Minki/linux
1
0
Fork 1
mirror of https://github.com/torvalds/linux.git synced 2024-11-01 01:31:44 +00:00
Code Issues Packages Projects Releases Wiki Activity
002ea9eefe
linux/arch/arm/vfp/vfphw.S

310 lines
8.2 KiB
ArmAsm
Raw Normal View History

Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-16 22:20:36 +00:00
/*
* linux/arch/arm/vfp/vfphw.S
*
* Copyright (C) 2004 ARM Limited.
* Written by Deep Blue Solutions Limited.
*
* This program is free software; you can redistribute it and/or modify
* it under the terms of the GNU General Public License version 2 as
* published by the Free Software Foundation.
*
* This code is called from the kernel's undefined instruction trap.
* r9 holds the return address for successful handling.
* lr holds the return address for unrecognised instructions.
* r10 points at the start of the private FP workspace in the thread structure
* sp points to a struct pt_regs (as defined in include/asm/proc/ptrace.h)
*/
#include <asm/thread_info.h>
#include <asm/vfpmacros.h>
#include "../kernel/entry-header.S"
.macro DBGSTR, str
#ifdef DEBUG
stmfd sp!, {r0-r3, ip, lr}
add r0, pc, #4
bl printk
b 1f
.asciz "<7>VFP: \str\n"
.balign 4
1: ldmfd sp!, {r0-r3, ip, lr}
#endif
.endm
.macro DBGSTR1, str, arg
#ifdef DEBUG
stmfd sp!, {r0-r3, ip, lr}
mov r1, \arg
add r0, pc, #4
bl printk
b 1f
.asciz "<7>VFP: \str\n"
.balign 4
1: ldmfd sp!, {r0-r3, ip, lr}
#endif
.endm
.macro DBGSTR3, str, arg1, arg2, arg3
#ifdef DEBUG
stmfd sp!, {r0-r3, ip, lr}
mov r3, \arg3
mov r2, \arg2
mov r1, \arg1
add r0, pc, #4
bl printk
b 1f
.asciz "<7>VFP: \str\n"
.balign 4
1: ldmfd sp!, {r0-r3, ip, lr}
#endif
.endm
@ VFP hardware support entry point.
@
@ r0 = faulted instruction
@ r2 = faulted PC+4
@ r9 = successful return
@ r10 = vfp_state union
[ARM] 4111/1: Allow VFP to work with thread migration on SMP The current lazy saving of the VFP registers is no longer possible with thread migration on SMP. This patch implements a per-CPU vfp-state pointer and the saving of the VFP registers at every context switch. The registers restoring is still performed in a lazy way. Signed-off-by: Catalin Marinas <catalin.marinas@arm.com> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2007-01-24 17:47:08 +00:00
@ r11 = CPU number
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-16 22:20:36 +00:00
@ lr = failure return
[ARM] 5227/1: Add the ENDPROC declarations to the .S files This declaration specifies the "function" type and size for various assembly functions, mainly needed for generating the correct branch instructions in Thumb-2. Signed-off-by: Catalin Marinas <catalin.marinas@arm.com> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2008-08-28 10:22:32 +00:00
ENTRY(vfp_support_entry)
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-16 22:20:36 +00:00
DBGSTR3 "instr %08x pc %08x state %p", r0, r2, r10
VFPFMRX r1, FPEXC @ Is the VFP enabled?
DBGSTR1 "fpexc %08x", r1
[ARM] vfp: make fpexc bit names less verbose Use the fpexc abbreviated names instead of long verbose names for fpexc bits. Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2007-07-18 08:37:10 +00:00
tst r1, #FPEXC_EN
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-16 22:20:36 +00:00
bne look_for_VFP_exceptions @ VFP is already enabled
DBGSTR1 "enable %x", r10
ARM: vfp: rename last_VFP_context to vfp_current_hw_state Rename the slightly confusing 'last_VFP_context' variable to be more descriptive of what it actually is. This variable stores a pointer to the current owner's vfpstate structure for the context held in the VFP hardware. Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2011-07-09 12:44:04 +00:00
ldr r3, vfp_current_hw_state_address
[ARM] vfp: make fpexc bit names less verbose Use the fpexc abbreviated names instead of long verbose names for fpexc bits. Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2007-07-18 08:37:10 +00:00
orr r1, r1, #FPEXC_EN @ user FPEXC has the enable bit set
ARM: vfp: rename last_VFP_context to vfp_current_hw_state Rename the slightly confusing 'last_VFP_context' variable to be more descriptive of what it actually is. This variable stores a pointer to the current owner's vfpstate structure for the context held in the VFP hardware. Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2011-07-09 12:44:04 +00:00
ldr r4, [r3, r11, lsl #2] @ vfp_current_hw_state pointer
[ARM] vfp: make fpexc bit names less verbose Use the fpexc abbreviated names instead of long verbose names for fpexc bits. Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2007-07-18 08:37:10 +00:00
bic r5, r1, #FPEXC_EX @ make sure exceptions are disabled
ARM: vfp: rename check_exception to vfp_hw_state_valid Rename this branch to more accurately reflect why its taken, rather than what the following code does. It is the only caller of this code. This helps to clarify following changes, yet this change results in no actual code change. Document the VFP hardware state at the target of this branch. Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2011-07-09 13:24:36 +00:00
cmp r4, r10 @ this thread owns the hw context?
ARM: vfp: fix a hole in VFP thread migration Fix a hole in the VFP thread migration. Lets define two threads. Thread 1, we'll call 'interesting_thread' which is a thread which is running on CPU0, using VFP (so vfp_current_hw_state[0] = &interesting_thread->vfpstate) and gets migrated off to CPU1, where it continues execution of VFP instructions. Thread 2, we'll call 'new_cpu0_thread' which is the thread which takes over on CPU0. This has also been using VFP, and last used VFP on CPU0, but doesn't use it again. The following code will be executed twice: cpu = thread->cpu; /* * On SMP, if VFP is enabled, save the old state in * case the thread migrates to a different CPU. The * restoring is done lazily. */ if ((fpexc & FPEXC_EN) && vfp_current_hw_state[cpu]) { vfp_save_state(vfp_current_hw_state[cpu], fpexc); vfp_current_hw_state[cpu]->hard.cpu = cpu; } /* * Thread migration, just force the reloading of the * state on the new CPU in case the VFP registers * contain stale data. */ if (thread->vfpstate.hard.cpu != cpu) vfp_current_hw_state[cpu] = NULL; The first execution will be on CPU0 to switch away from 'interesting_thread'. interesting_thread->cpu will be 0. So, vfp_current_hw_state[0] points at interesting_thread->vfpstate. The hardware state will be saved, along with the CPU number (0) that it was executing on. 'thread' will be 'new_cpu0_thread' with new_cpu0_thread->cpu = 0. Also, because it was executing on CPU0, new_cpu0_thread->vfpstate.hard.cpu = 0, and so the thread migration check is not triggered. This means that vfp_current_hw_state[0] remains pointing at interesting_thread. The second execution will be on CPU1 to switch _to_ 'interesting_thread'. So, 'thread' will be 'interesting_thread' and interesting_thread->cpu now will be 1. The previous thread executing on CPU1 is not relevant to this so we shall ignore that. We get to the thread migration check. Here, we discover that interesting_thread->vfpstate.hard.cpu = 0, yet interesting_thread->cpu is now 1, indicating thread migration. We set vfp_current_hw_state[1] to NULL. So, at this point vfp_current_hw_state[] contains the following: [0] = &interesting_thread->vfpstate [1] = NULL Our interesting thread now executes a VFP instruction, takes a fault which loads the state into the VFP hardware. Now, through the assembly we now have: [0] = &interesting_thread->vfpstate [1] = &interesting_thread->vfpstate CPU1 stops due to ptrace (and so saves its VFP state) using the thread switch code above), and CPU0 calls vfp_sync_hwstate(). if (vfp_current_hw_state[cpu] == &thread->vfpstate) { vfp_save_state(&thread->vfpstate, fpexc | FPEXC_EN); BANG, we corrupt interesting_thread's VFP state by overwriting the more up-to-date state saved by CPU1 with the old VFP state from CPU0. Fix this by ensuring that we have sane semantics for the various state describing variables: 1. vfp_current_hw_state[] points to the current owner of the context information stored in each CPUs hardware, or NULL if that state information is invalid. 