While tracking down a problem where constant expressions used by
BUILD_BUG_ON() suddenly stopped working[1], we found that an added static
initializer was convincing the compiler that it couldn't track the state
of the prior statically initialized value. Tracing this down found that
ffs() was used in the initializer macro, but since it wasn't marked with
__attribute__const__, the compiler had to assume the function might
change variable states as a side-effect (which is not true for ffs(),
which provides deterministic math results).
Add missing __attribute_const__ annotations to MIPS's implementations of
ffs(), __ffs(), fls(), and __fls() functions. These are pure mathematical
functions that always return the same result for the same input with no
side effects, making them eligible for compiler optimization.
Build tested ARCH=mips defconfig with GCC mipsel-linux-gnu 14.2.0.
Link: https://github.com/KSPP/linux/issues/364 [1]
Acked-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lore.kernel.org/r/20250804164417.1612371-12-kees@kernel.org
Signed-off-by: Kees Cook <kees@kernel.org>
Pull bitmap updates from Yury Norov:
- introduce for_each_set_bitrange()
- use find_first_*_bit() instead of find_next_*_bit() where possible
- unify for_each_bit() macros
* tag 'bitmap-5.17-rc1' of git://github.com/norov/linux:
vsprintf: rework bitmap_list_string
lib: bitmap: add performance test for bitmap_print_to_pagebuf
bitmap: unify find_bit operations
mm/percpu: micro-optimize pcpu_is_populated()
Replace for_each_*_bit_from() with for_each_*_bit() where appropriate
find: micro-optimize for_each_{set,clear}_bit()
include/linux: move for_each_bit() macros from bitops.h to find.h
cpumask: replace cpumask_next_* with cpumask_first_* where appropriate
tools: sync tools/bitmap with mother linux
all: replace find_next{,_zero}_bit with find_first{,_zero}_bit where appropriate
cpumask: use find_first_and_bit()
lib: add find_first_and_bit()
arch: remove GENERIC_FIND_FIRST_BIT entirely
include: move find.h from asm_generic to linux
bitops: move find_bit_*_le functions from le.h to find.h
bitops: protect find_first_{,zero}_bit properly
find_bit API and bitmap API are closely related, but inclusion paths
are different - include/asm-generic and include/linux, correspondingly.
In the past it made a lot of troubles due to circular dependencies
and/or undefined symbols. Fix this by moving find.h under include/linux.
Signed-off-by: Yury Norov <yury.norov@gmail.com>
Tested-by: Wolfram Sang <wsa+renesas@sang-engineering.com>
Acked-by: Geert Uytterhoeven <geert@linux-m68k.org>
all that "asm/llsc.h" does is just to help inline asm, which can be
stringifyed from "asm/asm.h"
+. Since "asm/asm.h" has all we need, retire "asm/llsc.h"
+. remove unused header file
Inspired-by: Maciej W. Rozycki <macro@orcam.me.uk>
Signed-off-by: Huang Pei <huangpei@loongson.cn>
Signed-off-by: Thomas Bogendoerfer <tsbogend@alpha.franken.de>
Solves the following repetitive warning when building with -Wshadow:
In file included from ./include/linux/bitops.h:32,
from ./include/linux/kernel.h:11,
from ./include/linux/skbuff.h:13,
from ./include/linux/if_ether.h:19,
from ./include/linux/etherdevice.h:20:
./arch/mips/include/asm/bitops.h: In function ‘test_and_set_bit_lock’:
./arch/mips/include/asm/bitops.h:46:16: warning: declaration of ‘orig’ shadows a previous local [-Wshadow]
46 | unsigned long orig, temp; \
| ^~~~
./arch/mips/include/asm/bitops.h:190:10: note: in expansion of macro ‘__test_bit_op’
190 | orig = __test_bit_op(*m, "%0",
| ^~~~~~~~~~~~~
./arch/mips/include/asm/bitops.h:185:21: note: shadowed declaration is here
185 | unsigned long res, orig;
| ^~~~
./arch/mips/include/asm/bitops.h: In function ‘test_and_clear_bit’:
