[deliverable/linux.git] / arch / powerpc / mm / hugepage-hash64.c

/*
 * Copyright IBM Corporation, 2013
 * Author Aneesh Kumar K.V <aneesh.kumar@linux.vnet.ibm.com>
 *
 * This program is free software; you can redistribute it and/or modify it
 * under the terms of version 2.1 of the GNU Lesser General Public License
 * as published by the Free Software Foundation.
 *
 * This program is distributed in the hope that it would be useful, but
 * WITHOUT ANY WARRANTY; without even the implied warranty of
 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.
 *
 */

/*
 * PPC64 THP Support for hash based MMUs
 */
#include <linux/mm.h>
#include <asm/machdep.h>

int __hash_page_thp(unsigned long ea, unsigned long access, unsigned long vsid,
		    pmd_t *pmdp, unsigned long trap, int local, int ssize,
		    unsigned int psize)
{
	unsigned int index, valid;
	unsigned char *hpte_slot_array;
	unsigned long rflags, pa, hidx;
	unsigned long old_pmd, new_pmd;
	int ret, lpsize = MMU_PAGE_16M;
	unsigned long vpn, hash, shift, slot;

	/*
	 * atomically mark the linux large page PMD busy and dirty
	 */
	do {
		old_pmd = pmd_val(*pmdp);
		/* If PMD busy, retry the access */
		if (unlikely(old_pmd & _PAGE_BUSY))
			return 0;
		/* If PMD is trans splitting retry the access */
		if (unlikely(old_pmd & _PAGE_SPLITTING))
			return 0;
		/* If PMD permissions don't match, take page fault */
		if (unlikely(access & ~old_pmd))
			return 1;
		/*
		 * Try to lock the PTE, add ACCESSED and DIRTY if it was
		 * a write access
		 */
		new_pmd = old_pmd | _PAGE_BUSY | _PAGE_ACCESSED;
		if (access & _PAGE_RW)
			new_pmd |= _PAGE_DIRTY;
	} while (old_pmd != __cmpxchg_u64((unsigned long *)pmdp,
					  old_pmd, new_pmd));
	/*
	 * PP bits. _PAGE_USER is already PP bit 0x2, so we only
	 * need to add in 0x1 if it's a read-only user page
	 */
	rflags = new_pmd & _PAGE_USER;
	if ((new_pmd & _PAGE_USER) && !((new_pmd & _PAGE_RW) &&
					   (new_pmd & _PAGE_DIRTY)))
		rflags |= 0x1;
	/*
	 * _PAGE_EXEC -> HW_NO_EXEC since it's inverted
	 */
	rflags |= ((new_pmd & _PAGE_EXEC) ? 0 : HPTE_R_N);

#if 0
	if (!cpu_has_feature(CPU_FTR_COHERENT_ICACHE)) {

		/*
		 * No CPU has hugepages but lacks no execute, so we
		 * don't need to worry about that case
		 */
		rflags = hash_page_do_lazy_icache(rflags, __pte(old_pte), trap);
	}
#endif
	/*
	 * Find the slot index details for this ea, using base page size.
	 */
	shift = mmu_psize_defs[psize].shift;
	index = (ea & ~HPAGE_PMD_MASK) >> shift;
	BUG_ON(index >= 4096);

	vpn = hpt_vpn(ea, vsid, ssize);
	hash = hpt_hash(vpn, shift, ssize);
	hpte_slot_array = get_hpte_slot_array(pmdp);

	valid = hpte_valid(hpte_slot_array, index);
	if (valid) {
		/* update the hpte bits */
		hidx =  hpte_hash_index(hpte_slot_array, index);
		if (hidx & _PTEIDX_SECONDARY)
			hash = ~hash;
		slot = (hash & htab_hash_mask) * HPTES_PER_GROUP;
		slot += hidx & _PTEIDX_GROUP_IX;

		ret = ppc_md.hpte_updatepp(slot, rflags, vpn,
					   psize, lpsize, ssize, local);
		/*
		 * We failed to update, try to insert a new entry.
		 */
		if (ret == -1) {
			/*
			 * large pte is marked busy, so we can be sure
			 * nobody is looking at hpte_slot_array. hence we can
			 * safely update this here.
			 */
			valid = 0;
			new_pmd &= ~_PAGE_HPTEFLAGS;
			hpte_slot_array[index] = 0;
		} else
			/* clear the busy bits and set the hash pte bits */
			new_pmd = (new_pmd & ~_PAGE_HPTEFLAGS) | _PAGE_HASHPTE;
	}

	if (!valid) {
		unsigned long hpte_group;

		/* insert new entry */
		pa = pmd_pfn(__pmd(old_pmd)) << PAGE_SHIFT;
repeat:
		hpte_group = ((hash & htab_hash_mask) * HPTES_PER_GROUP) & ~0x7UL;

