diff options
| author | Hugh Dickins <hugh@veritas.com> | 2005-11-07 14:08:46 -0800 |
|---|---|---|
| committer | David S. Miller <davem@davemloft.net> | 2005-11-07 14:08:46 -0800 |
| commit | b8ae48656db860d4c83a29aa7b0588fc89361935 (patch) | |
| tree | dbeee0739669c6e77799b5c3a88fa27326647e20 /arch/sparc64/kernel/setup.c | |
| parent | 5a820fa7e1a34f12fec4e6766e5c335ae9427028 (diff) | |
[SPARC64] mm: don't re-evaluate *ptep
sparc64 prom_callback and new_setup_frame32 each operates on a user page
table without holding lock, and no doubt they've good reason. But I'd
feel more confident if they were to do a "pte = *ptep" and then operate
on pte, rather than re-evaluating *ptep.
Signed-off-by: Hugh Dickins <hugh@veritas.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Diffstat (limited to 'arch/sparc64/kernel/setup.c')
| -rw-r--r-- | arch/sparc64/kernel/setup.c | 12 |
1 files changed, 8 insertions, 4 deletions
diff --git a/arch/sparc64/kernel/setup.c b/arch/sparc64/kernel/setup.c index c1f34237cdf..bf1849dd9c4 100644 --- a/arch/sparc64/kernel/setup.c +++ b/arch/sparc64/kernel/setup.c @@ -154,6 +154,7 @@ int prom_callback(long *args) pud_t *pudp; pmd_t *pmdp; pte_t *ptep; + pte_t pte; for_each_process(p) { mm = p->mm; @@ -178,8 +179,9 @@ int prom_callback(long *args) * being called from inside OBP. */ ptep = pte_offset_map(pmdp, va); - if (pte_present(*ptep)) { - tte = pte_val(*ptep); + pte = *ptep; + if (pte_present(pte)) { + tte = pte_val(pte); res = PROM_TRUE; } pte_unmap(ptep); @@ -218,6 +220,7 @@ int prom_callback(long *args) pud_t *pudp; pmd_t *pmdp; pte_t *ptep; + pte_t pte; int error; if ((va >= LOW_OBP_ADDRESS) && (va < HI_OBP_ADDRESS)) { @@ -240,8 +243,9 @@ int prom_callback(long *args) * being called from inside OBP. */ ptep = pte_offset_kernel(pmdp, va); - if (pte_present(*ptep)) { - tte = pte_val(*ptep); + pte = *ptep; + if (pte_present(pte)) { + tte = pte_val(pte); res = PROM_TRUE; } goto done; |