./arch/mips/include/asm/bitops.h:46:16: warning: declaration of ‘orig’ shadows a previous local [-Wshadow]
46 | unsigned long orig, temp; \
| ^~~~
./arch/mips/include/asm/bitops.h:236:9: note: in expansion of macro ‘__test_bit_op’
236 | res = __test_bit_op(*m, "%1",
| ^~~~~~~~~~~~~
./arch/mips/include/asm/bitops.h:229:21: note: shadowed declaration is here
229 | unsigned long res, orig;
| ^~~~
./arch/mips/include/asm/bitops.h:46:16: warning: declaration of ‘orig’ shadows a previous local [-Wshadow]
46 | unsigned long orig, temp; \
| ^~~~
./arch/mips/include/asm/bitops.h:241:10: note: in expansion of macro ‘__test_bit_op’
241 | orig = __test_bit_op(*m, "%0",
| ^~~~~~~~~~~~~
./arch/mips/include/asm/bitops.h:229:21: note: shadowed declaration is here
229 | unsigned long res, orig;
| ^~~~
./arch/mips/include/asm/bitops.h: In function ‘test_and_change_bit’:
./arch/mips/include/asm/bitops.h:46:16: warning: declaration of ‘orig’ shadows a previous local [-Wshadow]
46 | unsigned long orig, temp; \
| ^~~~
./arch/mips/include/asm/bitops.h:273:10: note: in expansion of macro ‘__test_bit_op’
273 | orig = __test_bit_op(*m, "%0",
| ^~~~~~~~~~~~~
./arch/mips/include/asm/bitops.h:266:21: note: shadowed declaration is here
266 | unsigned long res, orig;
| ^~~~
Signed-off-by: Alexander Lobakin <alobakin@pm.me>
Signed-off-by: Thomas Bogendoerfer <tsbogend@alpha.franken.de>
Use smp_mb__before_atomic() rather than smp_mb__before_llsc() in
test_and_set_bit(), test_and_clear_bit() & test_and_change_bit(). The
_atomic() versions make semantic sense in these cases, and will allow a
later patch to omit redundant barriers for Loongson3 systems that
already include a barrier within __test_bit_op().
Signed-off-by: Paul Burton <paul.burton@mips.com>
Cc: linux-mips@vger.kernel.org
Cc: Huacai Chen <chenhc@lemote.com>
Cc: Jiaxun Yang <jiaxun.yang@flygoat.com>
Cc: linux-kernel@vger.kernel.org
Generate the sync instructions required to workaround Loongson3 LL/SC
errata within inline asm blocks, which feels a little safer than doing
it from C where strictly speaking the compiler would be well within its
rights to insert a memory access between the separate asm statements we
previously had, containing sync & ll instructions respectively.
Signed-off-by: Paul Burton <paul.burton@mips.com>
Cc: linux-mips@vger.kernel.org
Cc: Huacai Chen <chenhc@lemote.com>
Cc: Jiaxun Yang <jiaxun.yang@flygoat.com>
Cc: linux-kernel@vger.kernel.org
Rather than using custom SZLONG_LOG & SZLONG_MASK macros to shift & mask
a bit index to form word & bit offsets respectively, make use of the
standard BIT_WORD() & BITS_PER_LONG macros for the same purpose.
volatile is added to the definition of pointers to the long-sized word
we'll operate on, in order to prevent the compiler complaining that we
cast away the volatile qualifier of the addr argument. This should have
no effect on generated code, which in the LL/SC case is inline asm
anyway & in the non-LLSC case access is constrained by compiler barriers
provided by raw_local_irq_{save,restore}().
Signed-off-by: Paul Burton <paul.burton@mips.com>
Cc: linux-mips@vger.kernel.org
Cc: Huacai Chen <chenhc@lemote.com>
Cc: Jiaxun Yang <jiaxun.yang@flygoat.com>
Cc: linux-kernel@vger.kernel.org
The logical operations or & xor used in the test_and_set_bit_lock(),
test_and_clear_bit() & test_and_change_bit() functions currently force
the value 1<<bit to be placed in a register. If the bit is compile-time
constant & fits within the immediate field of an or/xor instruction (ie.
16 bits) then we can make use of the ori/xori instruction variants &
avoid the use of an extra register. Add the extra "i" constraints in
order to allow use of these immediate encodings.
Signed-off-by: Paul Burton <paul.burton@mips.com>
Cc: linux-mips@vger.kernel.org
Cc: Huacai Chen <chenhc@lemote.com>
Cc: Jiaxun Yang <jiaxun.yang@flygoat.com>
Cc: linux-kernel@vger.kernel.org
The only difference between test_and_set_bit() & test_and_set_bit_lock()
is memory ordering barrier semantics - the former provides a full
barrier whilst the latter only provides acquire semantics.
We can therefore implement test_and_set_bit() in terms of
test_and_set_bit_lock() with the addition of the extra memory barrier.
Do this in order to avoid duplicating logic.
Signed-off-by: Paul Burton <paul.burton@mips.com>
Cc: linux-mips@vger.kernel.org
Cc: Huacai Chen <chenhc@lemote.com>
Cc: Jiaxun Yang <jiaxun.yang@flygoat.com>
Cc: linux-kernel@vger.kernel.org
The start position for an ins instruction is always encoded as an
immediate, so allowing registers to be used by the inline asm makes no
sense. It should never happen anyway since a bit index should always be
small enough to be treated as an immediate, but remove the nonsensical
"r" for sanity.
Signed-off-by: Paul Burton <paul.burton@mips.com>
Cc: linux-mips@vger.kernel.org
Cc: Huacai Chen <chenhc@lemote.com>
Cc: Jiaxun Yang <jiaxun.yang@flygoat.com>
Cc: linux-kernel@vger.kernel.org
set_bit() can set bits 0-15 using an ori instruction, rather than
loading the value -1 into a register & then using an ins instruction.
That is, rather than the following:
li t0, -1
ll t1, 0(t2)
ins t1, t0, 4, 1
sc t1, 0(t2)
We can have the simpler:
ll t1, 0(t2)
ori t1, t1, 0x10
sc t1, 0(t2)
The or path already allows immediates to be used, so simply restricting
the ins path to bits that don't fit in immediates is sufficient to take
advantage of this.
Signed-off-by: Paul Burton <paul.burton@mips.com>
Cc: linux-mips@vger.kernel.org
Cc: Huacai Chen <chenhc@lemote.com>
Cc: Jiaxun Yang <jiaxun.yang@flygoat.com>
Cc: linux-kernel@vger.kernel.org
Reorder conditions in our various bitops functions that check
kernel_uses_llsc such that they handle the !kernel_uses_llsc case first.
This allows us to avoid the need to duplicate the kernel_uses_llsc check
in all the other cases. For functions that don't involve barriers common
to the various implementations, we switch to returning from within each
if block making each case easier to read in isolation.
Signed-off-by: Paul Burton <paul.burton@mips.com>
Cc: linux-mips@vger.kernel.org
Cc: Huacai Chen <chenhc@lemote.com>
Cc: Jiaxun Yang <jiaxun.yang@flygoat.com>
Cc: linux-kernel@vger.kernel.org
Recent probing at the Linux Kernel Memory Model uncovered a
'surprise'. Strongly ordered architectures where the atomic RmW
primitive implies full memory ordering and
smp_mb__{before,after}_atomic() are a simple barrier() (such as MIPS
without WEAK_REORDERING_BEYOND_LLSC) fail for:
*x = 1;
atomic_inc(u);
smp_mb__after_atomic();
r0 = *y;
Because, while the atomic_inc() implies memory order, it
(surprisingly) does not provide a compiler barrier. This then allows
the compiler to re-order like so:
atomic_inc(u);
*x = 1;
smp_mb__after_atomic();
r0 = *y;
Which the CPU is then allowed to re-order (under TSO rules) like:
atomic_inc(u);
r0 = *y;
*x = 1;
And this very much was not intended. Therefore strengthen the atomic
RmW ops to include a compiler barrier.
Reported-by: Andrea Parri <andrea.parri@amarulasolutions.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Signed-off-by: Paul Burton <paul.burton@mips.com>
The comment describing the loongson_llsc_mb() reorder case doesn't
make any sense what so ever. Instruction re-ordering is not an SMP
artifact, but rather a CPU local phenomenon. Clarify the comment by
explaining that these issue cause a coherence fail.
For the branch speculation case; if futex_atomic_cmpxchg_inatomic()
needs one at the bne branch target, then surely the normal
__cmpxch_asm() implementation does too. We cannot rely on the
barriers from cmpxchg() because cmpxchg_local() is implemented with
the same macro, and branch prediction and speculation are, too, CPU
local.
Fixes: e02e07e312 ("MIPS: Loongson: Introduce and use loongson_llsc_mb()")
Cc: Huacai Chen <chenhc@lemote.com>
Cc: Huang Pei <huangpei@loongson.cn>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Signed-off-by: Paul Burton <paul.burton@mips.com>
On the Loongson-2G/2H/3A/3B there is a hardware flaw that ll/sc and
lld/scd is very weak ordering. We should add sync instructions "before
each ll/lld" and "at the branch-target between ll/sc" to workaround.
Otherwise, this flaw will cause deadlock occasionally (e.g. when doing
heavy load test with LTP).
Below is the explaination of CPU designer:
"For Loongson 3 family, when a memory access instruction (load, store,
or prefetch)'s executing occurs between the execution of LL and SC, the
success or failure of SC is not predictable. Although programmer would
not insert memory access instructions between LL and SC, the memory
instructions before LL in program-order, may dynamically executed
between the execution of LL/SC, so a memory fence (SYNC) is needed
before LL/LLD to avoid this situation.
Since Loongson-3A R2 (3A2000), we have improved our hardware design to
handle this case. But we later deduce a rarely circumstance that some
speculatively executed memory instructions due to branch misprediction
between LL/SC still fall into the above case, so a memory fence (SYNC)
at branch-target (if its target is not between LL/SC) is needed for
Loongson 3A1000, 3B1500, 3A2000 and 3A3000.
Our processor is continually evolving and we aim to to remove all these
workaround-SYNCs around LL/SC for new-come processor."
Here is an example:
Both cpu1 and cpu2 simutaneously run atomic_add by 1 on same atomic var,
this bug cause both 'sc' run by two cpus (in atomic_add) succeed at same
time('sc' return 1), and the variable is only *added by 1*, sometimes,
which is wrong and unacceptable(it should be added by 2).
Why disable fix-loongson3-llsc in compiler?
Because compiler fix will cause problems in kernel's __ex_table section.
This patch fix all the cases in kernel, but:
+. the fix at the end of futex_atomic_cmpxchg_inatomic is for branch-target
of 'bne', there other cases which smp_mb__before_llsc() and smp_llsc_mb() fix
the ll and branch-target coincidently such as atomic_sub_if_positive/
cmpxchg/xchg, just like this one.
+. Loongson 3 does support CONFIG_EDAC_ATOMIC_SCRUB, so no need to touch
edac.h
+. local_ops and cmpxchg_local should not be affected by this bug since
only the owner can write.
+. mips_atomic_set for syscall.c is deprecated and rarely used, just let
it go
Signed-off-by: Huacai Chen <chenhc@lemote.com>
Signed-off-by: Huang Pei <huangpei@loongson.cn>
[paul.burton@mips.com:
- Simplify the addition of -mno-fix-loongson3-llsc to cflags, and add
a comment describing why it's there.
- Make loongson_llsc_mb() a no-op when
CONFIG_CPU_LOONGSON3_WORKAROUNDS=n, rather than a compiler memory
barrier.
- Add a comment describing the bug & how loongson_llsc_mb() helps
in asm/barrier.h.]
Signed-off-by: Paul Burton <paul.burton@mips.com>
Cc: Ralf Baechle <ralf@linux-mips.org>
Cc: ambrosehua@gmail.com
Cc: Steven J . Hill <Steven.Hill@cavium.com>
Cc: linux-mips@linux-mips.org
Cc: Fuxin Zhang <zhangfx@lemote.com>
Cc: Zhangjin Wu <wuzhangjin@gmail.com>
Cc: Li Xuefeng <lixuefeng@loongson.cn>
Cc: Xu Chenghua <xuchenghua@loongson.cn>
Kees Cook
Matthew Wilcox (Oracle)andAndrew Morton
Linus Torvalds
Yury Norov
Huang PeiandThomas Bogendoerfer
Alexander LobakinandThomas Bogendoerfer
Geert UytterhoevenandThomas Bogendoerfer
Paul Burton
Peter ZijlstraandPaul Burton
Masahiro YamadaandLinus Torvalds
Huacai ChenandPaul Burton