Commit 58baac0a authored by Trond Myklebust's avatar Trond Myklebust

Merge branch 'flexfiles'

* flexfiles:
  pNFS/flexfiles: Ensure we record layoutstats even if RPC is terminated early
  pNFS: Add flag to track if we've called nfs4_ff_layout_stat_io_start_read/write
  pNFS/flexfiles: Fix a statistics gathering imbalance
  pNFS/flexfiles: Don't mark the entire layout as failed, when returning it
  pNFS/flexfiles: Don't prevent flexfiles client from retrying LAYOUTGET
  pnfs/flexfiles: count io stat in rpc_count_stats callback
  pnfs/flexfiles: do not mark delay-like status as DS failure
  NFS41: map NFS4ERR_LAYOUTUNAVAILABLE to ENODATA
  nfs: only remove page from mapping if launder_page fails
  nfs: handle request add failure properly
  nfs: centralize pgio error cleanup
  nfs: clean up rest of reqs when failing to add one
  NFS41: pop some layoutget errors to application
  pNFS/flexfiles: Support server-supplied layoutstats sampling period
parents 8d0ed0ca 4d0ac221
......@@ -670,6 +670,10 @@ static void nfs_direct_write_reschedule(struct nfs_direct_req *dreq)
req = nfs_list_entry(reqs.next);
nfs_direct_setup_mirroring(dreq, &desc, req);
if (desc.pg_error < 0) {
list_splice_init(&reqs, &failed);
goto out_failed;
}
list_for_each_entry_safe(req, tmp, &reqs, wb_list) {
if (!nfs_pageio_add_request(&desc, req)) {
......@@ -677,13 +681,17 @@ static void nfs_direct_write_reschedule(struct nfs_direct_req *dreq)
nfs_list_add_request(req, &failed);
spin_lock(cinfo.lock);
dreq->flags = 0;
dreq->error = -EIO;
if (desc.pg_error < 0)
dreq->error = desc.pg_error;
else
dreq->error = -EIO;
spin_unlock(cinfo.lock);
}
nfs_release_request(req);
}
nfs_pageio_complete(&desc);
out_failed:
while (!list_empty(&failed)) {
req = nfs_list_entry(failed.next);
nfs_list_remove_request(req);
......@@ -900,6 +908,11 @@ static ssize_t nfs_direct_write_schedule_iovec(struct nfs_direct_req *dreq,
}
nfs_direct_setup_mirroring(dreq, &desc, req);
if (desc.pg_error < 0) {
nfs_free_request(req);
result = desc.pg_error;
break;
}
nfs_lock_request(req);
req->wb_index = pos >> PAGE_SHIFT;
......
......@@ -545,7 +545,7 @@ static int nfs_launder_page(struct page *page)
inode->i_ino, (long long)page_offset(page));
nfs_fscache_wait_on_page_write(nfsi, page);
return nfs_wb_page(inode, page);
return nfs_wb_launder_page(inode, page);
}
static int nfs_swap_activate(struct swap_info_struct *sis, struct file *file,
......
......@@ -202,6 +202,7 @@ static int filelayout_async_handle_error(struct rpc_task *task,
task->tk_status);
nfs4_mark_deviceid_unavailable(devid);
pnfs_error_mark_layout_for_return(inode, lseg);
pnfs_set_lo_fail(lseg);
rpc_wake_up(&tbl->slot_tbl_waitq);
/* fall through */
default:
......@@ -883,13 +884,19 @@ static void
filelayout_pg_init_read(struct nfs_pageio_descriptor *pgio,
struct nfs_page *req)
{
if (!pgio->pg_lseg)
if (!pgio->pg_lseg) {
pgio->pg_lseg = pnfs_update_layout(pgio->pg_inode,
req->wb_context,
0,
NFS4_MAX_UINT64,
IOMODE_READ,
GFP_KERNEL);
if (IS_ERR(pgio->pg_lseg)) {
pgio->pg_error = PTR_ERR(pgio->pg_lseg);
pgio->pg_lseg = NULL;
return;
}
}
/* If no lseg, fall back to read through mds */
if (pgio->pg_lseg == NULL)
nfs_pageio_reset_read_mds(pgio);
......@@ -902,13 +909,20 @@ filelayout_pg_init_write(struct nfs_pageio_descriptor *pgio,
struct nfs_commit_info cinfo;
int status;
if (!pgio->pg_lseg)
if (!pgio->pg_lseg) {
pgio->pg_lseg = pnfs_update_layout(pgio->pg_inode,
req->wb_context,
0,
NFS4_MAX_UINT64,
IOMODE_RW,
GFP_NOFS);
if (IS_ERR(pgio->pg_lseg)) {
pgio->pg_error = PTR_ERR(pgio->pg_lseg);
pgio->pg_lseg = NULL;
return;
}
}
/* If no lseg, fall back to write through mds */
if (pgio->pg_lseg == NULL)
goto out_mds;
......
