* [PATCH 1/7] wave4: nfs: rearrange nfs_write_rpcsetup
2011-02-11 16:54 [PATCH 0/7] pnfs wave 4 proposal Fred Isaman
@ 2011-02-11 16:54 ` Fred Isaman
2011-02-14 17:22 ` Andy Adamson
2011-02-11 16:54 ` [PATCH 2/7] wave4: Send lseg down through nfs_write_rpcsetup into nfs_initiate_write Fred Isaman
` (6 subsequent siblings)
7 siblings, 1 reply; 15+ messages in thread
From: Fred Isaman @ 2011-02-11 16:54 UTC (permalink / raw)
To: linux-nfs
From: The pNFS Team <linux-nfs@vger.kernel.org>
Reorder nfs_write_rpcsetup, preparing for a pnfs switch.
FIXME: these signoffs are ridiculous for this simple patch.
Lets just pick an author and be done with it.
Reported-by: Alexandros Batsakis <batsakis@netapp.com>
Signed-off-by: Andy Adamson <andros@citi.umich.edu>
Signed-off-by: Boaz Harrosh <bharrosh@panasas.com>
Signed-off-by: Dean Hildebrand <dhildeb@us.ibm.com>
Signed-off-by: Fred Isaman <iisaman@citi.umich.edu>
Signed-off-by: J. Bruce Fields <bfields@fieldses.org>
Signed-off-by: Mike Sager <sager@netapp.com>
Signed-off-by: Ricardo Labiaga <Ricardo.Labiaga@netapp.com>
Signed-off-by: Tao Guo <guotao@nrchpc.ac.cn>
Signed-off-by: Andy Adamson <andros@netapp.com>
Signed-off-by: Benny Halevy <bhalevy@panasas.com>
Signed-off-by: Fred Isaman <iisaman@netapp.com>
---
fs/nfs/write.c | 82 +++++++++++++++++++++++++++++++------------------------
1 files changed, 46 insertions(+), 36 deletions(-)
diff --git a/fs/nfs/write.c b/fs/nfs/write.c
index aca0268..5604854 100644
--- a/fs/nfs/write.c
+++ b/fs/nfs/write.c
@@ -782,25 +782,21 @@ static int flush_task_priority(int how)
return RPC_PRIORITY_NORMAL;
}
-/*
- * Set up the argument/result storage required for the RPC call.
- */
-static int nfs_write_rpcsetup(struct nfs_page *req,
- struct nfs_write_data *data,
- const struct rpc_call_ops *call_ops,
- unsigned int count, unsigned int offset,
- int how)
+static int nfs_initiate_write(struct nfs_write_data *data,
+ struct rpc_clnt *clnt,
+ const struct rpc_call_ops *call_ops,
+ int how)
{
- struct inode *inode = req->wb_context->path.dentry->d_inode;
+ struct inode *inode = data->inode;
int priority = flush_task_priority(how);
struct rpc_task *task;
struct rpc_message msg = {
.rpc_argp = &data->args,
.rpc_resp = &data->res,
- .rpc_cred = req->wb_context->cred,
+ .rpc_cred = data->cred,
};
struct rpc_task_setup task_setup_data = {
- .rpc_client = NFS_CLIENT(inode),
+ .rpc_client = clnt,
.task = &data->task,
.rpc_message = &msg,
.callback_ops = call_ops,
@@ -811,12 +807,49 @@ static int nfs_write_rpcsetup(struct nfs_page *req,
};
int ret = 0;
+ /* Set up the initial task struct. */
+ NFS_PROTO(inode)->write_setup(data, &msg);
+
+ dprintk("NFS: %5u initiated write call "
+ "(req %s/%lld, %u bytes @ offset %llu)\n",
+ data->task.tk_pid,
+ inode->i_sb->s_id,
+ (long long)NFS_FILEID(inode),
+ data->args.count,
+ (unsigned long long)data->args.offset);
+
+ task = rpc_run_task(&task_setup_data);
+ if (IS_ERR(task)) {
+ ret = PTR_ERR(task);
+ goto out;
+ }
+ if (how & FLUSH_SYNC) {
+ ret = rpc_wait_for_completion_task(task);
+ if (ret == 0)
+ ret = task->tk_status;
+ }
+ rpc_put_task(task);
+out:
+ return ret;
+}
+
+/*
+ * Set up the argument/result storage required for the RPC call.
