acrn-hypervisor/hypervisor/arch/x86/io.c
Huihuang Shi 496e40072e HV:treewide:fix rest of violations related parameter changed
Misra c required parameter should not changed in the scope
of function,use local variable to replace it.

Signed-off-by: Huihuang Shi <huihuang.shi@intel.com>
Reviewed-by: Junjie Mao <junjie.mao@intel.com>
2018-07-27 16:42:19 +08:00

276 lines
6.7 KiB
C

/*
* Copyright (C) 2018 Intel Corporation. All rights reserved.
*
* SPDX-License-Identifier: BSD-3-Clause
*/
#include <hypervisor.h>
int dm_emulate_pio_post(struct vcpu *vcpu)
{
uint16_t cur = vcpu->vcpu_id;
int cur_context = vcpu->arch_vcpu.cur_context;
union vhm_request_buffer *req_buf = NULL;
struct io_request *io_req = &vcpu->req;
struct pio_request *pio_req = &io_req->reqs.pio;
uint64_t mask = 0xFFFFFFFFUL >> (32UL - 8UL * pio_req->size);
uint64_t *rax;
struct vhm_request *vhm_req;
req_buf = (union vhm_request_buffer *)(vcpu->vm->sw.io_shared_page);
vhm_req = &req_buf->req_queue[cur];
rax = &vcpu->arch_vcpu.contexts[cur_context].guest_cpu_regs.regs.rax;
io_req->processed = vhm_req->processed;
pio_req->value = vhm_req->reqs.pio.value;
/* VHM emulation data already copy to req, mark to free slot now */
vhm_req->valid = 0;
if (io_req->processed != REQ_STATE_SUCCESS) {
return -1;
}
if (pio_req->direction == REQUEST_READ) {
uint64_t value = (uint64_t)pio_req->value;
*rax = ((*rax) & ~mask) | (value & mask);
}
return 0;
}
static void
dm_emulate_pio_pre(struct vcpu *vcpu, uint64_t exit_qual, uint64_t req_value)
{
struct pio_request *pio_req = &vcpu->req.reqs.pio;
pio_req->value = req_value;
}
int io_instr_vmexit_handler(struct vcpu *vcpu)
{
uint64_t exit_qual;
uint64_t mask;
uint16_t port, size;
struct vm_io_handler *handler;
struct vm *vm = vcpu->vm;
struct io_request *io_req = &vcpu->req;
struct pio_request *pio_req = &io_req->reqs.pio;
int cur_context_idx = vcpu->arch_vcpu.cur_context;
struct run_context *cur_context;
int status = -EINVAL;
io_req->type = REQ_PORTIO;
io_req->processed = REQ_STATE_PENDING;
cur_context = &vcpu->arch_vcpu.contexts[cur_context_idx];
exit_qual = vcpu->arch_vcpu.exit_qualification;
pio_req->size = VM_EXIT_IO_INSTRUCTION_SIZE(exit_qual) + 1UL;
pio_req->address = VM_EXIT_IO_INSTRUCTION_PORT_NUMBER(exit_qual);
if (VM_EXIT_IO_INSTRUCTION_ACCESS_DIRECTION(exit_qual) == 0UL) {
pio_req->direction = REQUEST_WRITE;
} else {
pio_req->direction = REQUEST_READ;
}
size = (uint16_t)pio_req->size;
port = (uint16_t)pio_req->address;
mask = 0xffffffffUL >> (32U - 8U * size);
TRACE_4I(TRACE_VMEXIT_IO_INSTRUCTION,
(uint32_t)port,
(uint32_t)pio_req->direction,
(uint32_t)size,
(uint32_t)cur_context_idx);
/*
* Post-conditions of the loop:
*
* status == 0 : The access has been handled properly.
* status == -EIO : The access spans multiple devices and cannot
* be handled.
* status == -EINVAL : No valid handler found for this access.
*/
for (handler = vm->arch_vm.io_handler;
handler; handler = handler->next) {
if ((port >= (handler->desc.addr + handler->desc.len)) ||
(port + size <= handler->desc.addr)) {
continue;
} else if (!((port >= handler->desc.addr) && ((port + size)
<= (handler->desc.addr + handler->desc.len)))) {
pr_fatal("Err:IO, port 0x%04x, size=%hu spans devices",
port, size);
status = -EIO;
break;
} else {
struct cpu_gp_regs *regs =
&cur_context->guest_cpu_regs.regs;
if (pio_req->direction == REQUEST_WRITE) {
handler->desc.io_write(handler, vm, port, size,