......@@ -505,9 +505,17 @@ ff_layout_alloc_lseg(struct pnfs_layout_hdr *lh,
}
p = xdr_inline_decode(&stream, 4);
if (p)
fls->flags = be32_to_cpup(p);
if (!p)
goto out_sort_mirrors;
fls->flags = be32_to_cpup(p);
p = xdr_inline_decode(&stream, 4);
if (!p)
goto out_sort_mirrors;
for (i=0; i < fls->mirror_array_cnt; i++)
fls->mirror_array[i]->report_interval = be32_to_cpup(p);
out_sort_mirrors:
ff_layout_sort_mirrors(fls);
rc = ff_layout_check_layout(lgr);
if (rc)
......@@ -603,7 +611,9 @@ nfs4_ff_layoutstat_start_io(struct nfs4_ff_layout_mirror *mirror,
mirror->start_time = now;
if (ktime_equal(mirror->last_report_time, notime))
mirror->last_report_time = now;
if (layoutstats_timer != 0)
if (mirror->report_interval != 0)
report_interval = (s64)mirror->report_interval * 1000LL;
else if (layoutstats_timer != 0)
report_interval = (s64)layoutstats_timer * 1000LL;
if (ktime_to_ms(ktime_sub(now, mirror->last_report_time)) >=
report_interval) {
......@@ -785,13 +795,19 @@ ff_layout_pg_init_read(struct nfs_pageio_descriptor *pgio,
int ds_idx;
/* Use full layout for now */
if (!pgio->pg_lseg)
if (!pgio->pg_lseg) {
pgio->pg_lseg = pnfs_update_layout(pgio->pg_inode,
req->wb_context,
0,
NFS4_MAX_UINT64,
IOMODE_READ,
GFP_KERNEL);
if (IS_ERR(pgio->pg_lseg)) {
pgio->pg_error = PTR_ERR(pgio->pg_lseg);
pgio->pg_lseg = NULL;
return;
}
}
/* If no lseg, fall back to read through mds */
if (pgio->pg_lseg == NULL)
goto out_mds;
......@@ -825,13 +841,19 @@ ff_layout_pg_init_write(struct nfs_pageio_descriptor *pgio,
int i;
int status;
if (!pgio->pg_lseg)
if (!pgio->pg_lseg) {
pgio->pg_lseg = pnfs_update_layout(pgio->pg_inode,
req->wb_context,
0,
NFS4_MAX_UINT64,
IOMODE_RW,
GFP_NOFS);
if (IS_ERR(pgio->pg_lseg)) {
pgio->pg_error = PTR_ERR(pgio->pg_lseg);
pgio->pg_lseg = NULL;
return;
}
}
/* If no lseg, fall back to write through mds */
if (pgio->pg_lseg == NULL)
goto out_mds;
......@@ -867,18 +889,25 @@ static unsigned int
ff_layout_pg_get_mirror_count_write(struct nfs_pageio_descriptor *pgio,
struct nfs_page *req)
{
if (!pgio->pg_lseg)
if (!pgio->pg_lseg) {
pgio->pg_lseg = pnfs_update_layout(pgio->pg_inode,
req->wb_context,
0,
NFS4_MAX_UINT64,
IOMODE_RW,
GFP_NOFS);
if (IS_ERR(pgio->pg_lseg)) {
pgio->pg_error = PTR_ERR(pgio->pg_lseg);
pgio->pg_lseg = NULL;
goto out;
}
}
if (pgio->pg_lseg)
return FF_LAYOUT_MIRROR_COUNT(pgio->pg_lseg);
/* no lseg means that pnfs is not in use, so no mirroring here */
nfs_pageio_reset_write_mds(pgio);
out:
return 1;
}
......@@ -1101,7 +1130,7 @@ static int ff_layout_async_handle_error_v3(struct rpc_task *task,
return -NFS4ERR_RESET_TO_PNFS;
out_retry:
task->tk_status = 0;
rpc_restart_call(task);
rpc_restart_call_prepare(task);
rpc_delay(task, NFS_JUKEBOX_RETRY_TIME);
return -EAGAIN;
}
......@@ -1159,6 +1188,14 @@ static void ff_layout_io_track_ds_error(struct pnfs_layout_segment *lseg,
}
}
switch (status) {
case NFS4ERR_DELAY:
case NFS4ERR_GRACE:
return;
default:
break;
}
mirror = FF_LAYOUT_COMP(lseg, idx);
err = ff_layout_track_ds_error(FF_LAYOUT_FROM_HDR(lseg->pls_layout),
mirror, offset, length, status, opnum,
......@@ -1242,14 +1279,31 @@ ff_layout_reset_to_mds(struct pnfs_layout_segment *lseg, int idx)
return ff_layout_test_devid_unavailable(node);
}
static int ff_layout_read_prepare_common(struct rpc_task *task,
struct nfs_pgio_header *hdr)
static void ff_layout_read_record_layoutstats_start(struct rpc_task *task,
struct nfs_pgio_header *hdr)
{
if (test_and_set_bit(NFS_IOHDR_STAT, &hdr->flags))
return;
nfs4_ff_layout_stat_io_start_read(hdr->inode,
FF_LAYOUT_COMP(hdr->lseg, hdr->pgio_mirror_idx),
hdr->args.count,
task->tk_start);
}
static void ff_layout_read_record_layoutstats_done(struct rpc_task *task,
struct nfs_pgio_header *hdr)
{
if (!test_and_clear_bit(NFS_IOHDR_STAT, &hdr->flags))
return;
nfs4_ff_layout_stat_io_end_read(task,
FF_LAYOUT_COMP(hdr->lseg, hdr->pgio_mirror_idx),
hdr->args.count,
hdr->res.count);
}
static int ff_layout_read_prepare_common(struct rpc_task *task,
struct nfs_pgio_header *hdr)
{
if (unlikely(test_bit(NFS_CONTEXT_BAD, &hdr->args.context->flags))) {
rpc_exit(task, -EIO);
return -EIO;
......@@ -1265,6 +1319,7 @@ static int ff_layout_read_prepare_common(struct rpc_task *task,
}
hdr->pgio_done_cb = ff_layout_read_done_cb;
ff_layout_read_record_layoutstats_start(task, hdr);
return 0;
}
......@@ -1323,10 +1378,6 @@ static void ff_layout_read_call_done(struct rpc_task *task, void *data)
dprintk("--> %s task->tk_status %d\n", __func__, task->tk_status);
nfs4_ff_layout_stat_io_end_read(task,
FF_LAYOUT_COMP(hdr->lseg, hdr->pgio_mirror_idx),
hdr->args.count, hdr->res.count);
if (test_bit(NFS_IOHDR_REDO, &hdr->flags) &&
task->tk_status == 0) {
nfs4_sequence_done(task, &hdr->res.seq_res);
