| Message ID | 20230113171026.582290-12-david@redhat.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
| Series | mm: support __HAVE_ARCH_PTE_SWP_EXCLUSIVE on all architectures with swap PTEs | expand |
Hi David, On Fri, Jan 13, 2023 at 6:16 PM David Hildenbrand <david@redhat.com> wrote: > Let's support __HAVE_ARCH_PTE_SWP_EXCLUSIVE by stealing one bit > from the type. Generic MM currently only uses 5 bits for the type > (MAX_SWAPFILES_SHIFT), so the stolen bit is effectively unused. > > The shift by 2 when converting between PTE and arch-specific swap entry > makes the swap PTE layout a little bit harder to decipher. > > While at it, drop the comment from paulus---copy-and-paste leftover > from powerpc where we actually have _PAGE_HASHPTE---and mask the type in > __swp_entry_to_pte() as well. > > Cc: Michal Simek <monstr@monstr.eu> > Signed-off-by: David Hildenbrand <david@redhat.com> Thanks for your patch, which is now commit b5c88f21531c3457 ("microblaze/mm: support __HAVE_ARCH_PTE_SWP_EXCLUSIVE") in > arch/m68k/include/asm/mcf_pgtable.h | 4 +-- What is this m68k change doing here? Sorry for not noticing this earlier. Furthermore, several things below look strange to me... > arch/microblaze/include/asm/pgtable.h | 45 +++++++++++++++++++++------ > 2 files changed, 37 insertions(+), 12 deletions(-) > > diff --git a/arch/m68k/include/asm/mcf_pgtable.h b/arch/m68k/include/asm/mcf_pgtable.h > index 3f8f4d0e66dd..e573d7b649f7 100644 > --- a/arch/m68k/include/asm/mcf_pgtable.h > +++ b/arch/m68k/include/asm/mcf_pgtable.h > @@ -46,8 +46,8 @@ > #define _CACHEMASK040 (~0x060) > #define _PAGE_GLOBAL040 0x400 /* 68040 global bit, used for kva descs */ > > -/* We borrow bit 7 to store the exclusive marker in swap PTEs. */ > -#define _PAGE_SWP_EXCLUSIVE 0x080 > +/* We borrow bit 24 to store the exclusive marker in swap PTEs. */ > +#define _PAGE_SWP_EXCLUSIVE CF_PAGE_NOCACHE CF_PAGE_NOCACHE is 0x80, so this is still bit 7, thus the new comment is wrong? > > /* > * Externally used page protection values. > diff --git a/arch/microblaze/include/asm/pgtable.h b/arch/microblaze/include/asm/pgtable.h > index 42f5988e998b..7e3de54bf426 100644 > --- a/arch/microblaze/include/asm/pgtable.h > +++ b/arch/microblaze/include/asm/pgtable.h > @@ -131,10 +131,10 @@ extern pte_t *va_to_pte(unsigned long address); > * of the 16 available. Bit 24-26 of the TLB are cleared in the TLB > * miss handler. Bit 27 is PAGE_USER, thus selecting the correct > * zone. > - * - PRESENT *must* be in the bottom two bits because swap cache > - * entries use the top 30 bits. Because 4xx doesn't support SMP > - * anyway, M is irrelevant so we borrow it for PAGE_PRESENT. Bit 30 > - * is cleared in the TLB miss handler before the TLB entry is loaded. > + * - PRESENT *must* be in the bottom two bits because swap PTEs use the top > + * 30 bits. Because 4xx doesn't support SMP anyway, M is irrelevant so we > + * borrow it for PAGE_PRESENT. Bit 30 is cleared in the TLB miss handler > + * before the TLB entry is loaded. So the PowerPC 4xx comment is still here? > * - All other bits of the PTE are loaded into TLBLO without > * * modification, leaving us only the bits 20, 21, 24, 25, 26, 30 for > * software PTE bits. We actually use bits 21, 24, 25, and > @@ -155,6 +155,9 @@ extern pte_t *va_to_pte(unsigned long address); > #define _PAGE_ACCESSED 0x400 /* software: R: page referenced */ > #define _PMD_PRESENT PAGE_MASK > > +/* We borrow bit 24 to store the exclusive marker in swap PTEs. */ > +#define _PAGE_SWP_EXCLUSIVE _PAGE_DIRTY _PAGE_DIRTY is 0x80, so this is also bit 7, thus the new comment is wrong? Gr{oetje,eeting}s, Geert