2. thread->vfpstate.hard.cpu always contains the most recent CPU number which the state was loaded into or NR_CPUS if no CPU owns the state. So, for a particular CPU to be a valid owner of the VFP state for a particular thread t, two things must be true: vfp_current_hw_state[cpu] == &t->vfpstate && t->vfpstate.hard.cpu == cpu. and that is valid from the moment a CPU loads the saved VFP context into the hardware. This gives clear and consistent semantics to interpreting these variables. This patch also fixes thread copying, ensuring that t->vfpstate.hard.cpu is invalidated, otherwise CPU0 may believe it was the last owner. The hole can happen thus: - thread1 runs on CPU2 using VFP, migrates to CPU3, exits and thread_info freed. - New thread allocated from a previously running thread on CPU2, reusing memory for thread1 and copying vfp.hard.cpu. At this point, the following are true: new_thread1->vfpstate.hard.cpu == 2 &new_thread1->vfpstate == vfp_current_hw_state[2] Lastly, this also addresses thread flushing in a similar way to thread copying. Hole is: - thread runs on CPU0, using VFP, migrates to CPU1 but does not use VFP. - thread calls execve(), so thread flush happens, leaving vfp_current_hw_state[0] intact. This vfpstate is memset to 0 causing thread->vfpstate.hard.cpu = 0. - thread migrates back to CPU0 before using VFP. At this point, the following are true: thread->vfpstate.hard.cpu == 0 &thread->vfpstate == vfp_current_hw_state[0] Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2011-07-09 15:09:43 +00:00
#ifndef CONFIG_SMP
@ For UP, checking that this thread owns the hw context is
@ sufficient to determine that the hardware state is valid.
ARM: vfp: rename check_exception to vfp_hw_state_valid Rename this branch to more accurately reflect why its taken, rather than what the following code does. It is the only caller of this code. This helps to clarify following changes, yet this change results in no actual code change. Document the VFP hardware state at the target of this branch. Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2011-07-09 13:24:36 +00:00
beq vfp_hw_state_valid
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-16 22:20:36 +00:00
ARM: vfp: fix a hole in VFP thread migration Fix a hole in the VFP thread migration. Lets define two threads. Thread 1, we'll call 'interesting_thread' which is a thread which is running on CPU0, using VFP (so vfp_current_hw_state[0] = &interesting_thread->vfpstate) and gets migrated off to CPU1, where it continues execution of VFP instructions. Thread 2, we'll call 'new_cpu0_thread' which is the thread which takes over on CPU0. This has also been using VFP, and last used VFP on CPU0, but doesn't use it again. The following code will be executed twice: cpu = thread->cpu; /* * On SMP, if VFP is enabled, save the old state in * case the thread migrates to a different CPU. The * restoring is done lazily. */ if ((fpexc & FPEXC_EN) && vfp_current_hw_state[cpu]) { vfp_save_state(vfp_current_hw_state[cpu], fpexc); vfp_current_hw_state[cpu]->hard.cpu = cpu; } /* * Thread migration, just force the reloading of the * state on the new CPU in case the VFP registers * contain stale data. */ if (thread->vfpstate.hard.cpu != cpu) vfp_current_hw_state[cpu] = NULL; The first execution will be on CPU0 to switch away from 'interesting_thread'. interesting_thread->cpu will be 0. So, vfp_current_hw_state[0] points at interesting_thread->vfpstate. The hardware state will be saved, along with the CPU number (0) that it was executing on. 'thread' will be 'new_cpu0_thread' with new_cpu0_thread->cpu = 0. Also, because it was executing on CPU0, new_cpu0_thread->vfpstate.hard.cpu = 0, and so the thread migration check is not triggered. This means that vfp_current_hw_state[0] remains pointing at interesting_thread. The second execution will be on CPU1 to switch _to_ 'interesting_thread'. So, 'thread' will be 'interesting_thread' and interesting_thread->cpu now will be 1. The previous thread executing on CPU1 is not relevant to this so we shall ignore that. We get to the thread migration check. Here, we discover that interesting_thread->vfpstate.hard.cpu = 0, yet interesting_thread->cpu is now 1, indicating thread migration. We set vfp_current_hw_state[1] to NULL. So, at this point vfp_current_hw_state[] contains the following: [0] = &interesting_thread->vfpstate [1] = NULL Our interesting thread now executes a VFP instruction, takes a fault which loads the state into the VFP hardware. Now, through the assembly we now have: [0] = &interesting_thread->vfpstate [1] = &interesting_thread->vfpstate CPU1 stops due to ptrace (and so saves its VFP state) using the thread switch code above), and CPU0 calls vfp_sync_hwstate(). if (vfp_current_hw_state[cpu] == &thread->vfpstate) { vfp_save_state(&thread->vfpstate, fpexc | FPEXC_EN); BANG, we corrupt interesting_thread's VFP state by overwriting the more up-to-date state saved by CPU1 with the old VFP state from CPU0. Fix this by ensuring that we have sane semantics for the various state describing variables: 1. vfp_current_hw_state[] points to the current owner of the context information stored in each CPUs hardware, or NULL if that state information is invalid. 2. thread->vfpstate.hard.cpu always contains the most recent CPU number which the state was loaded into or NR_CPUS if no CPU owns the state. So, for a particular CPU to be a valid owner of the VFP state for a particular thread t, two things must be true: vfp_current_hw_state[cpu] == &t->vfpstate && t->vfpstate.hard.cpu == cpu. and that is valid from the moment a CPU loads the saved VFP context into the hardware. This gives clear and consistent semantics to interpreting these variables. This patch also fixes thread copying, ensuring that t->vfpstate.hard.cpu is invalidated, otherwise CPU0 may believe it was the last owner. The hole can happen thus: - thread1 runs on CPU2 using VFP, migrates to CPU3, exits and thread_info freed. - New thread allocated from a previously running thread on CPU2, reusing memory for thread1 and copying vfp.hard.cpu. At this point, the following are true: new_thread1->vfpstate.hard.cpu == 2 &new_thread1->vfpstate == vfp_current_hw_state[2] Lastly, this also addresses thread flushing in a similar way to thread copying. Hole is: - thread runs on CPU0, using VFP, migrates to CPU1 but does not use VFP. - thread calls execve(), so thread flush happens, leaving vfp_current_hw_state[0] intact. This vfpstate is memset to 0 causing thread->vfpstate.hard.cpu = 0. - thread migrates back to CPU0 before using VFP. At this point, the following are true: thread->vfpstate.hard.cpu == 0 &thread->vfpstate == vfp_current_hw_state[0] Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2011-07-09 15:09:43 +00:00
@ On UP, we lazily save the VFP context. As a different
@ thread wants ownership of the VFP hardware, save the old
@ state if there was a previous (valid) owner.