......@@ -1341,10 +1392,20 @@ static void ff_layout_read_count_stats(struct rpc_task *task, void *data)
{
struct nfs_pgio_header *hdr = data;
ff_layout_read_record_layoutstats_done(task, hdr);
rpc_count_iostats_metrics(task,
&NFS_CLIENT(hdr->inode)->cl_metrics[NFSPROC4_CLNT_READ]);
}
static void ff_layout_read_release(void *data)
{
struct nfs_pgio_header *hdr = data;
ff_layout_read_record_layoutstats_done(&hdr->task, hdr);
pnfs_generic_rw_release(data);
}
static int ff_layout_write_done_cb(struct rpc_task *task,
struct nfs_pgio_header *hdr)
{
......@@ -1362,15 +1423,12 @@ static int ff_layout_write_done_cb(struct rpc_task *task,
switch (err) {
case -NFS4ERR_RESET_TO_PNFS:
pnfs_set_retry_layoutget(hdr->lseg->pls_layout);
ff_layout_reset_write(hdr, true);
return task->tk_status;
case -NFS4ERR_RESET_TO_MDS:
pnfs_clear_retry_layoutget(hdr->lseg->pls_layout);
ff_layout_reset_write(hdr, false);
return task->tk_status;
case -EAGAIN:
rpc_restart_call_prepare(task);
return -EAGAIN;
}
......@@ -1402,11 +1460,9 @@ static int ff_layout_commit_done_cb(struct rpc_task *task,
switch (err) {
case -NFS4ERR_RESET_TO_PNFS:
pnfs_set_retry_layoutget(data->lseg->pls_layout);
pnfs_generic_prepare_to_resend_writes(data);
return -EAGAIN;
case -NFS4ERR_RESET_TO_MDS:
pnfs_clear_retry_layoutget(data->lseg->pls_layout);
pnfs_generic_prepare_to_resend_writes(data);
return -EAGAIN;
case -EAGAIN:
......@@ -1421,14 +1477,31 @@ static int ff_layout_commit_done_cb(struct rpc_task *task,
return 0;
}
static int ff_layout_write_prepare_common(struct rpc_task *task,
struct nfs_pgio_header *hdr)
static void ff_layout_write_record_layoutstats_start(struct rpc_task *task,
struct nfs_pgio_header *hdr)
{
if (test_and_set_bit(NFS_IOHDR_STAT, &hdr->flags))
return;
nfs4_ff_layout_stat_io_start_write(hdr->inode,
FF_LAYOUT_COMP(hdr->lseg, hdr->pgio_mirror_idx),
hdr->args.count,
task->tk_start);
}
static void ff_layout_write_record_layoutstats_done(struct rpc_task *task,
struct nfs_pgio_header *hdr)
{
if (!test_and_clear_bit(NFS_IOHDR_STAT, &hdr->flags))
return;
nfs4_ff_layout_stat_io_end_write(task,
FF_LAYOUT_COMP(hdr->lseg, hdr->pgio_mirror_idx),
hdr->args.count, hdr->res.count,
hdr->res.verf->committed);
}
static int ff_layout_write_prepare_common(struct rpc_task *task,
struct nfs_pgio_header *hdr)
{
if (unlikely(test_bit(NFS_CONTEXT_BAD, &hdr->args.context->flags))) {
rpc_exit(task, -EIO);
return -EIO;
......@@ -1445,6 +1518,7 @@ static int ff_layout_write_prepare_common(struct rpc_task *task,
return -EAGAIN;
}
ff_layout_write_record_layoutstats_start(task, hdr);
return 0;
}
......@@ -1480,11 +1554,6 @@ static void ff_layout_write_call_done(struct rpc_task *task, void *data)
{
struct nfs_pgio_header *hdr = data;
nfs4_ff_layout_stat_io_end_write(task,
FF_LAYOUT_COMP(hdr->lseg, hdr->pgio_mirror_idx),
hdr->args.count, hdr->res.count,
hdr->res.verf->committed);
if (test_bit(NFS_IOHDR_REDO, &hdr->flags) &&
task->tk_status == 0) {
nfs4_sequence_done(task, &hdr->res.seq_res);
......@@ -1499,18 +1568,53 @@ static void ff_layout_write_count_stats(struct rpc_task *task, void *data)
{
struct nfs_pgio_header *hdr = data;
ff_layout_write_record_layoutstats_done(task, hdr);
rpc_count_iostats_metrics(task,
&NFS_CLIENT(hdr->inode)->cl_metrics[NFSPROC4_CLNT_WRITE]);
}
static void ff_layout_commit_prepare_common(struct rpc_task *task,
static void ff_layout_write_release(void *data)
{
struct nfs_pgio_header *hdr = data;
ff_layout_write_record_layoutstats_done(&hdr->task, hdr);
pnfs_generic_rw_release(data);
}
static void ff_layout_commit_record_layoutstats_start(struct rpc_task *task,
struct nfs_commit_data *cdata)
{
if (test_and_set_bit(NFS_IOHDR_STAT, &cdata->flags))
return;
nfs4_ff_layout_stat_io_start_write(cdata->inode,
FF_LAYOUT_COMP(cdata->lseg, cdata->ds_commit_index),
0, task->tk_start);
}
static void ff_layout_commit_record_layoutstats_done(struct rpc_task *task,
struct nfs_commit_data *cdata)
{
struct nfs_page *req;
__u64 count = 0;
if (!test_and_clear_bit(NFS_IOHDR_STAT, &cdata->flags))
return;
if (task->tk_status == 0) {
list_for_each_entry(req, &cdata->pages, wb_list)
count += req->wb_bytes;
}
nfs4_ff_layout_stat_io_end_write(task,
FF_LAYOUT_COMP(cdata->lseg, cdata->ds_commit_index),
count, count, NFS_FILE_SYNC);
}
static void ff_layout_commit_prepare_common(struct rpc_task *task,
struct nfs_commit_data *cdata)
{
ff_layout_commit_record_layoutstats_start(task, cdata);
}
static void ff_layout_commit_prepare_v3(struct rpc_task *task, void *data)
{
ff_layout_commit_prepare_common(task, data);
......@@ -1531,19 +1635,6 @@ static void ff_layout_commit_prepare_v4(struct rpc_task *task, void *data)
static void ff_layout_commit_done(struct rpc_task *task, void *data)
{
struct nfs_commit_data *cdata = data;
