David Gibson <dwg@xxxxxxxxxxx> writes:
> On Thu, Apr 04, 2013 at 11:27:46AM +0530, Aneesh Kumar K.V wrote:
>> From: "Aneesh Kumar K.V" <aneesh.kumar@xxxxxxxxxxxxxxxxxx>
>>
>> We look at both the segment base page size and actual page size and store
>> the pte-lp-encodings in an array per base page size.
>>
>> We also update all relevant functions to take actual page size argument
>> so that we can use the correct PTE LP encoding in HPTE. This should also
>> get the basic Multiple Page Size per Segment (MPSS) support. This is needed
>> to enable THP on ppc64.
>>
....
>> +static inline int hpte_actual_psize(struct hash_pte *hptep, int psize)
>> +{
>> + int i, shift;
>> + unsigned int mask;
>> + /* Look at the 8 bit LP value */
>> + unsigned int lp = (hptep->r >> LP_SHIFT) & ((1 << LP_BITS) - 1);
>> +
>> + if (!(hptep->v & HPTE_V_VALID))
>> + return -1;
>
> Folding the validity check into the size check seems confusing to me.
We do end up with invalid hpte with which we call
hpte_actual_psize. So that check is needed. I can either move to caller,
but then i will have to replicate it in all the call sites.
>
>> + /* First check if it is large page */
>> + if (!(hptep->v & HPTE_V_LARGE))
>> + return MMU_PAGE_4K;
>> +
>> + /* start from 1 ignoring MMU_PAGE_4K */
>> + for (i = 1; i < MMU_PAGE_COUNT; i++) {
>> + /* valid entries have a shift value */
>> + if (!mmu_psize_defs[i].shift)
>> + continue;
>
> Isn't this check redundant with the one below?
Yes. I guess we can safely assume that if penc is valid then we do
support that specific large page.
I will drop this and keep the penc check. That is more correct check
>
>> + /* invalid penc */
>> + if (mmu_psize_defs[psize].penc[i] == -1)
>> + continue;
>> + /*
>> + * encoding bits per actual page size
>> + * PTE LP actual page size
>> + * rrrr rrrz >=8KB
>> + * rrrr rrzz >=16KB
>> + * rrrr rzzz >=32KB
>> + * rrrr zzzz >=64KB
>> + * .......
>> + */
>> + shift = mmu_psize_defs[i].shift - LP_SHIFT;
>> + if (shift > LP_BITS)
>> + shift = LP_BITS;
>> + mask = (1 << shift) - 1;
>> + if ((lp & mask) == mmu_psize_defs[psize].penc[i])
>> + return i;
>> + }
>
> Shouldn't we have a BUG() or something here. If we get here we've
> somehow created a PTE with LP bits we can't interpret, yes?
>
I don't know. Is BUG() the right thing to do ?
>> + return -1;
>> +}
>> +
>> static long native_hpte_updatepp(unsigned long slot, unsigned long newpp,
>> unsigned long vpn, int psize, int ssize,
>> int local)
>> @@ -251,6 +294,7 @@ static long native_hpte_updatepp(unsigned long slot, unsigned long newpp,
>> struct hash_pte *hptep = htab_address + slot;
>> unsigned long hpte_v, want_v;
>> int ret = 0;
>> + int actual_psize;
>>
>> want_v = hpte_encode_avpn(vpn, psize, ssize);
>>
>> @@ -260,9 +304,13 @@ static long native_hpte_updatepp(unsigned long slot, unsigned long newpp,
>> native_lock_hpte(hptep);
>>
>> hpte_v = hptep->v;
>> -
>> + actual_psize = hpte_actual_psize(hptep, psize);
>> + if (actual_psize < 0) {
>> + native_unlock_hpte(hptep);
>> + return -1;
>> + }
>
> Wouldn't it make more sense to only do the psize lookup once you've
> found a matching hpte?
But we need to do psize lookup even if V_COMPARE fail, because we want
to do tlbie in both the case.