		/* clear the busy bits and set the hash pte bits */
		new_pmd = (new_pmd & ~_PAGE_HPTEFLAGS) | _PAGE_HASHPTE;

		/* Add in WIMG bits */
		rflags |= (new_pmd & (_PAGE_WRITETHRU | _PAGE_NO_CACHE |
				      _PAGE_GUARDED));
		/*
		 * enable the memory coherence always
		 */
		rflags |= HPTE_R_M;

		/* Insert into the hash table, primary slot */
		slot = ppc_md.hpte_insert(hpte_group, vpn, pa, rflags, 0,
					  psize, lpsize, ssize);
		/*
		 * Primary is full, try the secondary
		 */
		if (unlikely(slot == -1)) {
			hpte_group = ((~hash & htab_hash_mask) *
				      HPTES_PER_GROUP) & ~0x7UL;
			slot = ppc_md.hpte_insert(hpte_group, vpn, pa,
						  rflags, HPTE_V_SECONDARY,
						  psize, lpsize, ssize);
			if (slot == -1) {
				if (mftb() & 0x1)
					hpte_group = ((hash & htab_hash_mask) *
						      HPTES_PER_GROUP) & ~0x7UL;

				ppc_md.hpte_remove(hpte_group);
				goto repeat;
			}
		}
		/*
		 * Hypervisor failure. Restore old pmd and return -1
		 * similar to __hash_page_*
		 */
		if (unlikely(slot == -2)) {
			*pmdp = __pmd(old_pmd);
			hash_failure_debug(ea, access, vsid, trap, ssize,
					   psize, lpsize, old_pmd);
			return -1;
		}
		/*
		 * large pte is marked busy, so we can be sure
		 * nobody is looking at hpte_slot_array. hence we can
		 * safely update this here.
		 */
		mark_hpte_slot_valid(hpte_slot_array, index, slot);
	}
	/*
	 * No need to use ldarx/stdcx here
	 */
	*pmdp = __pmd(new_pmd & ~_PAGE_BUSY);
	return 0;
}
Commit	Line	Data
6d492ecc AK	1	/*
	2	* Copyright IBM Corporation, 2013
	3	* Author Aneesh Kumar K.V <aneesh.kumar@linux.vnet.ibm.com>
	4	*
	5	* This program is free software; you can redistribute it and/or modify it
	6	* under the terms of version 2.1 of the GNU Lesser General Public License
	7	* as published by the Free Software Foundation.
	8	*
	9	* This program is distributed in the hope that it would be useful, but
	10	* WITHOUT ANY WARRANTY; without even the implied warranty of
	11	* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.
	12	*
	13	*/
	14
	15	/*
	16	* PPC64 THP Support for hash based MMUs
	17	*/
	18	#include <linux/mm.h>
	19	#include <asm/machdep.h>
	20
	21	int __hash_page_thp(unsigned long ea, unsigned long access, unsigned long vsid,
	22	pmd_t *pmdp, unsigned long trap, int local, int ssize,
	23	unsigned int psize)
	24	{
	25	unsigned int index, valid;
	26	unsigned char *hpte_slot_array;
	27	unsigned long rflags, pa, hidx;
	28	unsigned long old_pmd, new_pmd;
	29	int ret, lpsize = MMU_PAGE_16M;
	30	unsigned long vpn, hash, shift, slot;
	31
	32	/*
	33	* atomically mark the linux large page PMD busy and dirty
	34	*/
	35	do {
	36	old_pmd = pmd_val(*pmdp);
	37	/* If PMD busy, retry the access */
	38	if (unlikely(old_pmd & _PAGE_BUSY))
	39	return 0;
0ac52dd7 AK	40	/* If PMD is trans splitting retry the access */
	41	if (unlikely(old_pmd & _PAGE_SPLITTING))
	42	return 0;
6d492ecc AK	43	/* If PMD permissions don't match, take page fault */
	44	if (unlikely(access & ~old_pmd))
	45	return 1;
	46	/*
	47	* Try to lock the PTE, add ACCESSED and DIRTY if it was
	48	* a write access
	49	*/
	50	new_pmd = old_pmd \| _PAGE_BUSY \| _PAGE_ACCESSED;
	51	if (access & _PAGE_RW)
	52	new_pmd \|= _PAGE_DIRTY;
	53	} while (old_pmd != __cmpxchg_u64((unsigned long *)pmdp,
	54	old_pmd, new_pmd));
	55	/*
	56	* PP bits. _PAGE_USER is already PP bit 0x2, so we only
	57	* need to add in 0x1 if it's a read-only user page
	58	*/
	59	rflags = new_pmd & _PAGE_USER;
