* [PATCH 0/2] update pnfs-submit-wave3 @ 2011-02-14 19:31 andros 2011-02-14 19:31 ` [PATCH 1/2] SQUASHME pnfs_submit wave3 general cleanup andros 2011-02-14 22:16 ` [PATCH 0/2] update pnfs-submit-wave3 Benny Halevy 0 siblings, 2 replies; 4+ messages in thread From: andros @ 2011-02-14 19:31 UTC (permalink / raw) To: bhalevy; +Cc: linux-nfs These two patches bring Benny's git://linux-nfs.org/~bhalevy/linux-pnfs.git pnfs-submit-wave3-rev5 branch in line with the wave 3 code just submitted for review, caveat some author and patch comment differences. The 16 patch wave 3 patch set can be found at git://linux-nfs.org/projects/andros/nfs-2.6.git wave3-on-nfs-for-next-rev7 0001-SQUASHME-pnfs_submit-wave3-general-cleanup.patch 0002-SQUASHME-pnfs-submit-wave3-avoid-calculating-j-index.patch -->Andy ^ permalink raw reply [flat|nested] 4+ messages in thread
* [PATCH 1/2] SQUASHME pnfs_submit wave3 general cleanup 2011-02-14 19:31 [PATCH 0/2] update pnfs-submit-wave3 andros @ 2011-02-14 19:31 ` andros 2011-02-14 19:31 ` [PATCH 2/2] SQUASHME pnfs-submit wave3 avoid calculating j index twice per call andros 2011-02-14 22:16 ` [PATCH 0/2] update pnfs-submit-wave3 Benny Halevy 1 sibling, 1 reply; 4+ messages in thread From: andros @ 2011-02-14 19:31 UTC (permalink / raw) To: bhalevy; +Cc: linux-nfs, Andy Adamson From: Andy Adamson <andros@netapp.com> Signed-off-by: Andy Adamson <andros@netapp.com> --- fs/nfs/client.c | 11 ++++++----- fs/nfs/nfs4filelayoutdev.c | 2 +- fs/nfs/nfs4proc.c | 2 +- fs/nfs/pagelist.c | 5 +++-- fs/nfs/pnfs.h | 1 + fs/nfs/read.c | 20 +++++++------------- 6 files changed, 19 insertions(+), 22 deletions(-) diff --git a/fs/nfs/client.c b/fs/nfs/client.c index 17da633..78e6ebe 100644 --- a/fs/nfs/client.c +++ b/fs/nfs/client.c @@ -490,11 +490,12 @@ static struct nfs_client *nfs_match_client(const struct nfs_client_initdata *dat * Look up a client by IP address and protocol version * - creates a new record if one doesn't yet exist */ -static struct nfs_client *nfs_get_client(struct nfs_client_initdata *cl_init, - const struct rpc_timeout *timeparms, - const char *ip_addr, - rpc_authflavor_t authflavour, - int noresvport) +static struct nfs_client * +nfs_get_client(const struct nfs_client_initdata *cl_init, + const struct rpc_timeout *timeparms, + const char *ip_addr, + rpc_authflavor_t authflavour, + int noresvport) { struct nfs_client *clp, *new = NULL; int error; diff --git a/fs/nfs/nfs4filelayoutdev.c b/fs/nfs/nfs4filelayoutdev.c index b8c3681..1e1c536 100644 --- a/fs/nfs/nfs4filelayoutdev.c +++ b/fs/nfs/nfs4filelayoutdev.c @@ -586,8 +586,8 @@ nfs4_fl_prepare_ds(struct pnfs_layout_segment *lseg, u32 ds_idx) struct nfs_server *s = NFS_SERVER(lseg->pls_layout->plh_inode); int err; - /* Already tried to connect, don't try again */ if (dsaddr->deviceid.de_flags & NFS4_DEVICE_ID_NEG_ENTRY) { + /* Already tried to connect, don't try again */ dprintk("%s Deviceid marked out of use\n", __func__); return NULL; } diff --git