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-16 22:20:36 +00:00
VFPFMXR FPEXC, r5 @ enable VFP, disable any pending
@ exceptions, so we can get at the
@ rest of it
DBGSTR1 "save old state %p", r4
ARM: vfp: fix a hole in VFP thread migration Fix a hole in the VFP thread migration. Lets define two threads. Thread 1, we'll call 'interesting_thread' which is a thread which is running on CPU0, using VFP (so vfp_current_hw_state[0] = &interesting_thread->vfpstate) and gets migrated off to CPU1, where it continues execution of VFP instructions. Thread 2, we'll call 'new_cpu0_thread' which is the thread which takes over on CPU0. This has also been using VFP, and last used VFP on CPU0, but doesn't use it again. The following code will be executed twice: cpu = thread->cpu; /* * On SMP, if VFP is enabled, save the old state in * case the thread migrates to a different CPU. The * restoring is done lazily. */ if ((fpexc & FPEXC_EN) && vfp_current_hw_state[cpu]) { vfp_save_state(vfp_current_hw_state[cpu], fpexc); vfp_current_hw_state[cpu]->hard.cpu = cpu; } /* * Thread migration, just force the reloading of the * state on the new CPU in case the VFP registers * contain stale data. */ if (thread->vfpstate.hard.cpu != cpu) vfp_current_hw_state[cpu] = NULL; The first execution will be on CPU0 to switch away from 'interesting_thread'. interesting_thread->cpu will be 0. So, vfp_current_hw_state[0] points at interesting_thread->vfpstate. The hardware state will be saved, along with the CPU number (0) that it was executing on. 'thread' will be 'new_cpu0_thread' with new_cpu0_thread->cpu = 0. Also, because it was executing on CPU0, new_cpu0_thread->vfpstate.hard.cpu = 0, and so the thread migration check is not triggered. This means that vfp_current_hw_state[0] remains pointing at interesting_thread. The second execution will be on CPU1 to switch _to_ 'interesting_thread'. So, 'thread' will be 'interesting_thread' and interesting_thread->cpu now will be 1. The previous thread executing on CPU1 is not relevant to this so we shall ignore that. We get to the thread migration check. Here, we discover that interesting_thread->vfpstate.hard.cpu = 0, yet interesting_thread->cpu is now 1, indicating thread migration. We set vfp_current_hw_state[1] to NULL. So, at this point vfp_current_hw_state[] contains the following: [0] = &interesting_thread->vfpstate [1] = NULL Our interesting thread now executes a VFP instruction, takes a fault which loads the state into the VFP hardware. Now, through the assembly we now have: [0] = &interesting_thread->vfpstate [1] = &interesting_thread->vfpstate CPU1 stops due to ptrace (and so saves its VFP state) using the thread switch code above), and CPU0 calls vfp_sync_hwstate(). if (vfp_current_hw_state[cpu] == &thread->vfpstate) { vfp_save_state(&thread->vfpstate, fpexc | FPEXC_EN); BANG, we corrupt interesting_thread's VFP state by overwriting the more up-to-date state saved by CPU1 with the old VFP state from CPU0. Fix this by ensuring that we have sane semantics for the various state describing variables: 1. vfp_current_hw_state[] points to the current owner of the context information stored in each CPUs hardware, or NULL if that state information is invalid. 2. thread->vfpstate.hard.cpu always contains the most recent CPU number which the state was loaded into or NR_CPUS if no CPU owns the state. So, for a particular CPU to be a valid owner of the VFP state for a particular thread t, two things must be true: vfp_current_hw_state[cpu] == &t->vfpstate && t->vfpstate.hard.cpu == cpu. and that is valid from the moment a CPU loads the saved VFP context into the hardware. This gives clear and consistent semantics to interpreting these variables. This patch also fixes thread copying, ensuring that t->vfpstate.hard.cpu is invalidated, otherwise CPU0 may believe it was the last owner. The hole can happen thus: - thread1 runs on CPU2 using VFP, migrates to CPU3, exits and thread_info freed. - New thread allocated from a previously running thread on CPU2, reusing memory for thread1 and copying vfp.hard.cpu. At this point, the following are true: new_thread1->vfpstate.hard.cpu == 2 &new_thread1->vfpstate == vfp_current_hw_state[2] Lastly, this also addresses thread flushing in a similar way to thread copying. Hole is: - thread runs on CPU0, using VFP, migrates to CPU1 but does not use VFP. - thread calls execve(), so thread flush happens, leaving vfp_current_hw_state[0] intact. This vfpstate is memset to 0 causing thread->vfpstate.hard.cpu = 0. - thread migrates back to CPU0 before using VFP. At this point, the following are true: thread->vfpstate.hard.cpu == 0 &thread->vfpstate == vfp_current_hw_state[0] Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2011-07-09 15:09:43 +00:00
cmp r4, #0 @ if the vfp_current_hw_state is NULL
beq vfp_reload_hw @ then the hw state needs reloading
[ARM] 4583/1: ARMv7: Add VFPv3 support This patch adds the support for VFPv3 (the kernel currently supports VFPv2). The main difference is 32 double registers (compared to 16). Signed-off-by: Catalin Marinas <catalin.marinas@arm.com> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2007-09-25 14:22:24 +00:00
VFPFSTMIA r4, r5 @ save the working registers
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-16 22:20:36 +00:00
VFPFMRX r5, FPSCR @ current status
Fix the VFP handling on the Feroceon CPU This CPU generates synchronous VFP exceptions in a non-standard way - the FPEXC.EX bit set but without the FPSCR.IXE bit being set like in the VFP subarchitecture 1 or just the FPEXC.DEX bit like in VFP subarchitecture 2. The main problem is that the faulty instruction (which needs to be emulated in software) will be restarted several times (normally until a context switch disables the VFP). This patch ensures that the VFP exception is treated as synchronous. Signed-off-by: Catalin Marinas <catalin.marinas@arm.com> Cc: Nicolas Pitre <nico@cam.org>
2009-05-30 13:00:18 +00:00
#ifndef CONFIG_CPU_FEROCEON
[ARM] 4582/2: Add support for the common VFP subarchitecture This patch allows the VFP support code to run correctly on CPUs compatible with the common VFP subarchitecture specification (Appendix B in the ARM ARM v7-A and v7-R edition). It implements support for VFP subarchitecture 2 while being backwards compatible with subarchitecture 1. On VFP subarchitecture 1, the arithmetic exceptions are asynchronous (or imprecise as described in the old ARM ARM) unless the FPSCR.IXE bit is 1. The exceptional instructions can be read from FPINST and FPINST2 registers. With VFP subarchitecture 2, the arithmetic exceptions can also be synchronous and marked by the FPEXC.DEX bit (the FPEXC.EX bit is cleared). CPUs implementing the synchronous arithmetic exceptions don't have the FPINST and FPINST2 registers and accessing them would trigger and undefined exception. Note that FPEXC.EX bit has an additional meaning on subarchitecture 1 - if it isn't set, there is no additional information in FPINST and FPINST2 that needs to be saved at context switch or when lazy-loading the VFP state of a different thread. The patch also removes the clearing of the cumulative exception flags in FPSCR when additional exceptions were raised. It is up to the user application to clear these bits. Signed-off-by: Catalin Marinas <catalin.marinas@arm.com> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2007-11-22 17:32:01 +00:00
tst r1, #FPEXC_EX @ is there additional state to save?
