From b29a62d87cc0af3e9d134e9e0863b2cb053070b8 Mon Sep 17 00:00:00 2001 From: Nicolas Pitre Date: Sun, 7 Jul 2024 15:05:19 -0400 Subject: mul_u64_u64_div_u64: make it precise always MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Patch series "mul_u64_u64_div_u64: new implementation", v3. This provides an implementation for mul_u64_u64_div_u64() that always produces exact results. This patch (of 2): Library facilities must always return exact results. If the caller may be contented with approximations then it should do the approximation on its own. In this particular case the comment in the code says "the algorithm ... below might lose some precision". Well, if you try it with e.g.: a = 18446462598732840960 b = 18446462598732840960 c = 18446462598732840961 then the produced answer is 0 whereas the exact answer should be 18446462598732840959. This is _some_ precision lost indeed! Let's reimplement this function so it always produces the exact result regardless of its inputs while preserving existing fast paths when possible. Uwe said: : My personal interest is to get the calculations in pwm drivers right. : This function is used in several drivers below drivers/pwm/ . With the : errors in mul_u64_u64_div_u64(), pwm consumers might not get the : settings they request. Although I have to admit that I'm not aware it : breaks real use cases (because typically the periods used are too short : to make the involved multiplications overflow), but I pretty sure am : not aware of all usages and it breaks testing. : : Another justification is commits like : https://git.kernel.org/tip/77baa5bafcbe1b2a15ef9c37232c21279c95481c, : where people start to work around the precision shortcomings of : mul_u64_u64_div_u64(). Link: https://lkml.kernel.org/r/20240707190648.1982714-1-nico@fluxnic.net Link: https://lkml.kernel.org/r/20240707190648.1982714-2-nico@fluxnic.net Signed-off-by: Nicolas Pitre Tested-by: Uwe Kleine-König Reviewed-by: Uwe Kleine-König Tested-by: Biju Das Signed-off-by: Andrew Morton --- lib/math/div64.c | 108 +++++++++++++++++++++++++++++++++---------------------- 1 file changed, 65 insertions(+), 43 deletions(-) (limited to 'lib') diff --git a/lib/math/div64.c b/lib/math/div64.c index 191761b1b623..b7fc75246399 100644 --- a/lib/math/div64.c +++ b/lib/math/div64.c @@ -186,55 +186,77 @@ EXPORT_SYMBOL(iter_div_u64_rem); #ifndef mul_u64_u64_div_u64 u64 mul_u64_u64_div_u64(u64 a, u64 b, u64 c) { - u64 res = 0, div, rem; - int shift; + if (ilog2(a) + ilog2(b) <= 62) + return div64_u64(a * b, c); - /* can a * b overflow ? */ - if (ilog2(a) + ilog2(b) > 62) { - /* - * Note that the algorithm after the if block below might lose - * some precision and the result is more exact for b > a. So - * exchange a and b if a is bigger than b. - * - * For example with a = 43980465100800, b = 100000000, c = 1000000000 - * the below calculation doesn't modify b at all because div == 0 - * and then shift becomes 45 + 26 - 62 = 9 and so the result - * becomes 4398035251080. However with a and b swapped the exact - * result is calculated (i.e. 4398046510080). - */ - if (a > b) - swap(a, b); +#if defined(__SIZEOF_INT128__) + + /* native 64x64=128 bits multiplication */ + u128 prod = (u128)a * b; + u64 n_lo = prod, n_hi = prod >> 64; + +#else + + /* perform a 64x64=128 bits multiplication manually */ + u32 a_lo = a, a_hi = a >> 32, b_lo = b, b_hi = b >> 32; + u64 x, y, z; + + x = (u64)a_lo * b_lo; + y = (u64)a_lo * b_hi + (u32)(x >> 32); + z = (u64)a_hi * b_hi + (u32)(y >> 32); + y = (u64)a_hi * b_lo + (u32)y; + z += (u32)(y >> 32); + x = (y << 32) + (u32)x; + + u64 n_lo = x, n_hi = z; + +#endif + + int shift = __builtin_ctzll(c); + /* try reducing the fraction in case the dividend becomes <= 64 bits */ + if ((n_hi >> shift) == 0) { + u64 n = (n_lo >> shift) | (n_hi << (64 - shift)); + + return div64_u64(n, c >> shift); /* - * (b * a) / c is equal to - * - * (b / c) * a + - * (b % c) * a / c - * - * if nothing overflows. Can the 1st multiplication - * overflow? Yes, but we do not care: this can only - * happen if the end result can't fit in u64 anyway. - * - * So the code below does - * - * res = (b / c) * a; - * b = b % c; + * The remainder value if needed would be: + * res = div64_u64_rem(n, c >> shift, &rem); + * rem = (rem << shift) + (n_lo - (n << shift)); */ - div = div64_u64_rem(b, c, &rem); - res = div * a; - b = rem; - - shift = ilog2(a) + ilog2(b) - 62; - if (shift > 0) { - /* drop precision */ - b >>= shift; - c >>= shift; - if (!c) - return res; - } } - return res + div64_u64(a * b, c); + if (n_hi >= c) { + /* overflow: result is unrepresentable in a u64 */ + return -1; + } + + /* Do the full 128 by 64 bits division */ + + shift = __builtin_clzll(c); + c <<= shift; + + int p = 64 + shift; + u64 res = 0; + bool carry; + + do { + carry = n_hi >> 63; + shift = carry ? 1 : __builtin_clzll(n_hi); + if (p < shift) + break; + p -= shift; + n_hi <<= shift; + n_hi |= n_lo >> (64 - shift); + n_lo <<= shift; + if (carry || (n_hi >= c)) { + n_hi -= c; + res |= 1ULL << p; + } + } while (n_hi); + /* The remainder value if needed would be n_hi << p */ + + return res; } EXPORT_SYMBOL(mul_u64_u64_div_u64); #endif -- cgit v1.2.3 From 1635e62e75a7bbb1c6274f6b43911cedfe0da60a Mon Sep 17 00:00:00 2001 From: Nicolas Pitre Date: Sun, 7 Jul 2024 15:05:20 -0400 Subject: mul_u64_u64_div_u64: basic sanity test MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Verify that edge cases produce proper results, and some more. [npitre@baylibre.com: avoid undefined shift value] Link: https://lkml.kernel.org/r/7rrs9pn1-n266-3013-9q6n-1osp8r8s0rrn@syhkavp.arg Link: https://lkml.kernel.org/r/20240707190648.1982714-3-nico@fluxnic.net Signed-off-by: Nicolas Pitre Reviewed-by: Uwe Kleine-König Cc: Biju Das Signed-off-by: Andrew Morton --- lib/Kconfig.debug | 10 ++++ lib/math/Makefile | 1 + lib/math/div64.c | 9 +++- lib/math/test_mul_u64_u64_div_u64.c | 99 +++++++++++++++++++++++++++++++++++++ 4 files changed, 118 insertions(+), 1 deletion(-) create mode 100644 lib/math/test_mul_u64_u64_div_u64.c (limited to 'lib') diff --git a/lib/Kconfig.debug b/lib/Kconfig.debug index a30c03a66172..bf0995d328b3 100644 --- a/lib/Kconfig.debug +++ b/lib/Kconfig.debug @@ -2280,6 +2280,16 @@ config TEST_DIV64 If unsure, say N. +config TEST_MULDIV64 + tristate "mul_u64_u64_div_u64() test" + depends on DEBUG_KERNEL || m + help + Enable this to turn on 'mul_u64_u64_div_u64()' function test. + This test is executed only once during system boot (so affects + only boot time), or at module load time. + + If unsure, say N. + config TEST_IOV_ITER tristate "Test iov_iter operation" if !KUNIT_ALL_TESTS depends on KUNIT diff --git a/lib/math/Makefile b/lib/math/Makefile index 91fcdb0c9efe..981a26127e08 100644 --- a/lib/math/Makefile +++ b/lib/math/Makefile @@ -6,4 +6,5 @@ obj-$(CONFIG_PRIME_NUMBERS) += prime_numbers.o obj-$(CONFIG_RATIONAL) += rational.o obj-$(CONFIG_TEST_DIV64) += test_div64.o +obj-$(CONFIG_TEST_MULDIV64) += test_mul_u64_u64_div_u64.o obj-$(CONFIG_RATIONAL_KUNIT_TEST) += rational-test.o diff --git a/lib/math/div64.c b/lib/math/div64.c index b7fc75246399..5faa29208bdb 100644 --- a/lib/math/div64.c +++ b/lib/math/div64.c @@ -212,11 +212,18 @@ u64 mul_u64_u64_div_u64(u64 a, u64 b, u64 c) #endif + /* make sure c is not zero, trigger exception otherwise */ +#pragma GCC diagnostic push +#pragma GCC diagnostic ignored "-Wdiv-by-zero" + if (unlikely(c == 0)) + return 1/0; +#pragma GCC diagnostic pop + int shift = __builtin_ctzll(c); /* try reducing the fraction in case the dividend becomes <= 64 bits */ if ((n_hi >> shift) == 0) { - u64 n = (n_lo >> shift) | (n_hi << (64 - shift)); + u64 n = shift ? (n_lo >> shift) | (n_hi << (64 - shift)) : n_lo; return div64_u64(n, c >> shift); /* diff --git a/lib/math/test_mul_u64_u64_div_u64.c b/lib/math/test_mul_u64_u64_div_u64.c new file mode 100644 index 000000000000..58d058de4e73 --- /dev/null +++ b/lib/math/test_mul_u64_u64_div_u64.c @@ -0,0 +1,99 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * Copyright (C) 2024 BayLibre SAS + */ + +#define pr_fmt(fmt) KBUILD_MODNAME ": " fmt + +#include +#include +#include +#include + +typedef struct { u64 a; u64 b; u64 c; u64 result; } test_params; + +static test_params test_values[] = { +/* this contains many edge values followed by a couple random values */ +{ 0xb, 0x7, 