On 26.02.23 21:13, Geert Uytterhoeven wrote: > Hi David, Hi Geert, > > On Fri, Jan 13, 2023 at 6:16 PM David Hildenbrand <david@redhat.com> wrote: >> Let's support __HAVE_ARCH_PTE_SWP_EXCLUSIVE by stealing one bit >> from the type. Generic MM currently only uses 5 bits for the type >> (MAX_SWAPFILES_SHIFT), so the stolen bit is effectively unused. >> >> The shift by 2 when converting between PTE and arch-specific swap entry >> makes the swap PTE layout a little bit harder to decipher. >> >> While at it, drop the comment from paulus---copy-and-paste leftover >> from powerpc where we actually have _PAGE_HASHPTE---and mask the type in >> __swp_entry_to_pte() as well. >> >> Cc: Michal Simek <monstr@monstr.eu> >> Signed-off-by: David Hildenbrand <david@redhat.com> > > Thanks for your patch, which is now commit b5c88f21531c3457 > ("microblaze/mm: support __HAVE_ARCH_PTE_SWP_EXCLUSIVE") in > Right, it went upstream, so we can only fixup. >> arch/m68k/include/asm/mcf_pgtable.h | 4 +-- > > What is this m68k change doing here? > Sorry for not noticing this earlier. Thanks for the late review, still valuable :) That hunk should have gone into the previous patch, looks like I messed that up when reworking. > > Furthermore, several things below look strange to me... > >> arch/microblaze/include/asm/pgtable.h | 45 +++++++++++++++++++++------ >> 2 files changed, 37 insertions(+), 12 deletions(-) >> >> diff --git a/arch/m68k/include/asm/mcf_pgtable.h b/arch/m68k/include/asm/mcf_pgtable.h >> index 3f8f4d0e66dd..e573d7b649f7 100644 >> --- a/arch/m68k/include/asm/mcf_pgtable.h >> +++ b/arch/m68k/include/asm/mcf_pgtable.h >> @@ -46,8 +46,8 @@ >> #define _CACHEMASK040 (~0x060) >> #define _PAGE_GLOBAL040 0x400 /* 68040 global bit, used for kva descs */ >> >> -/* We borrow bit 7 to store the exclusive marker in swap PTEs. */ >> -#define _PAGE_SWP_EXCLUSIVE 0x080 >> +/* We borrow bit 24 to store the exclusive marker in swap PTEs. */ >> +#define _PAGE_SWP_EXCLUSIVE CF_PAGE_NOCACHE > > CF_PAGE_NOCACHE is 0x80, so this is still bit 7, thus the new comment > is wrong? You're right, it's still bit 7 (and we use LSB-0 bit numbering in that file). I'll send a fixup. > >> >> /* >> * Externally used page protection values. >> diff --git a/arch/microblaze/include/asm/pgtable.h b/arch/microblaze/include/asm/pgtable.h >> index 42f5988e998b..7e3de54bf426 100644 >> --- a/arch/microblaze/include/asm/pgtable.h >> +++ b/arch/microblaze/include/asm/pgtable.h >> @@ -131,10 +131,10 @@ extern pte_t *va_to_pte(unsigned long address); >> * of the 16 available. Bit 24-26 of the TLB are cleared in the TLB >> * miss handler. Bit 27 is PAGE_USER, thus selecting the correct >> * zone. >> - * - PRESENT *must* be in the bottom two bits because swap cache >> - * entries use the top 30 bits. Because 4xx doesn't support SMP >> - * anyway, M is irrelevant so we borrow it for PAGE_PRESENT. Bit 30 >> - * is cleared in the TLB miss handler before the TLB entry is loaded. >> + * - PRESENT *must* be in the bottom two bits because swap PTEs use the top >> + * 30 bits. Because 4xx doesn't support SMP anyway, M is irrelevant so we >> + * borrow it for PAGE_PRESENT. Bit 30 is cleared in the TLB miss handler >> + * before the TLB entry is loaded. > > So the PowerPC 4xx comment is still here? I