| Message ID | 20200131100052.58761-1-thomas_os@shipmail.org (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
| Series | mm/mapping_dirty_helpers: Update huge page-table entry callbacks | expand |
On Fri, Jan 31, 2020 at 10:00:52AM +0000, Thomas Hellström (VMware) wrote: > From: Thomas Hellstrom <thellstrom@vmware.com> > > Following the update of pagewalk code > commit a07984d48146 ("mm: pagewalk: add p4d_entry() and pgd_entry()") > we can modify the mapping_dirty_helpers' huge page-table entry callbacks > to avoid splitting when a huge pud or -pmd is encountered. > > Signed-off-by: Thomas Hellstrom <thellstrom@vmware.com> > Cc: Steven Price <steven.price@arm.com> LGTM Reviewed-by: Steven Price <steven.price@arm.com> > --- > mm/mapping_dirty_helpers.c | 42 ++++++++++++++++++++++++++++++++++---- > 1 file changed, 38 insertions(+), 4 deletions(-) > > diff --git a/mm/mapping_dirty_helpers.c b/mm/mapping_dirty_helpers.c > index 71070dda9643..2c7d03675903 100644 > --- a/mm/mapping_dirty_helpers.c > +++ b/mm/mapping_dirty_helpers.c > @@ -111,26 +111,60 @@ static int clean_record_pte(pte_t *pte, unsigned long addr, > return 0; > } > > -/* wp_clean_pmd_entry - The pagewalk pmd callback. */ > +/* > + * wp_clean_pmd_entry - The pagewalk pmd callback. > + * > + * Dirty-tracking should take place on the PTE level, so > + * WARN() if encountering a dirty huge pmd. > + * Furthermore, never split huge pmds, since that currently > + * causes dirty info loss. The pagefault handler should do > + * that if needed. > + */ > static int wp_clean_pmd_entry(pmd_t *pmd, unsigned long addr, unsigned long end, > struct mm_walk *walk) > { > - /* Dirty-tracking should be handled on the pte level */ > pmd_t pmdval = pmd_read_atomic(pmd); > > + if (!pmd_trans_unstable(&pmdval)) > + return 0; > + > + if (pmd_none(pmdval)) { > + walk->action = ACTION_AGAIN; > + return 0; > + } > + > + /* Huge pmd, present or migrated */ > + walk->action = ACTION_CONTINUE; > if (pmd_trans_huge(pmdval) || pmd_devmap(pmdval)) > WARN_ON(pmd_write(pmdval) || pmd_dirty(pmdval)); > > return 0; > } > > -/* wp_clean_pud_entry - The pagewalk pud callback. */ > +/* > + * wp_clean_pud_entry - The pagewalk pud callback. > + * > + * Dirty-tracking should take place on the PTE level, so > + * WARN() if encountering a dirty huge puds. > + * Furthermore, never split huge puds, since that currently > + * causes dirty info loss. The pagefault handler should do > + * that if needed. > + */ > static int wp_clean_pud_entry(pud_t *pud, unsigned long addr, unsigned long end, > struct mm_walk *walk) > { > - /* Dirty-tracking should be handled on the pte level */ > pud_t pudval = READ_ONCE(*pud); > > + if (!pud_trans_unstable(&pudval)) > + return 0; > + > + if (pud_none(pudval)) { > + walk->action = ACTION_AGAIN; > + return 0; > + } > + > + /* Huge pud */ > + walk->action = ACTION_CONTINUE; > if (pud_trans_huge(pudval) || pud_devmap(pudval)) > WARN_ON(pud_write(pudval) || pud_dirty(pudval)); > > -- > 2.21.1 >
On 1/31/20 12:33 PM, Steven Price wrote: > On Fri, Jan 31, 2020 at 10:00:52AM +0000, Thomas Hellström (VMware) wrote: >> From: Thomas Hellstrom <thellstrom@vmware.com> >> >> Following the update of pagewalk code >> commit a07984d48146 ("mm: pagewalk: add p4d_entry() and pgd_entry()") >> we can modify the mapping_dirty_helpers' huge page-table entry callbacks >> to avoid splitting when a huge pud or -pmd is encountered. >> >> Signed-off-by: Thomas Hellstrom <thellstrom@vmware.com> >> Cc: Steven Price <steven.price@arm.com> > LGTM > > Reviewed-by: Steven Price <steven.price@arm.com> Thanks for reviewing, Steven! /Thomas
diff --git a/mm/mapping_dirty_helpers.c b/mm/mapping_dirty_helpers.c index 71070dda9643..2c7d03675903 100644 --- a/mm/mapping_dirty_helpers.c +++ b/mm/mapping_dirty_helpers.c @@ -111,26 +111,60 @@ static int clean_record_pte(pte_t *pte, unsigned long addr, return 0; } -/* wp_clean_pmd_entry - The pagewalk pmd callback. */ +/* + * wp_clean_pmd_entry - The pagewalk pmd callback. + * + * Dirty-tracking should take place on the PTE level, so + * WARN() if encountering a dirty huge pmd. + * Furthermore, never split huge pmds, since that currently + * causes dirty info loss. The pagefault handler should do + * that if needed. + */ static int wp_clean_pmd_entry(pmd_t *pmd, unsigned long addr, unsigned long end, struct mm_walk *walk) { - /* Dirty-tracking should be handled on the pte level */ pmd_t pmdval = pmd_read_atomic(pmd); + if (!pmd_trans_unstable(&pmdval)) + return 0; + + if (pmd_none(pmdval)) { + walk->action = ACTION_AGAIN; + return 0; + } + + /* Huge pmd, present or migrated */ + walk->action = ACTION_CONTINUE; if (pmd_trans_huge(pmdval) || pmd_devmap(pmdval)) WARN_ON(pmd_write(pmdval) || pmd_dirty(pmdval)); return 0; } -/* wp_clean_pud_entry - The pagewalk pud callback. */ +/* + * wp_clean_pud_entry - The pagewalk pud callback. + * + * Dirty-tracking should take place on the PTE level, so + * WARN() if encountering a dirty huge puds. + * Furthermore, never split huge puds, since that currently + * causes dirty info loss. The pagefault handler should do + * that if needed. + */ static int wp_clean_pud_entry(pud_t *pud, unsigned long addr, unsigned long end, struct mm_walk *walk) { - /* Dirty-tracking should be handled on the pte level */ pud_t pudval = READ_ONCE(*pud); + if (!pud_trans_unstable(&pudval)) + return 0; + + if (pud_none(pudval)) { + walk->action = ACTION_AGAIN; + return 0; + } + + /* Huge pud */ + walk->action = ACTION_CONTINUE; if (pud_trans_huge(pudval) || pud_devmap(pudval)) WARN_ON(pud_write(pudval) || pud_dirty(pudval));