a/fs/nfs/nfs4proc.c b/fs/nfs/nfs4proc.c index aab3efd..0f73db0 100644 --- a/fs/nfs/nfs4proc.c +++ b/fs/nfs/nfs4proc.c @@ -3111,7 +3111,7 @@ static void nfs4_proc_read_setup(struct nfs_read_data *data, struct rpc_message /* Reset the the nfs_read_data to send the read to the MDS. */ void nfs4_reset_read(struct rpc_task *task, struct nfs_read_data *data) { - dprintk("%s Reset task for i/o through \n", __func__); + dprintk("%s Reset task for i/o through\n", __func__); /* offsets will differ in the dense stripe case */ data->args.offset = data->mds_offset; data->ds_clp = NULL; diff --git a/fs/nfs/pagelist.c b/fs/nfs/pagelist.c index f0aa20b..2c793a7 100644 --- a/fs/nfs/pagelist.c +++ b/fs/nfs/pagelist.c @@ -257,8 +257,9 @@ static int nfs_can_coalesce_requests(struct nfs_page *prev, return 0; if (prev->wb_pgbase + prev->wb_bytes != PAGE_CACHE_SIZE) return 0; - /* For non-whole file layouts, need to check that req is inside of - * pgio->pg_test. + /* + * Non-whole file layouts need to check that req is inside of + * pgio->pg_lseg. */ if (pgio->pg_test && !pgio->pg_test(pgio, prev, req)) return 0; diff --git a/fs/nfs/pnfs.h b/fs/nfs/pnfs.h index d2d708a..a760363 100644 --- a/fs/nfs/pnfs.h +++ b/fs/nfs/pnfs.h @@ -194,6 +194,7 @@ void pnfs_roc_release(struct inode *ino); void pnfs_roc_set_barrier(struct inode *ino, u32 barrier); bool pnfs_roc_drain(struct inode *ino, u32 *barrier); + static inline int lo_fail_bit(u32 iomode) { return iomode == IOMODE_RW ? diff --git a/fs/nfs/read.c b/fs/nfs/read.c index 896dc4e..9447156 100644 --- a/fs/nfs/read.c +++ b/fs/nfs/read.c @@ -22,7 +22,6 @@ #include <linux/module.h> #include <asm/system.h> -#include <linux/module.h> #include "pnfs.h" #include "nfs4_fs.h" @@ -185,7 +184,8 @@ int nfs_initiate_read(struct nfs_read_data *data, struct rpc_clnt *clnt, /* Set up the initial task struct. */ NFS_PROTO(inode)->read_setup(data, &msg); - dprintk("NFS: %5u initiated read call (req %s/%Ld, %u bytes @ offset %Lu)\n", + dprintk("NFS: %5u initiated read call (req %s/%lld, %u bytes @ " + "offset %llu)\n", data->task.tk_pid, inode->i_sb->s_id, (long long)NFS_FILEID(inode), @@ -200,16 +200,6 @@ int nfs_initiate_read(struct nfs_read_data *data, struct rpc_clnt *clnt, } EXPORT_SYMBOL(nfs_initiate_read); -static int pnfs_initiate_read(struct nfs_read_data *data, struct rpc_clnt *clnt, - const struct rpc_call_ops *call_ops) -{ - if (data->lseg && - (pnfs_try_to_read_data(data, call_ops) == PNFS_ATTEMPTED)) - return 0; - - return nfs_initiate_read(data, clnt, call_ops); -} - /* * Set up the NFS read request struct */ @@ -238,7 +228,11 @@ static int nfs_read_rpcsetup(struct nfs_page *req, struct nfs_read_data *data, data->res.eof = 0; nfs_fattr_init(&data->fattr); - return pnfs_initiate_read(data, NFS_CLIENT(inode), call_ops); + if (data->lseg && + (pnfs_try_to_read_data(data, call_ops) == PNFS_ATTEMPTED)) + return 0; + + return nfs_initiate_read(data, NFS_CLIENT(inode), call_ops); } static void -- 1.6.6 ^ permalink raw reply related [flat|nested] 4+ messages in thread