only dropped the comment above __swp_type(). I guess you mean that we could also drop the "Because 4xx doesn't support SMP anyway, M is irrelevant so we borrow it for PAGE_PRESENT." sentence, correct? Not sure about the "Bit 30 is cleared in the TLB miss handler" comment, if that can similarly be dropped. > >> * - All other bits of the PTE are loaded into TLBLO without >> * * modification, leaving us only the bits 20, 21, 24, 25, 26, 30 for >> * software PTE bits. We actually use bits 21, 24, 25, and >> @@ -155,6 +155,9 @@ extern pte_t *va_to_pte(unsigned long address); >> #define _PAGE_ACCESSED 0x400 /* software: R: page referenced */ >> #define _PMD_PRESENT PAGE_MASK >> >> +/* We borrow bit 24 to store the exclusive marker in swap PTEs. */ >> +#define _PAGE_SWP_EXCLUSIVE _PAGE_DIRTY > > _PAGE_DIRTY is 0x80, so this is also bit 7, thus the new comment is > wrong? In the example, I use MSB-0 bit numbering (which I determined to be correct in microblaze context eventually, but I got confused a couple a times because it's very inconsistent). That should be MSB-0 bit 24. Thanks!
Hi David, On Mon, Feb 27, 2023 at 2:31 PM David Hildenbrand <david@redhat.com> wrote: > On 26.02.23 21:13, Geert Uytterhoeven wrote: > > On Fri, Jan 13, 2023 at 6:16 PM David Hildenbrand <david@redhat.com> wrote: > >> Let's support __HAVE_ARCH_PTE_SWP_EXCLUSIVE by stealing one bit > >> from the type. Generic MM currently only uses 5 bits for the type > >> (MAX_SWAPFILES_SHIFT), so the stolen bit is effectively unused. > >> > >> The shift by 2 when converting between PTE and arch-specific swap entry > >> makes the swap PTE layout a little bit harder to decipher. > >> > >> While at it, drop the comment from paulus---copy-and-paste leftover > >> from powerpc where we actually have _PAGE_HASHPTE---and mask the type in > >> __swp_entry_to_pte() as well. > >> > >> Cc: Michal Simek <monstr@monstr.eu> > >> Signed-off-by: David Hildenbrand <david@redhat.com> > > > > Thanks for your patch, which is now commit b5c88f21531c3457 > > ("microblaze/mm: support __HAVE_ARCH_PTE_SWP_EXCLUSIVE") in > > > > Right, it went upstream, so we can only fixup. > > >> arch/m68k/include/asm/mcf_pgtable.h | 4 +-- > > > > What is this m68k change doing here? > > Sorry for not noticing this earlier. > > Thanks for the late review, still valuable :) > > That hunk should have gone into the previous patch, looks like I messed > that up when reworking. > > > > > Furthermore, several things below look strange to me... > > > >> arch/microblaze/include/asm/pgtable.h | 45 +++++++++++++++++++++------ > >> 2 files changed, 37 insertions(+), 12 deletions(-) > >> > >> diff --git a/arch/m68k/include/asm/mcf_pgtable.h b/arch/m68k/include/asm/mcf_pgtable.h > >> index 3f8f4d0e66dd..e573d7b649f7 100644 > >> --- a/arch/m68k/include/asm/mcf_pgtable.h > >> +++ b/arch/m68k/include/asm/mcf_pgtable.h > >> @@ -46,8 +46,8 @@ > >> #define _CACHEMASK040 (~0x060) > >> #define _PAGE_GLOBAL040 0x400 /* 68040 global bit, used for kva descs */ > >> > >> -/* We borrow bit 7 to store the exclusive marker in swap PTEs. */ > >> -#define _PAGE_SWP_EXCLUSIVE 0x080 > >> +/* We borrow bit 24 to store the exclusive marker in swap PTEs. */ > >> +#define _PAGE_SWP_EXCLUSIVE CF_PAGE_NOCACHE > > > > CF_PAGE_NOCACHE is 0x80, so this is still bit 7, thus the new comment > > is wrong? > > You're right, it's still bit 7 (and we use LSB-0 bit numbering in that > file). I'll send a fixup. OK. > >> /* > >> * Externally