struct nfs_page *req;
__u64 count = 0;
if (task->tk_status == 0) {
list_for_each_entry(req, &cdata->pages, wb_list)
count += req->wb_bytes;
}
nfs4_ff_layout_stat_io_end_write(task,
FF_LAYOUT_COMP(cdata->lseg, cdata->ds_commit_index),
count, count, NFS_FILE_SYNC);
pnfs_generic_write_commit_done(task, data);
}
......@@ -1551,50 +1642,59 @@ static void ff_layout_commit_count_stats(struct rpc_task *task, void *data)
{
struct nfs_commit_data *cdata = data;
ff_layout_commit_record_layoutstats_done(task, cdata);
rpc_count_iostats_metrics(task,
&NFS_CLIENT(cdata->inode)->cl_metrics[NFSPROC4_CLNT_COMMIT]);
}
static void ff_layout_commit_release(void *data)
{
struct nfs_commit_data *cdata = data;
ff_layout_commit_record_layoutstats_done(&cdata->task, cdata);
pnfs_generic_commit_release(data);
}
static const struct rpc_call_ops ff_layout_read_call_ops_v3 = {
.rpc_call_prepare = ff_layout_read_prepare_v3,
.rpc_call_done = ff_layout_read_call_done,
.rpc_count_stats = ff_layout_read_count_stats,
.rpc_release = pnfs_generic_rw_release,
.rpc_release = ff_layout_read_release,
};
static const struct rpc_call_ops ff_layout_read_call_ops_v4 = {
.rpc_call_prepare = ff_layout_read_prepare_v4,
.rpc_call_done = ff_layout_read_call_done,
.rpc_count_stats = ff_layout_read_count_stats,
.rpc_release = pnfs_generic_rw_release,
.rpc_release = ff_layout_read_release,
};
static const struct rpc_call_ops ff_layout_write_call_ops_v3 = {
.rpc_call_prepare = ff_layout_write_prepare_v3,
.rpc_call_done = ff_layout_write_call_done,
.rpc_count_stats = ff_layout_write_count_stats,
.rpc_release = pnfs_generic_rw_release,
.rpc_release = ff_layout_write_release,
};
static const struct rpc_call_ops ff_layout_write_call_ops_v4 = {
.rpc_call_prepare = ff_layout_write_prepare_v4,
.rpc_call_done = ff_layout_write_call_done,
.rpc_count_stats = ff_layout_write_count_stats,
.rpc_release = pnfs_generic_rw_release,
.rpc_release = ff_layout_write_release,
};
static const struct rpc_call_ops ff_layout_commit_call_ops_v3 = {
.rpc_call_prepare = ff_layout_commit_prepare_v3,
.rpc_call_done = ff_layout_commit_done,
.rpc_count_stats = ff_layout_commit_count_stats,
.rpc_release = pnfs_generic_commit_release,
.rpc_release = ff_layout_commit_release,
};
static const struct rpc_call_ops ff_layout_commit_call_ops_v4 = {
.rpc_call_prepare = ff_layout_commit_prepare_v4,
.rpc_call_done = ff_layout_commit_done,
.rpc_count_stats = ff_layout_commit_count_stats,
.rpc_release = pnfs_generic_commit_release,
.rpc_release = ff_layout_commit_release,
};
static enum pnfs_try_status
......
......@@ -85,6 +85,7 @@ struct nfs4_ff_layout_mirror {
struct nfs4_ff_layoutstat write_stat;
ktime_t start_time;
ktime_t last_report_time;
u32 report_interval;
};
struct nfs4_ff_layout_segment {
......
......@@ -429,22 +429,14 @@ nfs4_ff_layout_prepare_ds(struct pnfs_layout_segment *lseg, u32 ds_idx,
mirror, lseg->pls_range.offset,
lseg->pls_range.length, NFS4ERR_NXIO,
OP_ILLEGAL, GFP_NOIO);
if (fail_return) {
pnfs_error_mark_layout_for_return(ino, lseg);
if (ff_layout_has_available_ds(lseg))
pnfs_set_retry_layoutget(lseg->pls_layout);
else
pnfs_clear_retry_layoutget(lseg->pls_layout);
} else {
if (!fail_return) {
if (ff_layout_has_available_ds(lseg))
set_bit(NFS_LAYOUT_RETURN_BEFORE_CLOSE,
&lseg->pls_layout->plh_flags);
else {
else
pnfs_error_mark_layout_for_return(ino, lseg);
pnfs_clear_retry_layoutget(lseg->pls_layout);
}
}
} else
pnfs_error_mark_layout_for_return(ino, lseg);
}
out_update_creds:
if (ff_layout_update_mirror_cred(mirror, ds))
......
......@@ -912,6 +912,12 @@ void nfs_file_clear_open_context(struct file *filp)
if (ctx) {
struct inode *inode = d_inode(ctx->dentry);
/*
* We fatal error on write before. Try to writeback
* every page again.
*/
if (ctx->error < 0)
invalidate_inode_pages2(inode->i_mapping);
filp->private_data = NULL;
spin_lock(&inode->i_lock);
list_move_tail(&ctx->list, &NFS_I(inode)->open_files);
......
......@@ -711,3 +711,17 @@ static inline u32 nfs_stateid_hash(nfs4_stateid *stateid)
return 0;
}
#endif
static inline bool nfs_error_is_fatal(int err)
{
switch (err) {
case -ERESTARTSYS:
case -EIO:
case -ENOSPC:
case -EROFS:
case -E2BIG:
return true;
default:
return false;
}
}
......@@ -7811,6 +7811,15 @@ static void nfs4_layoutget_done(struct rpc_task *task, void *calldata)
switch (task->tk_status) {
case 0:
goto out;
/*
* NFS4ERR_LAYOUTUNAVAILABLE means we are not supposed to use pnfs
* on the file. set tk_status to -ENODATA to tell upper layer to
* retry go inband.
*/
case -NFS4ERR_LAYOUTUNAVAILABLE:
task->tk_status = -ENODATA;
goto out;
/*
* NFS4ERR_BADLAYOUT means the MDS cannot return a layout of
* length lgp->args.minlength != 0 (see RFC5661 section 18.43.3).
......
