rasize (ra_pages) should be set higher than read size by default to allow parallel reads when reading large files in order to improve performance (otherwise there is much dead time on the network when doing readahead of large files). Default rasize to twice readsize. See attached patch This improved performance for me (although there may be cases where setting it more than 2*read_size still could help). ceph IIRC sets ra_pages even higher (to 8MB by default). Any thoughts? -- Thanks, Steve
From 68aac85b9b697cfc1dc1286f596ebbddb9ae3c56 Mon Sep 17 00:00:00 2001 From: Steve French <stfrench@xxxxxxxxxxxxx> Date: Mon, 8 May 2023 00:45:45 -0500 Subject: [PATCH] smb3: improve parallel reads of large files rasize (ra_pages) should be set higher than read size by default to allow parallel reads when reading large files in order to improve performance (otherwise there is much dead time on the network when doing readahead of large files). Default rasize to twice readsize. Signed-off-by: Steve French <stfrench@xxxxxxxxxxxxx> --- fs/cifs/cifsfs.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/fs/cifs/cifsfs.c b/fs/cifs/cifsfs.c index 32f7c81a7b89..81430abacf93 100644 --- a/fs/cifs/cifsfs.c +++ b/fs/cifs/cifsfs.c @@ -246,7 +246,7 @@ cifs_read_super(struct super_block *sb) if (cifs_sb->ctx->rasize) sb->s_bdi->ra_pages = cifs_sb->ctx->rasize / PAGE_SIZE; else - sb->s_bdi->ra_pages = cifs_sb->ctx->rsize / PAGE_SIZE; + sb->s_bdi->ra_pages = 2 * (cifs_sb->ctx->rsize / PAGE_SIZE); sb->s_blocksize = CIFS_MAX_MSGSIZE; sb->s_blocksize_bits = 14; /* default 2**14 = CIFS_MAX_MSGSIZE */ -- 2.34.1
