memory.c

这是一个同样来自贝尔实验室的和UNIX有着渊源的操作系统, 其简洁的设计和实现易于我们学习和理解

			}while(pa % MB);
		}
		else{
			nvalid[MemUPA] += MB/BY2PG;
			mapfree(&rmapupa, pa, MB);

			*pte = 0;
			pa += MB;
		}
		/*
		 * Done with this 4MB chunk, review the options:
		 * 1) not physical memory and >=16MB - invalidate the PDB entry;
		 * 2) physical memory - use the 4MB page extension if possible;
		 * 3) not physical memory and <16MB - use the 4MB page extension
		 *    if possible;
		 * 4) mixed or no 4MB page extension - commit the already
		 *    initialised space for the page table.
		 */
		if(pa%(4*MB) == 0 && pa >= 32*MB && nvalid[MemUPA] == (4*MB)/BY2PG){
			/*
			 * If we encounter a 4MB chunk of missing memory
			 * at a sufficiently high offset, call it the end of
			 * memory.  Otherwise we run the risk of thinking
			 * that video memory is real RAM.
			 */
			break;
		}
		if(pa <= maxkpa && pa%(4*MB) == 0){
			table = &m->pdb[PDX(KADDR(pa - 4*MB))];
			if(nvalid[MemUPA] == (4*MB)/BY2PG)
				*table = 0;
			else if(nvalid[MemRAM] == (4*MB)/BY2PG && (m->cpuiddx & 0x08))
				*table = (pa - 4*MB)|PTESIZE|PTEWRITE|PTEVALID;
			else if(nvalid[MemUMB] == (4*MB)/BY2PG && (m->cpuiddx & 0x08))
				*table = (pa - 4*MB)|PTESIZE|PTEWRITE|PTEUNCACHED|PTEVALID;
			else{
				*table = map|PTEWRITE|PTEVALID;
				map = 0;
			}
		}
		mmuflushtlb(PADDR(m->pdb));
		x += 0x3141526;
	}
	/*
	 * If we didn't reach the end of the 4MB chunk, that part won't
	 * be mapped.  Commit the already initialised space for the page table.
	 */
	if(pa % (4*MB) && pa <= maxkpa){
		m->pdb[PDX(KADDR(pa))] = map|PTEWRITE|PTEVALID;
		map = 0;
	}
	if(map)
		mapfree(&rmapram, map, BY2PG);

	m->pdb[PDX(vbase)] = 0;
	mmuflushtlb(PADDR(m->pdb));

	mapfree(&rmapupa, pa, (u32int)-pa);
	*k0 = kzero;
}

/*
 * BIOS Int 0x15 E820 memory map.
 */
enum
{
	SMAP = ('S'<<24)|('M'<<16)|('A'<<8)|'P',
	Ememory = 1,
	Ereserved = 2,
	Carry = 1,
};

typedef struct Emap Emap;
struct Emap
{
	uvlong base;
	uvlong len;
	ulong type;
};
static Emap emap[16];
int nemap;

static char *etypes[] =
{
	"type=0",
	"memory",
	"reserved",
	"acpi reclaim",
	"acpi nvs",
};

static int
emapcmp(const void *va, const void *vb)
{
	Emap *a, *b;
	
	a = (Emap*)va;
	b = (Emap*)vb;
	if(a->base < b->base)
		return -1;
	if(a->base > b->base)
		return 1;
	if(a->len < b->len)
		return -1;
	if(a->len > b->len)
		return 1;
	return a->type - b->type;
}

static void
map(ulong base, ulong len, int type)
{
	ulong e, n;
	ulong *table, flags, maxkpa;
	
	/*
	 * Split any call crossing 4*MB to make below simpler.
	 */
	if(base < 4*MB && len > 4*MB-base){
		n = 4*MB - base;
		map(base, n, type);
		map(4*MB, len-n, type);
	}
	
	/*
	 * Let lowraminit and umbscan hash out the low 4MB.
	 */
	if(base < 4*MB)
		return;

