPU_MIPSR6' turns into
'#ifdef CONFIG_CPU_HAS_LOAD_STORE_LR'
and vice-versa.
Also, when this variable is NOT selected (e.g. MIPSR6),
CONFIG_GENERIC_CSUM will default to 'y', to compile generic
C checksum code (instead of special assembly code that uses the
unsupported instructions).
This commit should not affect any existing CPU, and is required
for future Lexra CPU support, that misses these instructions too.
Signed-off-by: Yasha Cherikovsky <yasha.che3@gmail.com>
Signed-off-by: Paul Burton <paul.burton@mips.com>
Patchwork: https://patchwork.linux-mips.org/patch/20808/
Cc: Ralf Baechle <ralf@linux-mips.org>
Cc: Paul Burton <paul.burton@mips.com>
Cc: James Hogan <jhogan@kernel.org>
Cc: linux-mips@linux-mips.org
Cc: linux-kernel@vger.kernel.org
Since commit 15f37e1588 ("MIPS: store the appended
dtb address in a variable"),
in kernels with MIPS_RAW_APPENDED_DTB=y, the early boot code detects
the dtb and stores it in the 'fw_passed_dtb' variable.
However, the dtb is not stored in 'fw_passed_dtb' in kernels with
MIPS_ELF_APPENDED_DTB=y.
Under MIPS_ELF_APPENDED_DTB=y, the dtb is also located in the
__appended_dtb section, so we just need to update the #ifdef.
This will allow to access the dtb in a more uniform way.
Fixes: 15f37e1588 ("MIPS: store the appended dtb address in a variable")
Signed-off-by: Yasha Cherikovsky <yasha.che3@gmail.com>
Signed-off-by: Paul Burton <paul.burton@mips.com>
Patchwork: https://patchwork.linux-mips.org/patch/20803/
Cc: Ralf Baechle <ralf@linux-mips.org>
Cc: Paul Burton <paul.burton@mips.com>
Cc: James Hogan <jhogan@kernel.org>
Cc: linux-mips@linux-mips.org
Cc: linux-kernel@vger.kernel.org
Out-of-tree platforms may not be based on Generic as shown in customer
communication. Share the prepare method with all using UHI boot protocol,
and put into machine_kexec.c.
The benefit is that, when having kexec_args related problems, developers
will naturally look into machine_kexec.c, where "CONFIG_UHI_BOOT" will be
found, prompting them to add "select UHI_BOOT" to the platform Kconfig. It
would otherwise require a lot debugging or online searching to be aware
that the solution is in Generic code.
Tested-by: Rachel Mozes <rachel.mozes@intel.com>
Reported-by: Rachel Mozes <rachel.mozes@intel.com>
Signed-off-by: Dengcheng Zhu <dzhu@wavecomp.com>
Signed-off-by: Paul Burton <paul.burton@mips.com>
Patchwork: https://patchwork.linux-mips.org/patch/20569/
Cc: pburton@wavecomp.com
Cc: ralf@linux-mips.org
Cc: linux-mips@linux-mips.org
The existing implementation lets machine_kexec() CPU jump to reboot code
buffer, whereas other CPUs to relocated_kexec_smp_wait. The natural way to
bring up an SMP new kernel would be to let CPU0 do it while others being
halted. For those failing to do so, fall back to the jumping method.
Signed-off-by: Dengcheng Zhu <dzhu@wavecomp.com>
[paul.burton@mips.com: Guard kexec_nonboot_cpu_jump with CONFIG_SMP]
Signed-off-by: Paul Burton <paul.burton@mips.com>
Patchwork: https://patchwork.linux-mips.org/patch/20570/
Cc: pburton@wavecomp.com
Cc: ralf@linux-mips.org
Cc: linux-mips@linux-mips.org
Cc: rachel.mozes@intel.com
New Loongson-3 (Loongson-3A R2, Loongson-3A R3, and newer) has SFB
(Store Fill Buffer) which can improve the performance of memory access.
Now, SFB enablement is controlled by CONFIG_LOONGSON3_ENHANCEMENT, and
the generic kernel has no benefit from SFB (even it is running on a new
Loongson-3 machine). With this patch, we can enable SFB at runtime by
detecting the CPU type (the expense is war_io_reorder_wmb() will always
be a 'sync', which will hurt the performance of old Loongson-3).
[paul.burton@mips.com: Further info from Huacai:
In practise, I found that sometimes there are boot failures if I
enable SFB/LPA in cpu_probe(). I don't know why because processor
designers also haven't give me an explaination, but I think this may
have some relationships to speculative execution.]