0x3, 0x19 }, +{ 0xffff0000, 0xffff0000, 0xf, 0x1110eeef00000000 }, +{ 0xffffffff, 0xffffffff, 0x1, 0xfffffffe00000001 }, +{ 0xffffffff, 0xffffffff, 0x2, 0x7fffffff00000000 }, +{ 0x1ffffffff, 0xffffffff, 0x2, 0xfffffffe80000000 }, +{ 0x1ffffffff, 0xffffffff, 0x3, 0xaaaaaaa9aaaaaaab }, +{ 0x1ffffffff, 0x1ffffffff, 0x4, 0xffffffff00000000 }, +{ 0xffff000000000000, 0xffff000000000000, 0xffff000000000001, 0xfffeffffffffffff }, +{ 0x3333333333333333, 0x3333333333333333, 0x5555555555555555, 0x1eb851eb851eb851 }, +{ 0x7fffffffffffffff, 0x2, 0x3, 0x5555555555555554 }, +{ 0xffffffffffffffff, 0x2, 0x8000000000000000, 0x3 }, +{ 0xffffffffffffffff, 0x2, 0xc000000000000000, 0x2 }, +{ 0xffffffffffffffff, 0x4000000000000004, 0x8000000000000000, 0x8000000000000007 }, +{ 0xffffffffffffffff, 0x4000000000000001, 0x8000000000000000, 0x8000000000000001 }, +{ 0xffffffffffffffff, 0x8000000000000001, 0xffffffffffffffff, 0x8000000000000001 }, +{ 0xfffffffffffffffe, 0x8000000000000001, 0xffffffffffffffff, 0x8000000000000000 }, +{ 0xffffffffffffffff, 0x8000000000000001, 0xfffffffffffffffe, 0x8000000000000001 }, +{ 0xffffffffffffffff, 0x8000000000000001, 0xfffffffffffffffd, 0x8000000000000002 }, +{ 0x7fffffffffffffff, 0xffffffffffffffff, 0xc000000000000000, 0xaaaaaaaaaaaaaaa8 }, +{ 0xffffffffffffffff, 0x7fffffffffffffff, 0xa000000000000000, 0xccccccccccccccca }, +{ 0xffffffffffffffff, 0x7fffffffffffffff, 0x9000000000000000, 0xe38e38e38e38e38b }, +{ 0x7fffffffffffffff, 0x7fffffffffffffff, 0x5000000000000000, 0xccccccccccccccc9 }, +{ 0xffffffffffffffff, 0xfffffffffffffffe, 0xffffffffffffffff, 0xfffffffffffffffe }, +{ 0xe6102d256d7ea3ae, 0x70a77d0be4c31201, 0xd63ec35ab3220357, 0x78f8bf8cc86c6e18 }, +{ 0xf53bae05cb86c6e1, 0x3847b32d2f8d32e0, 0xcfd4f55a647f403c, 0x42687f79d8998d35 }, +{ 0x9951c5498f941092, 0x1f8c8bfdf287a251, 0xa3c8dc5f81ea3fe2, 0x1d887cb25900091f }, +{ 0x374fee9daa1bb2bb, 0x0d0bfbff7b8ae3ef, 0xc169337bd42d5179, 0x03bb2dbaffcbb961 }, +{ 0xeac0d03ac10eeaf0, 0x89be05dfa162ed9b, 0x92bb1679a41f0e4b, 0xdc5f5cc9e270d216 }, +}; + +/* + * The above table can be verified with the following shell script: + * + * #!/bin/sh + * sed -ne 's/^{ \+\(.*\), \+\(.*\), \+\(.*\), \+\(.*\) },$/\1 \2 \3 \4/p' \ + * lib/math/test_mul_u64_u64_div_u64.c | + * while read a b c r; do + * expected=$( printf "obase=16; ibase=16; %X * %X / %X\n" $a $b $c | bc ) + * given=$( printf "%X\n" $r ) + * if [ "$expected" = "$given" ]; then + * echo "$a * $b / $c = $r OK" + * else + * echo "$a * $b / $c = $r is wrong" >&2 + * echo "should be equivalent to 0x$expected" >&2 + * exit 1 + * fi + * done + */ + +static int __init test_init(void) +{ + int i; + + pr_info("Starting mul_u64_u64_div_u64() test\n"); + + for (i = 0; i < ARRAY_SIZE(test_values); i++) { + u64 a = test_values[i].a; + u64 b = test_values[i].b; + u64 c = test_values[i].c; + u64 expected_result = test_values[i].result; + u64 result = mul_u64_u64_div_u64(a, b, c); + + if (result != expected_result) { + pr_err("ERROR: 0x%016llx * 0x%016llx / 0x%016llx\n", a, b, c); + pr_err("ERROR: expected result: %016llx\n", expected_result); + pr_err("ERROR: obtained result: %016llx\n", result); + } + } + + pr_info("Completed mul_u64_u64_div_u64() test\n"); + return 0; +} + +static void __exit test_exit(void) +{ +} + +module_init(test_init); +module_exit(test_exit); + +MODULE_AUTHOR("Nicolas Pitre"); +MODULE_LICENSE("GPL"); +MODULE_DESCRIPTION("mul_u64_u64_div_u64() test module"); -- cgit v1.2.3 From 053a5e4cbba88625ac6b53dea6371006237c34ba Mon Sep 17 00:00:00 2001 From: Jeff Johnson Date: Mon, 15 Jul 2024 07:18:56 -0700 Subject: lib: test_objpool: add missing MODULE_DESCRIPTION() macro make allmodconfig && make W=1 C=1 reports: WARNING: modpost: missing MODULE_DESCRIPTION() in lib/test_objpool.o Add the missing invocation of the MODULE_DESCRIPTION() macro. Link: https://lkml.kernel.org/r/20240715-md-lib-test_objpool-v2-1-5a2b9369c37e@quicinc.com Signed-off-by: Jeff Johnson Reviewed-by: Matt Wu Cc: Masami Hiramatsu Signed-off-by: Andrew Morton --- lib/test_objpool.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) (limited to 'lib') diff --git a/lib/test_objpool.c b/lib/test_objpool.c index bfdb81599832..5a3f6961a70f 100644 --- a/lib/test_objpool.c +++ b/lib/test_objpool.c @@ -687,4 +687,5 @@ static void __exit ot_mod_exit(void) module_init(ot_mod_init); module_exit(ot_mod_exit); -MODULE_LICENSE("GPL"); \ No newline at end of file +MODULE_DESCRIPTION("Test module for lockless object pool"); +MODULE_LICENSE("GPL"); -- cgit v1.2.3 From e24f4de8a72b50b67ea116b38152bb98360f81b3 Mon Sep 17 00:00:00 2001 From: Andrey Konovalov Date: Tue, 23 Jul 2024 00:37:26 +0200 Subject: kcov: don't instrument lib/find_bit.c This file produces large amounts of flaky coverage not useful for the KCOV's intended use case (guiding the fuzzing process). Link: https://lkml.kernel.org/r/20240722223726.194658-1-andrey.konovalov@linux.dev Signed-off-by: Andrey Konovalov Reviewed-by: Dmitry Vyukov Cc: Aleksandr Nogikh Cc: Alexander Potapenko Cc: Marco Elver Cc: Rasmus Villemoes Cc: Yury Norov Signed-off-by: Andrew Morton --- lib/Makefile | 1 + 1 file changed, 1 insertion(+) (limited to 'lib') diff --git a/lib/Makefile b/lib/Makefile index 322bb127b4dc..0fde1c360f32 100644 --- a/lib/Makefile +++ b/lib/Makefile @@ -14,6 +14,7 @@ KCOV_INSTRUMENT_list_debug.o := n KCOV_INSTRUMENT_debugobjects.o := n KCOV_INSTRUMENT_dynamic_debug.o := n KCOV_INSTRUMENT_fault-inject.o := n +KCOV_INSTRUMENT_find_bit.o := n # string.o implements standard library functions like memset/memcpy etc. # Use -ffreestanding to ensure that the compiler does not try to "optimize" -- cgit v1.2.3 From 836d13a6ef8a2eb0eab2bd2de06f2deabc62b060 Mon Sep 17 00:00:00 2001 From: Lasse Collin Date: Sun, 21 Jul 2024 16:36:18 +0300 Subject: xz: switch from public domain to BSD Zero Clause License (0BSD) Remove the public domain notices and add SPDX license identifiers. Change MODULE_LICENSE from "GPL" to "Dual BSD/GPL" because 0BSD should count as a BSD license variant here. The switch to 0BSD was done in the upstream XZ Embedded project because public domain has (real or perceived) legal issues in some jurisdictions. Link: https://lkml.kernel.org/r/20240721133633.47721-4-lasse.collin@tukaani.org Signed-off-by: Lasse Collin Reviewed-by: Sam James Cc: Thomas Gleixner Cc: Greg Kroah-Hartman Cc: Albert Ou Cc: Catalin Marinas Cc: Emil Renner Berthing Cc: Herbert Xu Cc: Joel Stanley Cc: Jonathan Corbet Cc: Jubin Zhong Cc: Jules Maselbas Cc: Krzysztof Kozlowski Cc: Michael Ellerman Cc: Palmer Dabbelt Cc: Paul Walmsley Cc: Randy Dunlap Cc: Rui Li Cc: Simon Glass Cc: Will Deacon Signed-off-by: Andrew Morton --- lib/decompress_unxz.c | 5 ++--- lib/xz/xz_crc32.c | 5 ++--- lib/xz/xz_dec_bcj.c | 5 ++--- lib/xz/xz_dec_lzma2.c | 5 ++--- lib/xz/xz_dec_stream.c | 5 ++--- lib/xz/xz_dec_syms.c | 12 +++--------- lib/xz/xz_dec_test.c | 12 +++--------- lib/xz/xz_lzma2.h | 5 ++--- lib/xz/xz_private.h | 5 ++--- lib/xz/xz_stream.h | 5 ++--- 10 files changed, 22 insertions(+), 42 deletions(-) (limited to 'lib') diff --git a/lib/decompress_unxz.c b/lib/decompress_unxz.c index 842894158944..34bb7efc0412 100644 --- a/lib/decompress_unxz.c +++ b/lib/decompress_unxz.c @@ -1,10 +1,9 @@ +// SPDX-License-Identifier: 0BSD + /* * Wrapper for decompressing XZ-compressed kernel, initramfs, and initrd * * Author: Lasse Collin - * - * This file has been put into the public domain. - * You can do whatever you want with this file. */ /* diff --git a/lib/xz/xz_crc32.c b/lib/xz/xz_crc32.c index 88a2c35e1b59..30b8a27110b1 100644 --- a/lib/xz/xz_crc32.c +++ b/lib/xz/xz_crc32.c @@ -1,11 +1,10 @@ +// SPDX-License-Identifier: 0BSD + /* * CRC32 using the polynomial from IEEE-802.3 * * Authors: Lasse Collin * Igor Pavlov - * - * This file has been put into the public domain. - * You can do whatever you want with this file. */ /* diff --git a/lib/xz/xz_dec_bcj.c b/lib/xz/xz_dec_bcj.c index ef449e97d1a1..ab9237ed6db8 100644 --- a/lib/xz/xz_dec_bcj.c +++ b/lib/xz/xz_dec_bcj.c @@ -1,11 +1,10 @@ +// SPDX-License-Identifier: 0BSD + /* * Branch/Call/Jump (BCJ) filter decoders * * Authors: Lasse Collin * Igor Pavlov - * - * This file has been put into the public domain. - * You can do whatever you want with this file. */ #include "xz_private.h" diff --git a/lib/xz/xz_dec_lzma2.c b/lib/xz/xz_dec_lzma2.c index 27ce34520e78..613939f5dd6c 