	/*
	 * Any non-memory below 16*MB is used as upper mem blocks.
	 */
	if(type == MemUPA && base < 16*MB && base+len > 16*MB){
		map(base, 16*MB-base, MemUMB);
		map(16*MB, len-(16*MB-base), MemUPA);
		return;
	}
	
	/*
	 * Memory below CPU0MACH is reserved for the kernel
	 * and already mapped.
	 */
	if(base < PADDR(CPU0MACH)+BY2PG){
		n = PADDR(CPU0MACH)+BY2PG - base;
		if(len <= n)
			return;
		map(PADDR(CPU0MACH), len-n, type);
		return;
	}
	
	/*
	 * Memory between KTZERO and end is the kernel itself
	 * and is already mapped.
	 */
	if(base < PADDR(KTZERO) && base+len > PADDR(KTZERO)){
		map(base, PADDR(KTZERO)-base, type);
		return;
	}
	if(PADDR(KTZERO) < base && base < PADDR(PGROUND((ulong)end))){
		n = PADDR(PGROUND((ulong)end));
		if(len <= n)
			return;
		map(PADDR(PGROUND((ulong)end)), len-n, type);
		return;
	}
	
	/*
	 * Now we have a simple case.
	 */
	// print("map %.8lux %.8lux %d\n", base, base+len, type);
	switch(type){
	case MemRAM:
		mapfree(&rmapram, base, len);
		flags = PTEWRITE|PTEVALID;
		break;
	case MemUMB:
		mapfree(&rmapumb, base, len);
		flags = PTEWRITE|PTEUNCACHED|PTEVALID;
		break;
	case MemUPA:
		mapfree(&rmapupa, base, len);
		flags = 0;
		break;
	default:
	case MemReserved:
		flags = 0;
		break;
	}
	
	/*
	 * bottom 4MB is already mapped - just twiddle flags.
	 * (not currently used - see above)
	 */
	if(base < 4*MB){
		table = KADDR(PPN(m->pdb[PDX(base)]));
		e = base+len;
		base = PPN(base);
		for(; base<e; base+=BY2PG)
			table[PTX(base)] |= flags;
		return;
	}
	
	/*
	 * Only map from KZERO to 2^32.
	 */
	if(flags){
		maxkpa = -KZERO;
		if(base >= maxkpa)
			return;
		if(len > maxkpa-base)
			len = maxkpa - base;
		pdbmap(m->pdb, base|flags, base+KZERO, len);
	}
}

static int
e820scan(void)
{
	int i;
	Ureg u;
	ulong cont, base, len;
	uvlong last;
	Emap *e;

	if(getconf("*norealmode") || getconf("*noe820scan"))
		return -1;

	cont = 0;
	for(i=0; i<nelem(emap); i++){
		memset(&u, 0, sizeof u);
		u.ax = 0xE820;
		u.bx = cont;
		u.cx = 20;
		u.dx = SMAP;
		u.es = (PADDR(RMBUF)>>4)&0xF000;
		u.di = PADDR(RMBUF)&0xFFFF;
		u.trap = 0x15;
		realmode(&u);
		cont = u.bx;
		if((u.flags&Carry) || u.ax != SMAP || u.cx != 20)
			break;
		e = &emap[nemap++];
		*e = *(Emap*)RMBUF;
		if(u.bx == 0)
			break;
	}
	if(nemap == 0)
		return -1;
	
	qsort(emap, nemap, sizeof emap[0], emapcmp);
	
	for(i=0; i<nemap; i++){
		e = &emap[i];
		print("E820: %.8llux %.8llux ", e->base, e->base+e->len);
		if(e->type < nelem(etypes))
			print("%s\n", etypes[e->type]);
		else
			print("type=%lud\n", e->type);
	}

	last = 0;
	for(i=0; i<nemap; i++){	
		e = &emap[i];
		/*
		 * pull out the info but only about the low 32 bits...
		 */
		if(e->base >= (1LL<<32))
			break;
		base = e->base;
		if(base+e->len > (1LL<<32))
			len = -base;
		else
			len = e->len;
		/*
		 * If the map skips addresses, mark them available.
		 */
		if(last < e->base)
			map(last, e->base-last, MemUPA);
		last = base+len;
		if(e->type == Ememory)
			map(base, len, MemRAM);
		else
			map(base, len, MemReserved);
	}
	if(last < (1LL<<32))
		map(last, (u32int)-last, MemUPA);
	return 0;
}

void
meminit(void)
{
	int i;
	Map *mp;
	Confmem *cm;
	ulong pa, *pte;
	ulong maxmem, lost;
	char *p;

	if(p = getconf("*maxmem"))
		maxmem = strtoul(p, 0, 0);
	else
		maxmem = 0;

