linux_dsm_epyc7002/arch/powerpc/kernel/fpu.S

246 lines
6.0 KiB
ArmAsm
Raw Normal View History

/*
* FPU support code, moved here from head.S so that it can be used
* by chips which use other head-whatever.S files.
*
* Copyright (C) 1995-1996 Gary Thomas (gdt@linuxppc.org)
* Copyright (C) 1996 Cort Dougan <cort@cs.nmt.edu>
* Copyright (C) 1996 Paul Mackerras.
* Copyright (C) 1997 Dan Malek (dmalek@jlc.net).
*
* This program is free software; you can redistribute it and/or
* modify it under the terms of the GNU General Public License
* as published by the Free Software Foundation; either version
* 2 of the License, or (at your option) any later version.
*
*/
#include <asm/reg.h>
#include <asm/page.h>
#include <asm/mmu.h>
#include <asm/pgtable.h>
#include <asm/cputable.h>
#include <asm/cache.h>
#include <asm/thread_info.h>
#include <asm/ppc_asm.h>
#include <asm/asm-offsets.h>
#include <asm/ptrace.h>
#ifdef CONFIG_VSX
#define __REST_32FPVSRS(n,c,base) \
BEGIN_FTR_SECTION \
b 2f; \
END_FTR_SECTION_IFSET(CPU_FTR_VSX); \
REST_32FPRS(n,base); \
b 3f; \
2: REST_32VSRS(n,c,base); \
3:
#define __REST_32FPVSRS_TRANSACT(n,c,base) \
BEGIN_FTR_SECTION \
b 2f; \
END_FTR_SECTION_IFSET(CPU_FTR_VSX); \
REST_32FPRS_TRANSACT(n,base); \
b 3f; \
2: REST_32VSRS_TRANSACT(n,c,base); \
3:
#define __SAVE_32FPVSRS(n,c,base) \
BEGIN_FTR_SECTION \
b 2f; \
END_FTR_SECTION_IFSET(CPU_FTR_VSX); \
SAVE_32FPRS(n,base); \
b 3f; \
2: SAVE_32VSRS(n,c,base); \
3:
#else
#define __REST_32FPVSRS(n,b,base) REST_32FPRS(n, base)
#define __REST_32FPVSRS_TRANSACT(n,b,base) REST_32FPRS(n, base)
#define __SAVE_32FPVSRS(n,b,base) SAVE_32FPRS(n, base)
#endif
#define REST_32FPVSRS(n,c,base) __REST_32FPVSRS(n,__REG_##c,__REG_##base)
#define REST_32FPVSRS_TRANSACT(n,c,base) \
__REST_32FPVSRS_TRANSACT(n,__REG_##c,__REG_##base)
#define SAVE_32FPVSRS(n,c,base) __SAVE_32FPVSRS(n,__REG_##c,__REG_##base)
#ifdef CONFIG_PPC_TRANSACTIONAL_MEM
/*
* Wrapper to call load_up_fpu from C.
* void do_load_up_fpu(struct pt_regs *regs);
*/
_GLOBAL(do_load_up_fpu)
mflr r0
std r0, 16(r1)
stdu r1, -112(r1)
subi r6, r3, STACK_FRAME_OVERHEAD
/* load_up_fpu expects r12=MSR, r13=PACA, and returns
* with r12 = new MSR.
*/
ld r12,_MSR(r6)
GET_PACA(r13)
bl load_up_fpu
std r12,_MSR(r6)
ld r0, 112+16(r1)
addi r1, r1, 112
mtlr r0
blr
/* void do_load_up_transact_fpu(struct thread_struct *thread)
*
* This is similar to load_up_fpu but for the transactional version of the FP
* register set. It doesn't mess with the task MSR or valid flags.
* Furthermore, we don't do lazy FP with TM currently.