100644 --- a/lib/xz/xz_dec_lzma2.c +++ b/lib/xz/xz_dec_lzma2.c @@ -1,11 +1,10 @@ +// SPDX-License-Identifier: 0BSD + /* * LZMA2 decoder * * Authors: Lasse Collin * Igor Pavlov - * - * This file has been put into the public domain. - * You can do whatever you want with this file. */ #include "xz_private.h" diff --git a/lib/xz/xz_dec_stream.c b/lib/xz/xz_dec_stream.c index 683570b93a8c..0058406ccd17 100644 --- a/lib/xz/xz_dec_stream.c +++ b/lib/xz/xz_dec_stream.c @@ -1,10 +1,9 @@ +// SPDX-License-Identifier: 0BSD + /* * .xz Stream decoder * * Author: Lasse Collin - * - * This file has been put into the public domain. - * You can do whatever you want with this file. */ #include "xz_private.h" diff --git a/lib/xz/xz_dec_syms.c b/lib/xz/xz_dec_syms.c index 61098c67a413..495d2cc2e6e8 100644 --- a/lib/xz/xz_dec_syms.c +++ b/lib/xz/xz_dec_syms.c @@ -1,10 +1,9 @@ +// SPDX-License-Identifier: 0BSD + /* * XZ decoder module information * * Author: Lasse Collin - * - * This file has been put into the public domain. - * You can do whatever you want with this file. */ #include @@ -25,9 +24,4 @@ EXPORT_SYMBOL(xz_dec_microlzma_end); MODULE_DESCRIPTION("XZ decompressor"); MODULE_VERSION("1.1"); MODULE_AUTHOR("Lasse Collin and Igor Pavlov"); - -/* - * This code is in the public domain, but in Linux it's simplest to just - * say it's GPL and consider the authors as the copyright holders. - */ -MODULE_LICENSE("GPL"); +MODULE_LICENSE("Dual BSD/GPL"); diff --git a/lib/xz/xz_dec_test.c b/lib/xz/xz_dec_test.c index da28a19d6c98..53d3600f2ddb 100644 --- a/lib/xz/xz_dec_test.c +++ b/lib/xz/xz_dec_test.c @@ -1,10 +1,9 @@ +// SPDX-License-Identifier: 0BSD + /* * XZ decoder tester * * Author: Lasse Collin - * - * This file has been put into the public domain. - * You can do whatever you want with this file. */ #include @@ -212,9 +211,4 @@ module_exit(xz_dec_test_exit); MODULE_DESCRIPTION("XZ decompressor tester"); MODULE_VERSION("1.0"); MODULE_AUTHOR("Lasse Collin "); - -/* - * This code is in the public domain, but in Linux it's simplest to just - * say it's GPL and consider the authors as the copyright holders. - */ -MODULE_LICENSE("GPL"); +MODULE_LICENSE("Dual BSD/GPL"); diff --git a/lib/xz/xz_lzma2.h b/lib/xz/xz_lzma2.h index 92d852d4f87a..d2632b7dfb9c 100644 --- a/lib/xz/xz_lzma2.h +++ b/lib/xz/xz_lzma2.h @@ -1,11 +1,10 @@ +/* SPDX-License-Identifier: 0BSD */ + /* * LZMA2 definitions * * Authors: Lasse Collin * Igor Pavlov - * - * This file has been put into the public domain. - * You can do whatever you want with this file. */ #ifndef XZ_LZMA2_H diff --git a/lib/xz/xz_private.h b/lib/xz/xz_private.h index bf1e94ec7873..2412a5d54801 100644 --- a/lib/xz/xz_private.h +++ b/lib/xz/xz_private.h @@ -1,10 +1,9 @@ +/* SPDX-License-Identifier: 0BSD */ + /* * Private includes and definitions * * Author: Lasse Collin - * - * This file has been put into the public domain. - * You can do whatever you want with this file. */ #ifndef XZ_PRIVATE_H diff --git a/lib/xz/xz_stream.h b/lib/xz/xz_stream.h index 430bb3a0d195..55f9f6f94b78 100644 --- a/lib/xz/xz_stream.h +++ b/lib/xz/xz_stream.h @@ -1,10 +1,9 @@ +/* SPDX-License-Identifier: 0BSD */ + /* * Definitions for handling the .xz file format * * Author: Lasse Collin - * - * This file has been put into the public domain. - * You can do whatever you want with this file. */ #ifndef XZ_STREAM_H -- cgit v1.2.3 From ff221153aafa08159f3dcc187c6f3a7a837e1c3d Mon Sep 17 00:00:00 2001 From: Lasse Collin Date: Sun, 21 Jul 2024 16:36:19 +0300 Subject: xz: fix comments and coding style - Fix comments that were no longer in sync with the code below them. - Fix language errors. - Fix coding style. Link: https://lkml.kernel.org/r/20240721133633.47721-5-lasse.collin@tukaani.org Signed-off-by: Lasse Collin Reviewed-by: Sam James Cc: Albert Ou Cc: Catalin Marinas Cc: Emil Renner Berthing Cc: Greg Kroah-Hartman Cc: Herbert Xu Cc: Joel Stanley Cc: Jonathan Corbet Cc: Jubin Zhong Cc: Jules Maselbas Cc: Krzysztof Kozlowski Cc: Michael Ellerman Cc: Palmer Dabbelt Cc: Paul Walmsley Cc: Randy Dunlap Cc: Rui Li Cc: Simon Glass Cc: Thomas Gleixner Cc: Will Deacon Signed-off-by: Andrew Morton --- lib/decompress_unxz.c | 20 ++++++++++---------- lib/xz/Kconfig | 3 ++- 2 files changed, 12 insertions(+), 11 deletions(-) (limited to 'lib') diff --git a/lib/decompress_unxz.c b/lib/decompress_unxz.c index 34bb7efc0412..46aa3be13fc5 100644 --- a/lib/decompress_unxz.c +++ b/lib/decompress_unxz.c @@ -102,7 +102,7 @@ #ifdef STATIC # define XZ_PREBOOT #else -#include +# include #endif #ifdef __KERNEL__ # include @@ -219,7 +219,7 @@ void *memmove(void *dest, const void *src, size_t size) #endif /* - * Since we need memmove anyway, would use it as memcpy too. + * Since we need memmove anyway, we could use it as memcpy too. * Commented out for now to avoid breaking things. */ /* @@ -389,17 +389,17 @@ error_alloc_state: } /* - * This macro is used by architecture-specific files to decompress + * This function is used by architecture-specific files to decompress * the kernel image. */ #ifdef XZ_PREBOOT -STATIC int INIT __decompress(unsigned char *buf, long len, - long (*fill)(void*, unsigned long), - long (*flush)(void*, unsigned long), - unsigned char *out_buf, long olen, - long *pos, - void (*error)(char *x)) +STATIC int INIT __decompress(unsigned char *in, long in_size, + long (*fill)(void *dest, unsigned long size), + long (*flush)(void *src, unsigned long size), + unsigned char *out, long out_size, + long *in_used, + void (*error)(char *x)) { - return unxz(buf, len, fill, flush, out_buf, pos, error); + return unxz(in, in_size, fill, flush, out, in_used, error); } #endif diff --git a/lib/xz/Kconfig b/lib/xz/Kconfig index aef086a6bf2f..6b80453d8f54 100644 --- a/lib/xz/Kconfig +++ b/lib/xz/Kconfig @@ -5,7 +5,8 @@ config XZ_DEC help LZMA2 compression algorithm and BCJ filters are supported using the .xz file format as the container. For integrity checking, - CRC32 is supported. See Documentation/staging/xz.rst for more information. + CRC32 is supported. See Documentation/staging/xz.rst for more + information. if XZ_DEC -- cgit v1.2.3 From 2ee96abef214550d9e92f5143ee3ac1fd1323e67 Mon Sep 17 00:00:00 2001 From: Lasse Collin Date: Sun, 21 Jul 2024 16:36:24 +0300 Subject: xz: cleanup CRC32 edits from 2018 In 2018, a dependency on was added to avoid duplicating the same constant in multiple files. Two months later it was found to be a bad idea and the definition of CRC32_POLY_LE macro was moved into xz_private.h to avoid including . xz_private.h is a wrong place for it too. Revert back to the upstream version which has the poly in xz_crc32_init() in xz_crc32.c. Link: https://lkml.kernel.org/r/20240721133633.47721-10-lasse.collin@tukaani.org Fixes: faa16bc404d7 ("lib: Use existing define with polynomial") Fixes: 242cdad873a7 ("lib/xz: Put CRC32_POLY_LE in xz_private.h") Signed-off-by: Lasse Collin Reviewed-by: Sam James Tested-by: Michael Ellerman (powerpc) Cc: Krzysztof Kozlowski Cc: Herbert Xu Cc: Joel Stanley Cc: Albert Ou Cc: Catalin Marinas Cc: Emil Renner Berthing Cc: Greg Kroah-Hartman Cc: Jonathan Corbet Cc: Jubin Zhong Cc: Jules Maselbas Cc: Palmer Dabbelt Cc: Paul Walmsley Cc: Randy Dunlap Cc: Rui Li Cc: Simon Glass Cc: Thomas Gleixner Cc: Will Deacon Signed-off-by: Andrew Morton --- lib/xz/xz_crc32.c | 2 +- lib/xz/xz_private.h | 4 ---- 2 files changed, 1 insertion(+), 5 deletions(-) (limited to 'lib') diff --git a/lib/xz/xz_crc32.c b/lib/xz/xz_crc32.c index 30b8a27110b1..effdf34ec48d 100644 --- a/lib/xz/xz_crc32.c +++ b/lib/xz/xz_crc32.c @@ -28,7 +28,7 @@ STATIC_RW_DATA uint32_t xz_crc32_table[256]; XZ_EXTERN void xz_crc32_init(void) { - const uint32_t poly = CRC32_POLY_LE; + const uint32_t poly = 0xEDB88320; uint32_t i; uint32_t j; diff --git a/lib/xz/xz_private.h b/lib/xz/xz_private.h index 2412a5d54801..811add814ae4 100644 --- a/lib/xz/xz_private.h +++ b/lib/xz/xz_private.h @@ -104,10 +104,6 @@ # endif #endif -#ifndef CRC32_POLY_LE -#define CRC32_POLY_LE 0xedb88320 -#endif - /* * Allocate memory for LZMA2 decoder. xz_dec_lzma2_reset() must be used * before calling xz_dec_lzma2_run(). -- cgit v1.2.3 From bdfc0411717d52b9d2f00e48c452a61389814693 Mon Sep 17 00:00:00 2001 From: Lasse Collin Date: Sun, 21 Jul 2024 16:36:25 +0300 Subject: xz: optimize for-loop conditions in the BCJ decoders Compilers cannot optimize