ARMv7: Branch over conditional undefined instructions in vfphw.S On ARMv7, conditional undefined instructions may generate exceptions even if the condition is not met. The vfphw.S contains the FPINST and FPINST2 access instructions which may not be present on processors with synchronous VFP exceptions. Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
2008-11-06 13:23:08 +00:00
beq 1f
VFPFMRX r6, FPINST @ FPINST (only if FPEXC.EX is set)
tst r1, #FPEXC_FP2V @ is there an FPINST2 to read?
beq 1f
VFPFMRX r8, FPINST2 @ FPINST2 if needed (and present)
1:
Fix the VFP handling on the Feroceon CPU This CPU generates synchronous VFP exceptions in a non-standard way - the FPEXC.EX bit set but without the FPSCR.IXE bit being set like in the VFP subarchitecture 1 or just the FPEXC.DEX bit like in VFP subarchitecture 2. The main problem is that the faulty instruction (which needs to be emulated in software) will be restarted several times (normally until a context switch disables the VFP). This patch ensures that the VFP exception is treated as synchronous. Signed-off-by: Catalin Marinas <catalin.marinas@arm.com> Cc: Nicolas Pitre <nico@cam.org>
2009-05-30 13:00:18 +00:00
#endif
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-16 22:20:36 +00:00
stmia r4, {r1, r5, r6, r8} @ save FPEXC, FPSCR, FPINST, FPINST2
ARM: vfp: fix a hole in VFP thread migration Fix a hole in the VFP thread migration. Lets define two threads. Thread 1, we'll call 'interesting_thread' which is a thread which is running on CPU0, using VFP (so vfp_current_hw_state[0] = &interesting_thread->vfpstate) and gets migrated off to CPU1, where it continues execution of VFP instructions. Thread 2, we'll call 'new_cpu0_thread' which is the thread which takes over on CPU0. This has also been using VFP, and last used VFP on CPU0, but doesn't use it again. The following code will be executed twice: cpu = thread->cpu; /* * On SMP, if VFP is enabled, save the old state in * case the thread migrates to a different CPU. The * restoring is done lazily. */ if ((fpexc & FPEXC_EN) && vfp_current_hw_state[cpu]) { vfp_save_state(vfp_current_hw_state[cpu], fpexc); vfp_current_hw_state[cpu]->hard.cpu = cpu; } /* * Thread migration, just force the reloading of the * state on the new CPU in case the VFP registers * contain stale data. */ if (thread->vfpstate.hard.cpu != cpu) vfp_current_hw_state[cpu] = NULL; The first execution will be on CPU0 to switch away from 'interesting_thread'. interesting_thread->cpu will be 0. So, vfp_current_hw_state[0] points at interesting_thread->vfpstate. The hardware state will be saved, along with the CPU number (0) that it was executing on. 'thread' will be 'new_cpu0_thread' with new_cpu0_thread->cpu = 0. Also, because it was executing on CPU0, new_cpu0_thread->vfpstate.hard.cpu = 0, and so the thread migration check is not triggered. This means that vfp_current_hw_state[0] remains pointing at interesting_thread. The second execution will be on CPU1 to switch _to_ 'interesting_thread'. So, 'thread' will be 'interesting_thread' and interesting_thread->cpu now will be 1. The previous thread executing on CPU1 is not relevant to this so we shall ignore that. We get to the thread migration check. Here, we discover that interesting_thread->vfpstate.hard.cpu = 0, yet interesting_thread->cpu is now 1, indicating thread migration. We set vfp_current_hw_state[1] to NULL. So, at this point vfp_current_hw_state[] contains the following: [0] = &interesting_thread->vfpstate [1] = NULL Our interesting thread now executes a VFP instruction, takes a fault which loads the state into the VFP hardware. Now, through the assembly we now have: [0] = &interesting_thread->vfpstate [1] = &interesting_thread->vfpstate CPU1 stops due to ptrace (and so saves its VFP state) using the thread switch code above), and CPU0 calls vfp_sync_hwstate(). if (vfp_current_hw_state[cpu] == &thread->vfpstate) { vfp_save_state(&thread->vfpstate, fpexc | FPEXC_EN); BANG, we corrupt interesting_thread's VFP state by overwriting the more up-to-date state saved by CPU1 with the old VFP state from CPU0. Fix this by ensuring that we have sane semantics for the various state describing variables: 1. vfp_current_hw_state[] points to the current owner of the context information stored in each CPUs hardware, or NULL if that state information is invalid. 2. thread->vfpstate.hard.cpu always contains the most recent CPU number which the state was loaded into or NR_CPUS if no CPU owns the state. So, for a particular CPU to be a valid owner of the VFP state for a particular thread t, two things must be true: vfp_current_hw_state[cpu] == &t->vfpstate && t->vfpstate.hard.cpu == cpu. and that is valid from the moment a CPU loads the saved VFP context into the hardware. This gives clear and consistent semantics to interpreting these variables. This patch also fixes thread copying, ensuring that t->vfpstate.hard.cpu is invalidated, otherwise CPU0 may believe it was the last owner. The hole can happen thus: - thread1 runs on CPU2 using VFP, migrates to CPU3, exits and thread_info freed. - New thread allocated from a previously running thread on CPU2, reusing memory for thread1 and copying vfp.hard.cpu. At this point, the following are true: new_thread1->vfpstate.hard.cpu == 2 &new_thread1->vfpstate == vfp_current_hw_state[2] Lastly, this also addresses thread flushing in a similar way to thread copying. Hole is: - thread runs on CPU0, using VFP, migrates to CPU1 but does not use VFP. - thread calls execve(), so thread flush happens, leaving vfp_current_hw_state[0] intact. This vfpstate is memset to 0 causing thread->vfpstate.hard.cpu = 0. - thread migrates back to CPU0 before using VFP. At this point, the following are true: thread->vfpstate.hard.cpu == 0 &thread->vfpstate == vfp_current_hw_state[0] Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2011-07-09 15:09:43 +00:00
vfp_reload_hw:
#else
@ For SMP, if this thread does not own the hw context, then we
@ need to reload it. No need to save the old state as on SMP,
@ we always save the state when we switch away from a thread.
bne vfp_reload_hw
@ This thread has ownership of the current hardware context.
@ However, it may have been migrated to another CPU, in which
@ case the saved state is newer than the hardware context.
@ Check this by looking at the CPU number which the state was
@ last loaded onto.
ldr ip, [r10, #VFP_CPU]
teq ip, r11
beq vfp_hw_state_valid
vfp_reload_hw:
@ We're loading this threads state into the VFP hardware. Update
@ the CPU number which contains the most up to date VFP context.
