Skip to content

mm: Fix various readahead quirks

Jan Kara linux-fsdevel linux-mm 1 version Applied

Status from this message: [merged mm-stable] readahead-simplify-gotos-in-page_cache_sync_ra.patch removed from -mm tree

Get series with b4
b4 am [email protected]

Versions newest first

  1. v1 Latest10 patches 4 repliesApplied 10/10

Patches in v1 All 10 applied

#Subject and filesSizeRepliesApplied
1/10readahead: Make sure sync readahead reads needed pagemm/readahead.c+1 -10 repliesApplied
2/10filemap: Fix page_cache_next_miss() when no hole foundmm/filemap.c+3 -30 repliesApplied
3/10readahead: Properly shorten readahead when falling back to do_page_cache_ra()mm/readahead.c+3 -20 repliesApplied
4/10readahead: Drop pointless index from force_page_cache_ra()mm/readahead.c+1 -40 repliesApplied
5/10readahead: Drop index argument of page_cache_async_readahead()fs/btrfs/relocation.c fs/btrfs/send.c include/linux/pagemap.h+5 -70 repliesApplied
6/10readahead: Drop dead code in page_cache_ra_order()mm/readahead.c+0 -50 repliesApplied
7/10readahead: Drop dead code in ondemand_readahead()mm/readahead.c+2 -190 repliesApplied
8/10readahead: Disentangle async and sync readaheadmm/readahead.c+77 -850 repliesApplied
9/10readahead: Fold try_context_readahead() into its single callermm/readahead.c+22 -620 repliesApplied
10/10readahead: Simplify gotos in page_cache_sync_ra()mm/readahead.c+9 -170 repliesApplied

Reviews trailers per patch

None of the 10 patches has a Reviewed-by or Acked-by of their own; the series has one for all of them.

PatchJosef BacikZhang Peng
SeriesR-b T-b
1/10
2/10
3/10
4/10
5/10
6/10
7/10
8/10
9/10
10/10

Trailers given in replies are filled; those carried in the patches from earlier versions are outlined with dashes.

Read the v1 thread

Cover letter [PATCH 0/10] mm: Fix various readahead quirks

Hello!

When we were internally testing performance of recent kernels, we have noticed quite variable performance of readahead arising from various quirks in readahead code. So I went on a cleaning spree there. This is a batch of patches resulting out of that. A quick testing in my test VM with the following fio job file:

[global]
direct=0
ioengine=sync
invalidate=1
blocksize=4k
size=10g
readwrite=read

[reader]
numjobs=1

shows that this patch series improves the throughput from variable one in 310-340 MB/s range to rather stable one at 350 MB/s. As a side effect these cleanups also address the issue noticed by Bruz Zhang [1].

Honza

[1] https://lore.kernel.org/all/[email protected]/