the addition "i + 4" away since theoretically it could overflow. Link: https://lkml.kernel.org/r/20240721133633.47721-11-lasse.collin@tukaani.org Signed-off-by: Lasse Collin Reviewed-by: Sam James Cc: Albert Ou Cc: Catalin Marinas Cc: Emil Renner Berthing Cc: Greg Kroah-Hartman Cc: Herbert Xu Cc: Joel Stanley Cc: Jonathan Corbet Cc: Jubin Zhong Cc: Jules Maselbas Cc: Krzysztof Kozlowski Cc: Michael Ellerman Cc: Palmer Dabbelt Cc: Paul Walmsley Cc: Randy Dunlap Cc: Rui Li Cc: Simon Glass Cc: Thomas Gleixner Cc: Will Deacon Signed-off-by: Andrew Morton --- lib/xz/xz_dec_bcj.c | 23 ++++++++++++++++++----- 1 file changed, 18 insertions(+), 5 deletions(-) (limited to 'lib') diff --git a/lib/xz/xz_dec_bcj.c b/lib/xz/xz_dec_bcj.c index ab9237ed6db8..e0b4bf4999c0 100644 --- a/lib/xz/xz_dec_bcj.c +++ b/lib/xz/xz_dec_bcj.c @@ -161,7 +161,9 @@ static size_t bcj_powerpc(struct xz_dec_bcj *s, uint8_t *buf, size_t size) size_t i; uint32_t instr; - for (i = 0; i + 4 <= size; i += 4) { + size &= ~(size_t)3; + + for (i = 0; i < size; i += 4) { instr = get_unaligned_be32(buf + i); if ((instr & 0xFC000003) == 0x48000001) { instr &= 0x03FFFFFC; @@ -218,7 +220,9 @@ static size_t bcj_ia64(struct xz_dec_bcj *s, uint8_t *buf, size_t size) /* Instruction normalized with bit_res for easier manipulation */ uint64_t norm; - for (i = 0; i + 16 <= size; i += 16) { + size &= ~(size_t)15; + + for (i = 0; i < size; i += 16) { mask = branch_table[buf[i] & 0x1F]; for (slot = 0, bit_pos = 5; slot < 3; ++slot, bit_pos += 41) { if (((mask >> slot) & 1) == 0) @@ -266,7 +270,9 @@ static size_t bcj_arm(struct xz_dec_bcj *s, uint8_t *buf, size_t size) size_t i; uint32_t addr; - for (i = 0; i + 4 <= size; i += 4) { + size &= ~(size_t)3; + + for (i = 0; i < size; i += 4) { if (buf[i + 3] == 0xEB) { addr = (uint32_t)buf[i] | ((uint32_t)buf[i + 1] << 8) | ((uint32_t)buf[i + 2] << 16); @@ -289,7 +295,12 @@ static size_t bcj_armthumb(struct xz_dec_bcj *s, uint8_t *buf, size_t size) size_t i; uint32_t addr; - for (i = 0; i + 4 <= size; i += 2) { + if (size < 4) + return 0; + + size -= 4; + + for (i = 0; i <= size; i += 2) { if ((buf[i + 1] & 0xF8) == 0xF0 && (buf[i + 3] & 0xF8) == 0xF8) { addr = (((uint32_t)buf[i + 1] & 0x07) << 19) @@ -317,7 +328,9 @@ static size_t bcj_sparc(struct xz_dec_bcj *s, uint8_t *buf, size_t size) size_t i; uint32_t instr; - for (i = 0; i + 4 <= size; i += 4) { + size &= ~(size_t)3; + + for (i = 0; i < size; i += 4) { instr = get_unaligned_be32(buf + i); if ((instr >> 22) == 0x100 || (instr >> 22) == 0x1FF) { instr <<= 2; -- cgit v1.2.3 From 4b62813f5e7d44a33ebd74f03da041712c702bf0 Mon Sep 17 00:00:00 2001 From: Lasse Collin Date: Sun, 21 Jul 2024 16:36:26 +0300 Subject: xz: Add ARM64 BCJ filter Also omit a duplicated check for XZ_DEC_ARM in xz_private.h. A later commit updates lib/decompress_unxz.c to enable this filter for kernel decompression. lib/decompress_unxz.c is already used if CONFIG_EFI_ZBOOT=y && CONFIG_KERNEL_XZ=y. This filter can be used by Squashfs without modifications to the Squashfs kernel code (only needs support in userspace Squashfs-tools). Link: https://lkml.kernel.org/r/20240721133633.47721-12-lasse.collin@tukaani.org Signed-off-by: Lasse Collin Reviewed-by: Sam James Cc: Albert Ou Cc: Catalin Marinas Cc: Emil Renner Berthing Cc: Greg Kroah-Hartman Cc: Herbert Xu Cc: Joel Stanley Cc: Jonathan Corbet Cc: Jubin Zhong Cc: Jules Maselbas Cc: Krzysztof Kozlowski Cc: Michael Ellerman Cc: Palmer Dabbelt Cc: Paul Walmsley Cc: Randy Dunlap Cc: Rui Li Cc: Simon Glass Cc: Thomas Gleixner Cc: Will Deacon Signed-off-by: Andrew Morton --- lib/xz/Kconfig | 5 +++++ lib/xz/xz_dec_bcj.c | 52 +++++++++++++++++++++++++++++++++++++++++++++++++++- lib/xz/xz_private.h | 7 +++++-- 3 files changed, 61 insertions(+), 3 deletions(-) (limited to 'lib') diff --git a/lib/xz/Kconfig b/lib/xz/Kconfig index 6b80453d8f54..1166627a87dc 100644 --- a/lib/xz/Kconfig +++ b/lib/xz/Kconfig @@ -30,6 +30,11 @@ config XZ_DEC_ARMTHUMB default y select XZ_DEC_BCJ +config XZ_DEC_ARM64 + bool "ARM64 BCJ filter decoder" if EXPERT + default y + select XZ_DEC_BCJ + config XZ_DEC_SPARC bool "SPARC BCJ filter decoder" if EXPERT default y diff --git a/lib/xz/xz_dec_bcj.c b/lib/xz/xz_dec_bcj.c index e0b4bf4999c0..941198a8a55b 100644 --- a/lib/xz/xz_dec_bcj.c +++ b/lib/xz/xz_dec_bcj.c @@ -23,7 +23,8 @@ struct xz_dec_bcj { BCJ_IA64 = 6, /* Big or little endian */ BCJ_ARM = 7, /* Little endian only */ BCJ_ARMTHUMB = 8, /* Little endian only */ - BCJ_SPARC = 9 /* Big or little endian */ + BCJ_SPARC = 9, /* Big or little endian */ + BCJ_ARM64 = 10 /* AArch64 */ } type; /* @@ -346,6 +347,47 @@ static size_t bcj_sparc(struct xz_dec_bcj *s, uint8_t *buf, size_t size) } #endif +#ifdef XZ_DEC_ARM64 +static size_t bcj_arm64(struct xz_dec_bcj *s, uint8_t *buf, size_t size) +{ + size_t i; + uint32_t instr; + uint32_t addr; + + size &= ~(size_t)3; + + for (i = 0; i < size; i += 4) { + instr = get_unaligned_le32(buf + i); + + if ((instr >> 26) == 0x25) { + /* BL instruction */ + addr = instr - ((s->pos + (uint32_t)i) >> 2); + instr = 0x94000000 | (addr & 0x03FFFFFF); + put_unaligned_le32(instr, buf + i); + + } else if ((instr & 0x9F000000) == 0x90000000) { + /* ADRP instruction */ + addr = ((instr >> 29) & 3) | ((instr >> 3) & 0x1FFFFC); + + /* Only convert values in the range +/-512 MiB. */ + if ((addr + 0x020000) & 0x1C0000) + continue; + + addr -= (s->pos + (uint32_t)i) >> 12; + + instr &= 0x9000001F; + instr |= (addr & 3) << 29; + instr |= (addr & 0x03FFFC) << 3; + instr |= (0U - (addr & 0x020000)) & 0xE00000; + + put_unaligned_le32(instr, buf + i); + } + } + + return i; +} +#endif + /* * Apply the selected BCJ filter. Update *pos and s->pos to match the amount * of data that got filtered. @@ -392,6 +434,11 @@ static void bcj_apply(struct xz_dec_bcj *s, case BCJ_SPARC: filtered = bcj_sparc(s, buf, size); break; +#endif +#ifdef XZ_DEC_ARM64 + case BCJ_ARM64: + filtered = bcj_arm64(s, buf, size); + break; #endif default: /* Never reached but silence compiler warnings. */ @@ -565,6 +612,9 @@ XZ_EXTERN enum xz_ret xz_dec_bcj_reset(struct xz_dec_bcj *s, uint8_t id) #endif #ifdef XZ_DEC_SPARC case BCJ_SPARC: +#endif +#ifdef XZ_DEC_ARM64 + case BCJ_ARM64: #endif break; diff --git a/lib/xz/xz_private.h b/lib/xz/xz_private.h index 811add814ae4..307e0de8c260 100644 --- a/lib/xz/xz_private.h +++ b/lib/xz/xz_private.h @@ -36,6 +36,9 @@ # ifdef CONFIG_XZ_DEC_SPARC # define XZ_DEC_SPARC # endif +# ifdef CONFIG_XZ_DEC_ARM64 +# define XZ_DEC_ARM64 +# endif # ifdef CONFIG_XZ_DEC_MICROLZMA # define XZ_DEC_MICROLZMA # endif @@ -97,9 +100,9 @@ */ #ifndef XZ_DEC_BCJ # if defined(XZ_DEC_X86) || defined(XZ_DEC_POWERPC) \ - || defined(XZ_DEC_IA64) || defined(XZ_DEC_ARM) \ + || defined(XZ_DEC_IA64) \ || defined(XZ_DEC_ARM) || defined(XZ_DEC_ARMTHUMB) \ - || defined(XZ_DEC_SPARC) + || defined(XZ_DEC_SPARC) || defined(XZ_DEC_ARM64) # define XZ_DEC_BCJ # endif #endif -- cgit v1.2.3 From 93d09773d1a5339160e23906c68c42644e13e3d8 Mon Sep 17 00:00:00 2001 From: Lasse Collin Date: Sun, 21 Jul 2024 16:36:27 +0300 Subject: xz: add RISC-V BCJ filter A later commit updates lib/decompress_unxz.c to enable this filter for kernel decompression. lib/decompress_unxz.c is already used if CONFIG_EFI_ZBOOT=y && CONFIG_KERNEL_XZ=y. This filter can be used by Squashfs without modifications to the Squashfs kernel code (only needs support in userspace Squashfs-tools). Link: https://lkml.kernel.org/r/20240721133633.47721-13-lasse.collin@tukaani.org Signed-off-by: Lasse Collin Reviewed-by: Sam James Cc: Albert Ou Cc: Catalin Marinas Cc: Emil Renner Berthing Cc: Greg Kroah-Hartman Cc: Herbert Xu Cc: Joel Stanley Cc: Jonathan Corbet Cc: Jubin Zhong Cc: Jules Maselbas Cc: Krzysztof Kozlowski Cc: Michael Ellerman Cc: Palmer Dabbelt Cc: Paul Walmsley Cc: Randy Dunlap Cc: Rui Li Cc: Simon Glass Cc: Thomas Gleixner Cc: Will Deacon Signed-off-by: Andrew Morton --- lib/xz/Kconfig | 5 +++ lib/xz/xz_dec_bcj.c | 104 ++++++++++++++++++++++++++++++++++++++++++++++++++- lib/xz/xz_dec_syms.c | 2 +- lib/xz/xz_private.h | 6 ++- 4 files changed, 114 insertions(+), 3 deletions(-) (limited to 'lib') diff --git a/lib/xz/Kconfig b/lib/xz/Kconfig index 1166627a87dc..20aa459bfb3e 100644 --- a/lib/xz/Kconfig +++ b/lib/xz/Kconfig @@ -40,6 +40,11 @@ config XZ_DEC_SPARC default y select XZ_DEC_BCJ +config XZ_DEC_RISCV + bool "RISC-V BCJ filter decoder" if EXPERT + default y + select XZ_DEC_BCJ + config XZ_DEC_MICROLZMA bool "MicroLZMA decoder" default n diff --git a/lib/xz/xz_dec_bcj.c b/lib/xz/xz_dec_bcj.c index 941198a8a55b..42d7f268726f 100644 --- a/lib/xz/xz_dec_bcj.c +++ b/lib/xz/xz_dec_bcj.c @@ -24,7 +24,8 @@ struct xz_dec_bcj { BCJ_ARM = 7, /* Little endian only */ BCJ_ARMTHUMB = 8, /* Little endian only */ BCJ_SPARC = 9, /* Big or little endian */ - BCJ_ARM64 = 10 /* AArch64 */ + BCJ_ARM64 = 10, /* AArch64 */ + BCJ_RISCV = 11 /* RV32GQC_Zfh, RV64GQC_Zfh */ } type; /* @@ -388,6 +389,99 @@ static size_t bcj_arm64(struct xz_dec_bcj *s, uint8_t *buf, size_t size) } #endif +#ifdef XZ_DEC_RISCV +static size_t bcj_riscv(struct xz_dec_bcj *s, uint8_t *buf, size_t size) +{ + size_t i; + uint32_t b1; + uint32_t b2; + uint32_t b3; + uint32_t instr; + uint32_t instr2; + uint32_t instr2_rs1; + uint32_t addr; + + if (size < 8) + return 0; + + size -= 8; + + for (i = 0; i <= size; i += 2) { + instr = buf[i]; + + if (instr == 0xEF) { + /* JAL */ + b1 = buf[i + 1]; + if ((b1 & 0x0D) != 0) + continue; + + b2 = buf[i + 2]; + b3 = buf[i + 3]; + + addr = ((b1 & 0xF0) << 13) | (b2 << 9) | (b3 << 1); + addr -= s->pos + (uint32_t)i; + + buf[i + 1] = (uint8_t)((b1 & 0x0F) + | ((addr >> 8) & 0xF0)); + + buf[i + 2] = (uint8_t)(((addr >> 16) & 0x0F) + | ((addr >> 7) & 0x10) + | ((addr << 4) & 0xE0)); + + buf[i + 3] = (uint8_t)(((addr >> 4) & 0x7F) + | ((addr >> 13) & 0x80)); + + i += 4 - 2; + + } else if ((instr & 0x7F) == 0x17) { + /* AUIPC */ + instr |= (uint32_t)buf[i + 1] << 8; + instr |= (uint32_t)buf[i + 2] << 16; + instr |= (uint32_t)buf[i + 3] << 24; + + if (instr & 0xE80) { + /* AUIPC's rd doesn't equal x0 or x2. */ + instr2 = get_unaligned_le32(buf + i + 4); + + if (((instr << 8) ^ (instr2 - 3)) & 0xF8003) { + i += 6 - 2; + continue; + } + + addr = (instr & 0xFFFFF000) + (instr2 >> 20); + + instr = 0x17 | (2 << 7) | (instr2 << 12); + instr2 = addr; + } else { + /* AUIPC's rd equals x0 or x2. */ + instr2_rs1 = instr >> 27; + + if ((uint32_t)((instr - 0x3117) << 18) + >= (instr2_rs1 & 0x1D)) { + i += 4 - 2; + continue; + } + + addr = get_unaligned_be32(buf + i + 4); + addr -= s->pos + (uint32_t)i; + + instr2 = (instr >> 12) | (addr << 20); + + instr = 0x17 | (instr2_rs1 << 7) + | ((addr + 0x800) & 0xFFFFF000); + } + + put_unaligned_le32(instr, buf + i); + put_unaligned_le32(instr2, buf + i + 4); + + i += 8 - 2; + } + } + + return i; +} +#endif + /* * Apply the selected BCJ filter. Update *pos and s->pos to match the amount * of data that got filtered. @@ -439,6 +533,11 @@ static void bcj_apply(struct xz_dec_bcj *s, case BCJ_ARM64: filtered = bcj_arm64(s, buf, size); break; +#endif +#ifdef XZ_DEC_RISCV + case BCJ_RISCV: + filtered = bcj_riscv(s, buf, size); + break; #endif default: /* Never reached but silence compiler warnings. */ @@ -615,6 +714,9 @@ XZ_EXTERN enum xz_ret xz_dec_bcj_reset(struct xz_dec_bcj *s, uint8_t id) #endif #ifdef XZ_DEC_ARM64 case BCJ_ARM64: +#endif +#ifdef XZ_DEC_RISCV + case BCJ_RISCV: #endif break; diff --git a/lib/xz/xz_dec_syms.c b/lib/xz/xz_dec_syms.c index 495d2cc2e6e8..f40817d65897 100644 --- a/lib/xz/xz_dec_syms.c +++ b/lib/xz/xz_dec_syms.c @@ -22,6 +22,6 @@ EXPORT_SYMBOL(xz_dec_microlzma_end); #endif MODULE_DESCRIPTION("XZ decompressor"); -MODULE_VERSION("1.1"); +MODULE_VERSION("1.2"); MODULE_AUTHOR("Lasse Collin and Igor Pavlov"); MODULE_LICENSE("Dual BSD/GPL"); diff --git a/lib/xz/xz_private.h b/lib/xz/xz_private.h index 307e0de8c260..a8b1cbe8d21d 100644 --- a/lib/xz/xz_private.h +++ b/lib/xz/xz_private.h @@ -39,6 +39,9 @@ # ifdef CONFIG_XZ_DEC_ARM64 # define XZ_DEC_ARM64 # endif +# ifdef CONFIG_XZ_DEC_RISCV +# define XZ_DEC_RISCV +# endif # ifdef CONFIG_XZ_DEC_MICROLZMA # define XZ_DEC_MICROLZMA # endif @@ -102,7 +105,8 @@ # if defined(XZ_DEC_X86) || defined(XZ_DEC_POWERPC) \ || defined(XZ_DEC_IA64) \ || defined(XZ_DEC_ARM) || defined(XZ_DEC_ARMTHUMB) \ - || defined(XZ_DEC_SPARC) || defined(XZ_DEC_ARM64) + || defined(XZ_DEC_SPARC) || defined(XZ_DEC_ARM64) \ + || defined(XZ_DEC_RISCV) # define XZ_DEC_BCJ # endif #endif -- cgit v1.2.3 From 7472ff8adad8655f38b060a602f66e59c93c4793 Mon Sep 17 00:00:00 2001 From: Lasse Collin Date: Sun, 21 Jul 2024 16:36:29 +0300 Subject: xz: adjust arch-specific options for better kernel compression Use LZMA2 options that match the arch-specific alignment of instructions. This change reduces compressed kernel size 0-2 % depending on the arch. On 1-byte-aligned x86 it makes no difference and on 4-byte-aligned archs it helps the most. Use the ARM-Thumb filter for ARM-Thumb2 kernels. This reduces compressed kernel size about 5 %.[1] Previously such kernels were compressed using the ARM filter which didn't do anything useful with ARM-Thumb2 code. Add BCJ filter support for ARM64 and RISC-V. Compared to unfiltered XZ or plain LZMA, the compressed kernel size is reduced about 5 % on ARM64 and 7 % on RISC-V. A new enough version of the xz tool is required: 5.4.0 for ARM64 and 5.6.0 for RISC-V. With an old xz version, a message is printed to standard error and the kernel is compressed without the filter. Update lib/decompress_unxz.c to match the changes to xz_wrap.sh. Update the CONFIG_KERNEL_XZ help text in init/Kconfig: - Add the RISC-V and ARM64 filters. - Clarify that the PowerPC filter is for big endian only. - Omit IA-64. Link: https://lore.kernel.org/lkml/1637379771-39449-1-git-send-email-zhongjubin@huawei.com/ [1] Link: https://lkml.kernel.org/r/20240721133633.47721-15-lasse.collin@tukaani.org Signed-off-by: Lasse Collin Reviewed-by: Sam James Cc: Simon Glass Cc: Catalin Marinas Cc: Will Deacon Cc: Paul Walmsley Cc: Palmer Dabbelt Cc: Albert Ou Cc: Jubin Zhong Cc: Jules Maselbas Cc: Emil Renner Berthing Cc: Greg Kroah-Hartman Cc: Herbert Xu Cc: Joel Stanley Cc: Jonathan Corbet Cc: Krzysztof Kozlowski Cc: Michael Ellerman Cc: Randy Dunlap Cc: Rui Li Cc: Thomas Gleixner Signed-off-by: Andrew Morton --- lib/decompress_unxz.c | 14 ++++++++++++-- 1 file changed, 12 insertions(+), 2 deletions(-) (limited to 'lib') diff --git a/lib/decompress_unxz.c b/lib/decompress_unxz.c index 46aa3be13fc5..cae00395d7a6 100644 --- a/lib/decompress_unxz.c +++ b/lib/decompress_unxz.c @@ -126,11 +126,21 @@ #ifdef CONFIG_X86 # define XZ_DEC_X86 #endif -#ifdef CONFIG_PPC +#if defined(CONFIG_PPC) && defined(CONFIG_CPU_BIG_ENDIAN) # define XZ_DEC_POWERPC #endif #ifdef CONFIG_ARM -# define XZ_DEC_ARM +# ifdef CONFIG_THUMB2_KERNEL +# define XZ_DEC_ARMTHUMB +# else +# define XZ_DEC_ARM +# endif +#endif +#ifdef CONFIG_ARM64 +# define XZ_DEC_ARM64 +#endif +#ifdef CONFIG_RISCV +# define XZ_DEC_RISCV #endif #ifdef CONFIG_SPARC # define XZ_DEC_SPARC -- cgit v1.2.3 From c6f371bab25edccd39caa5dd452b50d9dfdf4ff0 Mon Sep 17 00:00:00 2001 From: Lasse Collin Date: Wed, 24 Jul 2024 14:05:41 +0300 Subject: xz: remove XZ_EXTERN and extern from functions XZ_EXTERN was used to make internal functions static in the preboot code. However, in other decompressors this hasn't been done. On x86-64, this makes no difference to the kernel image size. Omit XZ_EXTERN and let some of the internal functions be extern in the preboot code. Omitting XZ_EXTERN from include/linux/xz.h fixes warnings in "make htmldocs" and makes the intradocument links to xz_dec functions work in Documentation/staging/xz.rst. The alternative would have been to add "XZ_EXTERN" to c_id_attributes in Documentation/conf.py but omitting XZ_EXTERN seemed cleaner. Link: https://lore.kernel.org/lkml/20240723205437.3c0664b0@kaneli/ Link: https://lkml.kernel.org/r/20240724110544.16430-1-lasse.collin@tukaani.org Signed-off-by: Lasse Collin Tested-by: Michael Ellerman (powerpc) Cc: Jonathan Corbet Cc: Sam James Cc: Albert Ou Cc: Catalin Marinas Cc: Emil Renner Berthing Cc: Greg Kroah-Hartman Cc: Herbert Xu Cc: Joel Stanley Cc: Jubin Zhong Cc: Jules Maselbas Cc: Krzysztof Kozlowski Cc: Palmer Dabbelt Cc: Paul Walmsley Cc: Randy Dunlap Cc: Rui Li Cc: Simon Glass Cc: Thomas Gleixner Cc: Will Deacon Signed-off-by: Andrew Morton --- lib/decompress_unxz.c | 1 - lib/xz/xz_crc32.c | 4 ++-- lib/xz/xz_dec_bcj.c | 9 ++++----- lib/xz/xz_dec_lzma2.c | 10 ++++------ lib/xz/xz_dec_stream.c | 8 ++++---- lib/xz/xz_private.h | 20 ++++++++------------ 6 files changed, 22 insertions(+), 30 deletions(-) (limited to 'lib') diff --git a/lib/decompress_unxz.c b/lib/decompress_unxz.c