minix/servers/vm/utility.c
Ben Gras 50e2064049 No more intel/minix segments.
This commit removes all traces of Minix segments (the text/data/stack
memory map abstraction in the kernel) and significance of Intel segments
(hardware segments like CS, DS that add offsets to all addressing before
page table translation). This ultimately simplifies the memory layout
and addressing and makes the same layout possible on non-Intel
architectures.

There are only two types of addresses in the world now: virtual
and physical; even the kernel and processes have the same virtual
address space. Kernel and user processes can be distinguished at a
glance as processes won't use 0xF0000000 and above.

No static pre-allocated memory sizes exist any more.

Changes to booting:
        . The pre_init.c leaves the kernel and modules exactly as
          they were left by the bootloader in physical memory
        . The kernel starts running using physical addressing,
          loaded at a fixed location given in its linker script by the
          bootloader.  All code and data in this phase are linked to
          this fixed low location.
        . It makes a bootstrap pagetable to map itself to a
          fixed high location (also in linker script) and jumps to
          the high address. All code and data then use this high addressing.
        . All code/data symbols linked at the low addresses is prefixed by
          an objcopy step with __k_unpaged_*, so that that code cannot
          reference highly-linked symbols (which aren't valid yet) or vice
          versa (symbols that aren't valid any more).
        . The two addressing modes are separated in the linker script by
          collecting the unpaged_*.o objects and linking them with low
          addresses, and linking the rest high. Some objects are linked
          twice, once low and once high.
        . The bootstrap phase passes a lot of information (e.g. free memory
          list, physical location of the modules, etc.) using the kinfo
          struct.
        . After this bootstrap the low-linked part is freed.
        . The kernel maps in VM into the bootstrap page table so that VM can
          begin executing. Its first job is to make page tables for all other
          boot processes. So VM runs before RS, and RS gets a fully dynamic,
          VM-managed address space. VM gets its privilege info from RS as usual
          but that happens after RS starts running.
        . Both the kernel loading VM and VM organizing boot processes happen
	  using the libexec logic. This removes the last reason for VM to
	  still know much about exec() and vm/exec.c is gone.

Further Implementation:
        . All segments are based at 0 and have a 4 GB limit.
        . The kernel is mapped in at the top of the virtual address
          space so as not to constrain the user processes.
        . Processes do not use segments from the LDT at all; there are
          no segments in the LDT any more, so no LLDT is needed.
        . The Minix segments T/D/S are gone and so none of the
          user-space or in-kernel copy functions use them. The copy
          functions use a process endpoint of NONE to realize it's
          a physical address, virtual otherwise.
        . The umap call only makes sense to translate a virtual address
          to a physical address now.
        . Segments-related calls like newmap and alloc_segments are gone.
        . All segments-related translation in VM is gone (vir2map etc).
        . Initialization in VM is simpler as no moving around is necessary.
        . VM and all other boot processes can be linked wherever they wish
          and will be mapped in at the right location by the kernel and VM
          respectively.

Other changes:
        . The multiboot code is less special: it does not use mb_print
          for its diagnostics any more but uses printf() as normal, saving
          the output into the diagnostics buffer, only printing to the
          screen using the direct print functions if a panic() occurs.
        . The multiboot code uses the flexible 'free memory map list'
          style to receive the list of free memory if available.
        . The kernel determines the memory layout of the processes to
          a degree: it tells VM where the kernel starts and ends and
          where the kernel wants the top of the process to be. VM then
          uses this entire range, i.e. the stack is right at the top,
          and mmap()ped bits of memory are placed below that downwards,
          and the break grows upwards.

Other Consequences:
        . Every process gets its own page table as address spaces
          can't be separated any more by segments.
        . As all segments are 0-based, there is no distinction between
          virtual and linear addresses, nor between userspace and
          kernel addresses.
        . Less work is done when context switching, leading to a net
          performance increase. (8% faster on my machine for 'make servers'.)
	. The layout and configuration of the GDT makes sysenter and syscall
	  possible.
2012-07-15 22:30:15 +02:00

