From 773e1c5fa4bf1faa25e119490b26ece2ef1bdb46 Mon Sep 17 00:00:00 2001 From: George Spelvin Date: Tue, 7 Jun 2016 19:45:06 -0400 Subject: [PATCH 1/3] parisc: Add PA-RISC is interesting; integer multiplies are implemented in the FPU, so are painful in the kernel. But it tries to be friendly to shift-and-add sequences for constant multiplies. __hash_32 is implemented using the same shift-and-add sequence as Microblaze, just scheduled for the PA7100. (It's 2-way superscalar but in-order, like the Pentium.) hash_64 was tricky, but a suggestion from Jason Thong allowed a good solution by breaking up the multiplier. After a lot of manual optimization, I found a 19-instruction sequence for the multiply that can be executed in 10 cycles using only 4 temporaries. (The PA8xxx can issue 4 instructions per cycle, but 2 must be ALU ops and 2 must be loads/stores. And the final add can't be paired.) An alternative considered, but ultimately not used, was Thomas Wang's 64-to-32-bit integer hash. At 12 instructions, it's smaller, but they're all sequentially dependent, so it has longer latency. https://web.archive.org/web/2011/http://www.concentric.net/~Ttwang/tech/inthash.htm http://burtleburtle.net/bob/hash/integer.html Signed-off-by: George Spelvin Cc: Helge Deller Cc: linux-parisc@vger.kernel.org Signed-off-by: Helge Deller --- arch/parisc/Kconfig | 1 + arch/parisc/include/asm/hash.h | 146 +++++++++++++++++++++++++++++++++ 2 files changed, 147 insertions(+) create mode 100644 arch/parisc/include/asm/hash.h diff --git a/arch/parisc/Kconfig b/arch/parisc/Kconfig index dc117385ce2e..cd8778103165 100644 --- a/arch/parisc/Kconfig +++ b/arch/parisc/Kconfig @@ -31,6 +31,7 @@ config PARISC select TTY # Needed for pdc_cons.c select HAVE_DEBUG_STACKOVERFLOW select HAVE_ARCH_AUDITSYSCALL + select HAVE_ARCH_HASH select HAVE_ARCH_SECCOMP_FILTER select HAVE_ARCH_TRACEHOOK select HAVE_UNSTABLE_SCHED_CLOCK if (SMP || !64BIT) diff --git a/arch/parisc/include/asm/hash.h b/arch/parisc/include/asm/hash.h new file mode 100644 index 000000000000..dbe93311aa26 --- /dev/null +++ b/arch/parisc/include/asm/hash.h @@ -0,0 +1,146 @@ +#ifndef _ASM_HASH_H +#define _ASM_HASH_H + +/* + * HP-PA only implements integer multiply in the FPU. However, for + * integer multiplies by constant, it has a number of shift-and-add + * (but no shift-and-subtract, sigh!) instructions that a compiler + * can synthesize a code sequence with. + * + * Unfortunately, GCC isn't very efficient at using them. For example + * it uses three instructions for "x *= 21" when only two are needed. + * But we can find a sequence manually. + */ + +#define HAVE_ARCH__HASH_32 1 + +/* + * This is a multiply by GOLDEN_RATIO_32 = 0x61C88647 optimized for the + * PA7100 pairing rules. This is an in-order 2-way superscalar processor. + * Only one instruction in a pair may be a shift (by more than 3 bits), + * but other than that, simple ALU ops (including shift-and-add by up + * to 3 bits) may be paired arbitrarily. + * + * PA8xxx processors also dual-issue ALU instructions, although with + * fewer constraints, so this schedule is good for them, too. + * + * This 6-step sequence was found by Yevgen Voronenko's implementation + * of the Hcub algorithm at http://spiral.ece.cmu.edu/mcm/gen.html. + */ +static inline u32 __attribute_const__ __hash_32(u32 x) +{ + u32 a, b, c; + + /* + * Phase 1: Compute a = (x << 19) + x, + * b = (x << 9) + a, c = (x << 23) + b. + */ + a = x << 19; /* Two shifts can't be paired */ + b = x << 9; a += x; + c = x << 23; b += a; + c += b; + /* Phase 2: Return (b<<11) + (c<<6) + (a<<3) - c */ + b <<= 11; + a += c << 3; b -= c; + return (a << 3) + b; +} + +#if BITS_PER_LONG == 64 + +#define HAVE_ARCH_HASH_64 1 + +/* + * Finding a good shift-and-add chain for GOLDEN_RATIO_64 is tricky, + * because available software for the purpose chokes on constants this + * large. (It's mostly designed for compiling FIR filter coefficients + * into FPGAs.) + * + * However, Jason Thong pointed out a work-around. The Hcub software + * (http://spiral.ece.cmu.edu/mcm/gen.html) is designed for *multiple* + * constant multiplication, and is good at finding shift-and-add chains + * which share common terms. + * + * Looking at 0x0x61C8864680B583EB in binary: + * 0110000111001000100001100100011010000000101101011000001111101011 + * \______________/ \__________/ \_______/ \________/ + * \____________________________/ \____________________/ + * you can see the non-zero bits are divided into several well-separated + * blocks. Hcub can find algorithms for those terms separately, which + * can then be shifted and added together. + * + * Dividing the input into 2, 3 or 4 blocks, Hcub can find solutions + * with 10, 9 or 8 adds, respectively, making a total of 11 for the + * whole number. + * + * Using just two large blocks, 0xC3910C8D << 31 in the high bits, + * and 0xB583EB in the low bits, produces as good an algorithm as any, + * and with one more small shift than alternatives. + * + * The high bits are a larger number and more work to compute, as well + * as needing one extra cycle to shift left 31 bits before the final + * addition, so they are the critical path for scheduling. The low bits + * can fit into the scheduling slots left over. + */ + + +/* + * This _ASSIGN(dst, src) macro performs "dst = src", but prevents GCC + * from inferring anything about the value assigned to "dest". + * + * This prevents it from mis-optimizing certain sequences. + * In particular, gcc is annoyingly eager to combine consecutive shifts. + * Given "x <<= 19; y += x; z += x << 1;", GCC will turn this into + * "y += x << 19; z += x << 20;" even though the latter sequence needs + * an additional instruction and temporary register. + * + * Because no actual assembly code is generated, this construct is + * usefully portable across all GCC platforms, and so can be test-compiled + * on non-PA systems. + * + * In two places, additional unused input dependencies are added. This + * forces GCC's scheduling so it does not rearrange instructions too much. + * Because the PA-8xxx is out of order, I'm not sure how much this matters, + * but why make it more difficult for the processor than necessary? + */ +#define _ASSIGN(dst, src, ...) asm("" : "=r" (dst) : "0" (src), ##__VA_ARGS__) + +/* + * Multiply by GOLDEN_RATIO_64 = 0x0x61C8864680B583EB using a heavily + * optimized shift-and-add sequence. + * + * Without the final shift, the multiply proper is 19 instructions, + * 10 cycles and uses only 4 temporaries. Whew! + * + * You are not expected to understand this. + */ +static __always_inline u32 __attribute_const__ +hash_64(u64 a, unsigned int bits) +{ + u64 b, c, d; + + /* + * Encourage GCC to move a dynamic shift to %sar early, + * thereby freeing up an additional temporary register. + */ + if (!__builtin_constant_p(bits)) + asm("" : "=q" (bits) : "0" (64 - bits)); + else + bits = 64 - bits; + + _ASSIGN(b, a*5); c = a << 13; + b = (b << 2) + a; _ASSIGN(d, a << 17); + a = b + (a << 1); c += d; + d = a << 10; _ASSIGN(a, a << 19); + d = a - d; _ASSIGN(a, a << 4, "X" (d)); + c += b; a += b; + d -= c; c += a << 1; + a += c << 3; _ASSIGN(b, b << (7+31), "X" (c), "X" (d)); + a <<= 31; b += d; + a += b; + return a >> bits; +} +#undef _ASSIGN /* We're a widely-used header file, so don't litter! */ + +#endif /* BITS_PER_LONG == 64 */ + +#endif /* _ASM_HASH_H */ From 49ea1480f433d7673715f471c10f05bb08d49205 Mon Sep 17 00:00:00 2001 From: Helge Deller Date: Tue, 2 Aug 2016 16:26:23 +0200 Subject: [PATCH 2/3] parisc: Add break statements to pdc_pat_io_pci_cfg_read() Dan Carpenter noticed that pdc_pat_io_pci_cfg_read() is problematic because it's missing some break statements so it copies 4 bytes regardless of whether you asked for only 1 or 2. Reported-by: Dan Carpenter Signed-off-by: Helge Deller --- arch/parisc/kernel/firmware.