str r11, [r10, #VFP_CPU]
VFPFMXR FPEXC, r5 @ enable VFP, disable any pending
@ exceptions, so we can get at the
@ rest of it
[ARM] 4111/1: Allow VFP to work with thread migration on SMP The current lazy saving of the VFP registers is no longer possible with thread migration on SMP. This patch implements a per-CPU vfp-state pointer and the saving of the VFP registers at every context switch. The registers restoring is still performed in a lazy way. Signed-off-by: Catalin Marinas <catalin.marinas@arm.com> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2007-01-24 17:47:08 +00:00
#endif
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-16 22:20:36 +00:00
DBGSTR1 "load state %p", r10
ARM: vfp: rename last_VFP_context to vfp_current_hw_state Rename the slightly confusing 'last_VFP_context' variable to be more descriptive of what it actually is. This variable stores a pointer to the current owner's vfpstate structure for the context held in the VFP hardware. Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2011-07-09 12:44:04 +00:00
str r10, [r3, r11, lsl #2] @ update the vfp_current_hw_state pointer
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-16 22:20:36 +00:00
@ Load the saved state back into the VFP
[ARM] 4583/1: ARMv7: Add VFPv3 support This patch adds the support for VFPv3 (the kernel currently supports VFPv2). The main difference is 32 double registers (compared to 16). Signed-off-by: Catalin Marinas <catalin.marinas@arm.com> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2007-09-25 14:22:24 +00:00
VFPFLDMIA r10, r5 @ reload the working registers while
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-16 22:20:36 +00:00
@ FPEXC is in a safe state
[ARM] 3398/1: Fix the VFP registers loading/storing base address Patch from Catalin Marinas The current VFP code corrupts the VFP registers (including the control ones) if more than one floating point application is executed at the same time. This patch fixes the updating of the load/store base addresses for the VFP registers. Signed-off-by: Catalin Marinas <catalin.marinas@arm.com> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2006-03-25 21:58:00 +00:00
ldmia r10, {r1, r5, r6, r8} @ load FPEXC, FPSCR, FPINST, FPINST2
Fix the VFP handling on the Feroceon CPU This CPU generates synchronous VFP exceptions in a non-standard way - the FPEXC.EX bit set but without the FPSCR.IXE bit being set like in the VFP subarchitecture 1 or just the FPEXC.DEX bit like in VFP subarchitecture 2. The main problem is that the faulty instruction (which needs to be emulated in software) will be restarted several times (normally until a context switch disables the VFP). This patch ensures that the VFP exception is treated as synchronous. Signed-off-by: Catalin Marinas <catalin.marinas@arm.com> Cc: Nicolas Pitre <nico@cam.org>
2009-05-30 13:00:18 +00:00
#ifndef CONFIG_CPU_FEROCEON
[ARM] 4582/2: Add support for the common VFP subarchitecture This patch allows the VFP support code to run correctly on CPUs compatible with the common VFP subarchitecture specification (Appendix B in the ARM ARM v7-A and v7-R edition). It implements support for VFP subarchitecture 2 while being backwards compatible with subarchitecture 1. On VFP subarchitecture 1, the arithmetic exceptions are asynchronous (or imprecise as described in the old ARM ARM) unless the FPSCR.IXE bit is 1. The exceptional instructions can be read from FPINST and FPINST2 registers. With VFP subarchitecture 2, the arithmetic exceptions can also be synchronous and marked by the FPEXC.DEX bit (the FPEXC.EX bit is cleared). CPUs implementing the synchronous arithmetic exceptions don't have the FPINST and FPINST2 registers and accessing them would trigger and undefined exception. Note that FPEXC.EX bit has an additional meaning on subarchitecture 1 - if it isn't set, there is no additional information in FPINST and FPINST2 that needs to be saved at context switch or when lazy-loading the VFP state of a different thread. The patch also removes the clearing of the cumulative exception flags in FPSCR when additional exceptions were raised. It is up to the user application to clear these bits. Signed-off-by: Catalin Marinas <catalin.marinas@arm.com> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2007-11-22 17:32:01 +00:00
tst r1, #FPEXC_EX @ is there additional state to restore?
ARMv7: Branch over conditional undefined instructions in vfphw.S On ARMv7, conditional undefined instructions may generate exceptions even if the condition is not met. The vfphw.S contains the FPINST and FPINST2 access instructions which may not be present on processors with synchronous VFP exceptions. Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
2008-11-06 13:23:08 +00:00
beq 1f
VFPFMXR FPINST, r6 @ restore FPINST (only if FPEXC.EX is set)
tst r1, #FPEXC_FP2V @ is there an FPINST2 to write?
beq 1f
VFPFMXR FPINST2, r8 @ FPINST2 if needed (and present)
1:
Fix the VFP handling on the Feroceon CPU This CPU generates synchronous VFP exceptions in a non-standard way - the FPEXC.EX bit set but without the FPSCR.IXE bit being set like in the VFP subarchitecture 1 or just the FPEXC.DEX bit like in VFP subarchitecture 2. The main problem is that the faulty instruction (which needs to be emulated in software) will be restarted several times (normally until a context switch disables the VFP). This patch ensures that the VFP exception is treated as synchronous. Signed-off-by: Catalin Marinas <catalin.marinas@arm.com> Cc: Nicolas Pitre <nico@cam.org>
2009-05-30 13:00:18 +00:00
#endif
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-16 22:20:36 +00:00
VFPFMXR FPSCR, r5 @ restore status
ARM: vfp: rename check_exception to vfp_hw_state_valid Rename this branch to more accurately reflect why its taken, rather than what the following code does. It is the only caller of this code. This helps to clarify following changes, yet this change results in no actual code change. Document the VFP hardware state at the target of this branch. Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2011-07-09 13:24:36 +00:00
@ The context stored in the VFP hardware is up to date with this thread
vfp_hw_state_valid:
[ARM] vfp: make fpexc bit names less verbose Use the fpexc abbreviated names instead of long verbose names for fpexc bits. Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2007-07-18 08:37:10 +00:00
tst r1, #FPEXC_EX
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-16 22:20:36 +00:00
bne process_exception @ might as well handle the pending
@ exception before retrying branch
@ out before setting an FPEXC that
@ stops us reading stuff
VFPFMXR FPEXC, r1 @ restore FPEXC last
sub r2, r2, #4
str r2, [sp, #S_PC] @ retry the instruction
[ARM] 5440/1: Fix VFP state corruption due to preemption during VFP exceptions We've observed that ARM VFP state can be corrupted during VFP exception handling when PREEMPT is enabled. The exact conditions are difficult to reproduce but appear to occur during VFP exception handling when a task causes a VFP exception which is handled via VFP_bounce and is then preempted by yet another task which in turn causes yet another VFP exception. Since the VFP_bounce code is not preempt safe, VFP state then becomes corrupt. In order to prevent preemption from occuring while handling a VFP exception, this patch disables preemption while handling VFP exceptions. Signed-off-by: George G. Davis <gdavis@mvista.com> Acked-by: Catalin Marinas <catalin.marinas@arm.com> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2009-04-01 19:27:18 +00:00
#ifdef CONFIG_PREEMPT
get_thread_info r10
ldr r4, [r10, #TI_PREEMPT] @ get preempt count
sub r11, r4, #1 @ decrement it
str r11, [r10, #TI_PREEMPT]
#endif
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-16 22:20:36 +00:00
mov pc, r9 @ we think we have handled things
look_for_VFP_exceptions:
[ARM] 4582/2: Add support for the common VFP subarchitecture This patch allows the VFP support code to run correctly on CPUs compatible with the common VFP subarchitecture specification (Appendix B in the ARM ARM v7-A and v7-R edition). It implements support for VFP subarchitecture 2 while being backwards compatible with subarchitecture 1. On VFP subarchitecture 1, the arithmetic exceptions are asynchronous (or imprecise as described in the old ARM ARM) unless the FPSCR.IXE bit is 1. The exceptional instructions can be read from FPINST and FPINST2 registers. With VFP subarchitecture 2, the arithmetic exceptions can also be synchronous and marked by the FPEXC.DEX bit (the FPEXC.EX bit is cleared). CPUs implementing the synchronous arithmetic exceptions don't have the FPINST and FPINST2 registers and accessing them would trigger and undefined exception. Note that FPEXC.EX bit has an additional meaning on subarchitecture 1 - if it isn't set, there is no additional information in FPINST and FPINST2 that needs to be saved at context switch or when lazy-loading the VFP state of a different thread. The patch also removes the clearing of the cumulative exception flags in FPSCR when additional exceptions were raised. It is up to the user application to clear these bits. Signed-off-by: Catalin Marinas <catalin.marinas@arm.com> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2007-11-22 17:32:01 +00:00
@ Check for synchronous or asynchronous exception
tst r1, #FPEXC_EX | FPEXC_DEX
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-16 22:20:36 +00:00
bne process_exception