*/
_GLOBAL(do_load_up_transact_fpu)
mfmsr r6
ori r5,r6,MSR_FP
#ifdef CONFIG_VSX
BEGIN_FTR_SECTION
oris r5,r5,MSR_VSX@h
END_FTR_SECTION_IFSET(CPU_FTR_VSX)
#endif
SYNC
MTMSRD(r5)
lfd fr0,THREAD_TRANSACT_FPSCR(r3)
MTFSF_L(fr0)
REST_32FPVSRS_TRANSACT(0, R4, R3)
/* FP/VSX off again */
MTMSRD(r6)
SYNC
blr
#endif /* CONFIG_PPC_TRANSACTIONAL_MEM */
/*
* This task wants to use the FPU now.
* On UP, disable FP for the task which had the FPU previously,
* and save its floating-point registers in its thread_struct.
* Load up this task's FP registers from its thread_struct,
* enable the FPU for the current task and return to the task.
*/
_GLOBAL(load_up_fpu)
mfmsr r5
ori r5,r5,MSR_FP
#ifdef CONFIG_VSX
BEGIN_FTR_SECTION
oris r5,r5,MSR_VSX@h
END_FTR_SECTION_IFSET(CPU_FTR_VSX)
#endif
SYNC
MTMSRD(r5) /* enable use of fpu now */
isync
/*
* For SMP, we don't do lazy FPU switching because it just gets too
* horrendously complex, especially when a task switches from one CPU
* to another. Instead we call giveup_fpu in switch_to.
*/
#ifndef CONFIG_SMP
LOAD_REG_ADDRBASE(r3, last_task_used_math)
toreal(r3)
PPC_LL r4,ADDROFF(last_task_used_math)(r3)
PPC_LCMPI 0,r4,0
beq 1f
toreal(r4)
addi r4,r4,THREAD /* want last_task_used_math->thread */
SAVE_32FPVSRS(0, R5, R4)
mffs fr0
[PATCH] powerpc: Fix handling of fpscr on 64-bit The recent merge of fpu.S broken the handling of fpscr for ARCH=powerpc and CONFIG_PPC64=y. FP registers could be corrupted, leading to strange random application crashes. The confusion arises, because the thread_struct has (and requires) a 64-bit area to save the fpscr, because we use load/store double instructions to get it in to/out of the FPU. However, only the low 32-bits are actually used, so we want to treat it as a 32-bit quantity when manipulating its bits to avoid extra load/stores on 32-bit. This patch replaces the current definition with a structure of two 32-bit quantities (pad and val), to clarify things as much as is possible. The 'val' field is used when manipulating bits, the structure itself is used when obtaining the address for loading/unloading the value from the FPU. While we're at it, consolidate the 4 (!) almost identical versions of cvt_fd() and cvt_df() (arch/ppc/kernel/misc.S, arch/ppc64/kernel/misc.S, arch/powerpc/kernel/misc_32.S, arch/powerpc/kernel/misc_64.S) into a single version in fpu.S. The new version takes a pointer to thread_struct and applies the correct offset itself, rather than a pointer to the fpscr field itself, again to avoid confusion as to which is the correct field to use. Finally, this patch makes ARCH=ppc64 also use the consolidated fpu.S code, which it previously did not. Built for G5 (ARCH=ppc64 and ARCH=powerpc), 32-bit powermac (ARCH=ppc and ARCH=powerpc) and Walnut (ARCH=ppc, CONFIG_MATH_EMULATION=y). Booted on G5 (ARCH=powerpc) and things which previously fell over no longer do. Signed-off-by: David Gibson <dwg@au1.ibm.com> Signed-off-by: Paul Mackerras <paulus@samba.org>
2005-10-27 13:27:25 +07:00
stfd fr0,THREAD_FPSCR(r4)
PPC_LL r5,PT_REGS(r4)
toreal(r5)
PPC_LL r4,_MSR-STACK_FRAME_OVERHEAD(r5)
li r10,MSR_FP|MSR_FE0|MSR_FE1
andc r4,r4,r10 /* disable FP for previous task */
PPC_STL r4,_MSR-STACK_FRAME_OVERHEAD(r5)
1:
#endif /* CONFIG_SMP */
/* enable use of FP after return */
#ifdef CONFIG_PPC32
mfspr r5,SPRN_SPRG_THREAD /* current task's THREAD (phys) */
lwz r4,THREAD_FPEXC_MODE(r5)
ori r9,r9,MSR_FP /* enable FP for current */