+ */
+static int nfs_write_rpcsetup(struct nfs_page *req,
+ struct nfs_write_data *data,
+ const struct rpc_call_ops *call_ops,
+ unsigned int count, unsigned int offset,
+ int how)
+{
+ struct inode *inode = req->wb_context->path.dentry->d_inode;
+
/* Set up the RPC argument and reply structs
* NB: take care not to mess about with data->commit et al. */
data->req = req;
data->inode = inode = req->wb_context->path.dentry->d_inode;
- data->cred = msg.rpc_cred;
+ data->cred = req->wb_context->cred;
data->args.fh = NFS_FH(inode);
data->args.offset = req_offset(req) + offset;
@@ -837,30 +870,7 @@ static int nfs_write_rpcsetup(struct nfs_page *req,
data->res.verf = &data->verf;
nfs_fattr_init(&data->fattr);
- /* Set up the initial task struct. */
- NFS_PROTO(inode)->write_setup(data, &msg);
-
- dprintk("NFS: %5u initiated write call "
- "(req %s/%lld, %u bytes @ offset %llu)\n",
- data->task.tk_pid,
- inode->i_sb->s_id,
- (long long)NFS_FILEID(inode),
- count,
- (unsigned long long)data->args.offset);
-
- task = rpc_run_task(&task_setup_data);
- if (IS_ERR(task)) {
- ret = PTR_ERR(task);
- goto out;
- }
- if (how & FLUSH_SYNC) {
- ret = rpc_wait_for_completion_task(task);
- if (ret == 0)
- ret = task->tk_status;
- }
- rpc_put_task(task);
-out:
- return ret;
+ return nfs_initiate_write(data, NFS_CLIENT(inode), call_ops, how);
}
/* If a nfs_flush_* function fails, it should remove reqs from @head and
--
1.7.2.1
^ permalink raw reply related [flat|nested] 15+ messages in thread
* Re: [PATCH 1/7] wave4: nfs: rearrange nfs_write_rpcsetup
2011-02-11 16:54 ` [PATCH 1/7] wave4: nfs: rearrange nfs_write_rpcsetup Fred Isaman
@ 2011-02-14 17:22 ` Andy Adamson
2011-02-14 17:27 ` Fred Isaman
0 siblings, 1 reply; 15+ messages in thread
From: Andy Adamson @ 2011-02-14 17:22 UTC (permalink / raw)
To: Fred Isaman; +Cc: linux-nfs
Since I'm the author for the read version of this code, I'll raise my hand to be the author for this version.
-->Andy
On Feb 11, 2011, at 11:54 AM, Fred Isaman wrote:
> From: The pNFS Team <linux-nfs@vger.kernel.org>
>
> Reorder nfs_write_rpcsetup, preparing for a pnfs switch.
>
> FIXME: these signoffs are ridiculous for this simple patch.
> Lets just pick an author and be done with it.
>
> Reported-by: Alexandros Batsakis <batsakis@netapp.com>
> Signed-off-by: Andy Adamson <andros@citi.umich.edu>
> Signed-off-by: Boaz Harrosh <bharrosh@panasas.com>
> Signed-off-by: Dean Hildebrand <dhildeb@us.ibm.com>
> Signed-off-by: Fred Isaman <iisaman@citi.umich.edu>
> Signed-off-by: J. Bruce Fields <bfields@fieldses.org>
> Signed-off-by: Mike Sager <sager@netapp.com>
> Signed-off-by: Ricardo Labiaga <Ricardo.Labiaga@netapp.com>
> Signed-off-by: Tao Guo <guotao@nrchpc.ac.cn>
> Signed-off-by: Andy Adamson <andros@netapp.com>
> Signed-off-by: Benny Halevy <bhalevy@panasas.com>
> Signed-off-by: Fred Isaman <iisaman@netapp.com>
> ---
> fs/nfs/write.c | 82 +++++++++++++++++++++++++++++++------------------------
> 1 files changed, 46 insertions(+), 36 deletions(-)
>
> diff --git a/fs/nfs/write.c b/fs/nfs/write.c
> index aca0268..5604854 100644
> --- a/fs/nfs/write.c
> +++ b/fs/nfs/write.c
> @@ -782,25 +782,21 @@ static int flush_task_priority(int how)
> return RPC_PRIORITY_NORMAL;
> }
>
> -/*
> - * Set up the argument/result storage required for the RPC call.