	/*
	 * Set special attributes for memory between 640KB and 1MB:
	 *   VGA memory is writethrough;
	 *   BIOS ROM's/UMB's are uncached;
	 * then scan for useful memory.
	 */
	for(pa = 0xA0000; pa < 0xC0000; pa += BY2PG){
		pte = mmuwalk(m->pdb, (ulong)KADDR(pa), 2, 0);
		*pte |= PTEWT;
	}
	for(pa = 0xC0000; pa < 0x100000; pa += BY2PG){
		pte = mmuwalk(m->pdb, (ulong)KADDR(pa), 2, 0);
		*pte |= PTEUNCACHED;
	}
	mmuflushtlb(PADDR(m->pdb));

	umbscan();
	lowraminit();
	if(e820scan() < 0)
		ramscan(maxmem);

	/*
	 * Set the conf entries describing banks of allocatable memory.
	 */
	for(i=0; i<nelem(mapram) && i<nelem(conf.mem); i++){
		mp = &rmapram.map[i];
		cm = &conf.mem[i];
		cm->base = mp->addr;
		cm->npage = mp->size/BY2PG;
	}
	
	lost = 0;
	for(; i<nelem(mapram); i++)
		lost += rmapram.map[i].size;
	if(lost)
		print("meminit - lost %lud bytes\n", lost);

	if(MEMDEBUG)
		memdebug();
}

/*
 * Allocate memory from the upper memory blocks.
 */
ulong
umbmalloc(ulong addr, int size, int align)
{
	ulong a;

	if(a = mapalloc(&rmapumb, addr, size, align))
		return (ulong)KADDR(a);

	return 0;
}

void
umbfree(ulong addr, int size)
{
	mapfree(&rmapumb, PADDR(addr), size);
}

ulong
umbrwmalloc(ulong addr, int size, int align)
{
	ulong a;
	uchar *p;

	if(a = mapalloc(&rmapumbrw, addr, size, align))
		return(ulong)KADDR(a);

	/*
	 * Perhaps the memory wasn't visible before
	 * the interface is initialised, so try again.
	 */
	if((a = umbmalloc(addr, size, align)) == 0)
		return 0;
	p = (uchar*)a;
	p[0] = 0xCC;
	p[size-1] = 0xCC;
	if(p[0] == 0xCC && p[size-1] == 0xCC)
		return a;
	umbfree(a, size);

	return 0;
}

void
umbrwfree(ulong addr, int size)
{
	mapfree(&rmapumbrw, PADDR(addr), size);
}

/*
 * Give out otherwise-unused physical address space
 * for use in configuring devices.  Note that unlike upamalloc
 * before it, upaalloc does not map the physical address
 * into virtual memory.  Call vmap to do that.
 */
ulong
upaalloc(int size, int align)
{
	ulong a;

	a = mapalloc(&rmapupa, 0, size, align);
	if(a == 0){
		print("out of physical address space allocating %d\n", size);
		mapprint(&rmapupa);
	}
	return a;
}

void
upafree(ulong pa, int size)
{
	mapfree(&rmapupa, pa, size);
}

void
upareserve(ulong pa, int size)
{
	ulong a;
	
	a = mapalloc(&rmapupa, pa, size, 0);
	if(a != pa){
		/*
		 * This can happen when we're using the E820
		 * map, which might have already reserved some
		 * of the regions claimed by the pci devices.
		 */
	//	print("upareserve: cannot reserve pa=%#.8lux size=%d\n", pa, size);
		if(a != 0)
			mapfree(&rmapupa, a, size);
	}
}

void
memorysummary(void)
{
	memdebug();
}