| Message ID | 20221125034502.1559986-1-anshuman.khandual@arm.com (mailing list archive) |
|---|---|
| State | New |
| Headers | show |
| Series | s390/mm: Use pmd_pgtable_page() helper in __gmap_segment_gaddr() | expand |
On Fri, Nov 25, 2022 at 09:15:02AM +0530, Anshuman Khandual wrote: Hi Anshuman, > In __gmap_segment_gaddr() pmd level page table page is being extracted from > the pmd pointer, similar to pmd_pgtable_page() implementation. This reduces > some redundancy by directly using pmd_pgtable_page() instead, though first > making it available. [...] > diff --git a/arch/s390/mm/gmap.c b/arch/s390/mm/gmap.c > index 02d15c8dc92e..8947451ae021 100644 > --- a/arch/s390/mm/gmap.c > +++ b/arch/s390/mm/gmap.c > @@ -336,12 +336,11 @@ static int gmap_alloc_table(struct gmap *gmap, unsigned long *table, > static unsigned long __gmap_segment_gaddr(unsigned long *entry) > { > struct page *page; > - unsigned long offset, mask; > + unsigned long offset; > > offset = (unsigned long) entry / sizeof(unsigned long); > offset = (offset & (PTRS_PER_PMD - 1)) * PMD_SIZE; > - mask = ~(PTRS_PER_PMD * sizeof(pmd_t) - 1); > - page = virt_to_page((void *)((unsigned long) entry & mask)); > + page = pmd_pgtable_page((pmd_t *) entry); > return page->index + offset; > } Looks okay to me. > diff --git a/include/linux/mm.h b/include/linux/mm.h > index e9e387caffac..5ead9e997510 100644 > --- a/include/linux/mm.h > +++ b/include/linux/mm.h > @@ -2403,7 +2403,7 @@ static inline void pgtable_pte_page_dtor(struct page *page) > > #if USE_SPLIT_PMD_PTLOCKS > > -static struct page *pmd_pgtable_page(pmd_t *pmd) > +static inline struct page *pmd_pgtable_page(pmd_t *pmd) > { > unsigned long mask = ~(PTRS_PER_PMD * sizeof(pmd_t) - 1); > return virt_to_page((void *)((unsigned long) pmd & mask)); This chunk does not appear to belong to this patch. Thanks!
On 11/28/22 11:37, Alexander Gordeev wrote: > On Fri, Nov 25, 2022 at 09:15:02AM +0530, Anshuman Khandual wrote: > > Hi Anshuman, > >> In __gmap_segment_gaddr() pmd level page table page is being extracted from >> the pmd pointer, similar to pmd_pgtable_page() implementation. This reduces >> some redundancy by directly using pmd_pgtable_page() instead, though first >> making it available. > > [...] > >> diff --git a/arch/s390/mm/gmap.c b/arch/s390/mm/gmap.c >> index 02d15c8dc92e..8947451ae021 100644 >> --- a/arch/s390/mm/gmap.c >> +++ b/arch/s390/mm/gmap.c >> @@ -336,12 +336,11 @@ static int gmap_alloc_table(struct gmap *gmap, unsigned long *table, >> static unsigned long __gmap_segment_gaddr(unsigned long *entry) >> { >> struct page *page; >> - unsigned long offset, mask; >> + unsigned long offset; >> >> offset = (unsigned long) entry / sizeof(unsigned long); >> offset = (offset & (PTRS_PER_PMD - 1)) * PMD_SIZE; >> - mask = ~(PTRS_PER_PMD * sizeof(pmd_t) - 1); >> - page = virt_to_page((void *)((unsigned long) entry & mask)); >> + page = pmd_pgtable_page((pmd_t *) entry); >> return page->index + offset; >> } > > Looks okay to me. > >> diff --git a/include/linux/mm.h b/include/linux/mm.h >> index e9e387caffac..5ead9e997510 100644 >> --- a/include/linux/mm.h >> +++ b/include/linux/mm.h >> @@ -2403,7 +2403,7 @@ static inline void pgtable_pte_page_dtor(struct page *page) >> >> #if USE_SPLIT_PMD_PTLOCKS >> >> -static struct page *pmd_pgtable_page(pmd_t *pmd) >> +static inline struct page *pmd_pgtable_page(pmd_t *pmd) >> { >> unsigned long mask = ~(PTRS_PER_PMD * sizeof(pmd_t) - 1); >> return virt_to_page((void *)((unsigned long) pmd & mask)); > > This chunk does not appear to belong to this patch. Should not this helper be made a 'static inline' before using it else where ?