> - */
> -static int nfs_write_rpcsetup(struct nfs_page *req,
> - struct nfs_write_data *data,
> - const struct rpc_call_ops *call_ops,
> - unsigned int count, unsigned int offset,
> - int how)
> +static int nfs_initiate_write(struct nfs_write_data *data,
> + struct rpc_clnt *clnt,
> + const struct rpc_call_ops *call_ops,
> + int how)
> {
> - struct inode *inode = req->wb_context->path.dentry->d_inode;
> + struct inode *inode = data->inode;
> int priority = flush_task_priority(how);
> struct rpc_task *task;
> struct rpc_message msg = {
> .rpc_argp = &data->args,
> .rpc_resp = &data->res,
> - .rpc_cred = req->wb_context->cred,
> + .rpc_cred = data->cred,
> };
> struct rpc_task_setup task_setup_data = {
> - .rpc_client = NFS_CLIENT(inode),
> + .rpc_client = clnt,
> .task = &data->task,
> .rpc_message = &msg,
> .callback_ops = call_ops,
> @@ -811,12 +807,49 @@ static int nfs_write_rpcsetup(struct nfs_page *req,
> };
> int ret = 0;
>
> + /* Set up the initial task struct. */
> + NFS_PROTO(inode)->write_setup(data, &msg);
> +
> + dprintk("NFS: %5u initiated write call "
> + "(req %s/%lld, %u bytes @ offset %llu)\n",
> + data->task.tk_pid,
> + inode->i_sb->s_id,
> + (long long)NFS_FILEID(inode),
> + data->args.count,
> + (unsigned long long)data->args.offset);
> +
> + task = rpc_run_task(&task_setup_data);
> + if (IS_ERR(task)) {
> + ret = PTR_ERR(task);
> + goto out;
> + }
> + if (how & FLUSH_SYNC) {
> + ret = rpc_wait_for_completion_task(task);
> + if (ret == 0)
> + ret = task->tk_status;
> + }
> + rpc_put_task(task);
> +out:
> + return ret;
> +}
> +
> +/*
> + * Set up the argument/result storage required for the RPC call.
> + */
> +static int nfs_write_rpcsetup(struct nfs_page *req,
> + struct nfs_write_data *data,
> + const struct rpc_call_ops *call_ops,
> + unsigned int count, unsigned int offset,
> + int how)
> +{
> + struct inode *inode = req->wb_context->path.dentry->d_inode;
> +
> /* Set up the RPC argument and reply structs
> * NB: take care not to mess about with data->commit et al. */
>
> data->req = req;
> data->inode = inode = req->wb_context->path.dentry->d_inode;
> - data->cred = msg.rpc_cred;
> + data->cred = req->wb_context->cred;
>
> data->args.fh = NFS_FH(inode);
> data->args.offset = req_offset(req) + offset;
> @@ -837,30 +870,7 @@ static int nfs_write_rpcsetup(struct nfs_page *req,
> data->res.verf = &data->verf;
> nfs_fattr_init(&data->fattr);
>
> - /* Set up the initial task struct. */
> - NFS_PROTO(inode)->write_setup(data, &msg);
> -
> - dprintk("NFS: %5u initiated write call "
> - "(req %s/%lld, %u bytes @ offset %llu)\n",
> - data->task.tk_pid,
> - inode->i_sb->s_id,
> - (long long)NFS_FILEID(inode),
> - count,
> - (unsigned long long)data->args.offset);
> -
> - task = rpc_run_task(&task_setup_data);
> - if (IS_ERR(task)) {
> - ret = PTR_ERR(task);
> - goto out;
> - }
> - if (how & FLUSH_SYNC) {
> - ret = rpc_wait_for_completion_task(task);
> - if (ret == 0)
> - ret = task->tk_status;
> - }
> - rpc_put_task(task);
> -out:
> - return ret;
> + return nfs_initiate_write(data, NFS_CLIENT(inode), call_ops, how);
> }
>
> /* If a nfs_flush_* function fails, it should remove reqs from @head and
> --
> 1.7.2.1
>
> --
> To unsubscribe from this list: send the line "unsubscribe linux-nfs" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at http://vger.kernel.org/majordomo-info.html
^ permalink raw reply [flat|nested] 15+ messages in thread
* Re: [PATCH 1/7] wave4: nfs: rearrange nfs_write_rpcsetup
2011-02-14 17:22 ` Andy Adamson
@ 2011-02-14 17:27 ` Fred Isaman
2011-02-14 18:11 ` William A. (Andy) Adamson
0 siblings, 1 reply; 15+ messages in thread
From: Fred Isaman @ 2011-02-14 17:27 UTC (permalink / raw)
To: Andy Adamson; +Cc: linux-nfs
On Feb 14, 2011, at 12:22 PM, Andy Adamson wrote:
> Since I'm the author for the read version of this code, I'll raise my hand to be the author for this version.