used page protection values. > >> diff --git a/arch/microblaze/include/asm/pgtable.h b/arch/microblaze/include/asm/pgtable.h > >> index 42f5988e998b..7e3de54bf426 100644 > >> --- a/arch/microblaze/include/asm/pgtable.h > >> +++ b/arch/microblaze/include/asm/pgtable.h > >> @@ -131,10 +131,10 @@ extern pte_t *va_to_pte(unsigned long address); > >> * of the 16 available. Bit 24-26 of the TLB are cleared in the TLB > >> * miss handler. Bit 27 is PAGE_USER, thus selecting the correct > >> * zone. > >> - * - PRESENT *must* be in the bottom two bits because swap cache > >> - * entries use the top 30 bits. Because 4xx doesn't support SMP > >> - * anyway, M is irrelevant so we borrow it for PAGE_PRESENT. Bit 30 > >> - * is cleared in the TLB miss handler before the TLB entry is loaded. > >> + * - PRESENT *must* be in the bottom two bits because swap PTEs use the top > >> + * 30 bits. Because 4xx doesn't support SMP anyway, M is irrelevant so we > >> + * borrow it for PAGE_PRESENT. Bit 30 is cleared in the TLB miss handler > >> + * before the TLB entry is loaded. > > > > So the PowerPC 4xx comment is still here? > > I only dropped the comment above __swp_type(). I guess you mean that we > could also drop the "Because 4xx doesn't support SMP anyway, M is > irrelevant so we borrow it for PAGE_PRESENT." sentence, correct? Not Yes, that's what I meant. > sure about the "Bit 30 is cleared in the TLB miss handler" comment, if > that can similarly be dropped. No idea, didn't check. But if it was copied from PPC, chances are high it's no longer true.... > >> * - All other bits of the PTE are loaded into TLBLO without > >> * * modification, leaving us only the bits 20, 21, 24, 25, 26, 30 for > >> * software PTE bits. We actually use bits 21, 24, 25, and > >> @@ -155,6 +155,9 @@ extern pte_t *va_to_pte(unsigned long address); > >> #define _PAGE_ACCESSED 0x400 /* software: R: page referenced */ > >> #define _PMD_PRESENT PAGE_MASK > >> > >> +/* We borrow bit 24 to store the exclusive marker in swap PTEs. */ > >> +#define _PAGE_SWP_EXCLUSIVE _PAGE_DIRTY > > > > _PAGE_DIRTY is 0x80, so this is also bit 7, thus the new comment is > > wrong? > > In the example, I use MSB-0 bit numbering (which I determined to be > correct in microblaze context eventually, but I got confused a couple a > times because it's very inconsistent). That should be MSB-0 bit 24. Thanks, TIL microblaze uses IBM bit numbering... Gr{oetje,eeting}s, Geert
>>>> /* >>>> * Externally used page protection values. >>>> diff --git a/arch/microblaze/include/asm/pgtable.h b/arch/microblaze/include/asm/pgtable.h >>>> index 42f5988e998b..7e3de54bf426 100644 >>>> --- a/arch/microblaze/include/asm/pgtable.h >>>> +++ b/arch/microblaze/include/asm/pgtable.h >>>> @@ -131,10 +131,10 @@ extern pte_t *va_to_pte(unsigned long address); >>>> * of the 16 available. Bit 24-26 of the TLB are cleared in the TLB >>>> * miss handler. Bit 27 is PAGE_USER, thus selecting the correct >>>> * zone. >>>> - * - PRESENT *must* be in the bottom two bits because swap cache >>>> - * entries use the top 30 bits. Because 4xx doesn't support SMP >>>> - * anyway, M is irrelevant so we borrow it for PAGE_PRESENT. Bit 30 >>>> - * is cleared in the TLB miss handler before the TLB entry is loaded. >>>> + * - PRESENT *must* be in the bottom two bits because swap PTEs use the top >>>> + * 30 bits. Because 4xx doesn't support SMP anyway, M is irrelevant so we >>>> + * borrow it for PAGE_PRESENT. Bit 30 