diff options
| author | Martin Schwidefsky <schwidefsky@de.ibm.com> | 2010-08-24 09:26:21 +0200 |
|---|---|---|
| committer | Martin Schwidefsky <sky@mschwide.boeblingen.de.ibm.com> | 2010-08-24 09:26:34 +0200 |
| commit | 050eef364ad700590a605a0749f825cab4834b1e (patch) | |
| tree | 2714c9cf7edcbf394971cc2c929e5ab2ea34d6a6 /arch/s390/include/asm/pgtable.h | |
| parent | 7af048dc7639db5202c56fecf2346c310647a218 (diff) | |
[S390] fix tlb flushing vs. concurrent /proc accesses
The tlb flushing code uses the mm_users field of the mm_struct to
decide if each page table entry needs to be flushed individually with
IPTE or if a global flush for the mm_struct is sufficient after all page
table updates have been done. The comment for mm_users says "How many
users with user space?" but the /proc code increases mm_users after it
found the process structure by pid without creating a new user process.
Which makes mm_users useless for the decision between the two tlb
flusing methods. The current code can be confused to not flush tlb
entries by a concurrent access to /proc files if e.g. a fork is in
progres. The solution for this problem is to make the tlb flushing
logic independent from the mm_users field.
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
Diffstat (limited to 'arch/s390/include/asm/pgtable.h')
| -rw-r--r-- | arch/s390/include/asm/pgtable.h | 6 |
1 files changed, 4 insertions, 2 deletions
diff --git a/arch/s390/include/asm/pgtable.h b/arch/s390/include/asm/pgtable.h index 89a504c3f12..3157441ee1d 100644 --- a/arch/s390/include/asm/pgtable.h +++ b/arch/s390/include/asm/pgtable.h @@ -880,7 +880,8 @@ static inline void ptep_invalidate(struct mm_struct *mm, #define ptep_get_and_clear(__mm, __address, __ptep) \ ({ \ pte_t __pte = *(__ptep); \ - if (atomic_read(&(__mm)->mm_users) > 1 || \ + (__mm)->context.flush_mm = 1; \ + if (atomic_read(&(__mm)->context.attach_count) > 1 || \ (__mm) != current->active_mm) \ ptep_invalidate(__mm, __address, __ptep); \ else \ @@ -923,7 +924,8 @@ static inline pte_t ptep_get_and_clear_full(struct mm_struct *mm, ({ \ pte_t __pte = *(__ptep); \ if (pte_write(__pte)) { \ - if (atomic_read(&(__mm)->mm_users) > 1 || \ + (__mm)->context.flush_mm = 1; \ + if (atomic_read(&(__mm)->context.attach_count) > 1 || \ (__mm) != current->active_mm) \ ptep_invalidate(__mm, __addr, __ptep); \ set_pte_at(__mm, __addr, __ptep, pte_wrprotect(__pte)); \ |