>
> -->Andy
>
OK, do you wan't patch 5 as well?...they were both split from the same "pNFS Team" patch.
Fred
> On Feb 11, 2011, at 11:54 AM, Fred Isaman wrote:
>
>> From: The pNFS Team <linux-nfs@vger.kernel.org>
>>
>> Reorder nfs_write_rpcsetup, preparing for a pnfs switch.
>>
>> FIXME: these signoffs are ridiculous for this simple patch.
>> Lets just pick an author and be done with it.
>>
>> Reported-by: Alexandros Batsakis <batsakis@netapp.com>
>> Signed-off-by: Andy Adamson <andros@citi.umich.edu>
>> Signed-off-by: Boaz Harrosh <bharrosh@panasas.com>
>> Signed-off-by: Dean Hildebrand <dhildeb@us.ibm.com>
>> Signed-off-by: Fred Isaman <iisaman@citi.umich.edu>
>> Signed-off-by: J. Bruce Fields <bfields@fieldses.org>
>> Signed-off-by: Mike Sager <sager@netapp.com>
>> Signed-off-by: Ricardo Labiaga <Ricardo.Labiaga@netapp.com>
>> Signed-off-by: Tao Guo <guotao@nrchpc.ac.cn>
>> Signed-off-by: Andy Adamson <andros@netapp.com>
>> Signed-off-by: Benny Halevy <bhalevy@panasas.com>
>> Signed-off-by: Fred Isaman <iisaman@netapp.com>
>> ---
>> fs/nfs/write.c | 82 +++++++++++++++++++++++++++++++------------------------
>> 1 files changed, 46 insertions(+), 36 deletions(-)
>>
>> diff --git a/fs/nfs/write.c b/fs/nfs/write.c
>> index aca0268..5604854 100644
>> --- a/fs/nfs/write.c
>> +++ b/fs/nfs/write.c
>> @@ -782,25 +782,21 @@ static int flush_task_priority(int how)
>> return RPC_PRIORITY_NORMAL;
>> }
>>
>> -/*
>> - * Set up the argument/result storage required for the RPC call.