is cleared in the TLB miss handler >>>> + * before the TLB entry is loaded. >>> >>> So the PowerPC 4xx comment is still here? >> >> I only dropped the comment above __swp_type(). I guess you mean that we >> could also drop the "Because 4xx doesn't support SMP anyway, M is >> irrelevant so we borrow it for PAGE_PRESENT." sentence, correct? Not > > Yes, that's what I meant. > >> sure about the "Bit 30 is cleared in the TLB miss handler" comment, if >> that can similarly be dropped. > > No idea, didn't check. But if it was copied from PPC, chances are > high it's no longer true.... I'll have a look. > >>>> * - All other bits of the PTE are loaded into TLBLO without >>>> * * modification, leaving us only the bits 20, 21, 24, 25, 26, 30 for >>>> * software PTE bits. We actually use bits 21, 24, 25, and >>>> @@ -155,6 +155,9 @@ extern pte_t *va_to_pte(unsigned long address); >>>> #define _PAGE_ACCESSED 0x400 /* software: R: page referenced */ >>>> #define _PMD_PRESENT PAGE_MASK >>>> >>>> +/* We borrow bit 24 to store the exclusive marker in swap PTEs. */ >>>> +#define _PAGE_SWP_EXCLUSIVE _PAGE_DIRTY >>> >>> _PAGE_DIRTY is 0x80, so this is also bit 7, thus the new comment is >>> wrong? >> >> In the example, I use MSB-0 bit numbering (which I determined to be >> correct in microblaze context eventually, but I got confused a couple a >> times because it's very inconsistent). That should be MSB-0 bit 24. > > Thanks, TIL microblaze uses IBM bit numbering... I assume IBM bit numbering corresponds to MSB-0 bit numbering, correct? I recall that I used the comment above "/* Definitions for MicroBlaze. */" as an orientation. 0 1 2 3 4 ... 18 19 20 21 22 23 24 25 26 27 28 29 30 31 RPN..................... 0 0 EX WR ZSEL....... W I M G So ... either we adjust both or we leave it as is. (again, depends on what the right thing to to is -- which I don't know :) )
Hi David, On Mon, Feb 27, 2023 at 6:01 PM David Hildenbrand <david@redhat.com> wrote: > >>>> /* > >>>> * Externally used page protection values. > >>>> diff --git a/arch/microblaze/include/asm/pgtable.h b/arch/microblaze/include/asm/pgtable.h > >>>> index 42f5988e998b..7e3de54bf426 100644 > >>>> --- a/arch/microblaze/include/asm/pgtable.h > >>>> +++ b/arch/microblaze/include/asm/pgtable.h > >>>> * - All other bits of the PTE are loaded into TLBLO without > >>>> * * modification, leaving us only the bits 20, 21, 24, 25, 26, 30 for > >>>> * software PTE bits. We actually use bits 21, 24, 25, and > >>>> @@ -155,6 +155,9 @@ extern pte_t *va_to_pte(unsigned long address); > >>>> #define _PAGE_ACCESSED 0x400 /* software: R: page referenced */ > >>>> #define _PMD_PRESENT PAGE_MASK > >>>> > >>>> +/* We borrow bit 24 to store the exclusive marker in swap PTEs. */ > >>>> +#define _PAGE_SWP_EXCLUSIVE _PAGE_DIRTY > >>> > >>> _PAGE_DIRTY is 0x80, so this is also bit 7, thus the new comment is > >>> wrong? > >> > >> In the example, I use MSB-0 bit numbering (which I determined to be > >> correct in microblaze context eventually, but I got confused a couple a > >> times because it's very inconsistent). That should be MSB-0 bit 24. > > > > Thanks, TIL microblaze uses IBM bit numbering... > > I assume IBM bit numbering corresponds to MSB-0 bit numbering, correct? Correct, as seen in s370 and PowerPC manuals... > I recall that I used the comment above "/* Definitions for MicroBlaze. > */" as an orientation. > > 0 1 2 3 4 ... 18 19 20 21 22 23 24 25 26 27 28 29 30 31 > RPN..................... 0 0 EX WR ZSEL....... W I M G Indeed, that's where I noticed the "unconventional" numbering... > So ... either we adjust both or we leave it as is. (again, depends on > what the right thing to to is -- which I don't know :) ) It depends whether you want to match the hardware documentation, or the Linux BIT() macro and friends... Gr{oetje,eeting}s, Geert