or r9,r9,r4
#else
ld r4,PACACURRENT(r13)
addi r5,r4,THREAD /* Get THREAD */
lwz r4,THREAD_FPEXC_MODE(r5)
ori r12,r12,MSR_FP
or r12,r12,r4
std r12,_MSR(r1)
#endif
[PATCH] powerpc: Fix handling of fpscr on 64-bit The recent merge of fpu.S broken the handling of fpscr for ARCH=powerpc and CONFIG_PPC64=y. FP registers could be corrupted, leading to strange random application crashes. The confusion arises, because the thread_struct has (and requires) a 64-bit area to save the fpscr, because we use load/store double instructions to get it in to/out of the FPU. However, only the low 32-bits are actually used, so we want to treat it as a 32-bit quantity when manipulating its bits to avoid extra load/stores on 32-bit. This patch replaces the current definition with a structure of two 32-bit quantities (pad and val), to clarify things as much as is possible. The 'val' field is used when manipulating bits, the structure itself is used when obtaining the address for loading/unloading the value from the FPU. While we're at it, consolidate the 4 (!) almost identical versions of cvt_fd() and cvt_df() (arch/ppc/kernel/misc.S, arch/ppc64/kernel/misc.S, arch/powerpc/kernel/misc_32.S, arch/powerpc/kernel/misc_64.S) into a single version in fpu.S. The new version takes a pointer to thread_struct and applies the correct offset itself, rather than a pointer to the fpscr field itself, again to avoid confusion as to which is the correct field to use. Finally, this patch makes ARCH=ppc64 also use the consolidated fpu.S code, which it previously did not. Built for G5 (ARCH=ppc64 and ARCH=powerpc), 32-bit powermac (ARCH=ppc and ARCH=powerpc) and Walnut (ARCH=ppc, CONFIG_MATH_EMULATION=y). Booted on G5 (ARCH=powerpc) and things which previously fell over no longer do. Signed-off-by: David Gibson <dwg@au1.ibm.com> Signed-off-by: Paul Mackerras <paulus@samba.org>
2005-10-27 13:27:25 +07:00
lfd fr0,THREAD_FPSCR(r5)
MTFSF_L(fr0)
REST_32FPVSRS(0, R4, R5)
#ifndef CONFIG_SMP
subi r4,r5,THREAD
fromreal(r4)
PPC_STL r4,ADDROFF(last_task_used_math)(r3)
#endif /* CONFIG_SMP */
/* restore registers and return */
/* we haven't used ctr or xer or lr */
blr
/*
* giveup_fpu(tsk)
* Disable FP for the task given as the argument,
* and save the floating-point registers in its thread_struct.
* Enables the FPU for use in the kernel on return.
*/
_GLOBAL(giveup_fpu)
mfmsr r5
ori r5,r5,MSR_FP
#ifdef CONFIG_VSX
BEGIN_FTR_SECTION
oris r5,r5,MSR_VSX@h
END_FTR_SECTION_IFSET(CPU_FTR_VSX)
#endif
SYNC_601
ISYNC_601
MTMSRD(r5) /* enable use of fpu now */
SYNC_601
isync
PPC_LCMPI 0,r3,0
beqlr- /* if no previous owner, done */
addi r3,r3,THREAD /* want THREAD of task */
PPC_LL r5,PT_REGS(r3)
PPC_LCMPI 0,r5,0
SAVE_32FPVSRS(0, R4 ,R3)
mffs fr0
[PATCH] powerpc: Fix handling of fpscr on 64-bit The recent merge of fpu.S broken the handling of fpscr for ARCH=powerpc and CONFIG_PPC64=y. FP registers could be corrupted, leading to strange random application crashes. The confusion arises, because the thread_struct has (and requires) a 64-bit area to save the fpscr, because we use load/store double instructions to get it in to/out of the FPU. However, only the low 32-bits are actually used, so we want to treat it as a 32-bit quantity when manipulating its bits to avoid extra load/stores on 32-bit. This patch replaces the current definition with a structure of two 32-bit quantities (pad and val), to clarify things as much as is possible. The 'val' field is used when manipulating bits, the structure itself is used when obtaining the address for loading/unloading the value from the FPU. While we're at it, consolidate the 4 (!) almost identical versions of cvt_fd() and cvt_df() (arch/ppc/kernel/misc.S, arch/ppc64/kernel/misc.S, arch/powerpc/kernel/misc_32.S, arch/powerpc/kernel/misc_64.S) into a single version in fpu.S. The new version takes a pointer to thread_struct and applies the correct offset itself, rather than a pointer to the fpscr field itself, again to avoid confusion as to which is the correct field to use. Finally, this patch makes ARCH=ppc64 also use the consolidated fpu.S code, which it previously did not. Built for G5 (ARCH=ppc64 and ARCH=powerpc), 32-bit powermac (ARCH=ppc and ARCH=powerpc) and Walnut (ARCH=ppc, CONFIG_MATH_EMULATION=y). Booted on G5 (ARCH=powerpc) and things which previously fell over no longer do. Signed-off-by: David Gibson <dwg@au1.ibm.com> Signed-off-by: Paul Mackerras <paulus@samba.org>
2005-10-27 13:27:25 +07:00
stfd fr0,THREAD_FPSCR(r3)
beq 1f
PPC_LL r4,_MSR-STACK_FRAME_OVERHEAD(r5)
li r3,MSR_FP|MSR_FE0|MSR_FE1
#ifdef CONFIG_VSX
BEGIN_FTR_SECTION
oris r3,r3,MSR_VSX@h
END_FTR_SECTION_IFSET(CPU_FTR_VSX)
#endif
andc r4,r4,r3 /* disable FP for previous task */
PPC_STL r4,_MSR-STACK_FRAME_OVERHEAD(r5)
1:
#ifndef CONFIG_SMP
li r5,0
LOAD_REG_ADDRBASE(r4,last_task_used_math)
PPC_STL r5,ADDROFF(last_task_used_math)(r4)
#endif /* CONFIG_SMP */
blr
[PATCH] powerpc: Fix handling of fpscr on 64-bit The recent merge of fpu.S broken the handling of fpscr for ARCH=powerpc and CONFIG_PPC64=y. FP registers could be corrupted, leading to strange random application crashes. The confusion arises, because the thread_struct has (and requires) a 64-bit area to save the fpscr, because we use load/store double instructions to get it in to/out of the FPU. However, only the low 32-bits are actually used, so we want to treat it as a 32-bit quantity when manipulating its bits to avoid extra load/stores on 32-bit. This patch replaces the current definition with a structure of two 32-bit quantities (pad and val), to clarify things as much as is possible. The 'val' field is used when manipulating bits, the structure itself is used when obtaining the address for loading/unloading the value from the FPU. While we're at it, consolidate the 4 (!) almost identical versions of cvt_fd() and cvt_df() (arch/ppc/kernel/misc.S, arch/ppc64/kernel/misc.S, arch/powerpc/kernel/misc_32.S, arch/powerpc/kernel/misc_64.S) into a single version in fpu.S. The new version takes a pointer to thread_struct and applies the correct offset itself, rather than a pointer to the fpscr field itself, again to avoid confusion as to which is the correct field to use. Finally, this patch makes ARCH=ppc64 also use the consolidated fpu.S code, which it previously did not. Built for G5 (ARCH=ppc64 and ARCH=powerpc), 32-bit powermac (ARCH=ppc and ARCH=powerpc) and Walnut (ARCH=ppc, CONFIG_MATH_EMULATION=y). Booted on G5 (ARCH=powerpc) and things which previously fell over no longer do. Signed-off-by: David Gibson <dwg@au1.ibm.com> Signed-off-by: Paul Mackerras <paulus@samba.org>
2005-10-27 13:27:25 +07:00
/*
* These are used in the alignment trap handler when emulating
* single-precision loads and stores.
*/
_GLOBAL(cvt_fd)
lfs 0,0(r3)
stfd 0,0(r4)
blr
_GLOBAL(cvt_df)
lfd 0,0(r3)
stfs 0,0(r4)
blr