Signed-off-by: Huacai Chen <chenhc@lemote.com>
Signed-off-by: Paul Burton <paul.burton@mips.com>
Patchwork: https://patchwork.linux-mips.org/patch/20426/
Cc: Ralf Baechle <ralf@linux-mips.org>
Cc: James Hogan <jhogan@kernel.org>
Cc: linux-mips@linux-mips.org
Cc: Fuxin Zhang <zhangfx@lemote.com>
Cc: Zhangjin Wu <wuzhangjin@gmail.com>
Cc: Huacai Chen <chenhuacai@gmail.com>
The comment describing arch_mem_init() was separated from the definition
of arch_mem_init() by commit a09fc446fb ("[MIPS] setup.c: use
early_param() for early command line parsing"). Move the comment such
that it's next to the definition again for ease of reading.
Signed-off-by: Paul Burton <paul.burton@mips.com>
Clean up instances of casts to the type that a value already has, since
they are effectively no-ops and only serve to complicate the code.
This is the result of the following semantic patch:
@identitycast@
type T;
T *A;
@@
- (T *)(A)
+ A
Signed-off-by: Paul Burton <paul.burton@mips.com>
Patchwork: https://patchwork.linux-mips.org/patch/19599/
arch/mips appears to have inherited SLOW_DOWN_IO from arch/x86 in
antiquity, but we never define CONF_SLOWDOWN_IO so this is unused code.
Perhaps it was once useful to keep the MIPS header close to the x86
version to ease comparisons or porting changes, but they've diverged
significantly at this point & x86 does this differently now anyway.
Delete the dead code.
Signed-off-by: Paul Burton <paul.burton@mips.com>
Patchwork: https://patchwork.linux-mips.org/patch/20343/
Cc: linux-mips@linux-mips.org
MIPS has a copy of lib/iomap.c with minor alterations, none of which are
necessary given appropriate definitions of PIO_OFFSET, PIO_MASK &
PIO_RESERVED. Provide such definitions, select GENERIC_IOMAP & remove
arch/mips/lib/iomap.c to cut back on the needless duplication.
The one change this does make is to our mmio_{in,out}s[bwl] functions,
which began to deviate from their generic counterparts with commit
0845bb721e ("MIPS: iomap: Use __mem_{read,write}{b,w,l} for MMIO"). I
suspect that this commit was incorrect, and that the SEAD-3 platform
should have instead selected CONFIG_SWAP_IO_SPACE. Since the SEAD-3
platform code is now gone & the board is instead supported by the
generic platform (CONFIG_MIPS_GENERIC) which selects
CONFIG_SWAP_IO_SPACE anyway, this shouldn't be a problem any more.
Signed-off-by: Paul Burton <paul.burton@mips.com>
Patchwork: https://patchwork.linux-mips.org/patch/20342/
Cc: linux-mips@linux-mips.org
For a long time arch/mips/Makefile used nm to discover the kernel entry
point by looking for the address of the kernel_entry symbol. This
doesn't work for systems which make use of bit 0 of the PC to reflect
the ISA mode - ie. microMIPS (and MIPS16, but we don't support building
kernels that target MIPS16 anyway).
So for a while with commit 5fc9484f5e ("MIPS: Set ISA bit in entry-y
for microMIPS kernels") we manually modified the last nibble of the
output from nm, which worked but wasn't particularly pretty.
Commit 27c524d174 ("MIPS: Use the entry point from the ELF file
header") then cleaned this up by using objdump to print the ELF entry
point which includes the ISA bit, rather than using nm to print the
address of the kernel_entry symbol which doesn't. That removed the ugly
replacement of the last nibble, but added its own ugliness by needing to
manually sign extend in the 32 bit case.
Unfortunately it has been pointed out that objdump's output is
localised, and therefore grepping for its "start address" output doesn't
work when the user's language settings are such that objdump doesn't
print in English.
We could simply revert commit 27c524d174 ("MIPS: Use the entry point
from the ELF file header") and return to the manual replacement of the
last nibble of entry-y, but it seems that was found sufficiently
unpalatable to avoid. We could attempt to force the language used by
objdump by setting an environment variable such as LC_ALL, but that
seems fragile. Instead we add a small tool named elf-entry which simply
prints out the entry point of the kernel in the format we require.