On 27.02.23 20:46, Geert Uytterhoeven wrote: > Hi David, > > On Mon, Feb 27, 2023 at 6:01 PM David Hildenbrand <david@redhat.com> wrote: >>>>>> /* >>>>>> * Externally used page protection values. >>>>>> diff --git a/arch/microblaze/include/asm/pgtable.h b/arch/microblaze/include/asm/pgtable.h >>>>>> index 42f5988e998b..7e3de54bf426 100644 >>>>>> --- a/arch/microblaze/include/asm/pgtable.h >>>>>> +++ b/arch/microblaze/include/asm/pgtable.h > >>>>>> * - All other bits of the PTE are loaded into TLBLO without >>>>>> * * modification, leaving us only the bits 20, 21, 24, 25, 26, 30 for >>>>>> * software PTE bits. We actually use bits 21, 24, 25, and >>>>>> @@ -155,6 +155,9 @@ extern pte_t *va_to_pte(unsigned long address); >>>>>> #define _PAGE_ACCESSED 0x400 /* software: R: page referenced */ >>>>>> #define _PMD_PRESENT PAGE_MASK >>>>>> >>>>>> +/* We borrow bit 24 to store the exclusive marker in swap PTEs. */ >>>>>> +#define _PAGE_SWP_EXCLUSIVE _PAGE_DIRTY >>>>> >>>>> _PAGE_DIRTY is 0x80, so this is also bit 7, thus the new comment is >>>>> wrong? >>>> >>>> In the example, I use MSB-0 bit numbering (which I determined to be >>>> correct in microblaze context eventually, but I got confused a couple a >>>> times because it's very inconsistent). That should be MSB-0 bit 24. >>> >>> Thanks, TIL microblaze uses IBM bit numbering... >> >> I assume IBM bit numbering corresponds to MSB-0 bit numbering, correct? > > Correct, as seen in s370 and PowerPC manuals... Good, I have some solid s390x background, but thinking about the term "IBM PC" made me double-check that we're talking about the same thing ;) > >> I recall that I used the comment above "/* Definitions for MicroBlaze. >> */" as an orientation. >> >> 0 1 2 3 4 ... 18 19 20 21 22 23 24 25 26 27 28 29 30 31 >> RPN..................... 0 0 EX WR ZSEL....... W I M G > > Indeed, that's where I noticed the "unconventional" numbering... > >> So ... either we adjust both or we leave it as is. (again, depends on >> what the right thing to to is -- which I don't know :) ) > > It depends whether you want to match the hardware documentation, > or the Linux BIT() macro and friends... The hardware documentation, so we should be good. Thanks!
diff --git a/arch/m68k/include/asm/mcf_pgtable.h b/arch/m68k/include/asm/mcf_pgtable.h index 3f8f4d0e66dd..e573d7b649f7 100644 --- a/arch/m68k/include/asm/mcf_pgtable.h +++ b/arch/m68k/include/asm/mcf_pgtable.h @@ -46,8 +46,8 @@ #define _CACHEMASK040 (~0x060) #define _PAGE_GLOBAL040 0x400 /* 68040 global bit, used for kva descs */ -/* We borrow bit 7 to store the exclusive marker in swap PTEs. */ -#define _PAGE_SWP_EXCLUSIVE 0x080 +/* We borrow bit 24 to store the exclusive marker in swap PTEs. */ +#define _PAGE_SWP_EXCLUSIVE CF_PAGE_NOCACHE /* * Externally used page protection values. diff --git a/arch/microblaze/include/asm/pgtable.h b/arch/microblaze/include/asm/pgtable.h index 42f5988e998b..7e3de54bf426 100644 --- a/arch/microblaze/include/asm/pgtable.h +++ b/arch/microblaze/include/asm/pgtable.h @@ -131,10 +131,10 @@ extern pte_t *va_to_pte(unsigned long address); * of