[ARM] 4582/2: Add support for the common VFP subarchitecture This patch allows the VFP support code to run correctly on CPUs compatible with the common VFP subarchitecture specification (Appendix B in the ARM ARM v7-A and v7-R edition). It implements support for VFP subarchitecture 2 while being backwards compatible with subarchitecture 1. On VFP subarchitecture 1, the arithmetic exceptions are asynchronous (or imprecise as described in the old ARM ARM) unless the FPSCR.IXE bit is 1. The exceptional instructions can be read from FPINST and FPINST2 registers. With VFP subarchitecture 2, the arithmetic exceptions can also be synchronous and marked by the FPEXC.DEX bit (the FPEXC.EX bit is cleared). CPUs implementing the synchronous arithmetic exceptions don't have the FPINST and FPINST2 registers and accessing them would trigger and undefined exception. Note that FPEXC.EX bit has an additional meaning on subarchitecture 1 - if it isn't set, there is no additional information in FPINST and FPINST2 that needs to be saved at context switch or when lazy-loading the VFP state of a different thread. The patch also removes the clearing of the cumulative exception flags in FPSCR when additional exceptions were raised. It is up to the user application to clear these bits. Signed-off-by: Catalin Marinas <catalin.marinas@arm.com> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2007-11-22 17:32:01 +00:00
@ On some implementations of the VFP subarch 1, setting FPSCR.IXE
@ causes all the CDP instructions to be bounced synchronously without
@ setting the FPEXC.EX bit
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-16 22:20:36 +00:00
VFPFMRX r5, FPSCR
[ARM] 4582/2: Add support for the common VFP subarchitecture This patch allows the VFP support code to run correctly on CPUs compatible with the common VFP subarchitecture specification (Appendix B in the ARM ARM v7-A and v7-R edition). It implements support for VFP subarchitecture 2 while being backwards compatible with subarchitecture 1. On VFP subarchitecture 1, the arithmetic exceptions are asynchronous (or imprecise as described in the old ARM ARM) unless the FPSCR.IXE bit is 1. The exceptional instructions can be read from FPINST and FPINST2 registers. With VFP subarchitecture 2, the arithmetic exceptions can also be synchronous and marked by the FPEXC.DEX bit (the FPEXC.EX bit is cleared). CPUs implementing the synchronous arithmetic exceptions don't have the FPINST and FPINST2 registers and accessing them would trigger and undefined exception. Note that FPEXC.EX bit has an additional meaning on subarchitecture 1 - if it isn't set, there is no additional information in FPINST and FPINST2 that needs to be saved at context switch or when lazy-loading the VFP state of a different thread. The patch also removes the clearing of the cumulative exception flags in FPSCR when additional exceptions were raised. It is up to the user application to clear these bits. Signed-off-by: Catalin Marinas <catalin.marinas@arm.com> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2007-11-22 17:32:01 +00:00
tst r5, #FPSCR_IXE
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-16 22:20:36 +00:00
bne process_exception
@ Fall into hand on to next handler - appropriate coproc instr
@ not recognised by VFP
DBGSTR "not VFP"
[ARM] 5440/1: Fix VFP state corruption due to preemption during VFP exceptions We've observed that ARM VFP state can be corrupted during VFP exception handling when PREEMPT is enabled. The exact conditions are difficult to reproduce but appear to occur during VFP exception handling when a task causes a VFP exception which is handled via VFP_bounce and is then preempted by yet another task which in turn causes yet another VFP exception. Since the VFP_bounce code is not preempt safe, VFP state then becomes corrupt. In order to prevent preemption from occuring while handling a VFP exception, this patch disables preemption while handling VFP exceptions. Signed-off-by: George G. Davis <gdavis@mvista.com> Acked-by: Catalin Marinas <catalin.marinas@arm.com> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2009-04-01 19:27:18 +00:00
#ifdef CONFIG_PREEMPT
get_thread_info r10
ldr r4, [r10, #TI_PREEMPT] @ get preempt count
sub r11, r4, #1 @ decrement it
str r11, [r10, #TI_PREEMPT]
#endif
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-16 22:20:36 +00:00
mov pc, lr
process_exception:
DBGSTR "bounce"
mov r2, sp @ nothing stacked - regdump is at TOS
mov lr, r9 @ setup for a return to the user code.
@ Now call the C code to package up the bounce to the support code
@ r0 holds the trigger instruction
@ r1 holds the FPEXC value
@ r2 pointer to register dump
[ARM] 4582/2: Add support for the common VFP subarchitecture This patch allows the VFP support code to run correctly on CPUs compatible with the common VFP subarchitecture specification (Appendix B in the ARM ARM v7-A and v7-R edition). It implements support for VFP subarchitecture 2 while being backwards compatible with subarchitecture 1. On VFP subarchitecture 1, the arithmetic exceptions are asynchronous (or imprecise as described in the old ARM ARM) unless the FPSCR.IXE bit is 1. The exceptional instructions can be read from FPINST and FPINST2 registers. With VFP subarchitecture 2, the arithmetic exceptions can also be synchronous and marked by the FPEXC.DEX bit (the FPEXC.EX bit is cleared). CPUs implementing the synchronous arithmetic exceptions don't have the FPINST and FPINST2 registers and accessing them would trigger and undefined exception. Note that FPEXC.EX bit has an additional meaning on subarchitecture 1 - if it isn't set, there is no additional information in FPINST and FPINST2 that needs to be saved at context switch or when lazy-loading the VFP state of a different thread. The patch also removes the clearing of the cumulative exception flags in FPSCR when additional exceptions were raised. It is up to the user application to clear these bits. Signed-off-by: Catalin Marinas <catalin.marinas@arm.com> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2007-11-22 17:32:01 +00:00
b VFP_bounce @ we have handled this - the support
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-16 22:20:36 +00:00
@ code will raise an exception if
@ required. If not, the user code will
@ retry the faulted instruction
[ARM] 5227/1: Add the ENDPROC declarations to the .S files This declaration specifies the "function" type and size for various assembly functions, mainly needed for generating the correct branch instructions in Thumb-2. Signed-off-by: Catalin Marinas <catalin.marinas@arm.com> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2008-08-28 10:22:32 +00:00
ENDPROC(vfp_support_entry)
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-16 22:20:36 +00:00
[ARM] 5227/1: Add the ENDPROC declarations to the .S files This declaration specifies the "function" type and size for various assembly functions, mainly needed for generating the correct branch instructions in Thumb-2. Signed-off-by: Catalin Marinas <catalin.marinas@arm.com> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2008-08-28 10:22:32 +00:00
ENTRY(vfp_save_state)
[ARM] 4111/1: Allow VFP to work with thread migration on SMP The current lazy saving of the VFP registers is no longer possible with thread migration on SMP. This patch implements a per-CPU vfp-state pointer and the saving of the VFP registers at every context switch. The registers restoring is still performed in a lazy way. Signed-off-by: Catalin Marinas <catalin.marinas@arm.com> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2007-01-24 17:47:08 +00:00
@ Save the current VFP state
@ r0 - save location
@ r1 - FPEXC
DBGSTR1 "save VFP state %p", r0
[ARM] 4583/1: ARMv7: Add VFPv3 support This patch adds the support for VFPv3 (the kernel currently supports VFPv2). The main difference is 32 double registers (compared to 16). Signed-off-by: Catalin Marinas <catalin.marinas@arm.com> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2007-09-25 14:22:24 +00:00
VFPFSTMIA r0, r2 @ save the working registers
[ARM] 4111/1: Allow VFP to work with thread migration on SMP The current lazy saving of the VFP registers is no longer possible with thread migration on SMP. This patch implements a per-CPU vfp-state pointer and the saving of the VFP registers at every context switch. The registers restoring is still performed in a lazy way. Signed-off-by: Catalin Marinas <catalin.marinas@arm.com> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2007-01-24 17:47:08 +00:00
VFPFMRX r2, FPSCR @ current status
[ARM] 4582/2: Add support for the common VFP subarchitecture This patch allows the VFP support code to run correctly on CPUs compatible with the common VFP subarchitecture specification (Appendix B in the ARM ARM v7-A and v7-R edition). It implements support for VFP subarchitecture 2 while being backwards compatible with subarchitecture 1. On VFP subarchitecture 1, the arithmetic exceptions are asynchronous (or imprecise as described in the old ARM ARM) unless the FPSCR.IXE bit is 1. The exceptional instructions can be read from FPINST and FPINST2 registers. With VFP subarchitecture 2, the arithmetic exceptions can also be synchronous and marked by the FPEXC.DEX bit (the FPEXC.EX bit is cleared). CPUs implementing the synchronous arithmetic exceptions don't have the FPINST and FPINST2 registers and accessing them would trigger and undefined exception. Note that FPEXC.EX bit has an additional meaning on subarchitecture 1 - if it isn't set, there is no additional information in FPINST and FPINST2 that needs to be saved at context switch or when lazy-loading the VFP state of a different thread. The patch also removes the clearing of the cumulative exception flags in FPSCR when additional exceptions were raised. It is up to the user application to clear these bits. Signed-off-by: Catalin Marinas <catalin.marinas@arm.com> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2007-11-22 17:32:01 +00:00
tst r1, #FPEXC_EX @ is there additional state to save?
ARMv7: Branch over conditional undefined instructions in vfphw.S On ARMv7, conditional undefined instructions may generate exceptions even if the condition is not met. The vfphw.S contains the FPINST and FPINST2 access instructions which may not be present on processors with synchronous VFP exceptions. Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
2008-11-06 13:23:08 +00:00
beq 1f
VFPFMRX r3, FPINST @ FPINST (only if FPEXC.EX is set)
tst r1, #FPEXC_FP2V @ is there an FPINST2 to read?
beq 1f
VFPFMRX r12, FPINST2 @ FPINST2 if needed (and present)
1:
[ARM] 4111/1: Allow VFP to work with thread migration on SMP The current lazy saving of the VFP registers is no longer possible with thread migration on SMP. This patch implements a per-CPU vfp-state pointer and the saving of the VFP registers at every context switch. The registers restoring is still performed in a lazy way. Signed-off-by: Catalin Marinas <catalin.marinas@arm.com> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2007-01-24 17:47:08 +00:00
stmia r0, {r1, r2, r3, r12} @ save FPEXC, FPSCR, FPINST, FPINST2
mov pc, lr
[ARM] 5227/1: Add the ENDPROC declarations to the .S files This declaration specifies the "function" type and size for various assembly functions, mainly needed for generating the correct branch instructions in Thumb-2. Signed-off-by: Catalin Marinas <catalin.marinas@arm.com> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2008-08-28 10:22:32 +00:00
ENDPROC(vfp_save_state)
[ARM] 4111/1: Allow VFP to work with thread migration on SMP The current lazy saving of the VFP registers is no longer possible with thread migration on SMP. This patch implements a per-CPU vfp-state pointer and the saving of the VFP registers at every context switch. The registers restoring is still performed in a lazy way. Signed-off-by: Catalin Marinas <catalin.marinas@arm.com> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2007-01-24 17:47:08 +00:00
ARM: 6498/1: vfp: Correct data alignment for CONFIG_THUMB2_KERNEL Directives such as .long and .word do not magically cause the assembler location counter to become aligned in gas. As a result, using these directives in code sections can result in misaligned data words when building a Thumb-2 kernel (CONFIG_THUMB2_KERNEL). This is a Bad Thing, since the ABI permits the compiler to assume that fundamental types of word size or above are word- aligned when accessing them from C. If the data is not really word-aligned, this can cause impaired performance and stray alignment faults in some circumstances. In general, the following rules should be applied when using data word declaration directives inside code sections: * .quad and .double: .align 3 * .long, .word, .single, .float: .align (or .align 2) * .short: No explicit alignment required, since Thumb-2 instructions are always 2 or 4 bytes in size. immediately after an instruction. Reviewed-by: Will Deacon <will.deacon@arm.com> Signed-off-by: Dave Martin <dave.martin@linaro.org> Acked-by: Catalin Marinas <catalin.marinas@arm.com> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2010-11-29 18:43:22 +00:00
.align
ARM: vfp: rename last_VFP_context to vfp_current_hw_state Rename the slightly confusing 'last_VFP_context' variable to be more descriptive of what it actually is. This variable stores a pointer to the current owner's vfpstate structure for the context held in the VFP hardware. Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2011-07-09 12:44:04 +00:00
vfp_current_hw_state_address:
.word vfp_current_hw_state
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-16 22:20:36 +00:00
Thumb-2: Implement the unified VFP support This patch modifies the VFP files for the ARM/Thumb-2 unified assembly syntax. Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
2009-07-24 11:32:57 +00:00
.macro tbl_branch, base, tmp, shift
#ifdef CONFIG_THUMB2_KERNEL
adr \tmp, 1f
add \tmp, \tmp, \base, lsl \shift
mov pc, \tmp
#else
add pc, pc, \base, lsl \shift
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-16 22:20:36 +00:00
mov r0, r0
Thumb-2: Implement the unified VFP support This patch modifies the VFP files for the ARM/Thumb-2 unified assembly syntax. Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
2009-07-24 11:32:57 +00:00
#endif
1:
.endm
ENTRY(vfp_get_float)
tbl_branch r0, r3, #3
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-16 22:20:36 +00:00
.irp dr,0,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15
Thumb-2: Implement the unified VFP support This patch modifies the VFP files for the ARM/Thumb-2 unified assembly syntax. Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
2009-07-24 11:32:57 +00:00
1: mrc p10, 0, r0, c\dr, c0, 0 @ fmrs r0, s0
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-16 22:20:36 +00:00
mov pc, lr
Thumb-2: Implement the unified VFP support This patch modifies the VFP files for the ARM/Thumb-2 unified assembly syntax. Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
2009-07-24 11:32:57 +00:00
.org 1b + 8
1: mrc p10, 0, r0, c\dr, c0, 4 @ fmrs r0, s1
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-16 22:20:36 +00:00
mov pc, lr
Thumb-2: Implement the unified VFP support This patch modifies the VFP files for the ARM/Thumb-2 unified assembly syntax. Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
2009-07-24 11:32:57 +00:00
.org 1b + 8
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-16 22:20:36 +00:00
.endr
[ARM] 5227/1: Add the ENDPROC declarations to the .S files This declaration specifies the "function" type and size for various assembly functions, mainly needed for generating the correct branch instructions in Thumb-2. Signed-off-by: Catalin Marinas <catalin.marinas@arm.com> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2008-08-28 10:22:32 +00:00
ENDPROC(vfp_get_float)
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-16 22:20:36 +00:00
[ARM] 5227/1: Add the ENDPROC declarations to the .S files This declaration specifies the "function" type and size for various assembly functions, mainly needed for generating the correct branch instructions in Thumb-2. Signed-off-by: Catalin Marinas <catalin.marinas@arm.com> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2008-08-28 10:22:32 +00:00
ENTRY(vfp_put_float)
Thumb-2: Implement the unified VFP support This patch modifies the VFP files for the ARM/Thumb-2 unified assembly syntax. Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
2009-07-24 11:32:57 +00:00
tbl_branch r1, r3, #3
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-16 22:20:36 +00:00
.irp dr,0,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15
Thumb-2: Implement the unified VFP support This patch modifies the VFP files for the ARM/Thumb-2 unified assembly syntax. Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
2009-07-24 11:32:57 +00:00
1: mcr p10, 0, r0, c\dr, c0, 0 @ fmsr r0, s0
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-16 22:20:36 +00:00
mov pc, lr
Thumb-2: Implement the unified VFP support This patch modifies the VFP files for the ARM/Thumb-2 unified assembly syntax. Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
2009-07-24 11:32:57 +00:00
.org 1b + 8
1: mcr p10, 0, r0, c\dr, c0, 4 @ fmsr r0, s1
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-16 22:20:36 +00:00
mov pc, lr
Thumb-2: Implement the unified VFP support This patch modifies the VFP files for the ARM/Thumb-2 unified assembly syntax. Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
2009-07-24 11:32:57 +00:00
.org 1b + 8
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-16 22:20:36 +00:00
.endr