Signed-off-by: Paul Burton <paul.burton@mips.com>
Reported-by: Philippe Reynes <philippe.reynes@softathome.com>
Tested-by: Philippe Reynes <philippe.reynes@softathome.com>
Fixes: 27c524d174 ("MIPS: Use the entry point from the ELF file header")
Patchwork: https://patchwork.linux-mips.org/patch/20322/
Cc: James Hogan <jhogan@kernel.org>
Cc: Ralf Baechle <ralf@linux-mips.org>
Cc: linux-mips@linux-mips.org
Pull timer update from Thomas Gleixner:
"New defines for the compat time* types so they can be shared between
32bit and 64bit builds. Not used yet, but merging them now allows the
actual conversions to be merged through different maintainer trees
without dependencies
We still have compat interfaces for 32bit on 64bit even with the new
2038 safe timespec/val variants because pointer size is different. And
for the old style timespec/val interfaces we need yet another 'compat'
interface for both 32bit native and 32bit on 64bit"
* 'timers-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
y2038: Provide aliases for compat helpers
Pull IDA updates from Matthew Wilcox:
"A better IDA API:
id = ida_alloc(ida, GFP_xxx);
ida_free(ida, id);
rather than the cumbersome ida_simple_get(), ida_simple_remove().
The new IDA API is similar to ida_simple_get() but better named. The
internal restructuring of the IDA code removes the bitmap
preallocation nonsense.
I hope the net -200 lines of code is convincing"
* 'ida-4.19' of git://git.infradead.org/users/willy/linux-dax: (29 commits)
ida: Change ida_get_new_above to return the id
ida: Remove old API
test_ida: check_ida_destroy and check_ida_alloc
test_ida: Convert check_ida_conv to new API
test_ida: Move ida_check_max
test_ida: Move ida_check_leaf
idr-test: Convert ida_check_nomem to new API
ida: Start new test_ida module
target/iscsi: Allocate session IDs from an IDA
iscsi target: fix session creation failure handling
drm/vmwgfx: Convert to new IDA API
dmaengine: Convert to new IDA API
ppc: Convert vas ID allocation to new IDA API
media: Convert entity ID allocation to new IDA API
ppc: Convert mmu context allocation to new IDA API
Convert net_namespace to new IDA API
cb710: Convert to new IDA API
rsxx: Convert to new IDA API
osd: Convert to new IDA API
sd: Convert to new IDA API
...
Pull gcc plugin fix from Kees Cook:
"Lift gcc test into Kconfig. This is for better behavior when the
kernel is built with Clang, reported by Stefan Agner"
* tag 'gcc-plugins-v4.19-rc1-fix' of git://git.kernel.org/pub/scm/linux/kernel/git/kees/linux:
gcc-plugins: Disable when building under Clang
Pull perf updates from Thomas Gleixner:
"Kernel:
- Improve kallsyms coverage
- Add x86 entry trampolines to kcore
- Fix ARM SPE handling
- Correct PPC event post processing
Tools:
- Make the build system more robust
- Small fixes and enhancements all over the place
- Update kernel ABI header copies
- Preparatory work for converting libtraceevnt to a shared library
- License cleanups"
* 'perf-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip: (100 commits)
tools arch: Update arch/x86/lib/memcpy_64.S copy used in 'perf bench mem memcpy'
tools arch x86: Update tools's copy of cpufeatures.h
perf python: Fix pyrf_evlist__read_on_cpu() interface
perf mmap: Store real cpu number in 'struct perf_mmap'
perf tools: Remove ext from struct kmod_path
perf tools: Add gzip_is_compressed function
perf tools: Add lzma_is_compressed function
perf tools: Add is_compressed callback to compressions array
perf tools: Move the temp file processing into decompress_kmodule
perf tools: Use compression id in decompress_kmodule()
perf tools: Store compression id into struct dso
perf tools: Add compression id into 'struct kmod_path'
perf tools: Make is_supported_compression() static
perf tools: Make decompress_to_file() function static
perf tools: Get rid of dso__needs_decompress() call in __open_dso()
perf tools: Get rid of dso__needs_decompress() call in symbol__disassemble()
perf tools: Get rid of dso__needs_decompress() call in read_object_code()
tools lib traceevent: Change to SPDX License format
perf llvm: Allow passing options to llc in addition to clang
perf parser: Improve error message for PMU address filters
...
Pull x86 fixes from Thomas Gleixner:
- Correct the L1TF fallout on 32bit and the off by one in the 'too much
RAM for protection' calculation.
- Add a helpful kernel message for the 'too much RAM' case
- Unbreak the VDSO in case that the compiler desides to use indirect
jumps/calls and emits retpolines which cannot be resolved because the
kernel uses its own thunks, which does not work for the VDSO. Make it
use the builtin thunks.