>
>> /* Even if we miss, we need to invalidate the TLB */
>> - if (!HPTE_V_COMPARE(hpte_v, want_v) || !(hpte_v & HPTE_V_VALID)) {
>> + if (!HPTE_V_COMPARE(hpte_v, want_v)) {
>> DBG_LOW(" -> miss\n");
>> ret = -1;
>> } else {
>> @@ -274,7 +322,7 @@ static long native_hpte_updatepp(unsigned long slot, unsigned long newpp,
>> native_unlock_hpte(hptep);
>>
>> /* Ensure it is out of the tlb too. */
>> - tlbie(vpn, psize, ssize, local);
>> + tlbie(vpn, psize, actual_psize, ssize, local);
>>
>> return ret;
>> }
>> @@ -315,6 +363,7 @@ static long native_hpte_find(unsigned long vpn, int psize, int ssize)
>> static void native_hpte_updateboltedpp(unsigned long newpp, unsigned long ea,
>> int psize, int ssize)
>> {
>> + int actual_psize;
>> unsigned long vpn;
>> unsigned long vsid;
>> long slot;
>> @@ -327,13 +376,16 @@ static void native_hpte_updateboltedpp(unsigned long newpp, unsigned long ea,
>> if (slot == -1)
>> panic("could not find page to bolt\n");
>> hptep = htab_address + slot;
>> + actual_psize = hpte_actual_psize(hptep, psize);
>> + if (actual_psize < 0)
>> + return;
>>
>> /* Update the HPTE */
>> hptep->r = (hptep->r & ~(HPTE_R_PP | HPTE_R_N)) |
>> (newpp & (HPTE_R_PP | HPTE_R_N));
>>
>> /* Ensure it is out of the tlb too. */
>> - tlbie(vpn, psize, ssize, 0);
>> + tlbie(vpn, psize, actual_psize, ssize, 0);
>> }
>>
>> static void native_hpte_invalidate(unsigned long slot, unsigned long vpn,
>> @@ -343,6 +395,7 @@ static void native_hpte_invalidate(unsigned long slot, unsigned long vpn,
>> unsigned long hpte_v;
>> unsigned long want_v;
>> unsigned long flags;
>> + int actual_psize;
>>
>> local_irq_save(flags);
>>
>> @@ -352,35 +405,38 @@ static void native_hpte_invalidate(unsigned long slot, unsigned long vpn,
>> native_lock_hpte(hptep);
>> hpte_v = hptep->v;
>>
>> + actual_psize = hpte_actual_psize(hptep, psize);
>> + if (actual_psize < 0) {
>> + native_unlock_hpte(hptep);
>> + local_irq_restore(flags);
>> + return;
>> + }
>> /* Even if we miss, we need to invalidate the TLB */
>> - if (!HPTE_V_COMPARE(hpte_v, want_v) || !(hpte_v & HPTE_V_VALID))
>> + if (!HPTE_V_COMPARE(hpte_v, want_v))
>> native_unlock_hpte(hptep);
>> else
>> /* Invalidate the hpte. NOTE: this also unlocks it */
>> hptep->v = 0;
>>
>> /* Invalidate the TLB */
>> - tlbie(vpn, psize, ssize, local);
>> + tlbie(vpn, psize, actual_psize, ssize, local);
>>
>> local_irq_restore(flags);
>> }
>>
>> -#define LP_SHIFT 12
>> -#define LP_BITS 8
>> -#define LP_MASK(i) ((0xFF >> (i)) << LP_SHIFT)
>> -
>> static void hpte_decode(struct hash_pte *hpte, unsigned long slot,
>> - int *psize, int *ssize, unsigned long *vpn)
>> + int *psize, int *apsize, int *ssize, unsigned long *vpn)
>> {
>> unsigned long avpn, pteg, vpi;
>> unsigned long hpte_r = hpte->r;
>> unsigned long hpte_v = hpte->v;
>> unsigned long vsid, seg_off;
>> - int i, size, shift, penc;
>> + int i, size, a_size, shift, penc;
>>
>> - if (!(hpte_v & HPTE_V_LARGE))
>> - size = MMU_PAGE_4K;
>> - else {
>> + if (!(hpte_v & HPTE_V_LARGE)) {
>> + size = MMU_PAGE_4K;
>> + a_size = MMU_PAGE_4K;
>> + } else {
>> for (i = 0; i < LP_BITS; i++) {
>> if ((hpte_r & LP_MASK(i+1)) == LP_MASK(i+1))
>> break;
>> @@ -388,19 +444,26 @@ static void hpte_decode(struct hash_pte *hpte, unsigned long slot,
>> penc = LP_MASK(i+1) >> LP_SHIFT;
>> for (size = 0; size < MMU_PAGE_COUNT; size++) {
>
>>
>> - /* 4K pages are not represented by LP */
>> - if (size == MMU_PAGE_4K)
>> - continue;
>> -
>> /* valid entries have a shift value */
>> if (!mmu_psize_defs[size].shift)
>> continue;
>> + for (a_size = 0; a_size < MMU_PAGE_COUNT; a_size++) {
>
> Can't you resize hpte_actual_psize() here instead of recoding the
> lookup?
I thought about that, but re-coding avoided some repeated check. But
then, if I follow your review comments of avoiding hpte valid check etc, may
be I can reuse the hpte_actual_psize. Will try this.
>
>> - if (penc == mmu_psize_defs[size].penc)
>> - break;
>> + /* 4K pages are not represented by LP */
>> + if (a_size == MMU_PAGE_4K)
>> + continue;
>> +
>> + /* valid entries have a shift value */
>> + if (!mmu_psize_defs[a_size].shift)
>> + continue;
>> +
>> + if (penc == mmu_psize_defs[size].penc[a_size])
>> + goto out;
>> + }
>> }
>> }
>>
>> +out:
-aneesh
--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@xxxxxxxxx. For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@xxxxxxxxx";> email@xxxxxxxxx </a>