the 16 available. Bit 24-26 of the TLB are cleared in the TLB * miss handler. Bit 27 is PAGE_USER, thus selecting the correct * zone. - * - PRESENT *must* be in the bottom two bits because swap cache - * entries use the top 30 bits. Because 4xx doesn't support SMP - * anyway, M is irrelevant so we borrow it for PAGE_PRESENT. Bit 30 - * is cleared in the TLB miss handler before the TLB entry is loaded. + * - PRESENT *must* be in the bottom two bits because swap PTEs use the top + * 30 bits. Because 4xx doesn't support SMP anyway, M is irrelevant so we + * borrow it for PAGE_PRESENT. Bit 30 is cleared in the TLB miss handler + * before the TLB entry is loaded. * - All other bits of the PTE are loaded into TLBLO without * * modification, leaving us only the bits 20, 21, 24, 25, 26, 30 for * software PTE bits. We actually use bits 21, 24, 25, and @@ -155,6 +155,9 @@ extern pte_t *va_to_pte(unsigned long address); #define _PAGE_ACCESSED 0x400 /* software: R: page referenced */ #define _PMD_PRESENT PAGE_MASK +/* We borrow bit 24 to store the exclusive marker in swap PTEs. */ +#define _PAGE_SWP_EXCLUSIVE _PAGE_DIRTY + /* * Some bits are unused... */ @@ -393,18 +396,40 @@ static inline unsigned long pmd_page_vaddr(pmd_t pmd) extern pgd_t swapper_pg_dir[PTRS_PER_PGD]; /* - * Encode and decode a swap entry. - * Note that the bits we use in a PTE for representing a swap entry - * must not include the _PAGE_PRESENT bit, or the _PAGE_HASHPTE bit - * (if used). -- paulus + * Encode/decode swap entries and swap PTEs. Swap PTEs are all PTEs that + * are !pte_none() && !pte_present(). + * + * 1 1 1 1 1 1 1 1 1 2 2 2 2 2 2 2 2 2 2 3 3 + * 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 + * <------------------ offset -------------------> E < type -> 0 0 + * + * E is the exclusive marker that is not stored in swap entries. */ -#define __swp_type(entry) ((entry).val & 0x3f) +#define __swp_type(entry) ((entry).val & 0x1f) #define __swp_offset(entry) ((entry).val >> 6) #define __swp_entry(type, offset) \ - ((swp_entry_t) { (type) | ((offset) << 6) }) + ((swp_entry_t) { ((type) & 0x1f) | ((offset) << 6) }) #define __pte_to_swp_entry(pte) ((swp_entry_t) { pte_val(pte) >> 2 }) #define __swp_entry_to_pte(x) ((pte_t) { (x).val << 2 }) +#define __HAVE_ARCH_PTE_SWP_EXCLUSIVE +static inline int pte_swp_exclusive(pte_t pte) +{ + return pte_val(pte) & _PAGE_SWP_EXCLUSIVE; +} + +static inline pte_t pte_swp_mkexclusive(pte_t pte) +{ + pte_val(pte) |= _PAGE_SWP_EXCLUSIVE; + return pte; +} + +static inline pte_t pte_swp_clear_exclusive(pte_t pte) +{ + pte_val(pte) &= ~_PAGE_SWP_EXCLUSIVE; + return pte; +} + extern unsigned long iopa(unsigned long addr); /* Values for nocacheflag and cmode */
Let's support __HAVE_ARCH_PTE_SWP_EXCLUSIVE by stealing one bit from the type. Generic MM currently only uses 5 bits for the type (MAX_SWAPFILES_SHIFT), so the stolen bit is effectively unused. The shift by 2 when converting between PTE and arch-specific swap entry makes the swap PTE layout a little bit harder to decipher. While at it, drop the comment from paulus---copy-and-paste leftover from powerpc where we actually have _PAGE_HASHPTE---and mask the type in __swp_entry_to_pte() as well. Cc: Michal Simek <monstr@monstr.eu> Signed-off-by: David Hildenbrand <david@redhat.com> --- arch/m68k/include/asm/mcf_pgtable.h | 4 +-- arch/microblaze/include/asm/pgtable.h | 45 +++++++++++++++++++++------ 2 files changed, 37 insertions(+), 12 deletions(-)