summaryrefslogtreecommitdiffstats
path: root/arch/alpha/math-emu/math.c
diff options
context:
space:
mode:
authorSang-Heon Jeon <ekffu200098@gmail.com>2026-08-24 01:18:14 +0900
committerJulia Lawall <Julia.Lawall@inria.fr>2026-08-23 22:12:11 +0200
commit2bd30b8dc75b823dd7bbef9d7bc68e1fd7ecee5c (patch)
treedcc4da1d99f17541c5590ac63c77da4bf372a4ab /arch/alpha/math-emu/math.c
downloadlinux-stable-2bd30b8dc75b823dd7bbef9d7bc68e1fd7ecee5c.tar.gz
linux-stable-2bd30b8dc75b823dd7bbef9d7bc68e1fd7ecee5c.zip
coccinelle: alloc_cast: drop removed allocatorsgrafted
- pci_alloc_consistent() and pci_zalloc_consistent() were removed by commit 7968778914e5 ("PCI: Remove the deprecated "pci-dma-compat.h" API") - kmem_alloc() was removed by commit f078d4ea8276 ("xfs: convert kmem_alloc() to kmalloc()") - kmem_zalloc() was removed by commit 10634530f7ba ("xfs: convert kmem_zalloc() to kzalloc()") - kmem_zone_alloc() and kmem_zone_zalloc() were removed by commit bae633a4a283 ("xfs: remove xfs_zone_{alloc,zalloc} helpers") So drop them from the rules. No functional change. Signed-off-by: Sang-Heon Jeon <ekffu200098@gmail.com> Signed-off-by: Julia Lawall <Julia.Lawall@inria.fr>
Diffstat (limited to 'arch/alpha/math-emu/math.c')
-rw-r--r--arch/alpha/math-emu/math.c466
1 files changed, 466 insertions, 0 deletions
diff --git a/arch/alpha/math-emu/math.c b/arch/alpha/math-emu/math.c
new file mode 100644
index 000000000..e3f2df372
--- /dev/null
+++ b/arch/alpha/math-emu/math.c
@@ -0,0 +1,466 @@
+// SPDX-License-Identifier: GPL-2.0-only
+#include <linux/module.h>
+#include <linux/types.h>
+#include <linux/kernel.h>
+#include <linux/sched.h>
+#include <asm/ptrace.h>
+#include <asm/fpu.h>
+
+#include <linux/uaccess.h>
+
+#include "sfp-util.h"
+#include <math-emu/soft-fp.h>
+#include <math-emu/single.h>
+#include <math-emu/double.h>
+
+#define OPC_PAL 0x00
+#define OPC_INTA 0x10
+#define OPC_INTL 0x11
+#define OPC_INTS 0x12
+#define OPC_INTM 0x13
+#define OPC_FLTC 0x14
+#define OPC_FLTV 0x15
+#define OPC_FLTI 0x16
+#define OPC_FLTL 0x17
+#define OPC_MISC 0x18
+#define OPC_JSR 0x1a
+
+#define FOP_SRC_S 0
+#define FOP_SRC_T 2
+#define FOP_SRC_Q 3
+
+#define FOP_FNC_ADDx 0
+#define FOP_FNC_CVTQL 0
+#define FOP_FNC_SUBx 1
+#define FOP_FNC_MULx 2
+#define FOP_FNC_DIVx 3
+#define FOP_FNC_CMPxUN 4
+#define FOP_FNC_CMPxEQ 5
+#define FOP_FNC_CMPxLT 6
+#define FOP_FNC_CMPxLE 7
+#define FOP_FNC_SQRTx 11
+#define FOP_FNC_CVTxS 12
+#define FOP_FNC_CVTxT 14
+#define FOP_FNC_CVTxQ 15
+
+#define MISC_TRAPB 0x0000
+#define MISC_EXCB 0x0400
+
+#ifdef MODULE
+
+MODULE_DESCRIPTION("FP Software completion module");
+MODULE_LICENSE("GPL v2");
+
+extern long (*alpha_fp_emul_imprecise)(struct pt_regs *, unsigned long);