* [PATCH 2/2] SQUASHME pnfs-submit wave3 avoid calculating j index twice per call 2011-02-14 19:31 ` [PATCH 1/2] SQUASHME pnfs_submit wave3 general cleanup andros @ 2011-02-14 19:31 ` andros 0 siblings, 0 replies; 4+ messages in thread From: andros @ 2011-02-14 19:31 UTC (permalink / raw) To: bhalevy; +Cc: linux-nfs, Fred Isaman From: Fred Isaman <iisaman@netapp.com> Calculating the j index is expensive, so save and reuse it instead of recalculating. Signed-off-by: Fred Isaman <iisaman@netapp.com> --- fs/nfs/nfs4filelayout.c | 7 ++++--- fs/nfs/nfs4filelayout.h | 5 +++-- fs/nfs/nfs4filelayoutdev.c | 16 +++++++--------- 3 files changed, 14 insertions(+), 14 deletions(-) diff --git a/fs/nfs/nfs4filelayout.c b/fs/nfs/nfs4filelayout.c index 80e7a66..3768377 100644 --- a/fs/nfs/nfs4filelayout.c +++ b/fs/nfs/nfs4filelayout.c @@ -220,7 +220,7 @@ filelayout_read_pagelist(struct nfs_read_data *data) struct pnfs_layout_segment *lseg = data->lseg; struct nfs4_pnfs_ds *ds; loff_t offset = data->args.offset; - u32 idx; + u32 j, idx; struct nfs_fh *fh; dprintk("--> %s ino %lu pgbase %u req %Zu@%llu\n", @@ -228,7 +228,8 @@ filelayout_read_pagelist(struct nfs_read_data *data) 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); + j = nfs4_fl_calc_j_index(lseg, offset); + idx = nfs4_fl_calc_ds_index(lseg, j); ds = nfs4_fl_prepare_ds(lseg, idx); if (!ds) { /* Either layout fh index faulty, or ds connect failed */ @@ -241,7 +242,7 @@ filelayout_read_pagelist(struct nfs_read_data *data) /* No multipath support. Use first DS */ data->ds_clp = ds->ds_clp; - fh = nfs4_fl_select_ds_fh(lseg, offset); + fh = nfs4_fl_select_ds_fh(lseg, j); if (fh) data->args.fh = fh; diff --git a/fs/nfs/nfs4filelayout.h b/fs/nfs/nfs4filelayout.h index 7e33bd8..1809aa6 100644 --- a/fs/nfs/nfs4filelayout.h +++ b/fs/nfs/nfs4filelayout.h @@ -84,12 +84,13 @@ FILELAYOUT_LSEG(struct pnfs_layout_segment *lseg) } extern struct nfs_fh * -nfs4_fl_select_ds_fh(struct pnfs_layout_segment *lseg, loff_t offset); +nfs4_fl_select_ds_fh(struct pnfs_layout_segment *lseg, u32 j); extern void nfs4_fl_free_deviceid_callback(struct pnfs_deviceid_node *); extern void print_ds(struct nfs4_pnfs_ds *ds); extern void print_deviceid(struct nfs4_deviceid *dev_id); -u32 nfs4_fl_calc_ds_index(struct pnfs_layout_segment *lseg, loff_t offset); +u32 nfs4_fl_calc_j_index(struct pnfs_layout_segment *lseg, loff_t offset); +u32 nfs4_fl_calc_ds_index(struct pnfs_layout_segment *lseg, u32 j); struct nfs4_pnfs_ds *nfs4_fl_prepare_ds(struct pnfs_layout_segment *lseg, u32 ds_idx); extern struct nfs4_file_layout_dsaddr * diff --git a/fs/nfs/nfs4filelayoutdev.c b/fs/nfs/nfs4filelayoutdev.c index 1e1c536..2780d97 100644 --- a/fs/nfs/nfs4filelayoutdev.c +++ b/fs/nfs/nfs4filelayoutdev.c @@ -517,8 +517,8 @@ nfs4_fl_find_get_deviceid(struct nfs_client *clp, struct nfs4_deviceid *id) * Want res = (offset - layout->pattern_offset)/ layout->stripe_unit * Then: ((res + fsi) % dsaddr->stripe_count) */ -static u32 -_nfs4_fl_calc_j_index(struct pnfs_layout_segment *lseg, loff_t offset) +u32 +nfs4_fl_calc_j_index(struct pnfs_layout_segment *lseg, loff_t offset) { struct nfs4_filelayout_segment *flseg = FILELAYOUT_LSEG(lseg); u64 tmp; @@ -530,16 +530,13 @@ _nfs4_fl_calc_j_index(struct pnfs_layout_segment *lseg, loff_t offset) } u32 -nfs4_fl_calc_ds_index(struct pnfs_layout_segment *lseg, loff_t offset) +nfs4_fl_calc_ds_index(struct pnfs_layout_segment *lseg, u32 j) { - u32 j; - - j = _nfs4_fl_calc_j_index(lseg, offset); return FILELAYOUT_LSEG(lseg)->dsaddr->stripe_indices[j]; } struct nfs_fh * -nfs4_fl_select_ds_fh(struct pnfs_layout_segment *lseg, loff_t offset) +nfs4_fl_select_ds_fh(struct pnfs_layout_segment *lseg, u32 j) { struct nfs4_filelayout_segment *flseg = FILELAYOUT_LSEG(lseg); u32 i; @@ -551,11 +548,12 @@ nfs4_fl_select_ds_fh(struct pnfs_layout_segment *lseg, loff_t offset) /* Use the MDS OPEN fh set in nfs_read_rpcsetup */ return NULL; else - i = nfs4_fl_calc_ds_index(lseg, offset); + i = nfs4_fl_calc_ds_index(lseg, j); } else - i = _nfs4_fl_calc_j_index(lseg, offset); + i = j; return flseg->fh_array[i]; } + void filelayout_mark_devid_negative(struct nfs_client *mds_clp, struct pnfs_deviceid_node *devid, -- 1.6.6 ^ permalink raw reply related [flat|nested] 4+ messages in thread
* Re: [PATCH 0/2] update pnfs-submit-wave3 2011-02-14 19:31 [PATCH 0/2] update pnfs-submit-wave3 andros 2011-02-14 19:31 ` [PATCH 1/2] SQUASHME pnfs_submit wave3 general cleanup andros @ 2011-02-14 22:16 ` Benny Halevy 1 sibling, 0 replies; 4+ messages in thread From: Benny Halevy @ 2011-02-14 22:16 UTC (permalink / raw) To: andros; +Cc: linux-nfs On 2011-02-14 14:31, andros@netapp.com wrote: > These two patches bring Benny's git://linux-nfs.org/~bhalevy/linux-pnfs.git > pnfs-submit-wave3-rev5 branch in line with the wave 3 code just submitted for > review, caveat some author and patch comment differences. > > The 16 patch wave 3 patch set can be found at > git://linux-nfs.org/projects/andros/nfs-2.6.git wave3-on-nfs-for-next-rev7 > > 0001-SQUASHME-pnfs_submit-wave3-general-cleanup.patch > 0002-SQUASHME-pnfs-submit-wave3-avoid-calculating-j-index.patch > > > -->Andy Looks good to me. Merged to pnfs-submit-wave3, thanks! Benny ^ permalink raw reply [flat|nested] 4+ messages in thread
end of thread, other threads:[~2011-02-14 22:16 UTC | newest] Thread overview: 4+ messages (download: mbox.gz follow: Atom feed -- links below jump to the message on this page -- 2011-02-14 19:31 [PATCH 0/2] update pnfs-submit-wave3 andros 2011-02-14 19:31 ` [PATCH 1/2] SQUASHME pnfs_submit wave3 general cleanup andros 2011-02-14 19:31 ` [PATCH 2/2] SQUASHME pnfs-submit wave3 avoid calculating j index twice per call andros 2011-02-14 22:16 ` [PATCH 0/2] update pnfs-submit-wave3 Benny Halevy
This is an external index of several public inboxes, see mirroring instructions on how to clone and mirror all data and code used by this external index.