index cae00395d7a6..32138bb8ef77 100644 --- a/lib/decompress_unxz.c +++ b/lib/decompress_unxz.c @@ -107,7 +107,6 @@ #ifdef __KERNEL__ # include #endif -#define XZ_EXTERN STATIC #ifndef XZ_PREBOOT # include diff --git a/lib/xz/xz_crc32.c b/lib/xz/xz_crc32.c index effdf34ec48d..6a7906a328ba 100644 --- a/lib/xz/xz_crc32.c +++ b/lib/xz/xz_crc32.c @@ -26,7 +26,7 @@ STATIC_RW_DATA uint32_t xz_crc32_table[256]; -XZ_EXTERN void xz_crc32_init(void) +void xz_crc32_init(void) { const uint32_t poly = 0xEDB88320; @@ -45,7 +45,7 @@ XZ_EXTERN void xz_crc32_init(void) return; } -XZ_EXTERN uint32_t xz_crc32(const uint8_t *buf, size_t size, uint32_t crc) +uint32_t xz_crc32(const uint8_t *buf, size_t size, uint32_t crc) { crc = ~crc; diff --git a/lib/xz/xz_dec_bcj.c b/lib/xz/xz_dec_bcj.c index 42d7f268726f..8237db17eee3 100644 --- a/lib/xz/xz_dec_bcj.c +++ b/lib/xz/xz_dec_bcj.c @@ -572,9 +572,8 @@ static void bcj_flush(struct xz_dec_bcj *s, struct xz_buf *b) * data in chunks of 1-16 bytes. To hide this issue, this function does * some buffering. */ -XZ_EXTERN enum xz_ret xz_dec_bcj_run(struct xz_dec_bcj *s, - struct xz_dec_lzma2 *lzma2, - struct xz_buf *b) +enum xz_ret xz_dec_bcj_run(struct xz_dec_bcj *s, struct xz_dec_lzma2 *lzma2, + struct xz_buf *b) { size_t out_start; @@ -682,7 +681,7 @@ XZ_EXTERN enum xz_ret xz_dec_bcj_run(struct xz_dec_bcj *s, return s->ret; } -XZ_EXTERN struct xz_dec_bcj *xz_dec_bcj_create(bool single_call) +struct xz_dec_bcj *xz_dec_bcj_create(bool single_call) { struct xz_dec_bcj *s = kmalloc(sizeof(*s), GFP_KERNEL); if (s != NULL) @@ -691,7 +690,7 @@ XZ_EXTERN struct xz_dec_bcj *xz_dec_bcj_create(bool single_call) return s; } -XZ_EXTERN enum xz_ret xz_dec_bcj_reset(struct xz_dec_bcj *s, uint8_t id) +enum xz_ret xz_dec_bcj_reset(struct xz_dec_bcj *s, uint8_t id) { switch (id) { #ifdef XZ_DEC_X86 diff --git a/lib/xz/xz_dec_lzma2.c b/lib/xz/xz_dec_lzma2.c index 613939f5dd6c..83bb66b6016d 100644 --- a/lib/xz/xz_dec_lzma2.c +++ b/lib/xz/xz_dec_lzma2.c @@ -960,8 +960,7 @@ static bool lzma2_lzma(struct xz_dec_lzma2 *s, struct xz_buf *b) * Take care of the LZMA2 control layer, and forward the job of actual LZMA * decoding or copying of uncompressed chunks to other functions. */ -XZ_EXTERN enum xz_ret xz_dec_lzma2_run(struct xz_dec_lzma2 *s, - struct xz_buf *b) +enum xz_ret xz_dec_lzma2_run(struct xz_dec_lzma2 *s, struct xz_buf *b) { uint32_t tmp; @@ -1137,8 +1136,7 @@ XZ_EXTERN enum xz_ret xz_dec_lzma2_run(struct xz_dec_lzma2 *s, return XZ_OK; } -XZ_EXTERN struct xz_dec_lzma2 *xz_dec_lzma2_create(enum xz_mode mode, - uint32_t dict_max) +struct xz_dec_lzma2 *xz_dec_lzma2_create(enum xz_mode mode, uint32_t dict_max) { struct xz_dec_lzma2 *s = kmalloc(sizeof(*s), GFP_KERNEL); if (s == NULL) @@ -1161,7 +1159,7 @@ XZ_EXTERN struct xz_dec_lzma2 *xz_dec_lzma2_create(enum xz_mode mode, return s; } -XZ_EXTERN enum xz_ret xz_dec_lzma2_reset(struct xz_dec_lzma2 *s, uint8_t props) +enum xz_ret xz_dec_lzma2_reset(struct xz_dec_lzma2 *s, uint8_t props) { /* This limits dictionary size to 3 GiB to keep parsing simpler. */ if (props > 39) @@ -1197,7 +1195,7 @@ XZ_EXTERN enum xz_ret xz_dec_lzma2_reset(struct xz_dec_lzma2 *s, uint8_t props) return XZ_OK; } -XZ_EXTERN void xz_dec_lzma2_end(struct xz_dec_lzma2 *s) +void xz_dec_lzma2_end(struct xz_dec_lzma2 *s) { if (DEC_IS_MULTI(s->dict.mode)) vfree(s->dict.buf); diff --git a/lib/xz/xz_dec_stream.c b/lib/xz/xz_dec_stream.c index 0058406ccd17..f9d003684d56 100644 --- a/lib/xz/xz_dec_stream.c +++ b/lib/xz/xz_dec_stream.c @@ -746,7 +746,7 @@ static enum xz_ret dec_main(struct xz_dec *s, struct xz_buf *b) * actually succeeds (that's the price to pay of using the output buffer as * the workspace). */ -XZ_EXTERN enum xz_ret xz_dec_run(struct xz_dec *s, struct xz_buf *b) +enum xz_ret xz_dec_run(struct xz_dec *s, struct xz_buf *b) { size_t in_start; size_t out_start; @@ -782,7 +782,7 @@ XZ_EXTERN enum xz_ret xz_dec_run(struct xz_dec *s, struct xz_buf *b) return ret; } -XZ_EXTERN struct xz_dec *xz_dec_init(enum xz_mode mode, uint32_t dict_max) +struct xz_dec *xz_dec_init(enum xz_mode mode, uint32_t dict_max) { struct xz_dec *s = kmalloc(sizeof(*s), GFP_KERNEL); if (s == NULL) @@ -812,7 +812,7 @@ error_bcj: return NULL; } -XZ_EXTERN void xz_dec_reset(struct xz_dec *s) +void xz_dec_reset(struct xz_dec *s) { s->sequence = SEQ_STREAM_HEADER; s->allow_buf_error = false; @@ -824,7 +824,7 @@ XZ_EXTERN void xz_dec_reset(struct xz_dec *s) s->temp.size = STREAM_HEADER_SIZE; } -XZ_EXTERN void xz_dec_end(struct xz_dec *s) +void xz_dec_end(struct xz_dec *s) { if (s != NULL) { xz_dec_lzma2_end(s->lzma2); diff --git a/lib/xz/xz_private.h b/lib/xz/xz_private.h index a8b1cbe8d21d..5f1294a1408c 100644 --- a/lib/xz/xz_private.h +++ b/lib/xz/xz_private.h @@ -115,8 +115,7 @@ * Allocate memory for LZMA2 decoder. xz_dec_lzma2_reset() must be used * before calling xz_dec_lzma2_run(). */ -XZ_EXTERN struct xz_dec_lzma2 *xz_dec_lzma2_create(enum xz_mode mode, - uint32_t dict_max); +struct xz_dec_lzma2 *xz_dec_lzma2_create(enum xz_mode mode, uint32_t dict_max); /* * Decode the LZMA2 properties (one byte) and reset the decoder. Return @@ -124,22 +123,20 @@ XZ_EXTERN struct xz_dec_lzma2 *xz_dec_lzma2_create(enum xz_mode mode, * big enough, and XZ_OPTIONS_ERROR if props indicates something that this * decoder doesn't support. */ -XZ_EXTERN enum xz_ret xz_dec_lzma2_reset(struct xz_dec_lzma2 *s, - uint8_t props); +enum xz_ret xz_dec_lzma2_reset(struct xz_dec_lzma2 *s, uint8_t props); /* Decode raw LZMA2 stream from b->in to b->out. */ -XZ_EXTERN enum xz_ret xz_dec_lzma2_run(struct xz_dec_lzma2 *s, - struct xz_buf *b); +enum xz_ret xz_dec_lzma2_run(struct xz_dec_lzma2 *s, struct xz_buf *b); /* Free the memory allocated for the LZMA2 decoder. */ -XZ_EXTERN void xz_dec_lzma2_end(struct xz_dec_lzma2 *s); +void xz_dec_lzma2_end(struct xz_dec_lzma2 *s); #ifdef XZ_DEC_BCJ /* * Allocate memory for BCJ decoders. xz_dec_bcj_reset() must be used before * calling xz_dec_bcj_run(). */ -XZ_EXTERN struct xz_dec_bcj *xz_dec_bcj_create(bool single_call); +struct xz_dec_bcj *xz_dec_bcj_create(bool single_call); /* * Decode the Filter ID of a BCJ filter. This implementation doesn't @@ -147,16 +144,15 @@ XZ_EXTERN struct xz_dec_bcj *xz_dec_bcj_create(bool single_call); * is needed. Returns XZ_OK if the given Filter ID is supported. * Otherwise XZ_OPTIONS_ERROR is returned. */ -XZ_EXTERN enum xz_ret xz_dec_bcj_reset(struct xz_dec_bcj *s, uint8_t id); +enum xz_ret xz_dec_bcj_reset(struct xz_dec_bcj *s, uint8_t id); /* * Decode raw BCJ + LZMA2 stream. This must be used only if there actually is * a BCJ filter in the chain. If the chain has only LZMA2, xz_dec_lzma2_run() * must be called directly. */ -XZ_EXTERN enum xz_ret xz_dec_bcj_run(struct xz_dec_bcj *s, - struct xz_dec_lzma2 *lzma2, - struct xz_buf *b); +enum xz_ret xz_dec_bcj_run(struct xz_dec_bcj *s, struct xz_dec_lzma2 *lzma2, + struct xz_buf *b); /* Free the memory allocated for the BCJ filters. */ #define xz_dec_bcj_end(s) kfree(s) -- cgit v1.2.3 From 7b76689a021d19a016310bd5da35450641b67966 Mon Sep 17 00:00:00 2001 From: Markus Elfring Date: Sun, 14 Jul 2024 19:33:07 +0200 Subject: dyndbg: use seq_putc() in ddebug_proc_show() Single characters should be put into a sequence. Thus use the corresponding function "seq_putc". This issue was transformed by using the Coccinelle software. Link: https://lkml.kernel.org/r/375b5b4b-6295-419e-bae9-da724a7a682d@web.de Signed-off-by: Markus Elfring Cc: Jason Baron Cc: Jim Cromie Signed-off-by: Andrew Morton --- lib/dynamic_debug.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'lib') diff --git a/lib/dynamic_debug.c b/lib/dynamic_debug.c index f2c5e7910bb1..5a007952f7f2 100644 --- a/lib/dynamic_debug.c +++ b/lib/dynamic_debug.c @@ -1147,7 +1147,7 @@ static int ddebug_proc_show(struct seq_file *m, void *p) iter->table->mod_name, dp->function, ddebug_describe_flags(dp->flags, &flags)); seq_escape_str(m, dp->format, ESCAPE_SPACE, "\t\r\n\""); - seq_puts(m, "\""); + seq_putc(m, '"'); if (dp->class_id != _DPRINTK_CLASS_DFLT) { class = ddebug_class_name(iter, dp); @@ -1156,7 +1156,7 @@ static int ddebug_proc_show(struct seq_file *m, void *p) else seq_printf(m, " class unknown, _id:%d", dp->class_id); } - seq_puts(m, "\n"); + seq_putc(m, '\n'); return 0; } -- cgit v1.2.3 From fbe617af697c336db7630762158127eaa5a1d223 Mon Sep 17 00:00:00 2001 From: Markus Elfring Date: Sun, 14 Jul 2024 19:15:09 +0200 Subject: closures: use seq_putc() in debug_show() A single line break should be put into a sequence. Thus use the corresponding function "seq_putc". This issue was transformed by using the Coccinelle software. Link: https://lkml.kernel.org/r/e7faa2c4-9590-44b4-8669-69ef810277b1@web.de Signed-off-by: Markus Elfring Cc: Kent Overstreet Signed-off-by: Andrew Morton --- lib/closure.