	60	if ((new_pmd & _PAGE_USER) && !((new_pmd & _PAGE_RW) &&
	61	(new_pmd & _PAGE_DIRTY)))
	62	rflags \|= 0x1;
	63	/*
	64	* _PAGE_EXEC -> HW_NO_EXEC since it's inverted
	65	*/
	66	rflags \|= ((new_pmd & _PAGE_EXEC) ? 0 : HPTE_R_N);
	67
	68	#if 0
	69	if (!cpu_has_feature(CPU_FTR_COHERENT_ICACHE)) {
	70
	71	/*
	72	* No CPU has hugepages but lacks no execute, so we
	73	* don't need to worry about that case
	74	*/
	75	rflags = hash_page_do_lazy_icache(rflags, __pte(old_pte), trap);
	76	}
	77	#endif
	78	/*
	79	* Find the slot index details for this ea, using base page size.
	80	*/
	81	shift = mmu_psize_defs[psize].shift;
	82	index = (ea & ~HPAGE_PMD_MASK) >> shift;
	83	BUG_ON(index >= 4096);
	84
	85	vpn = hpt_vpn(ea, vsid, ssize);
	86	hash = hpt_hash(vpn, shift, ssize);
	87	hpte_slot_array = get_hpte_slot_array(pmdp);
	88
	89	valid = hpte_valid(hpte_slot_array, index);
	90	if (valid) {
	91	/* update the hpte bits */
	92	hidx = hpte_hash_index(hpte_slot_array, index);
	93	if (hidx & _PTEIDX_SECONDARY)
	94	hash = ~hash;
	95	slot = (hash & htab_hash_mask) * HPTES_PER_GROUP;
	96	slot += hidx & _PTEIDX_GROUP_IX;
	97
	98	ret = ppc_md.hpte_updatepp(slot, rflags, vpn,
	99	psize, lpsize, ssize, local);
	100	/*
	101	* We failed to update, try to insert a new entry.
	102	*/
	103	if (ret == -1) {
	104	/*
	105	* large pte is marked busy, so we can be sure
	106	* nobody is looking at hpte_slot_array. hence we can
107	* safely update this here.
108	*/
109	valid = 0;
110	new_pmd &= ~_PAGE_HPTEFLAGS;
111	hpte_slot_array[index] = 0;
112	} else
113	/* clear the busy bits and set the hash pte bits */
114	new_pmd = (new_pmd & ~_PAGE_HPTEFLAGS) \| _PAGE_HASHPTE;
115	}
116
117	if (!valid) {
118	unsigned long hpte_group;
119
120	/* insert new entry */
121	pa = pmd_pfn(__pmd(old_pmd)) << PAGE_SHIFT;
122	repeat:
123	hpte_group = ((hash & htab_hash_mask) * HPTES_PER_GROUP) & ~0x7UL;
124
125	/* clear the busy bits and set the hash pte bits */
126	new_pmd = (new_pmd & ~_PAGE_HPTEFLAGS) \| _PAGE_HASHPTE;
127
128	/* Add in WIMG bits */
129	rflags \|= (new_pmd & (_PAGE_WRITETHRU \| _PAGE_NO_CACHE \|
c8c06f5a AK	130	_PAGE_GUARDED));
	131	/*
	132	* enable the memory coherence always
	133	*/
	134	rflags \|= HPTE_R_M;
6d492ecc AK	135
	136	/* Insert into the hash table, primary slot */
	137	slot = ppc_md.hpte_insert(hpte_group, vpn, pa, rflags, 0,
	138	psize, lpsize, ssize);
	139	/*
	140	* Primary is full, try the secondary
	141	*/
	142	if (unlikely(slot == -1)) {
	143	hpte_group = ((~hash & htab_hash_mask) *
	144	HPTES_PER_GROUP) & ~0x7UL;
	145	slot = ppc_md.hpte_insert(hpte_group, vpn, pa,
	146	rflags, HPTE_V_SECONDARY,
	147	psize, lpsize, ssize);
	148	if (slot == -1) {
	149	if (mftb() & 0x1)
	150	hpte_group = ((hash & htab_hash_mask) *
	151	HPTES_PER_GROUP) & ~0x7UL;
	152
	153	ppc_md.hpte_remove(hpte_group);
	154	goto repeat;
	155	}
	156	}
	157	/*
	158	* Hypervisor failure. Restore old pmd and return -1
	159	* similar to __hash_page_*
	160	*/
	161	if (unlikely(slot == -2)) {
	162	*pmdp = __pmd(old_pmd);
	163	hash_failure_debug(ea, access, vsid, trap, ssize,
	164	psize, lpsize, old_pmd);
	165	return -1;
	166	}
	167	/*
	168	* large pte is marked busy, so we can be sure
	169	* nobody is looking at hpte_slot_array. hence we can
	170	* safely update this here.
	171	*/
	172	mark_hpte_slot_valid(hpte_slot_array, index, slot);
	173	}
	174	/*
	175	* No need to use ldarx/stdcx here
	176	*/
	177	*pmdp = __pmd(new_pmd & ~_PAGE_BUSY);
	178	return 0;
	179	}