332 lines
9.5 KiB
C

/* This file contains some utility routines for VM. */
#define _SYSTEM 1
#define _MINIX 1 /* To get the brk() prototype (as _brk()). */
#include <minix/callnr.h>
#include <minix/com.h>
#include <minix/config.h>
#include <minix/const.h>
#include <minix/ds.h>
#include <minix/endpoint.h>
#include <minix/minlib.h>
#include <minix/type.h>
#include <minix/ipc.h>
#include <minix/sysutil.h>
#include <minix/syslib.h>
#include <minix/type.h>
#include <minix/bitmap.h>
#include <string.h>
#include <errno.h>
#include <env.h>
#include <unistd.h>
#include <memory.h>
#include <assert.h>
#include <sys/param.h>
#include "proto.h"
#include "glo.h"
#include "util.h"
#include "region.h"
#include "sanitycheck.h"
#include <machine/archtypes.h>
#include "kernel/const.h"
#include "kernel/config.h"
#include "kernel/type.h"
#include "kernel/proc.h"
/*===========================================================================*
* get_mem_chunks *
*===========================================================================*/
void get_mem_chunks(mem_chunks)
struct memory *mem_chunks; /* store mem chunks here */
{
/* Initialize the free memory list from the 'memory' boot variable. Translate
* the byte offsets and sizes in this list to clicks, properly truncated.
*/
phys_bytes base, size, limit;
int i;
struct memory *memp;
/* Obtain and parse memory from system environment. */
if(env_memory_parse(mem_chunks, NR_MEMS) != OK)
panic("couldn't obtain memory chunks");
/* Round physical memory to clicks. Round start up, round end down. */
for (i = 0; i < NR_MEMS; i++) {
memp = &mem_chunks[i]; /* next mem chunk is stored here */
base = mem_chunks[i].base;
size = mem_chunks[i].size;
limit = base + size;
base = (phys_bytes) (CLICK_CEIL(base));
limit = (phys_bytes) (CLICK_FLOOR(limit));
if (limit <= base) {
memp->base = memp->size = 0;
} else {
memp->base = base >> CLICK_SHIFT;
memp->size = (limit - base) >> CLICK_SHIFT;
}
}
}
#if 0
/*===========================================================================*
* reserve_proc_mem *
*===========================================================================*/
void reserve_proc_mem(mem_chunks, map_ptr)
struct memory *mem_chunks; /* store mem chunks here */
struct mem_map *map_ptr; /* memory to remove */
{
/* Remove server memory from the free memory list.
*/
struct memory *memp;
for (memp = mem_chunks; memp < &mem_chunks[NR_MEMS]; memp++) {
if(memp->base <= map_ptr[T].mem_phys
&& memp->base+memp->size >= map_ptr[T].mem_phys)
{
phys_bytes progsz = map_ptr[S].mem_phys
- map_ptr[T].mem_phys;
phys_bytes progend = map_ptr[S].mem_phys;
if (memp->base == map_ptr[T].mem_phys) {
memp->base += progsz;
memp->size -= progsz;
} else {
struct memory *mempr;
/* have to split mem_chunks */
if(mem_chunks[NR_MEMS-1].size>0)
panic("reserve_proc_mem: can't find free mem_chunks to map: 0x%lx",
map_ptr[T].mem_phys);
for(mempr=&mem_chunks[NR_MEMS-1];mempr>memp;mempr--) {
*mempr=*(mempr-1);
}
assert(memp < &mem_chunks[NR_MEMS-1]);
(memp+1)->base = progend;
(memp+1)->size = memp->base + memp->size
- progend;
memp->size = map_ptr[T].mem_phys - memp->base;
}
break;
}
}
if (memp >= &mem_chunks[NR_MEMS]) {
panic("reserve_proc_mem: can't find map in mem_chunks: 0x%lx",
map_ptr[T].mem_phys);
}
}
#endif
/*===========================================================================*
* vm_isokendpt *
*===========================================================================*/
int vm_isokendpt(endpoint_t endpoint, int *proc)
{
*proc = _ENDPOINT_P(endpoint);
if(*proc < 0 || *proc >= NR_PROCS)
return EINVAL;
if(*proc >= 0 && endpoint != vmproc[*proc].vm_endpoint)
return EDEADEPT;
if(*proc >= 0 && !(vmproc[*proc].vm_flags & VMF_INUSE))
return EDEADEPT;
return OK;
}
struct proc mytmpproc;
/*===========================================================================*
* get_stack_ptr *
*===========================================================================*/
int get_stack_ptr(proc_nr_e, sp)
int proc_nr_e; /* process to get sp of */
vir_bytes *sp; /* put stack pointer here */
{
int s;
if ((s=sys_getproc(&mytmpproc, proc_nr_e)) != OK)
return(s);
*sp = mytmpproc.p_reg.sp;
return(OK);
}
/*===========================================================================*
* do_info *
*===========================================================================*/
