From: The pNFS Team <linux-nfs@xxxxxxxxxxxxxxx> Signed-off-by: Andy Adamson <andros@xxxxxxxxxxxxxx> Signed-off-by: Dean Hildebrand <dhildeb@xxxxxxxxxx> Signed-off-by: Fred Isaman <iisaman@xxxxxxxxxxxxxx> Signed-off-by: Fred Isaman <iisaman@xxxxxxxxxx> Signed-off-by: Mingyang Guo <guomingyang@xxxxxxxxxxxx> Signed-off-by: Oleg Drokin <green@xxxxxxxxxxxxxx> Signed-off-by: Ricardo Labiaga <Ricardo.Labiaga@xxxxxxxxxx> Tested-by: Guo Mingyang <guomingyang@xxxxxxxxxxxx> Signed-off-by: Andy Adamson <andros@xxxxxxxxxx> Signed-off-by: Benny Halevy <bhalevy@xxxxxxxxxxx> --- fs/nfs/nfs4filelayout.c | 88 +++++++++++++++++++++++++++++++++++++++++++++++ 1 files changed, 88 insertions(+), 0 deletions(-) diff --git a/fs/nfs/nfs4filelayout.c b/fs/nfs/nfs4filelayout.c index 6ec9957..3daf351 100644 --- a/fs/nfs/nfs4filelayout.c +++ b/fs/nfs/nfs4filelayout.c @@ -99,6 +99,93 @@ filelayout_get_dserver_offset(struct pnfs_layout_segment *lseg, loff_t offset) } /* + * Call ops for the async read/write cases + * In the case of dense layouts, the offset needs to be reset to its + * original value. + */ +static void filelayout_read_call_done(struct rpc_task *task, void *data) +{ + struct nfs_read_data *rdata = (struct nfs_read_data *)data; + + if (rdata->fldata.orig_offset) { + dprintk("%s new off %llu orig offset %llu\n", __func__, + rdata->args.offset, rdata->fldata.orig_offset); + rdata->args.offset = rdata->fldata.orig_offset; + } + + /* Note this may cause RPC to be resent */ + rdata->pdata.call_ops->rpc_call_done(task, data); +} + +static void filelayout_read_release(void *data) +{ + struct nfs_read_data *rdata = (struct nfs_read_data *)data; + + put_lseg(rdata->pdata.lseg); + rdata->pdata.lseg = NULL; + rdata->pdata.call_ops->rpc_release(data); +} + +struct rpc_call_ops filelayout_read_call_ops = { + .rpc_call_prepare = nfs_read_prepare, + .rpc_call_done = filelayout_read_call_done, + .rpc_release = filelayout_read_release, +}; + +/* Perform sync or async reads. + * + * An optimization for the NFS file layout driver + * allows the original read/write data structs to be passed in the + * last argument. + * + * TODO: join with write_pagelist? + */ +static enum pnfs_try_status +filelayout_read_pagelist(struct nfs_read_data *data, unsigned nr_pages) +{ + struct pnfs_layout_segment *lseg = data->pdata.lseg; + struct nfs4_pnfs_ds *ds; + loff_t offset = data->args.offset; + u32 idx; + struct nfs_fh *fh; + + dprintk("--> %s ino %lu nr_pages %d pgbase %u req %Zu@%llu\n", + __func__, data->inode->i_ino, nr_pages, + data->args.pgbase, (size_t)data->args.count, offset); + + /* Retrieve the correct rpc_client for the byte range */ + idx = nfs4_fl_calc_ds_index(lseg, offset); + ds = nfs4_fl_prepare_ds(lseg, idx); + if (!ds) { + printk(KERN_ERR "%s: prepare_ds failed, use MDS\n", __func__); + return PNFS_NOT_ATTEMPTED; + } + dprintk("%s USE DS:ip %x %hu\n", __func__, + ntohl(ds->ds_ip_addr), ntohs(ds->ds_port)); + + /* just try the first data server for the index..*/ + data->fldata.ds_nfs_client = ds->ds_clp; + fh = nfs4_fl_select_ds_fh(lseg, offset); + if (fh) + data->args.fh = fh; + + /* + * Now get the file offset on the dserver + * Set the read offset to this offset, and + * save the original offset in orig_offset + * In the case of aync reads, the offset will be reset in the + * call_ops->rpc_call_done() routine. + */ + data->args.offset = filelayout_get_dserver_offset(lseg, offset); + data->fldata.orig_offset = offset; + + /* Perform an asynchronous read */ + nfs_initiate_read(data, ds->ds_clp->cl_rpcclient, + &filelayout_read_call_ops); + return PNFS_ATTEMPTED; +} + +/* * filelayout_check_layout() * * Make sure layout segment parameters are sane WRT the device. @@ -320,6 +407,7 @@ static struct pnfs_layoutdriver_type filelayout_type = { .alloc_lseg = filelayout_alloc_lseg, .free_lseg = filelayout_free_lseg, .pg_test = filelayout_pg_test, + .read_pagelist = filelayout_read_pagelist, }; static int __init nfs4filelayout_init(void) -- 1.6.6 -- To unsubscribe from this list: send the line "unsubscribe linux-nfs" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html