| Message ID | 20240621144246.11148-1-jack@suse.cz (mailing list archive) |
|---|---|
| State | New |
| Headers | show |
| Series | mm: Avoid possible overflows in dirty throttling | expand |
On Fri, Jun 21, 2024 at 7:42 AM Jan Kara <jack@suse.cz> wrote: > > This reverts commit 9319b647902cbd5cc884ac08a8a6d54ce111fc78. > > The commit is broken in several ways. Firstly, the removed (u64) cast > from the multiplication will introduce a multiplication overflow on > 32-bit archs if wb_thresh * bg_thresh >= 1<<32 (which is actually common > - the default settings with 4GB of RAM will trigger this). Secondly, the > div64_u64() is unnecessarily expensive on 32-bit archs. We have > div64_ul() in case we want to be safe & cheap. Thirdly, if dirty > thresholds are larger than 1<<32 pages, then dirty balancing is > going to blow up in many other spectacular ways anyway so trying to fix > one possible overflow is just moot. > > CC: stable@vger.kernel.org > Fixes: 9319b647902c ("mm/writeback: fix possible divide-by-zero in wb_dirty_limits(), again") > Signed-off-by: Jan Kara <jack@suse.cz> > --- > mm/page-writeback.c | 2 +- > 1 file changed, 1 insertion(+), 1 deletion(-) > > diff --git a/mm/page-writeback.c b/mm/page-writeback.c > index 12c9297ed4a7..2573e2d504af 100644 > --- a/mm/page-writeback.c > +++ b/mm/page-writeback.c > @@ -1660,7 +1660,7 @@ static inline void wb_dirty_limits(struct dirty_throttle_control *dtc) > */ > dtc->wb_thresh = __wb_calc_thresh(dtc, dtc->thresh); > dtc->wb_bg_thresh = dtc->thresh ? > - div64_u64(dtc->wb_thresh * dtc->bg_thresh, dtc->thresh) : 0; > + div_u64((u64)dtc->wb_thresh * dtc->bg_thresh, dtc->thresh) : 0; > > /* > * In order to avoid the stacked BDI deadlock we need > -- > 2.35.3 > > Thanks Jan, Reviewed-By: Zach O'Keefe <zokeefe@google.com>
diff --git a/mm/page-writeback.c b/mm/page-writeback.c index 12c9297ed4a7..2573e2d504af 100644 --- a/mm/page-writeback.c +++ b/mm/page-writeback.c @@ -1660,7 +1660,7 @@ static inline void wb_dirty_limits(struct dirty_throttle_control *dtc) */ dtc->wb_thresh = __wb_calc_thresh(dtc, dtc->thresh); dtc->wb_bg_thresh = dtc->thresh ? - div64_u64(dtc->wb_thresh * dtc->bg_thresh, dtc->thresh) : 0; + div_u64((u64)dtc->wb_thresh * dtc->bg_thresh, dtc->thresh) : 0; /* * In order to avoid the stacked BDI deadlock we need
This reverts commit 9319b647902cbd5cc884ac08a8a6d54ce111fc78. The commit is broken in several ways. Firstly, the removed (u64) cast from the multiplication will introduce a multiplication overflow on 32-bit archs if wb_thresh * bg_thresh >= 1<<32 (which is actually common - the default settings with 4GB of RAM will trigger this). Secondly, the div64_u64() is unnecessarily expensive on 32-bit archs. We have div64_ul() in case we want to be safe & cheap. Thirdly, if dirty thresholds are larger than 1<<32 pages, then dirty balancing is going to blow up in many other spectacular ways anyway so trying to fix one possible overflow is just moot. CC: stable@vger.kernel.org Fixes: 9319b647902c ("mm/writeback: fix possible divide-by-zero in wb_dirty_limits(), again") Signed-off-by: Jan Kara <jack@suse.cz> --- mm/page-writeback.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)