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'lib') diff --git a/lib/closure.c b/lib/closure.c index 116afae2eed9..2bfe7d2a0048 100644 --- a/lib/closure.c +++ b/lib/closure.c @@ -278,7 +278,7 @@ static int debug_show(struct seq_file *f, void *data) seq_printf(f, " W %pS\n", (void *) cl->waiting_on); - seq_puts(f, "\n"); + seq_putc(f, '\n'); } spin_unlock_irq(&closure_list_lock); -- cgit v1.2.3 From 9a42bfd255b288dad2d1a9df0a1fe58394d5da12 Mon Sep 17 00:00:00 2001 From: Deshan Zhang Date: Thu, 25 Jul 2024 17:30:45 +0800 Subject: lib/lru_cache: fix spelling mistake "colision"->"collision" MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit There is a spelling mistake in a literal string and in cariable names. Fix these. Link: https://lkml.kernel.org/r/20240725093044.1742842-1-deshan@nfschina.com Signed-off-by: Deshan Zhang Cc: Christoph Böhmwalder Cc: Lars Ellenberg Cc: Philipp Reisner Signed-off-by: Andrew Morton --- lib/lru_cache.c | 10 +++++----- 1 file changed, 5 insertions(+), 5 deletions(-) (limited to 'lib') diff --git a/lib/lru_cache.c b/lib/lru_cache.c index b3d9187611de..9e0d469c7658 100644 --- a/lib/lru_cache.c +++ b/lib/lru_cache.c @@ -243,7 +243,7 @@ static struct lc_element *__lc_find(struct lru_cache *lc, unsigned int enr, BUG_ON(!lc); BUG_ON(!lc->nr_elements); - hlist_for_each_entry(e, lc_hash_slot(lc, enr), colision) { + hlist_for_each_entry(e, lc_hash_slot(lc, enr), collision) { /* "about to be changed" elements, pending transaction commit, * are hashed by their "new number". "Normal" elements have * lc_number == lc_new_number. */ @@ -303,7 +303,7 @@ void lc_del(struct lru_cache *lc, struct lc_element *e) BUG_ON(e->refcnt); e->lc_number = e->lc_new_number = LC_FREE; - hlist_del_init(&e->colision); + hlist_del_init(&e->collision); list_move(&e->list, &lc->free); RETURN(); } @@ -324,9 +324,9 @@ static struct lc_element *lc_prepare_for_change(struct lru_cache *lc, unsigned n PARANOIA_LC_ELEMENT(lc, e); e->lc_new_number = new_number; - if (!hlist_unhashed(&e->colision)) - __hlist_del(&e->colision); - hlist_add_head(&e->colision, lc_hash_slot(lc, new_number)); + if (!hlist_unhashed(&e->collision)) + __hlist_del(&e->collision); + hlist_add_head(&e->collision, lc_hash_slot(lc, new_number)); list_move(&e->list, &lc->to_be_changed); return e; -- cgit v1.2.3 From b6e21b71208f289a796d786bd695ec25eae4ed9a Mon Sep 17 00:00:00 2001 From: Thorsten Blum Date: Fri, 26 Jul 2024 17:49:46 +0200 Subject: lib: checksum: use ARRAY_SIZE() to improve assert_setup_correct() Use ARRAY_SIZE() to simplify the assert_setup_correct() function and improve its readability. Link: https://lkml.kernel.org/r/20240726154946.230928-1-thorsten.blum@toblux.com Signed-off-by: Thorsten Blum Signed-off-by: Andrew Morton --- lib/checksum_kunit.c | 9 +++------ 1 file changed, 3 insertions(+), 6 deletions(-) (limited to 'lib') diff --git a/lib/checksum_kunit.c b/lib/checksum_kunit.c index 4e4d081a1d3b..be04aa42125c 100644 --- a/lib/checksum_kunit.c +++ b/lib/checksum_kunit.c @@ -468,12 +468,9 @@ static __wsum to_wsum(u32 x) static void assert_setup_correct(struct kunit *test) { - CHECK_EQ(sizeof(random_buf) / sizeof(random_buf[0]), MAX_LEN); - CHECK_EQ(sizeof(expected_results) / sizeof(expected_results[0]), - MAX_LEN); - CHECK_EQ(sizeof(init_sums_no_overflow) / - sizeof(init_sums_no_overflow[0]), - MAX_LEN); + CHECK_EQ(ARRAY_SIZE(random_buf), MAX_LEN); + CHECK_EQ(ARRAY_SIZE(expected_results), MAX_LEN); + CHECK_EQ(ARRAY_SIZE(init_sums_no_overflow), MAX_LEN); } /* -- cgit v1.2.3 From e0ba72e3a4422c4255fa80191a637d7c65ef4c59 Mon Sep 17 00:00:00 2001 From: "J. R. Okajima" Date: Tue, 23 Jul 2024 16:40:17 +0000 Subject: lockdep: upper limit LOCKDEP_CHAINS_BITS CONFIG_LOCKDEP_CHAINS_BITS value decides the size of chain_hlocks[] in kernel/locking/lockdep.c, and it is checked by add_chain_cache() with BUILD_BUG_ON((1UL << 24) <= ARRAY_SIZE(chain_hlocks)); This patch is just to silence BUILD_BUG_ON(). See also https://lore.kernel.org/all/30795.1620913191@jrobl/ [cmllamas@google.com: fix minor checkpatch issues in commit log] Link: https://lkml.kernel.org/r/20240723164018.2489615-1-cmllamas@google.com Signed-off-by: J. R. Okajima Signed-off-by: Carlos Llamas Acked-by: Tetsuo Handa Cc: Peter Zijlstra Cc: Boqun Feng Cc: Ingo Molnar Cc: Waiman Long Cc: Will Deacon Signed-off-by: Andrew Morton --- lib/Kconfig.debug | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'lib') diff --git a/lib/Kconfig.debug b/lib/Kconfig.debug index bf0995d328b3..a81d452941ce 100644 --- a/lib/Kconfig.debug +++ b/lib/Kconfig.debug @@ -1515,7 +1515,7 @@ config LOCKDEP_BITS config LOCKDEP_CHAINS_BITS int "Bitsize for MAX_LOCKDEP_CHAINS" depends on LOCKDEP && !LOCKDEP_SMALL - range 10 30 + range 10 21 default 16 help Try increasing this value if you hit "BUG: MAX_LOCKDEP_CHAINS too low!" message. -- cgit v1.2.3 From a15bec6a8f2f177e6c1388f23d02436e27994299 Mon Sep 17 00:00:00 2001 From: Davidlohr Bueso Date: Tue, 6 Aug 2024 08:39:27 -0700 Subject: lib/rhashtable: cleanup fallback check in bucket_table_alloc() Upon allocation failure, the current check with the nofail bits is unnecessary, and further stands in the way of discouraging direct use of __GFP_NOFAIL. Remove this and replace with the proper way of determining if doing a non-blocking allocation for the nested table case. Link: https://lkml.kernel.org/r/20240806153927.184515-1-dave@stgolabs.net Signed-off-by: Davidlohr Bueso Suggested-by: Michal Hocko Cc: Davidlohr Bueso Cc: Herbert Xu Signed-off-by: Andrew Morton --- lib/rhashtable.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'lib') diff --git a/lib/rhashtable.c b/lib/rhashtable.c index dbbed19f8fff..6c902639728b 100644 --- a/lib/rhashtable.c +++ b/lib/rhashtable.c @@ -189,7 +189,7 @@ static struct bucket_table *bucket_table_alloc(struct rhashtable *ht, size = nbuckets; - if (tbl == NULL && (gfp & ~__GFP_NOFAIL) != GFP_KERNEL) { + if (tbl == NULL && !gfpflags_allow_blocking(gfp)) { tbl = nested_bucket_table_alloc(ht, nbuckets, gfp); nbuckets = 0; } -- cgit v1.2.3 From 6ce2082fd3a25d5a8c756120959237cace0379f1 Mon Sep 17 00:00:00 2001 From: Jani Nikula Date: Tue, 13 Aug 2024 15:12:35 +0300 Subject: fault-inject: improve build for CONFIG_FAULT_INJECTION=n MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit The fault-inject.h users across the kernel need to add a lot of #ifdef CONFIG_FAULT_INJECTION to cater for shortcomings in the header. Make fault-inject.h self-contained for CONFIG_FAULT_INJECTION=n, and add stubs for DECLARE_FAULT_ATTR(), setup_fault_attr(), should_fail_ex(), and should_fail() to allow removal of conditional compilation. [akpm@linux-foundation.org: repair fallout from no longer including debugfs.h into fault-inject.h] [akpm@linux-foundation.org: fix drivers/misc/xilinx_tmr_inject.c] [akpm@linux-foundation.org: Add debugfs.h inclusion to more files, per Stephen] Link: https://lkml.kernel.org/r/20240813121237.2382534-1-jani.nikula@intel.com Fixes: 6ff1cb355e62 ("[PATCH] fault-injection capabilities infrastructure") Signed-off-by: Jani Nikula Cc: Akinobu Mita Cc: Abhinav Kumar Cc: Dmitry Baryshkov Cc: Himal Prasad Ghimiray Cc: Lucas De Marchi Cc: Rob Clark Cc: Rodrigo Vivi Cc: Thomas Hellström Cc: Stephen Rothwell Signed-off-by: Andrew Morton --- lib/fault-inject.c | 1 + 1 file changed, 1 insertion(+) (limited to 'lib') diff --git a/lib/fault-inject.c b/lib/fault-inject.c index d608f9b48c10..52eb6ba29698 100644 --- a/lib/fault-inject.c +++ b/lib/fault-inject.c @@ -2,6 +2,7 @@ #include #include #include +#include #include #include #include -- cgit v1.2.3 From cbf164cd44e06c78938b4a4a4479d3541779c319 Mon Sep 17 00:00:00 2001 From: Kuan-Wei Chiu Date: Tue, 13 Aug 2024 01:02:29 +0800 Subject: lib/bcd: optimize _bin2bcd() for improved performance The original _bin2bcd() function used / 10 and % 10 operations for conversion. Although GCC optimizes these operations and