On Mon, Nov 28, 2022 at 12:01:06PM +0530, Anshuman Khandual wrote: > >> diff --git a/include/linux/mm.h b/include/linux/mm.h > >> index e9e387caffac..5ead9e997510 100644 > >> --- a/include/linux/mm.h > >> +++ b/include/linux/mm.h > >> @@ -2403,7 +2403,7 @@ static inline void pgtable_pte_page_dtor(struct page *page) > >> > >> #if USE_SPLIT_PMD_PTLOCKS > >> > >> -static struct page *pmd_pgtable_page(pmd_t *pmd) > >> +static inline struct page *pmd_pgtable_page(pmd_t *pmd) > >> { > >> unsigned long mask = ~(PTRS_PER_PMD * sizeof(pmd_t) - 1); > >> return virt_to_page((void *)((unsigned long) pmd & mask)); > > > > This chunk does not appear to belong to this patch. > > Should not this helper be made a 'static inline' before using it else where ? Probably yes, but it is not kvm s390-specific change. I guess, you wanted to make it together or as a follow-up to this one? https://lore.kernel.org/all/20221124131641.1523772-1-anshuman.khandual@arm.com/
On 11/28/22 12:43, Alexander Gordeev wrote: > On Mon, Nov 28, 2022 at 12:01:06PM +0530, Anshuman Khandual wrote: >>>> diff --git a/include/linux/mm.h b/include/linux/mm.h >>>> index e9e387caffac..5ead9e997510 100644 >>>> --- a/include/linux/mm.h >>>> +++ b/include/linux/mm.h >>>> @@ -2403,7 +2403,7 @@ static inline void pgtable_pte_page_dtor(struct page *page) >>>> >>>> #if USE_SPLIT_PMD_PTLOCKS >>>> >>>> -static struct page *pmd_pgtable_page(pmd_t *pmd) >>>> +static inline struct page *pmd_pgtable_page(pmd_t *pmd) >>>> { >>>> unsigned long mask = ~(PTRS_PER_PMD * sizeof(pmd_t) - 1); >>>> return virt_to_page((void *)((unsigned long) pmd & mask)); >>> >>> This chunk does not appear to belong to this patch. >> >> Should not this helper be made a 'static inline' before using it else where ? > > Probably yes, but it is not kvm s390-specific change. Right, just that the s390 change here is the first instance where this helper is being used outside the header, hence kept them together. > > I guess, you wanted to make it together or as a follow-up to this one? > https://lore.kernel.org/all/20221124131641.1523772-1-anshuman.khandual@arm.com/ > Unless too much trouble, could we just keep it here as proposed.
On Mon, Nov 28, 2022 at 01:13:45PM +0530, Anshuman Khandual wrote: > >>>> index e9e387caffac..5ead9e997510 100644 > >>>> --- a/include/linux/mm.h > >>>> +++ b/include/linux/mm.h > >>>> @@ -2403,7 +2403,7 @@ static inline void pgtable_pte_page_dtor(struct page *page) > >>>> > >>>> #if USE_SPLIT_PMD_PTLOCKS > >>>> > >>>> -static struct page *pmd_pgtable_page(pmd_t *pmd) > >>>> +static inline struct page *pmd_pgtable_page(pmd_t *pmd) > >>>> { > >>>> unsigned long mask = ~(PTRS_PER_PMD * sizeof(pmd_t) - 1); > >>>> return virt_to_page((void *)((unsigned long) pmd & mask)); > >>> > >>> This chunk does not appear to belong to this patch. > >> > >> Should not this helper be made a 'static inline' before using it else where ? > > > > Probably yes, but it is not kvm s390-specific change. > > Right, just that the s390 change here is the first instance where this helper > is being used outside the header, hence kept them together. > > > > > I guess, you wanted to make it together or as a follow-up to this one? > > https://lore.kernel.org/all/20221124131641.1523772-1-anshuman.khandual@arm.com/ > > > Unless too much trouble, could we just keep it here as proposed. For s390 part: Acked-by: Alexander Gordeev <agordeev@linux.ibm.com> Thanks!
diff --git a/arch/s390/mm/gmap.c b/arch/s390/mm/gmap.c index 02d15c8dc92e..8947451ae021 100644 --- a/arch/s390/mm/gmap.c +++ b/arch/s390/mm/gmap.c @@ -336,12 +336,11 @@ static int gmap_alloc_table(struct gmap *gmap, unsigned long *table, static unsigned long __gmap_segment_gaddr(unsigned long *entry) { struct page *page; - unsigned long offset, mask; + unsigned long offset; offset = (unsigned long) entry / sizeof(unsigned long); offset = (offset & (PTRS_PER_PMD - 1)) * PMD_SIZE; - mask = ~(PTRS_PER_PMD * sizeof(pmd_t) - 1); - page = virt_to_page((void *)((unsigned long) entry & mask)); + page = pmd_pgtable_page((pmd_t *) entry); return page->index + offset; } diff --git a/include/linux/mm.h b/include/linux/mm.h index e9e387caffac..5ead9e997510 100644 --- a/include/linux/mm.h +++ b/include/linux/mm.h @@ -2403,7 +2403,7 @@ static inline void pgtable_pte_page_dtor(struct page *page) #if USE_SPLIT_PMD_PTLOCKS -static struct page *pmd_pgtable_page(pmd_t *pmd) +static inline struct page *pmd_pgtable_page(pmd_t *pmd) { unsigned long mask = ~(PTRS_PER_PMD * sizeof(pmd_t) - 1); return virt_to_page((void *)((unsigned long) pmd & mask));
In __gmap_segment_gaddr() pmd level page table page is being extracted from the pmd pointer, similar to pmd_pgtable_page() implementation. This reduces some redundancy by directly using pmd_pgtable_page() instead, though first making it available. Cc: Christian Borntraeger <borntraeger@linux.ibm.com> Cc: David Hildenbrand <david@redhat.com> Cc: Heiko Carstens <hca@linux.ibm.com> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: kvm@vger.kernel.org Cc: linux-s390@vger.kernel.org Cc: linux-kernel@vger.kernel.org Cc: linux-mm@kvack.org Signed-off-by: Anshuman Khandual <anshuman.khandual@arm.com> --- This applies on v6.1-rc6 but after the following patch. Build tested for s390 plaform (defconfig). https://lore.kernel.org/all/20221124131641.1523772-1-anshuman.khandual@arm.com/ arch/s390/mm/gmap.c | 5 ++--- include/linux/mm.h | 2 +- 2 files changed, 3 insertions(+), 4 deletions(-)