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) diff --git a/arch/parisc/kernel/firmware.c b/arch/parisc/kernel/firmware.c index 22395901d47b..e5d71905cad5 100644 --- a/arch/parisc/kernel/firmware.c +++ b/arch/parisc/kernel/firmware.c @@ -1354,9 +1354,9 @@ int pdc_pat_io_pci_cfg_read(unsigned long pci_addr, int pci_size, u32 *mem_addr) retval = mem_pdc_call(PDC_PAT_IO, PDC_PAT_IO_PCI_CONFIG_READ, __pa(pdc_result), pci_addr, pci_size); switch(pci_size) { - case 1: *(u8 *) mem_addr = (u8) pdc_result[0]; - case 2: *(u16 *)mem_addr = (u16) pdc_result[0]; - case 4: *(u32 *)mem_addr = (u32) pdc_result[0]; + case 1: *(u8 *) mem_addr = (u8) pdc_result[0]; break; + case 2: *(u16 *)mem_addr = (u16) pdc_result[0]; break; + case 4: *(u32 *)mem_addr = (u32) pdc_result[0]; break; } spin_unlock_irqrestore(&pdc_lock, flags); From a549c45a22de800e00045959b2d4f1b15b9f815a Mon Sep 17 00:00:00 2001 From: Amitoj Kaur Chawla Date: Tue, 14 Jun 2016 12:13:06 +0530 Subject: [PATCH 3/3] parisc: Change structure intialisation to C99 style in iomap.c Replace the in order struct initialisation style with explicit field style. Signed-off-by: Amitoj Kaur Chawla Signed-off-by: Helge Deller --- arch/parisc/lib/iomap.c | 64 ++++++++++++++++++++--------------------- 1 file changed, 32 insertions(+), 32 deletions(-) diff --git a/arch/parisc/lib/iomap.c b/arch/parisc/lib/iomap.c index fb8e10a4fb39..eaffbb90aa14 100644 --- a/arch/parisc/lib/iomap.c +++ b/arch/parisc/lib/iomap.c @@ -125,22 +125,22 @@ static void ioport_write32r(void __iomem *addr, const void *s, unsigned long n) } static const struct iomap_ops ioport_ops = { - ioport_read8, - ioport_read16, - ioport_read16, - ioport_read32, - ioport_read32, - ioport_write8, - ioport_write16, - ioport_write16, - ioport_write32, - ioport_write32, - ioport_read8r, - ioport_read16r, - ioport_read32r, - ioport_write8r, - ioport_write16r, - ioport_write32r, + .read8 = ioport_read8, + .read16 = ioport_read16, + .read16be = ioport_read16, + .read32 = ioport_read32, + .read32be = ioport_read32, + .write8 = ioport_write8, + .write16 = ioport_write16, + .write16be = ioport_write16, + .write32 = ioport_write32, + .write32be = ioport_write32, + .read8r = ioport_read8r, + .read16r = ioport_read16r, + .read32r = ioport_read32r, + .write8r = ioport_write8r, + .write16r = ioport_write16r, + .write32r = ioport_write32r, }; /* Legacy I/O memory ops */ @@ -244,22 +244,22 @@ static void iomem_write32r(void __iomem *addr, const void *s, unsigned long n) } static const struct iomap_ops iomem_ops = { - iomem_read8, - iomem_read16, - iomem_read16be, - iomem_read32, - iomem_read32be, - iomem_write8, - iomem_write16, - iomem_write16be, - iomem_write32, - iomem_write32be, - iomem_read8r, - iomem_read16r, - iomem_read32r, - iomem_write8r, - iomem_write16r, - iomem_write32r, + .read8 = iomem_read8, + .read16 = iomem_read16, + .read16be = iomem_read16be, + .read32 = iomem_read32, + .read32be = iomem_read32be, + .write8 = iomem_write8, + .write16 = iomem_write16, + .write16be = iomem_write16be, + .write32 = iomem_write32, + .write32be = iomem_write32be, + .read8r = iomem_read8r, + .read16r = iomem_read16r, + .read32r = iomem_read32r, + .write8r = iomem_write8r, + .write16r = iomem_write16r, + .write32r = iomem_write32r, }; static const struct iomap_ops *iomap_ops[8] = {