- Re-export start_thread() which was unexported when the 32/64bit
implementation was unified. start_thread() is required by modular
binfmt handlers.
- Trivial cleanups
* 'x86-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
x86/speculation/l1tf: Suggest what to do on systems with too much RAM
x86/speculation/l1tf: Fix off-by-one error when warning that system has too much RAM
x86/kvm/vmx: Remove duplicate l1d flush definitions
x86/speculation/l1tf: Fix overflow in l1tf_pfn_limit() on 32bit
x86/process: Re-export start_thread()
x86/mce: Add notifier_block forward declaration
x86/vdso: Fix vDSO build if a retpoline is emitted
Pull irq update from Thomas Gleixner:
"A small set of updats/fixes for the irq subsystem:
- Allow GICv3 interrupts to be configured as wake-up sources to
enable wakeup from suspend
- Make the error handling of the STM32 irqchip init function work
- A set of small cleanups and improvements"
* 'irq-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
irqchip/gic-v3: Allow interrupt to be configured as wake-up sources
irqchip/tango: Set irq handler and data in one go
dt-bindings: irqchip: renesas-irqc: Document r8a774a1 support
irqchip/s3c24xx: Remove unneeded comparison of unsigned long to 0
irqchip/stm32: Fix init error handling
irqchip/bcm7038-l1: Hide cpu offline callback when building for !SMP
Pull licking update from Thomas Gleixner:
"Mark the switch cases which fall through to the next case with the
proper comment so the fallthrough compiler checks can be enabled"
* 'locking-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
futex: Mark expected switch fall-throughs
Pull libnvdimm memory-failure update from Dave Jiang:
"As it stands, memory_failure() gets thoroughly confused by dev_pagemap
backed mappings. The recovery code has specific enabling for several
possible page states and needs new enabling to handle poison in dax
mappings.
In order to support reliable reverse mapping of user space addresses:
1/ Add new locking in the memory_failure() rmap path to prevent races
that would typically be handled by the page lock.
2/ Since dev_pagemap pages are hidden from the page allocator and the
"compound page" accounting machinery, add a mechanism to determine
the size of the mapping that encompasses a given poisoned pfn.
3/ Given pmem errors can be repaired, change the speculatively
accessed poison protection, mce_unmap_kpfn(), to be reversible and
otherwise allow ongoing access from the kernel.
A side effect of this enabling is that MADV_HWPOISON becomes usable
for dax mappings, however the primary motivation is to allow the
system to survive userspace consumption of hardware-poison via dax.
Specifically the current behavior is:
mce: Uncorrected hardware memory error in user-access at af34214200
{1}[Hardware Error]: It has been corrected by h/w and requires no further action
mce: [Hardware Error]: Machine check events logged
{1}[Hardware Error]: event severity: corrected
Memory failure: 0xaf34214: reserved kernel page still referenced by 1 users
[..]
Memory failure: 0xaf34214: recovery action for reserved kernel page: Failed
mce: Memory error not recovered
<reboot>
...and with these changes:
Injecting memory failure for pfn 0x20cb00 at process virtual address 0x7f763dd00000
Memory failure: 0x20cb00: Killing dax-pmd:5421 due to hardware memory corruption
Memory failure: 0x20cb00: recovery action for dax page: Recovered
Given all the cross dependencies I propose taking this through
nvdimm.git with acks from Naoya, x86/core, x86/RAS, and of course dax
folks"
* tag 'libnvdimm-for-4.19_dax-memory-failure' of gitolite.kernel.org:pub/scm/linux/kernel/git/nvdimm/nvdimm:
libnvdimm, pmem: Restore page attributes when clearing errors
x86/memory_failure: Introduce {set, clear}_mce_nospec()
x86/mm/pat: Prepare {reserve, free}_memtype() for "decoy" addresses
mm, memory_failure: Teach memory_failure() about dev_pagemap pages
filesystem-dax: Introduce dax_lock_mapping_entry()
mm, memory_failure: Collect mapping size in collect_procs()
mm, madvise_inject_error: Let memory_failure() optionally take a page reference
mm, dev_pagemap: Do not clear ->mapping on final put
mm, madvise_inject_error: Disable MADV_SOFT_OFFLINE for ZONE_DEVICE pages
filesystem-dax: Set page->index
device-dax: Set page->index
device-dax: Enable page_mapping()
device-dax: Convert to vmf_insert_mixed and vm_fault_t