......@@ -664,22 +664,11 @@ EXPORT_SYMBOL_GPL(nfs_initiate_pgio);
* @desc: IO descriptor
* @hdr: pageio header
*/
static int nfs_pgio_error(struct nfs_pageio_descriptor *desc,
struct nfs_pgio_header *hdr)
static void nfs_pgio_error(struct nfs_pgio_header *hdr)
{
struct nfs_pgio_mirror *mirror;
u32 midx;
set_bit(NFS_IOHDR_REDO, &hdr->flags);
nfs_pgio_data_destroy(hdr);
hdr->completion_ops->completion(hdr);
/* TODO: Make sure it's right to clean up all mirrors here
* and not just hdr->pgio_mirror_idx */
for (midx = 0; midx < desc->pg_mirror_count; midx++) {
mirror = &desc->pg_mirrors[midx];
desc->pg_completion_ops->error_cleanup(&mirror->pg_list);
}
return -ENOMEM;
}
/**
......@@ -800,8 +789,11 @@ int nfs_generic_pgio(struct nfs_pageio_descriptor *desc,
unsigned int pagecount, pageused;
pagecount = nfs_page_array_len(mirror->pg_base, mirror->pg_count);
if (!nfs_pgarray_set(&hdr->page_array, pagecount))
return nfs_pgio_error(desc, hdr);
if (!nfs_pgarray_set(&hdr->page_array, pagecount)) {
nfs_pgio_error(hdr);
desc->pg_error = -ENOMEM;
return desc->pg_error;
}
nfs_init_cinfo(&cinfo, desc->pg_inode, desc->pg_dreq);
pages = hdr->page_array.pagevec;
......@@ -819,8 +811,11 @@ int nfs_generic_pgio(struct nfs_pageio_descriptor *desc,
*pages++ = last_page = req->wb_page;
}
}
if (WARN_ON_ONCE(pageused != pagecount))
return nfs_pgio_error(desc, hdr);
if (WARN_ON_ONCE(pageused != pagecount)) {
nfs_pgio_error(hdr);
desc->pg_error = -EINVAL;
return desc->pg_error;
}
if ((desc->pg_ioflags & FLUSH_COND_STABLE) &&
(desc->pg_moreio || nfs_reqs_to_commit(&cinfo)))
......@@ -843,10 +838,8 @@ static int nfs_generic_pg_pgios(struct nfs_pageio_descriptor *desc)
hdr = nfs_pgio_header_alloc(desc->pg_rw_ops);
if (!hdr) {
/* TODO: make sure this is right with mirroring - or
* should it back out all mirrors? */
desc->pg_completion_ops->error_cleanup(&mirror->pg_list);
return -ENOMEM;
desc->pg_error = -ENOMEM;
return desc->pg_error;
}
nfs_pgheader_init(desc, hdr, nfs_pgio_header_free);
ret = nfs_generic_pgio(desc, hdr);
......@@ -874,6 +867,9 @@ static int nfs_pageio_setup_mirroring(struct nfs_pageio_descriptor *pgio,
mirror_count = pgio->pg_ops->pg_get_mirror_count(pgio, req);
if (pgio->pg_error < 0)
return pgio->pg_error;
if (!mirror_count || mirror_count > NFS_PAGEIO_DESCRIPTOR_MIRROR_MAX)
return -EINVAL;
......@@ -982,6 +978,8 @@ static int nfs_pageio_do_add_request(struct nfs_pageio_descriptor *desc,
} else {
if (desc->pg_ops->pg_init)
desc->pg_ops->pg_init(desc, req);
if (desc->pg_error < 0)
return 0;
mirror->pg_base = req->wb_pgbase;
}
if (!nfs_can_coalesce_requests(prev, req, desc))
......@@ -1147,6 +1145,8 @@ int nfs_pageio_add_request(struct nfs_pageio_descriptor *desc,
bytes = req->wb_bytes;
nfs_pageio_setup_mirroring(desc, req);
if (desc->pg_error < 0)
goto out_failed;
for (midx = 0; midx < desc->pg_mirror_count; midx++) {
if (midx) {
......@@ -1163,7 +1163,8 @@ int nfs_pageio_add_request(struct nfs_pageio_descriptor *desc,
if (IS_ERR(dupreq)) {
nfs_page_group_unlock(req);
return 0;
desc->pg_error = PTR_ERR(dupreq);
goto out_failed;
}
nfs_lock_request(dupreq);
......@@ -1176,10 +1177,32 @@ int nfs_pageio_add_request(struct nfs_pageio_descriptor *desc,
if (nfs_pgio_has_mirroring(desc))
desc->pg_mirror_idx = midx;
if (!nfs_pageio_add_request_mirror(desc, dupreq))
return 0;
goto out_failed;
}
return 1;
out_failed:
/*
* We might have failed before sending any reqs over wire.
* Clean up rest of the reqs in mirror pg_list.
*/
if (desc->pg_error) {
struct nfs_pgio_mirror *mirror;
void (*func)(struct list_head *);
/* remember fatal errors */
if (nfs_error_is_fatal(desc->pg_error))
mapping_set_error(desc->pg_inode->i_mapping,
desc->pg_error);
func = desc->pg_completion_ops->error_cleanup;
for (midx = 0; midx < desc->pg_mirror_count; midx++) {
mirror = &desc->pg_mirrors[midx];
func(&mirror->pg_list);
}
}
return 0;
}
/*
......@@ -1232,7 +1255,7 @@ int nfs_pageio_resend(struct nfs_pageio_descriptor *desc,
nfs_pageio_complete(desc);
if (!list_empty(&failed)) {
list_move(&failed, &hdr->pages);
return -EIO;
return desc->pg_error < 0 ? desc->pg_error : -EIO;
}
return 0;
}
......