int do_info(message *m)
{
struct vm_stats_info vsi;
struct vm_usage_info vui;
static struct vm_region_info vri[MAX_VRI_COUNT];
struct vmproc *vmp;
vir_bytes addr, size, next, ptr;
int r, pr, dummy, count, free_pages, largest_contig;
if (vm_isokendpt(m->m_source, &pr) != OK)
return EINVAL;
vmp = &vmproc[pr];
ptr = (vir_bytes) m->VMI_PTR;
switch(m->VMI_WHAT) {
case VMIW_STATS:
vsi.vsi_pagesize = VM_PAGE_SIZE;
vsi.vsi_total = total_pages;
memstats(&dummy, &free_pages, &largest_contig);
vsi.vsi_free = free_pages;
vsi.vsi_largest = largest_contig;
get_stats_info(&vsi);
addr = (vir_bytes) &vsi;
size = sizeof(vsi);
break;
case VMIW_USAGE:
if (vm_isokendpt(m->VMI_EP, &pr) != OK)
return EINVAL;
get_usage_info(&vmproc[pr], &vui);
addr = (vir_bytes) &vui;
size = sizeof(vui);
break;
case VMIW_REGION:
if (vm_isokendpt(m->VMI_EP, &pr) != OK)
return EINVAL;
count = MIN(m->VMI_COUNT, MAX_VRI_COUNT);
next = m->VMI_NEXT;
count = get_region_info(&vmproc[pr], vri, count, &next);
m->VMI_COUNT = count;
m->VMI_NEXT = next;
addr = (vir_bytes) vri;
size = sizeof(vri[0]) * count;
break;
default:
return EINVAL;
}
if (size == 0)
return OK;
/* Make sure that no page faults can occur while copying out. A page
* fault would cause the kernel to send a notify to us, while we would
* be waiting for the result of the copy system call, resulting in a
* deadlock. Note that no memory mapping can be undone without the
* involvement of VM, so we are safe until we're done.
*/
r = handle_memory(vmp, ptr, size, 1 /*wrflag*/);
if (r != OK) return r;
/* Now that we know the copy out will succeed, perform the actual copy
* operation.
*/
return sys_datacopy(SELF, addr,
(vir_bytes) vmp->vm_endpoint, ptr, size);
}
/*===========================================================================*
* swap_proc_slot *
*===========================================================================*/
int swap_proc_slot(struct vmproc *src_vmp, struct vmproc *dst_vmp)
{
struct vmproc orig_src_vmproc, orig_dst_vmproc;
#if LU_DEBUG
printf("VM: swap_proc: swapping %d (%d) and %d (%d)\n",
src_vmp->vm_endpoint, src_vmp->vm_slot,
dst_vmp->vm_endpoint, dst_vmp->vm_slot);
#endif
/* Save existing data. */
orig_src_vmproc = *src_vmp;
orig_dst_vmproc = *dst_vmp;
/* Swap slots. */
*src_vmp = orig_dst_vmproc;
*dst_vmp = orig_src_vmproc;
/* Preserve endpoints and slot numbers. */
src_vmp->vm_endpoint = orig_src_vmproc.vm_endpoint;
src_vmp->vm_slot = orig_src_vmproc.vm_slot;
dst_vmp->vm_endpoint = orig_dst_vmproc.vm_endpoint;
dst_vmp->vm_slot = orig_dst_vmproc.vm_slot;
#if LU_DEBUG
printf("VM: swap_proc: swapped %d (%d) and %d (%d)\n",
src_vmp->vm_endpoint, src_vmp->vm_slot,
dst_vmp->vm_endpoint, dst_vmp->vm_slot);
#endif
return OK;
}
/*===========================================================================*
* swap_proc_dyn_data *
*===========================================================================*/
int swap_proc_dyn_data(struct vmproc *src_vmp, struct vmproc *dst_vmp)
{
int is_vm;
int r;
is_vm = (dst_vmp->vm_endpoint == VM_PROC_NR);
/* For VM, transfer memory regions above the stack first. */
if(is_vm) {
#if LU_DEBUG
printf("VM: swap_proc_dyn_data: tranferring regions above the stack from old VM (%d) to new VM (%d)\n",
src_vmp->vm_endpoint, dst_vmp->vm_endpoint);
#endif
r = pt_map_in_range(src_vmp, dst_vmp, VM_STACKTOP, 0);
if(r != OK) {
printf("swap_proc_dyn_data: pt_map_in_range failed\n");
return r;
}
}
#if LU_DEBUG
printf("VM: swap_proc_dyn_data: swapping regions' parents for %d (%d) and %d (%d)\n",
src_vmp->vm_endpoint, src_vmp->vm_slot,
dst_vmp->vm_endpoint, dst_vmp->vm_slot);
#endif
/* Swap vir_regions' parents. */
map_setparent(src_vmp);
map_setparent(dst_vmp);
/* For regular processes, transfer regions above the stack now.
* In case of rollback, we need to skip this step. To sandbox the
* new instance and prevent state corruption on rollback, we share all
* the regions between the two instances as COW.
*/
if(!is_vm) {
struct vir_region *vr;
vr = map_lookup(dst_vmp, VM_STACKTOP);
if(vr && !map_lookup(src_vmp, VM_STACKTOP)) {
#if LU_DEBUG
printf("VM: swap_proc_dyn_data: tranferring regions above the stack from %d to %d\n",
src_vmp->vm_endpoint, dst_vmp->vm_endpoint);
#endif
r = map_proc_copy_from(src_vmp, dst_vmp, vr);
if(r != OK) {
return r;
}
}
}
return OK;
}