regs->rax);
pr_dbg("IO write on port %04x, data %08x", port,
regs->rax & mask);
} else {
uint32_t data = handler->desc.io_read(handler,
vm, port, size);
regs->rax &= ~mask;
regs->rax |= data & mask;
pr_dbg("IO read on port %04x, data %08x",
port, data);
}
status = 0;
break;
}
}
/* Go for VHM */
if (status == -EINVAL) {
uint64_t rax = cur_context->guest_cpu_regs.regs.rax;
dm_emulate_pio_pre(vcpu, exit_qual, rax);
status = acrn_insert_request_wait(vcpu, io_req);
if (status != 0) {
pr_fatal("Err:IO %s access to port 0x%04x, size=%u",
(pio_req->direction != REQUEST_READ) ? "read" : "write",
port, size);
}
}
return status;
}
static void register_io_handler(struct vm *vm, struct vm_io_handler *hdlr)
{
if (vm->arch_vm.io_handler != NULL) {
hdlr->next = vm->arch_vm.io_handler;
}
vm->arch_vm.io_handler = hdlr;
}
static void empty_io_handler_list(struct vm *vm)
{
struct vm_io_handler *handler = vm->arch_vm.io_handler;
struct vm_io_handler *tmp;
while (handler != NULL) {
tmp = handler;
handler = tmp->next;
free(tmp);
}
vm->arch_vm.io_handler = NULL;
}
void free_io_emulation_resource(struct vm *vm)
{
empty_io_handler_list(vm);
/* Free I/O emulation bitmaps */
free(vm->arch_vm.iobitmap[0]);
free(vm->arch_vm.iobitmap[1]);
}
void allow_guest_io_access(struct vm *vm, uint32_t address_arg, uint32_t nbytes)
{
uint32_t address = address_arg;
uint32_t *b;
uint32_t i;
uint32_t a;
b = vm->arch_vm.iobitmap[0];
for (i = 0U; i < nbytes; i++) {
if ((address & 0x8000U) != 0U) {
b = vm->arch_vm.iobitmap[1];
}
a = address & 0x7fffU;
b[a >> 5] &= ~(1 << (a & 0x1fU));
address++;
}
}
static void deny_guest_io_access(struct vm *vm, uint32_t address_arg, uint32_t nbytes)
{
uint32_t address = address_arg;
uint32_t *b;
uint32_t i;
uint32_t a;
b = vm->arch_vm.iobitmap[0];
for (i = 0U; i < nbytes; i++) {
if ((address & 0x8000U) != 0U) {
b = vm->arch_vm.iobitmap[1];
}
a = address & 0x7fffU;
b[a >> 5U] |= (1U << (a & 0x1fU));
address++;
}
}
static struct vm_io_handler *create_io_handler(uint32_t port, uint32_t len,
io_read_fn_t io_read_fn_ptr,
io_write_fn_t io_write_fn_ptr)
{
struct vm_io_handler *handler;
handler = calloc(1, sizeof(struct vm_io_handler));
if (handler != NULL) {
handler->desc.addr = port;
handler->desc.len = len;
handler->desc.io_read = io_read_fn_ptr;
handler->desc.io_write = io_write_fn_ptr;
} else {
pr_err("Error: out of memory");
}
return handler;
}
void setup_io_bitmap(struct vm *vm)
{
/* Allocate VM architecture state and IO bitmaps A and B */
vm->arch_vm.iobitmap[0] = alloc_page();
vm->arch_vm.iobitmap[1] = alloc_page();
ASSERT((vm->arch_vm.iobitmap[0] != NULL) &&
(vm->arch_vm.iobitmap[1] != NULL), "");
if (is_vm0(vm)) {
(void)memset(vm->arch_vm.iobitmap[0], 0x00, CPU_PAGE_SIZE);
(void)memset(vm->arch_vm.iobitmap[1], 0x00, CPU_PAGE_SIZE);
} else {
/* block all IO port access from Guest */
(void)memset(vm->arch_vm.iobitmap[0], 0xFF, CPU_PAGE_SIZE);
(void)memset(vm->arch_vm.iobitmap[1], 0xFF, CPU_PAGE_SIZE);
}
}
void register_io_emulation_handler(struct vm *vm, struct vm_io_range *range,
io_read_fn_t io_read_fn_ptr,
io_write_fn_t io_write_fn_ptr)
{
struct vm_io_handler *handler = NULL;
if (io_read_fn_ptr == NULL || io_write_fn_ptr == NULL) {
pr_err("Invalid IO handler.");
return;
}
if (is_vm0(vm)) {
deny_guest_io_access(vm, range->base, range->len);
}
handler = create_io_handler(range->base,
range->len, io_read_fn_ptr, io_write_fn_ptr);
register_io_handler(vm, handler);
}