......@@ -618,7 +618,6 @@ pnfs_destroy_layout(struct nfs_inode *nfsi)
pnfs_get_layout_hdr(lo);
pnfs_layout_clear_fail_bit(lo, NFS_LAYOUT_RO_FAILED);
pnfs_layout_clear_fail_bit(lo, NFS_LAYOUT_RW_FAILED);
pnfs_clear_retry_layoutget(lo);
spin_unlock(&nfsi->vfs_inode.i_lock);
pnfs_free_lseg_list(&tmp_list);
pnfs_put_layout_hdr(lo);
......@@ -904,17 +903,9 @@ send_layoutget(struct pnfs_layout_hdr *lo,
lseg = nfs4_proc_layoutget(lgp, gfp_flags);
} while (lseg == ERR_PTR(-EAGAIN));
if (IS_ERR(lseg)) {
switch (PTR_ERR(lseg)) {
case -ENOMEM:
case -ERESTARTSYS:
break;
default:
/* remember that LAYOUTGET failed and suspend trying */
pnfs_layout_io_set_failed(lo, range->iomode);
}
return NULL;
} else
if (IS_ERR(lseg) && !nfs_error_is_fatal(PTR_ERR(lseg)))
lseg = NULL;
else
pnfs_layout_clear_fail_bit(lo,
pnfs_iomode_to_fail_bit(range->iomode));
......@@ -1102,7 +1093,6 @@ bool pnfs_roc(struct inode *ino)
&lo->plh_flags))
layoutreturn = pnfs_prepare_layoutreturn(lo);
pnfs_clear_retry_layoutget(lo);
list_for_each_entry_safe(lseg, tmp, &lo->plh_segs, pls_list)
/* If we are sending layoutreturn, invalidate all valid lsegs */
if (layoutreturn || test_bit(NFS_LSEG_ROC, &lseg->pls_flags)) {
......@@ -1465,25 +1455,15 @@ static bool pnfs_within_mdsthreshold(struct nfs_open_context *ctx,
return ret;
}
/* stop waiting if someone clears NFS_LAYOUT_RETRY_LAYOUTGET bit. */
static int pnfs_layoutget_retry_bit_wait(struct wait_bit_key *key, int mode)
{
if (!test_bit(NFS_LAYOUT_RETRY_LAYOUTGET, key->flags))
return 1;
return nfs_wait_bit_killable(key, mode);
}
static bool pnfs_prepare_to_retry_layoutget(struct pnfs_layout_hdr *lo)
{
if (!pnfs_should_retry_layoutget(lo))
return false;
/*
* send layoutcommit as it can hold up layoutreturn due to lseg
* reference
*/
pnfs_layoutcommit_inode(lo->plh_inode, false);
return !wait_on_bit_action(&lo->plh_flags, NFS_LAYOUT_RETURN,
pnfs_layoutget_retry_bit_wait,
nfs_wait_bit_killable,
TASK_UNINTERRUPTIBLE);
}
......@@ -1558,8 +1538,7 @@ pnfs_update_layout(struct inode *ino,
}
/* if LAYOUTGET already failed once we don't try again */
if (pnfs_layout_io_test_failed(lo, iomode) &&
!pnfs_should_retry_layoutget(lo)) {
if (pnfs_layout_io_test_failed(lo, iomode)) {
trace_pnfs_update_layout(ino, pos, count, iomode, lo,
PNFS_UPDATE_LAYOUT_IO_TEST_FAIL);
goto out_unlock;
......@@ -1636,7 +1615,6 @@ pnfs_update_layout(struct inode *ino,
arg.length = PAGE_CACHE_ALIGN(arg.length);
lseg = send_layoutget(lo, ctx, &arg, gfp_flags);
pnfs_clear_retry_layoutget(lo);
atomic_dec(&lo->plh_outstanding);
trace_pnfs_update_layout(ino, pos, count, iomode, lo,
PNFS_UPDATE_LAYOUT_SEND_LAYOUTGET);
......@@ -1649,7 +1627,7 @@ pnfs_update_layout(struct inode *ino,
"(%s, offset: %llu, length: %llu)\n",
__func__, ino->i_sb->s_id,
(unsigned long long)NFS_FILEID(ino),
lseg == NULL ? "not found" : "found",
IS_ERR_OR_NULL(lseg) ? "not found" : "found",
iomode==IOMODE_RW ? "read/write" : "read-only",
(unsigned long long)pos,
(unsigned long long)count);
......@@ -1785,7 +1763,6 @@ void pnfs_error_mark_layout_for_return(struct inode *inode,
struct pnfs_layout_segment *lseg)
{
struct pnfs_layout_hdr *lo = NFS_I(inode)->layout;
int iomode = pnfs_iomode_to_fail_bit(lseg->pls_range.iomode);
struct pnfs_layout_range range = {
.iomode = lseg->pls_range.iomode,
.offset = 0,
......@@ -1794,8 +1771,6 @@ void pnfs_error_mark_layout_for_return(struct inode *inode,
LIST_HEAD(free_me);
spin_lock(&inode->i_lock);
/* set failure bit so that pnfs path will be retried later */
pnfs_layout_set_fail_bit(lo, iomode);
if (lo->plh_return_iomode == 0)
lo->plh_return_iomode = range.iomode;
else if (lo->plh_return_iomode != range.iomode)
......@@ -1828,6 +1803,11 @@ pnfs_generic_pg_init_read(struct nfs_pageio_descriptor *pgio, struct nfs_page *r
rd_size,
IOMODE_READ,
GFP_KERNEL);
if (IS_ERR(pgio->pg_lseg)) {
pgio->pg_error = PTR_ERR(pgio->pg_lseg);
pgio->pg_lseg = NULL;
return;
}
}
/* If no lseg, fall back to read through mds */
if (pgio->pg_lseg == NULL)
......@@ -1840,13 +1820,19 @@ void
pnfs_generic_pg_init_write(struct nfs_pageio_descriptor *pgio,
struct nfs_page *req, u64 wb_size)
{
if (pgio->pg_lseg == NULL)
if (pgio->pg_lseg == NULL) {
pgio->pg_lseg = pnfs_update_layout(pgio->pg_inode,
req->wb_context,
req_offset(req),
wb_size,
IOMODE_RW,
GFP_NOFS);
if (IS_ERR(pgio->pg_lseg)) {
pgio->pg_error = PTR_ERR(pgio->pg_lseg);
pgio->pg_lseg = NULL;
return;
}
}
/* If no lseg, fall back to write through mds */
if (pgio->pg_lseg == NULL)
nfs_pageio_reset_write_mds(pgio);
......@@ -2014,15 +2000,13 @@ static void pnfs_writehdr_free(struct nfs_pgio_header *hdr)
int
pnfs_generic_pg_writepages(struct nfs_pageio_descriptor *desc)
{
struct nfs_pgio_mirror *mirror = nfs_pgio_current_mirror(desc);
struct nfs_pgio_header *hdr;
int ret;
hdr = nfs_pgio_header_alloc(desc->pg_rw_ops);
if (!hdr) {
desc->pg_completion_ops->error_cleanup(&mirror->pg_list);
return -ENOMEM;
desc->pg_error = -ENOMEM;
return desc->pg_error;
}
nfs_pgheader_init(desc, hdr, pnfs_writehdr_free);
......@@ -2145,15 +2129,13 @@ static void pnfs_readhdr_free(struct nfs_pgio_header *hdr)
int
pnfs_generic_pg_readpages(struct nfs_pageio_descriptor *desc)
{
struct nfs_pgio_mirror *mirror = nfs_pgio_current_mirror(desc);
struct nfs_pgio_header *hdr;
int ret;
hdr = nfs_pgio_header_alloc(desc->pg_rw_ops);
if (!hdr) {
desc->pg_completion_ops->error_cleanup(&mirror->pg_list);
return -ENOMEM;
desc->pg_error = -ENOMEM;
return desc->pg_error;
}
nfs_pgheader_init(desc, hdr, pnfs_readhdr_free);
hdr->lseg = pnfs_get_lseg(desc->pg_lseg);
......