does not generate division or modulus instructions, the new implementation reduces the number of mov instructions in the generated code for both x86-64 and ARM architectures. This optimization calculates the tens digit using (val * 103) >> 10, which is accurate for values of 'val' in the range [0, 178]. Given that the valid input range is [0, 99], this method ensures correctness while simplifying the generated code. Link: https://lkml.kernel.org/r/20240812170229.229380-1-visitorckw@gmail.com Signed-off-by: Kuan-Wei Chiu Cc: Ching-Chun (Jim) Huang Signed-off-by: Andrew Morton --- lib/bcd.c | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-) (limited to 'lib') diff --git a/lib/bcd.c b/lib/bcd.c index 7e4750b6e801..c5e79ba9cd7b 100644 --- a/lib/bcd.c +++ b/lib/bcd.c @@ -10,6 +10,8 @@ EXPORT_SYMBOL(_bcd2bin); unsigned char _bin2bcd(unsigned val) { - return ((val / 10) << 4) + val % 10; + const unsigned int t = (val * 103) >> 10; + + return (t << 4) | (val - t * 10); } EXPORT_SYMBOL(_bin2bcd); -- cgit v1.2.3 From 16d9691ad4b562ea19271f0788738f649c02cf3c Mon Sep 17 00:00:00 2001 From: Uros Bizjak Date: Wed, 14 Aug 2024 08:44:13 +0200 Subject: lib/percpu_counter: add missing __percpu qualifier to a cast Add missing __percpu qualifier to a (void *) cast to fix percpu_counter.c:212:36: warning: cast removes address space '__percpu' of expression percpu_counter.c:212:33: warning: incorrect type in assignment (different address spaces) percpu_counter.c:212:33: expected signed int [noderef] [usertype] __percpu *counters percpu_counter.c:212:33: got void * sparse warnings. Found by GCC's named address space checks. There were no changes in the resulting object file. Link: https://lkml.kernel.org/r/20240814064437.940162-1-ubizjak@gmail.com Signed-off-by: Uros Bizjak Cc: Dennis Zhou Cc: Tejun Heo Cc: Christoph Lameter Signed-off-by: Andrew Morton --- lib/percpu_counter.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'lib') diff --git a/lib/percpu_counter.c b/lib/percpu_counter.c index 51bc5246986d..2891f94a11c6 100644 --- a/lib/percpu_counter.c +++ b/lib/percpu_counter.c @@ -209,7 +209,7 @@ int __percpu_counter_init_many(struct percpu_counter *fbc, s64 amount, INIT_LIST_HEAD(&fbc[i].list); #endif fbc[i].count = amount; - fbc[i].counters = (void *)counters + (i * counter_size); + fbc[i].counters = (void __percpu *)counters + i * counter_size; debug_percpu_counter_activate(&fbc[i]); } -- cgit v1.2.3 From fb54ea1ee84534cab6a15515c73a0811bdcbc973 Mon Sep 17 00:00:00 2001 From: Andy Shevchenko Date: Wed, 21 Aug 2024 18:51:04 +0300 Subject: dimlib: use *-y instead of *-objs in Makefile *-objs suffix is reserved rather for (user-space) host programs while usually *-y suffix is used for kernel drivers (although *-objs works for that purpose for now). Let's correct the old usages of *-objs in Makefiles. Link: https://lkml.kernel.org/r/20240821155140.611514-1-andriy.shevchenko@linux.intel.com Signed-off-by: Andy Shevchenko Acked-by: Florian Fainelli Reviewed-by: Alexander Lobakin Cc: Rasmus Villemoes Cc: Tal Gilboa Signed-off-by: Andrew Morton --- lib/dim/Makefile | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'lib') diff --git a/lib/dim/Makefile b/lib/dim/Makefile index c4cc4026c451..5b9bfaac7ac1 100644 --- a/lib/dim/Makefile +++ b/lib/dim/Makefile @@ -4,4 +4,4 @@ obj-$(CONFIG_DIMLIB) += dimlib.o -dimlib-objs := dim.o net_dim.o rdma_dim.o +dimlib-y := dim.o net_dim.o rdma_dim.o -- cgit v1.2.3 From 38676d9e33133c0c39951b812b19cc5b9ff1978a Mon Sep 17 00:00:00 2001 From: Yang Ruibin <11162571@vivo.com> Date: Wed, 21 Aug 2024 03:34:40 -0400 Subject: lib: fix the NULL vs IS_ERR() bug for debugfs_create_dir() debugfs_create_dir() returns error pointers. It never returns NULL. So use IS_ERR() to check it. Link: https://lkml.kernel.org/r/20240821073441.9701-1-11162571@vivo.com Signed-off-by: Yang Ruibin <11162571@vivo.com> Signed-off-by: Andrew Morton --- lib/test_fpu_glue.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'lib') diff --git a/lib/test_fpu_glue.c b/lib/test_fpu_glue.c index 074f30301f29..c0596426370a 100644 --- a/lib/test_fpu_glue.c +++ b/lib/test_fpu_glue.c @@ -42,7 +42,7 @@ static int __init test_fpu_init(void) return -EINVAL; selftest_dir = debugfs_create_dir("selftest_helpers", NULL); - if (!selftest_dir) + if (IS_ERR(selftest_dir)) return -ENOMEM; debugfs_create_file_unsafe("test_fpu", 0444, selftest_dir, NULL, -- cgit v1.2.3 From 7b0a5b666959719043123a8882bae49ec699d948 Mon Sep 17 00:00:00 2001 From: Alok Swaminathan Date: Mon, 26 Aug 2024 11:57:09 -0400 Subject: lib: glob.c: added null check for character class Add null check for character class. Previously, an inverted character class could result in a nul byte being matched and lead to the function reading past the end of the inputted string. Link: https://lkml.kernel.org/r/20240826155709.12383-1-swaminathanalok@gmail.com Signed-off-by: Alok Swaminathan Signed-off-by: Andrew Morton --- lib/glob.c | 2 ++ 1 file changed, 2 insertions(+) (limited to 'lib') diff --git a/lib/glob.c b/lib/glob.c index 15b73f490720..aa57900d2062 100644 --- a/lib/glob.c +++ b/lib/glob.c @@ -68,6 +68,8 @@ bool __pure glob_match(char const *pat, char const *str) back_str = --str; /* Allow zero-length match */ break; case '[': { /* Character class */ + if (c == '\0') /* No possible match */ + return false; bool match = false, inverted = (*pat == '!'); char const *class = pat + inverted; unsigned char a = *class++; -- cgit v1.2.3 From e620799c414a035dea1208bcb51c869744931dbb Mon Sep 17 00:00:00 2001 From: I Hsin Cheng Date: Tue, 10 Sep 2024 12:35:31 +0800 Subject: list: test: fix tests for list_cut_position() Fix test for list_cut_position*() for the missing check of integer "i" after the second loop. The variable should be checked for second time to make sure both lists after the cut operation are formed as expected. Link: https://lkml.kernel.org/r/20240910043531.71343-1-richard120310@gmail.com Signed-off-by: I Hsin Cheng Cc: David Gow Signed-off-by: Andrew Morton --- lib/list-test.c | 6 ++++++ 1 file changed, 6 insertions(+) (limited to 'lib') diff --git a/lib/list-test.c b/lib/list-test.c index 37cbc33e9fdb..8d1d47a9fe9e 100644 --- a/lib/list-test.c +++ b/lib/list-test.c @@ -404,10 +404,13 @@ static void list_test_list_cut_position(struct kunit *test) KUNIT_EXPECT_EQ(test, i, 2); + i = 0; list_for_each(cur, &list1) { KUNIT_EXPECT_PTR_EQ(test, cur, &entries[i]); i++; } + + KUNIT_EXPECT_EQ(test, i, 1); } static void list_test_list_cut_before(struct kunit *test) @@ -432,10 +435,13 @@ static void list_test_list_cut_before(struct kunit *test) KUNIT_EXPECT_EQ(test, i, 1); + i = 0; list_for_each(cur, &list1) { KUNIT_EXPECT_PTR_EQ(test, cur, &entries[i]); i++; } + + KUNIT_EXPECT_EQ(test, i, 2); } static void list_test_list_splice(struct kunit *test) -- cgit v1.2.3 From 5e06e08939df1cafef97a8e04f4b88c2806b538a Mon Sep 17 00:00:00 2001 From: I Hsin Cheng Date: Tue, 10 Sep 2024 12:08:18 +0800 Subject: list: test: increase coverage of list_test_list_replace*() Increase the test coverage of list_test_list_replace*() by adding the checks to compare the pointer of "a_new.next" and "a_new.prev" to make sure a perfect circular doubly linked list is formed after the replacement. Link: https://lkml.kernel.org/r/20240910040818.65723-1-richard120310@gmail.com Signed-off-by: I Hsin Cheng Cc: David Gow Signed-off-by: Andrew Morton --- lib/list-test.c | 4 ++++ 1 file changed, 4 insertions(+) (limited to 'lib') diff --git a/lib/list-test.c b/lib/list-test.c index 8d1d47a9fe9e..4f3dc75baec1 100644 --- a/lib/list-test.c +++ b/lib/list-test.c @@ -102,6 +102,8 @@ static void list_test_list_replace(struct kunit *test) /* now: [list] -> a_new -> b */ KUNIT_EXPECT_PTR_EQ(test, list.next, &a_new); KUNIT_EXPECT_PTR_EQ(test, b.prev, &a_new); + KUNIT_EXPECT_PTR_EQ(test, a_new.next, &b); + KUNIT_EXPECT_PTR_EQ(test, a_new.prev, &list); } static void list_test_list_replace_init(struct kunit *test) @@ -118,6 +120,8 @@ static void list_test_list_replace_init(struct kunit *test) /* now: [list] -> a_new -> b */ KUNIT_EXPECT_PTR_EQ(test, list.next, &a_new); KUNIT_EXPECT_PTR_EQ(test, b.prev, &a_new); + KUNIT_EXPECT_PTR_EQ(test, a_new.next, &b); + KUNIT_EXPECT_PTR_EQ(test, a_new.prev, &list); /* check a_old is empty (initialized) */ KUNIT_EXPECT_TRUE(test, list_empty_careful(&a_old)); -- cgit v1.2.3