+extern long (*alpha_fp_emul) (unsigned long pc, unsigned long summary);
+
+static long (*save_emul_imprecise)(struct pt_regs *, unsigned long);
+static long (*save_emul) (unsigned long pc, unsigned long summary);
+
+long do_alpha_fp_emul_imprecise(struct pt_regs *, unsigned long);
+long do_alpha_fp_emul(unsigned long, unsigned long);
+
+static int alpha_fp_emul_init_module(void)
+{
+ save_emul_imprecise = alpha_fp_emul_imprecise;
+ save_emul = alpha_fp_emul;
+ alpha_fp_emul_imprecise = do_alpha_fp_emul_imprecise;
+ alpha_fp_emul = do_alpha_fp_emul;
+ return 0;
+}
+module_init(alpha_fp_emul_init_module);
+
+static void alpha_fp_emul_cleanup_module(void)
+{
+ alpha_fp_emul_imprecise = save_emul_imprecise;
+ alpha_fp_emul = save_emul;
+}
+module_exit(alpha_fp_emul_cleanup_module);
+
+#undef alpha_fp_emul_imprecise
+#define alpha_fp_emul_imprecise do_alpha_fp_emul_imprecise
+#undef alpha_fp_emul
+#define alpha_fp_emul do_alpha_fp_emul
+
+#endif /* MODULE */
+
+
+/*
+ * Exception bits of the exception summary register (EXC_SUM). Bit 0 is the
+ * software completion bit; bits 1 through 5 report the exceptions the
+ * hardware attributed to the trapping instruction, and lie at the same
+ * positions as the corresponding IEEE_TRAP_ENABLE_* bits.
+ */
+#define EXC_SUM_INV (1UL << 1)
+#define EXC_SUM_DZE (1UL << 2)
+#define EXC_SUM_OVF (1UL << 3)
+#define EXC_SUM_UNF (1UL << 4)
+#define EXC_SUM_INE (1UL << 5)
+#define EXC_SUM_MASK (EXC_SUM_INV | EXC_SUM_DZE | EXC_SUM_OVF \
+ | EXC_SUM_UNF | EXC_SUM_INE)
+
+/*
+ * Emulate the floating point instruction at address PC. SUMMARY is the
+ * exception summary register the trap was delivered with. Returns -1 if the
+ * instruction to be emulated is illegal (such as with the opDEC trap), else
+ * the SI_CODE for a SIGFPE signal, else 0 if everything's ok.
+ *
+ * Notice that the kernel does not and cannot use FP regs. This is good
+ * because it means that instead of saving/restoring all fp regs, we simply
+ * stick the result of the operation into the appropriate register.
+ */
+long
+alpha_fp_emul (unsigned long pc, unsigned long summary)
+{
+ FP_DECL_EX;
+ FP_DECL_S(SA); FP_DECL_S(SB); FP_DECL_S(SR);
+ FP_DECL_D(DA); FP_DECL_D(DB); FP_DECL_D(DR);
+
+ unsigned long fa, fb, fc, func, mode, src;
+ unsigned long res, va, vb, vc, swcr, fpcr;
+ __u32 insn;
+ long si_code;
+
+ get_user(insn, (__u32 __user *)pc);
+ fc = (insn >> 0) & 0x1f; /* destination register */
+ fb = (insn >> 16) & 0x1f;
+ fa = (insn >> 21) & 0x1f;
+ func = (insn >> 5) & 0xf;
+ src = (insn >> 9) & 0x3;
+ mode = (insn >> 11) & 0x3;
+
+ fpcr = rdfpcr();
+ swcr = swcr_update_status(current_thread_info()->ieee_state, fpcr);