......@@ -98,7 +98,6 @@ enum {
NFS_LAYOUT_RETURN_BEFORE_CLOSE, /* Return this layout before close */
NFS_LAYOUT_INVALID_STID, /* layout stateid id is invalid */
NFS_LAYOUT_FIRST_LAYOUTGET, /* Serialize first layoutget */
NFS_LAYOUT_RETRY_LAYOUTGET, /* Retry layoutget */
};
enum layoutdriver_policy_flags {
......@@ -379,26 +378,6 @@ nfs4_get_deviceid(struct nfs4_deviceid_node *d)
return d;
}
static inline void pnfs_set_retry_layoutget(struct pnfs_layout_hdr *lo)
{
if (!test_and_set_bit(NFS_LAYOUT_RETRY_LAYOUTGET, &lo->plh_flags))
atomic_inc(&lo->plh_refcount);
}
static inline void pnfs_clear_retry_layoutget(struct pnfs_layout_hdr *lo)
{
if (test_and_clear_bit(NFS_LAYOUT_RETRY_LAYOUTGET, &lo->plh_flags)) {
atomic_dec(&lo->plh_refcount);
/* wake up waiters for LAYOUTRETURN as that is not needed */
wake_up_bit(&lo->plh_flags, NFS_LAYOUT_RETURN);
}
}
static inline bool pnfs_should_retry_layoutget(struct pnfs_layout_hdr *lo)
{
return test_bit(NFS_LAYOUT_RETRY_LAYOUTGET, &lo->plh_flags);
}
static inline struct pnfs_layout_segment *
pnfs_get_lseg(struct pnfs_layout_segment *lseg)
{
......
......@@ -85,6 +85,23 @@ void nfs_pageio_reset_read_mds(struct nfs_pageio_descriptor *pgio)
}
EXPORT_SYMBOL_GPL(nfs_pageio_reset_read_mds);
static void nfs_readpage_release(struct nfs_page *req)
{
struct inode *inode = d_inode(req->wb_context->dentry);
dprintk("NFS: read done (%s/%llu %d@%lld)\n", inode->i_sb->s_id,
(unsigned long long)NFS_FILEID(inode), req->wb_bytes,
(long long)req_offset(req));
if (nfs_page_group_sync_on_bit(req, PG_UNLOCKPAGE)) {
if (PageUptodate(req->wb_page))
nfs_readpage_to_fscache(inode, req->wb_page, 0);
unlock_page(req->wb_page);
}
nfs_release_request(req);
}
int nfs_readpage_async(struct nfs_open_context *ctx, struct inode *inode,
struct page *page)
{
......@@ -106,7 +123,10 @@ int nfs_readpage_async(struct nfs_open_context *ctx, struct inode *inode,
nfs_pageio_init_read(&pgio, inode, false,
&nfs_async_read_completion_ops);
nfs_pageio_add_request(&pgio, new);
if (!nfs_pageio_add_request(&pgio, new)) {
nfs_list_remove_request(new);
nfs_readpage_release(new);
}
nfs_pageio_complete(&pgio);
/* It doesn't make sense to do mirrored reads! */
......@@ -115,24 +135,7 @@ int nfs_readpage_async(struct nfs_open_context *ctx, struct inode *inode,
pgm = &pgio.pg_mirrors[0];
NFS_I(inode)->read_io += pgm->pg_bytes_written;
return 0;
}
static void nfs_readpage_release(struct nfs_page *req)
{
struct inode *inode = d_inode(req->wb_context->dentry);
dprintk("NFS: read done (%s/%llu %d@%lld)\n", inode->i_sb->s_id,
(unsigned long long)NFS_FILEID(inode), req->wb_bytes,
(long long)req_offset(req));
if (nfs_page_group_sync_on_bit(req, PG_UNLOCKPAGE)) {
if (PageUptodate(req->wb_page))
nfs_readpage_to_fscache(inode, req->wb_page, 0);
unlock_page(req->wb_page);
}
nfs_release_request(req);
return pgio.pg_error < 0 ? pgio.pg_error : 0;
}
static void nfs_page_group_set_uptodate(struct nfs_page *req)
......@@ -361,6 +364,8 @@ readpage_async_filler(void *data, struct page *page)
if (len < PAGE_CACHE_SIZE)
zero_user_segment(page, len, PAGE_CACHE_SIZE);
if (!nfs_pageio_add_request(desc->pgio, new)) {
nfs_list_remove_request(new);
nfs_readpage_release(new);
error = desc->pgio->pg_error;
goto out_unlock;
}
......