[ARM] 5227/1: Add the ENDPROC declarations to the .S files This declaration specifies the "function" type and size for various assembly functions, mainly needed for generating the correct branch instructions in Thumb-2. Signed-off-by: Catalin Marinas <catalin.marinas@arm.com> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2008-08-28 10:22:32 +00:00
ENDPROC(vfp_put_float)
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-16 22:20:36 +00:00
[ARM] 5227/1: Add the ENDPROC declarations to the .S files This declaration specifies the "function" type and size for various assembly functions, mainly needed for generating the correct branch instructions in Thumb-2. Signed-off-by: Catalin Marinas <catalin.marinas@arm.com> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2008-08-28 10:22:32 +00:00
ENTRY(vfp_get_double)
Thumb-2: Implement the unified VFP support This patch modifies the VFP files for the ARM/Thumb-2 unified assembly syntax. Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
2009-07-24 11:32:57 +00:00
tbl_branch r0, r3, #3
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-16 22:20:36 +00:00
.irp dr,0,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15
Thumb-2: Implement the unified VFP support This patch modifies the VFP files for the ARM/Thumb-2 unified assembly syntax. Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
2009-07-24 11:32:57 +00:00
1: fmrrd r0, r1, d\dr
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-16 22:20:36 +00:00
mov pc, lr
Thumb-2: Implement the unified VFP support This patch modifies the VFP files for the ARM/Thumb-2 unified assembly syntax. Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
2009-07-24 11:32:57 +00:00
.org 1b + 8
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-16 22:20:36 +00:00
.endr
[ARM] 4583/1: ARMv7: Add VFPv3 support This patch adds the support for VFPv3 (the kernel currently supports VFPv2). The main difference is 32 double registers (compared to 16). Signed-off-by: Catalin Marinas <catalin.marinas@arm.com> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2007-09-25 14:22:24 +00:00
#ifdef CONFIG_VFPv3
@ d16 - d31 registers
.irp dr,0,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15
Thumb-2: Implement the unified VFP support This patch modifies the VFP files for the ARM/Thumb-2 unified assembly syntax. Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
2009-07-24 11:32:57 +00:00
1: mrrc p11, 3, r0, r1, c\dr @ fmrrd r0, r1, d\dr
[ARM] 4583/1: ARMv7: Add VFPv3 support This patch adds the support for VFPv3 (the kernel currently supports VFPv2). The main difference is 32 double registers (compared to 16). Signed-off-by: Catalin Marinas <catalin.marinas@arm.com> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2007-09-25 14:22:24 +00:00
mov pc, lr
Thumb-2: Implement the unified VFP support This patch modifies the VFP files for the ARM/Thumb-2 unified assembly syntax. Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
2009-07-24 11:32:57 +00:00
.org 1b + 8
[ARM] 4583/1: ARMv7: Add VFPv3 support This patch adds the support for VFPv3 (the kernel currently supports VFPv2). The main difference is 32 double registers (compared to 16). Signed-off-by: Catalin Marinas <catalin.marinas@arm.com> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2007-09-25 14:22:24 +00:00
.endr
#endif
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-16 22:20:36 +00:00
[ARM] 4583/1: ARMv7: Add VFPv3 support This patch adds the support for VFPv3 (the kernel currently supports VFPv2). The main difference is 32 double registers (compared to 16). Signed-off-by: Catalin Marinas <catalin.marinas@arm.com> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2007-09-25 14:22:24 +00:00
@ virtual register 16 (or 32 if VFPv3) for compare with zero
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-16 22:20:36 +00:00
mov r0, #0
mov r1, #0
mov pc, lr
[ARM] 5227/1: Add the ENDPROC declarations to the .S files This declaration specifies the "function" type and size for various assembly functions, mainly needed for generating the correct branch instructions in Thumb-2. Signed-off-by: Catalin Marinas <catalin.marinas@arm.com> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2008-08-28 10:22:32 +00:00
ENDPROC(vfp_get_double)
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-16 22:20:36 +00:00
[ARM] 5227/1: Add the ENDPROC declarations to the .S files This declaration specifies the "function" type and size for various assembly functions, mainly needed for generating the correct branch instructions in Thumb-2. Signed-off-by: Catalin Marinas <catalin.marinas@arm.com> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2008-08-28 10:22:32 +00:00
ENTRY(vfp_put_double)
Thumb-2: Implement the unified VFP support This patch modifies the VFP files for the ARM/Thumb-2 unified assembly syntax. Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
2009-07-24 11:32:57 +00:00
tbl_branch r2, r3, #3
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-16 22:20:36 +00:00
.irp dr,0,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15
Thumb-2: Implement the unified VFP support This patch modifies the VFP files for the ARM/Thumb-2 unified assembly syntax. Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
2009-07-24 11:32:57 +00:00
1: fmdrr d\dr, r0, r1
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-16 22:20:36 +00:00
mov pc, lr
Thumb-2: Implement the unified VFP support This patch modifies the VFP files for the ARM/Thumb-2 unified assembly syntax. Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
2009-07-24 11:32:57 +00:00
.org 1b + 8
Linux-2.6.12-rc2 Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
2005-04-16 22:20:36 +00:00
.endr
[ARM] 4583/1: ARMv7: Add VFPv3 support This patch adds the support for VFPv3 (the kernel currently supports VFPv2). The main difference is 32 double registers (compared to 16). Signed-off-by: Catalin Marinas <catalin.marinas@arm.com> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2007-09-25 14:22:24 +00:00
#ifdef CONFIG_VFPv3
@ d16 - d31 registers
.irp dr,0,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15
ARM: VFP: Fix vfp_put_double() for d16-d31 vfp_put_double() takes the double value in r0,r1 not r1,r2. Reported-by: Tarun Kanti DebBarma <tarun.kanti@ti.com> Cc: <stable@kernel.org> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2010-05-27 07:23:29 +00:00
1: mcrr p11, 3, r0, r1, c\dr @ fmdrr r0, r1, d\dr
[ARM] 4583/1: ARMv7: Add VFPv3 support This patch adds the support for VFPv3 (the kernel currently supports VFPv2). The main difference is 32 double registers (compared to 16). Signed-off-by: Catalin Marinas <catalin.marinas@arm.com> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2007-09-25 14:22:24 +00:00
mov pc, lr
Thumb-2: Implement the unified VFP support This patch modifies the VFP files for the ARM/Thumb-2 unified assembly syntax. Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
2009-07-24 11:32:57 +00:00
.org 1b + 8
[ARM] 4583/1: ARMv7: Add VFPv3 support This patch adds the support for VFPv3 (the kernel currently supports VFPv2). The main difference is 32 double registers (compared to 16). Signed-off-by: Catalin Marinas <catalin.marinas@arm.com> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2007-09-25 14:22:24 +00:00
.endr
#endif
[ARM] 5227/1: Add the ENDPROC declarations to the .S files This declaration specifies the "function" type and size for various assembly functions, mainly needed for generating the correct branch instructions in Thumb-2. Signed-off-by: Catalin Marinas <catalin.marinas@arm.com> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2008-08-28 10:22:32 +00:00
ENDPROC(vfp_put_double)
Reference in New Issue Copy Permalink