+
+ if (mode == 3) {
+ /* Dynamic -- get rounding mode from fpcr. */
+ mode = (fpcr >> FPCR_DYN_SHIFT) & 3;
+ }
+
+ switch (src) {
+ case FOP_SRC_S:
+ va = alpha_read_fp_reg_s(fa);
+ vb = alpha_read_fp_reg_s(fb);
+
+ FP_UNPACK_SP(SA, &va);
+ FP_UNPACK_SP(SB, &vb);
+
+ switch (func) {
+ case FOP_FNC_SUBx:
+ FP_SUB_S(SR, SA, SB);
+ goto pack_s;
+
+ case FOP_FNC_ADDx:
+ FP_ADD_S(SR, SA, SB);
+ goto pack_s;
+
+ case FOP_FNC_MULx:
+ FP_MUL_S(SR, SA, SB);
+ goto pack_s;
+
+ case FOP_FNC_DIVx:
+ FP_DIV_S(SR, SA, SB);
+ goto pack_s;
+
+ case FOP_FNC_SQRTx:
+ FP_SQRT_S(SR, SB);
+ goto pack_s;
+ }
+ goto bad_insn;
+
+ case FOP_SRC_T:
+ va = alpha_read_fp_reg(fa);
+ vb = alpha_read_fp_reg(fb);
+
+ if ((func & ~3) == FOP_FNC_CMPxUN) {
+ FP_UNPACK_RAW_DP(DA, &va);
+ FP_UNPACK_RAW_DP(DB, &vb);
+ if (!DA_e && !_FP_FRAC_ZEROP_1(DA)) {
+ FP_SET_EXCEPTION(FP_EX_DENORM);
+ if (FP_DENORM_ZERO)
+ _FP_FRAC_SET_1(DA, _FP_ZEROFRAC_1);
+ }
+ if (!DB_e && !_FP_FRAC_ZEROP_1(DB)) {
+ FP_SET_EXCEPTION(FP_EX_DENORM);
+ if (FP_DENORM_ZERO)
+ _FP_FRAC_SET_1(DB, _FP_ZEROFRAC_1);
+ }
+ FP_CMP_D(res, DA, DB, 3);
+ vc = 0x4000000000000000UL;
+ /* CMPTEQ, CMPTUN don't trap on QNaN,
+ while CMPTLT and CMPTLE do */
+ if (res == 3
+ && ((func & 3) >= 2
+ || FP_ISSIGNAN_D(DA)
+ || FP_ISSIGNAN_D(DB))) {
+ FP_SET_EXCEPTION(FP_EX_INVALID);
+ }
+ switch (func) {
+ case FOP_FNC_CMPxUN: if (res != 3) vc = 0; break;
+ case FOP_FNC_CMPxEQ: if (res) vc = 0; break;
+ case FOP_FNC_CMPxLT: if (res != -1) vc = 0; break;
+ case FOP_FNC_CMPxLE: if ((long)res > 0) vc = 0; break;
+ }
+ goto done_d;
+ }
+
+ FP_UNPACK_DP(DA, &va);
+ FP_UNPACK_DP(DB, &vb);
+
+ switch (func) {
+ case FOP_FNC_SUBx:
+ FP_SUB_D(DR, DA, DB);
+ goto pack_d;
+
+ case FOP_FNC_ADDx:
+ FP_ADD_D(DR, DA, DB);
+ goto pack_d;
+
+ case FOP_FNC_MULx:
+ FP_MUL_D(DR, DA, DB);
+ goto pack_d;
+
+ case FOP_FNC_DIVx:
+ FP_DIV_D(DR, DA, DB);
+ goto pack_d;
+
+ case FOP_FNC_SQRTx:
+ FP_SQRT_D(DR, DB);
+ goto pack_d;
+
+ case FOP_FNC_CVTxS:
+ /* It is irritating that DEC encoded CVTST with
+ SRC == T_floating. It is also interesting that
+ the bit used to tell the two apart is /U... */
+ if (insn & 0x2000) {
+ FP_CONV(S,D,1,1,SR,DB);
+ goto pack_s;
+ } else {
+ vb = alpha_read_fp_reg_s(fb);
+ FP_UNPACK_SP(SB, &vb);
+ DR_c = DB_c;
+ DR_s = DB_s;
+ DR_e = DB_e + (1024 - 128);
+ DR_f = SB_f << (52 - 23);
+ goto pack_d;
+ }
+
+ case FOP_FNC_CVTxQ:
+ if (DB_c == FP_CLS_NAN
+ && (_FP_FRAC_HIGH_RAW_D(DB) & _FP_QNANBIT_D)) {
+ /* AAHB Table B-2 says QNaN should not trigger INV */
+ vc = 0;
+ } else
+ FP_TO_INT_ROUND_D(vc, DB, 64, 2);
+ goto done_d;
+ }
+ goto bad_insn;
+
+ case FOP_SRC_Q:
+ vb = alpha_read_fp_reg(fb);
+
+ switch (func) {
+ case FOP_FNC_CVTQL:
+ /* Notice: We can get here only due to an integer
+ overflow. Such overflows are reported as invalid
+ ops. We return the result the hw would have
+ computed. */
+ vc = ((vb & 0xc0000000) << 32 | /* sign and msb */
+ (vb & 0x3fffffff) << 29); /* rest of the int */
+ FP_SET_EXCEPTION (FP_EX_INVALID);
+ goto done_d;
+
+ case FOP_FNC_CVTxS:
+ FP_FROM_INT_S(SR, ((long)vb), 64, long);
+ goto pack_s;
+
+ case FOP_FNC_CVTxT:
+ FP_FROM_INT_D(DR, ((long)vb), 64, long);
+ goto pack_d;
+ }
+ goto bad_insn;
+ }
+ goto bad_insn;
+
+pack_s:
+ FP_PACK_SP(&vc, SR);
+ if ((_fex & FP_EX_UNDERFLOW) && (swcr & IEEE_MAP_UMZ))
+ vc = 0;
+ alpha_write_fp_reg_s(fc, vc);
+ goto done;
+
+pack_d:
+ FP_PACK_DP(&vc, DR);
+ if ((_fex & FP_EX_UNDERFLOW) && (swcr & IEEE_MAP_UMZ))
+ vc = 0;
+done_d:
+ alpha_write_fp_reg(fc, vc);
+ goto done;
+
+ /*
+ * Take the appropriate action for each possible
+ * floating-point result:
+ *
+ * - Set the appropriate bits in the FPCR
+ * - If the specified exception is enabled in the FPCR,
+ * return. The caller (entArith) will dispatch
+ * the appropriate signal to the translated program.
+ *
+ * In addition, properly track the exception state in software
+ * as described in the Alpha Architecture Handbook section 4.7.7.3.
+ */
+done:
+ if (_fex) {
+ /* Record exceptions in software control word. */
+ swcr |= (_fex << IEEE_STATUS_TO_EXCSUM_SHIFT);
+ current_thread_info()->ieee_state
+ |= (_fex << IEEE_STATUS_TO_EXCSUM_SHIFT);
+ }
+
+ /*
+ * EV6 records exception status bits in the FPCR before delivering the
+ * software completion trap, and swcr_update_status() above merged them
+ * into SWCR. Some can be wrong for the instruction we just emulated:
+ * a CVTTS of a value exactly representable as a subnormal sets FPCR_UNF
+ * even though the result is exact. Clear the exceptions the trap
+ * reported but that soft-fp did not raise.
+ */
+ if (implver() == IMPLVER_EV6) {
+ unsigned long spurious = summary & EXC_SUM_MASK;
+
+ if (spurious & (EXC_SUM_UNF | EXC_SUM_OVF)) {
+ /*
+ * EXC_SUM reports only the underflow or overflow,
+ * but the hardware sets INE alongside it in the FPCR.
+ */
+ spurious |= EXC_SUM_INE;
+ } else if (!spurious) {
+ /*
+ * No exception reported, so this was a denormal
+ * operand trap, for which INE and UNF can be
+ * fabricated as well.
+ */
+ spurious = EXC_SUM_INE | EXC_SUM_UNF;
+ }
+
+ /*
+ * Never clear an exception software has confirmed. Every
+ * instruction that genuinely raises one traps for software
+ * completion and is recorded in ieee_state above, so a bit
+ * found there -- including one just set from _fex -- belongs
+ * to this or an earlier instruction and must survive.
+ */
+ spurious &= ~(current_thread_info()->ieee_state
+ >> IEEE_STATUS_TO_EXCSUM_SHIFT);
+
+ swcr &= ~(spurious << IEEE_STATUS_TO_EXCSUM_SHIFT);
+ }
+
+ /*
+ * Update hardware control register. This has to happen even when
+ * soft-fp raised nothing, to clear any fabricated bits.
+ */
+ fpcr &= (~FPCR_MASK | FPCR_DYN_MASK);
+ fpcr |= ieee_swcr_to_fpcr(swcr);
+ wrfpcr(fpcr);
+
+ if (_fex) {
+ /* Do we generate a signal? */
+ _fex = _fex & swcr & IEEE_TRAP_ENABLE_MASK;
+ si_code = 0;
+ if (_fex) {
+ if (_fex & IEEE_TRAP_ENABLE_DNO) si_code = FPE_FLTUND;
+ if (_fex & IEEE_TRAP_ENABLE_INE) si_code = FPE_FLTRES;
+ if (_fex & IEEE_TRAP_ENABLE_UNF) si_code = FPE_FLTUND;
+ if (_fex & IEEE_TRAP_ENABLE_OVF) si_code = FPE_FLTOVF;
+ if (_fex & IEEE_TRAP_ENABLE_DZE) si_code = FPE_FLTDIV;
+ if (_fex & IEEE_TRAP_ENABLE_INV) si_code = FPE_FLTINV;
+ }
+
+ return si_code;
+ }
+
+ /* We used to write the destination register here, but DEC FORTRAN
+ requires that the result *always* be written... so we do the write
+ immediately after the operations above. */
+
+ return 0;
+
+bad_insn:
+ printk(KERN_ERR "alpha_fp_emul: Invalid FP insn %#x at %#lx\n",
+ insn, pc);
+ return -1;
+}
+
+long
+alpha_fp_emul_imprecise (struct pt_regs *regs, unsigned long write_mask)
+{
+ unsigned long trigger_pc = regs->pc - 4;
+ unsigned long insn, opcode, rc, si_code = 0;
+
+ /*
+ * Turn off the bits corresponding to registers that are the
+ * target of instructions that set bits in the exception
+ * summary register. We have some slack doing this because a
+ * register that is the target of a trapping instruction can
+ * be written at most once in the trap shadow.
+ *
+ * Branches, jumps, TRAPBs, EXCBs and calls to PALcode all
+ * bound the trap shadow, so we need not look any further than
+ * up to the first occurrence of such an instruction.
+ */
+ while (write_mask) {
+ get_user(insn, (__u32 __user *)(trigger_pc));
+ opcode = insn >> 26;
+ rc = insn & 0x1f;
+
+ switch (opcode) {
+ case OPC_PAL:
+ case OPC_JSR:
+ case 0x30 ... 0x3f: /* branches */
+ goto egress;
+
+ case OPC_MISC:
+ switch (insn & 0xffff) {
+ case MISC_TRAPB:
+ case MISC_EXCB:
+ goto egress;
+
+ default:
+ break;
+ }
+ break;
+
+ case OPC_INTA:
+ case OPC_INTL:
+ case OPC_INTS:
+ case OPC_INTM:
+ write_mask &= ~(1UL << rc);
+ break;
+
+ case OPC_FLTC:
+ case OPC_FLTV:
+ case OPC_FLTI:
+ case OPC_FLTL:
+ write_mask &= ~(1UL << (rc + 32));
+ break;
+ }
+ if (!write_mask) {
+ /*
+ * Re-execute insns in the trap-shadow. Pass no
+ * exception summary: it describes the trap, which
+ * was taken anywhere in the shadow, and so is not
+ * attribution for this instruction. Nothing is
+ * lost, since only EV6 -- which traps precisely and
+ * never comes this way -- needs it.
+ */
+ regs->pc = trigger_pc + 4;
+ si_code = alpha_fp_emul(trigger_pc, 0);
+ goto egress;
+ }
+ trigger_pc -= 4;
+ }
+
+egress:
+ return si_code;
+}