......@@ -545,12 +545,22 @@ nfs_lock_and_join_requests(struct page *page, bool nonblock)
return head;
}
static void nfs_write_error_remove_page(struct nfs_page *req)
{
nfs_unlock_request(req);
nfs_end_page_writeback(req);
nfs_release_request(req);
generic_error_remove_page(page_file_mapping(req->wb_page),
req->wb_page);
}
/*
* Find an associated nfs write request, and prepare to flush it out
* May return an error if the user signalled nfs_wait_on_request().
*/
static int nfs_page_async_flush(struct nfs_pageio_descriptor *pgio,
struct page *page, bool nonblock)
struct page *page, bool nonblock,
bool launder)
{
struct nfs_page *req;
int ret = 0;
......@@ -567,8 +577,21 @@ static int nfs_page_async_flush(struct nfs_pageio_descriptor *pgio,
ret = 0;
if (!nfs_pageio_add_request(pgio, req)) {
nfs_redirty_request(req);
ret = pgio->pg_error;
/*
* Remove the problematic req upon fatal errors
* in launder case, while other dirty pages can
* still be around until they get flushed.
*/
if (nfs_error_is_fatal(ret)) {
nfs_context_set_write_error(req->wb_context, ret);
if (launder) {
nfs_write_error_remove_page(req);
goto out;
}
}
nfs_redirty_request(req);
ret = -EAGAIN;
} else
nfs_add_stats(page_file_mapping(page)->host,
NFSIOS_WRITEPAGES, 1);
......@@ -576,12 +599,14 @@ static int nfs_page_async_flush(struct nfs_pageio_descriptor *pgio,
return ret;
}
static int nfs_do_writepage(struct page *page, struct writeback_control *wbc, struct nfs_pageio_descriptor *pgio)
static int nfs_do_writepage(struct page *page, struct writeback_control *wbc,
struct nfs_pageio_descriptor *pgio, bool launder)
{
int ret;
nfs_pageio_cond_complete(pgio, page_file_index(page));
ret = nfs_page_async_flush(pgio, page, wbc->sync_mode == WB_SYNC_NONE);
ret = nfs_page_async_flush(pgio, page, wbc->sync_mode == WB_SYNC_NONE,
launder);
if (ret == -EAGAIN) {
redirty_page_for_writepage(wbc, page);
ret = 0;
......@@ -592,7 +617,9 @@ static int nfs_do_writepage(struct page *page, struct writeback_control *wbc, st
/*
* Write an mmapped page to the server.
*/
static int nfs_writepage_locked(struct page *page, struct writeback_control *wbc)
static int nfs_writepage_locked(struct page *page,
struct writeback_control *wbc,
bool launder)
{
struct nfs_pageio_descriptor pgio;
struct inode *inode = page_file_mapping(page)->host;
......@@ -601,7 +628,7 @@ static int nfs_writepage_locked(struct page *page, struct writeback_control *wbc
nfs_inc_stats(inode, NFSIOS_VFSWRITEPAGE);
nfs_pageio_init_write(&pgio, inode, wb_priority(wbc),
false, &nfs_async_write_completion_ops);
err = nfs_do_writepage(page, wbc, &pgio);
err = nfs_do_writepage(page, wbc, &pgio, launder);
nfs_pageio_complete(&pgio);
if (err < 0)
return err;
......@@ -614,7 +641,7 @@ int nfs_writepage(struct page *page, struct writeback_control *wbc)
{
int ret;
ret = nfs_writepage_locked(page, wbc);
ret = nfs_writepage_locked(page, wbc, false);
unlock_page(page);
return ret;
}
......@@ -623,7 +650,7 @@ static int nfs_writepages_callback(struct page *page, struct writeback_control *
{
int ret;
ret = nfs_do_writepage(page, wbc, data);
ret = nfs_do_writepage(page, wbc, data, false);
unlock_page(page);
return ret;
}
......@@ -1911,7 +1938,7 @@ int nfs_wb_page_cancel(struct inode *inode, struct page *page)
/*
* Write back all requests on one page - we do this before reading it.
*/
int nfs_wb_page(struct inode *inode, struct page *page)
int nfs_wb_single_page(struct inode *inode, struct page *page, bool launder)
{
loff_t range_start = page_file_offset(page);
loff_t range_end = range_start + (loff_t)(PAGE_CACHE_SIZE - 1);
......@@ -1928,7 +1955,7 @@ int nfs_wb_page(struct inode *inode, struct page *page)
for (;;) {
wait_on_page_writeback(page);
if (clear_page_dirty_for_io(page)) {
ret = nfs_writepage_locked(page, &wbc);
ret = nfs_writepage_locked(page, &wbc, launder);
if (ret < 0)
goto out_error;
continue;
......
......@@ -517,12 +517,24 @@ extern int nfs_updatepage(struct file *, struct page *, unsigned int, unsigned
*/
extern int nfs_sync_inode(struct inode *inode);
extern int nfs_wb_all(struct inode *inode);
extern int nfs_wb_page(struct inode *inode, struct page* page);
extern int nfs_wb_single_page(struct inode *inode, struct page *page, bool launder);
extern int nfs_wb_page_cancel(struct inode *inode, struct page* page);
extern int nfs_commit_inode(struct inode *, int);
extern struct nfs_commit_data *nfs_commitdata_alloc(void);
extern void nfs_commit_free(struct nfs_commit_data *data);
static inline int
nfs_wb_launder_page(struct inode *inode, struct page *page)
{
return nfs_wb_single_page(inode, page, true);
}
static inline int
nfs_wb_page(struct inode *inode, struct page *page)
{
return nfs_wb_single_page(inode, page, false);
}
static inline int
nfs_have_writebacks(struct inode *inode)
{
......
......@@ -1375,6 +1375,7 @@ enum {
NFS_IOHDR_ERROR = 0,
NFS_IOHDR_EOF,
NFS_IOHDR_REDO,
NFS_IOHDR_STAT,
};
struct nfs_pgio_header {
......@@ -1454,6 +1455,7 @@ struct nfs_commit_data {
const struct rpc_call_ops *mds_ops;
const struct nfs_commit_completion_ops *completion_ops;
int (*commit_done_cb) (struct rpc_task *task, struct nfs_commit_data *data);
unsigned long flags;
};
struct nfs_pgio_completion_ops {
......
Markdown is supported
0%
or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment