| Message ID | 20240429034738.138609-3-shikemeng@huaweicloud.com (mailing list archive) |
|---|---|
| State | New |
| Headers | show |
| Series | Add helper functions to remove repeated code and | expand |
Hello, On Mon, Apr 29, 2024 at 11:47:30AM +0800, Kemeng Shi wrote: > +/* > + * Dirty background will ignore pages being written as we're trying to > + * decide whether to put more under writeback. > + */ > +static void domain_dirty_avail(struct dirty_throttle_control *dtc, bool bg) I wonder whether it'd be better if the bool arg is flipped to something like `bool include_writeback` so that it's clear what the difference is between the two. Also, do global_domain_dirty_avail() and wb_domain_dirty_avail() have to be separate functions? They seem trivial enough to include into the body of domain_dirty_avail(). Are they used directly elsewhere? > +{ > + struct dirty_throttle_control *gdtc = mdtc_gdtc(dtc); > + > + if (gdtc) I know this test is used elsewhere but it isn't the most intuitive. Would it make sense to add dtc_is_global() (or dtc_is_gdtc()) helper instead? > + wb_domain_dirty_avail(dtc, bg); > + else > + global_domain_dirty_avail(dtc, bg); > +} Thanks.
on 5/2/2024 12:49 AM, Tejun Heo wrote: > Hello, > > On Mon, Apr 29, 2024 at 11:47:30AM +0800, Kemeng Shi wrote: >> +/* >> + * Dirty background will ignore pages being written as we're trying to >> + * decide whether to put more under writeback. >> + */ >> +static void domain_dirty_avail(struct dirty_throttle_control *dtc, bool bg) > > I wonder whether it'd be better if the bool arg is flipped to something like > `bool include_writeback` so that it's clear what the difference is between Sure, I rename 'bool bg' to 'bool include_writeback'. > the two. Also, do global_domain_dirty_avail() and wb_domain_dirty_avail() > have to be separate functions? They seem trivial enough to include into the > body of domain_dirty_avail(). Are they used directly elsewhere? I will fold global_domain_dirty_avail() and wb_domain_dirty_avail() and just use domain_dirty_avail. > >> +{ >> + struct dirty_throttle_control *gdtc = mdtc_gdtc(dtc); >> + >> + if (gdtc) > > I know this test is used elsewhere but it isn't the most intuitive. Would it > make sense to add dtc_is_global() (or dtc_is_gdtc()) helper instead? Will add helper dtc_is_global(). Thanks. Kemeng > >> + wb_domain_dirty_avail(dtc, bg); >> + else >> + global_domain_dirty_avail(dtc, bg); >> +} > > Thanks. >
diff --git a/mm/page-writeback.c b/mm/page-writeback.c index e1f73643aca1..e4f181d52035 100644 --- a/mm/page-writeback.c +++ b/mm/page-writeback.c @@ -837,6 +837,41 @@ static void mdtc_calc_avail(struct dirty_throttle_control *mdtc, mdtc->avail = filepages + min(headroom, other_clean); } +static inline void +global_domain_dirty_avail(struct dirty_throttle_control *dtc, bool bg) +{ + dtc->avail = global_dirtyable_memory(); + dtc->dirty = global_node_page_state(NR_FILE_DIRTY); + if (!bg) + dtc->dirty += global_node_page_state(NR_WRITEBACK); +} + +static inline void +wb_domain_dirty_avail(struct dirty_throttle_control *dtc, bool bg) +{ + unsigned long filepages = 0, headroom = 0, writeback = 0; + + mem_cgroup_wb_stats(dtc->wb, &filepages, &headroom, &dtc->dirty, + &writeback); + if (!bg) + dtc->dirty += writeback; + mdtc_calc_avail(dtc, filepages, headroom); +} + +/* + * Dirty background will ignore pages being written as we're trying to + * decide whether to put more under writeback. + */ +static void domain_dirty_avail(struct dirty_throttle_control *dtc, bool bg) +{ + struct dirty_throttle_control *gdtc = mdtc_gdtc(dtc); + + if (gdtc) + wb_domain_dirty_avail(dtc, bg); + else + global_domain_dirty_avail(dtc, bg); +} + /** * __wb_calc_thresh - @wb's share of dirty threshold * @dtc: dirty_throttle_context of interest @@ -2119,14 +2154,8 @@ bool wb_over_bg_thresh(struct bdi_writeback *wb) struct dirty_throttle_control * const mdtc = mdtc_valid(&mdtc_stor) ? &mdtc_stor : NULL; - /* - * Similar to balance_dirty_pages() but ignores pages being written - * as we're trying to decide whether to put more under writeback. - */ - gdtc->avail = global_dirtyable_memory(); - gdtc->dirty = global_node_page_state(NR_FILE_DIRTY); + domain_dirty_avail(gdtc, true); domain_dirty_limits(gdtc); - if (gdtc->dirty > gdtc->bg_thresh) return true; @@ -2135,13 +2164,8 @@ bool wb_over_bg_thresh(struct bdi_writeback *wb) return true; if (mdtc) { - unsigned long filepages, headroom, writeback; - - mem_cgroup_wb_stats(wb, &filepages, &headroom, &mdtc->dirty, - &writeback); - mdtc_calc_avail(mdtc, filepages, headroom); + domain_dirty_avail(mdtc, true); domain_dirty_limits(mdtc); /* ditto, ignore writeback */ - if (mdtc->dirty > mdtc->bg_thresh) return true;
Add general function domain_dirty_avail to calculate dirty and avail for either dirty limit or background writeback in either global domain or wb domain. Signed-off-by: Kemeng Shi <shikemeng@huaweicloud.com> --- mm/page-writeback.c | 50 +++++++++++++++++++++++++++++++++------------ 1 file changed, 37 insertions(+), 13 deletions(-)