>> - */
>> -static int nfs_write_rpcsetup(struct nfs_page *req,
>> - struct nfs_write_data *data,
>> - const struct rpc_call_ops *call_ops,
>> - unsigned int count, unsigned int offset,
>> - int how)
>> +static int nfs_initiate_write(struct nfs_write_data *data,
>> + struct rpc_clnt *clnt,
>> + const struct rpc_call_ops *call_ops,
>> + int how)
>> {
>> - struct inode *inode = req->wb_context->path.dentry->d_inode;
>> + struct inode *inode = data->inode;
>> int priority = flush_task_priority(how);
>> struct rpc_task *task;
>> struct rpc_message msg = {
>> .rpc_argp = &data->args,
>> .rpc_resp = &data->res,
>> - .rpc_cred = req->wb_context->cred,
>> + .rpc_cred = data->cred,
>> };
>> struct rpc_task_setup task_setup_data = {
>> - .rpc_client = NFS_CLIENT(inode),
>> + .rpc_client = clnt,
>> .task = &data->task,
>> .rpc_message = &msg,
>> .callback_ops = call_ops,
>> @@ -811,12 +807,49 @@ static int nfs_write_rpcsetup(struct nfs_page *req,
>> };
>> int ret = 0;
>>
>> + /* Set up the initial task struct. */
>> + NFS_PROTO(inode)->write_setup(data, &msg);
>> +
>> + dprintk("NFS: %5u initiated write call "
>> + "(req %s/%lld, %u bytes @ offset %llu)\n",
>> + data->task.tk_pid,
>> + inode->i_sb->s_id,
>> + (long long)NFS_FILEID(inode),
>> + data->args.count,
>> + (unsigned long long)data->args.offset);
>> +
>> + task = rpc_run_task(&task_setup_data);
>> + if (IS_ERR(task)) {
>> + ret = PTR_ERR(task);
>> + goto out;
>> + }
>> + if (how & FLUSH_SYNC) {
>> + ret = rpc_wait_for_completion_task(task);
>> + if (ret == 0)
>> + ret = task->tk_status;
>> + }
>> + rpc_put_task(task);
>> +out:
>> + return ret;
>> +}
>> +
>> +/*
>> + * Set up the argument/result storage required for the RPC call.
>> + */
>> +static int nfs_write_rpcsetup(struct nfs_page *req,
>> + struct nfs_write_data *data,
>> + const struct rpc_call_ops *call_ops,
>> + unsigned int count, unsigned int offset,
>> + int how)
>> +{
>> + struct inode *inode = req->wb_context->path.dentry->d_inode;
>> +
>> /* Set up the RPC argument and reply structs
>> * NB: take care not to mess about with data->commit et al. */
>>
>> data->req = req;
>> data->inode = inode = req->wb_context->path.dentry->d_inode;
>> - data->cred = msg.rpc_cred;
>> + data->cred = req->wb_context->cred;
>>
>> data->args.fh = NFS_FH(inode);
>> data->args.offset = req_offset(req) + offset;
>> @@ -837,30 +870,7 @@ static int nfs_write_rpcsetup(struct nfs_page *req,
>> data->res.verf = &data->verf;
>> nfs_fattr_init(&data->fattr);
>>
>> - /* Set up the initial task struct. */
>> - NFS_PROTO(inode)->write_setup(data, &msg);
>> -
>> - dprintk("NFS: %5u initiated write call "
>> - "(req %s/%lld, %u bytes @ offset %llu)\n",
>> - data->task.tk_pid,
>> - inode->i_sb->s_id,
>> - (long long)NFS_FILEID(inode),
>> - count,
>> - (unsigned long long)data->args.offset);
>> -
>> - task = rpc_run_task(&task_setup_data);
>> - if (IS_ERR(task)) {
>> - ret = PTR_ERR(task);
>> - goto out;
>> - }
>> - if (how & FLUSH_SYNC) {
>> - ret = rpc_wait_for_completion_task(task);
>> - if (ret == 0)
>> - ret = task->tk_status;
>> - }
>> - rpc_put_task(task);
>> -out:
>> - return ret;
>> + return nfs_initiate_write(data, NFS_CLIENT(inode), call_ops, how);
>> }
>>
>> /* If a nfs_flush_* function fails, it should remove reqs from @head and
>> --
>> 1.7.2.1
>>
>> --
>> To unsubscribe from this list: send the line "unsubscribe linux-nfs" in
>> the body of a message to majordomo@vger.kernel.org
>> More majordomo info at http://vger.kernel.org/majordomo-info.html
>
^ permalink raw reply [flat|nested] 15+ messages in thread
* Re: [PATCH 1/7] wave4: nfs: rearrange nfs_write_rpcsetup
2011-02-14 17:27 ` Fred Isaman
@ 2011-02-14 18:11 ` William A. (Andy) Adamson
0 siblings, 0 replies; 15+ messages in thread
From: William A. (Andy) Adamson @ 2011-02-14 18:11 UTC (permalink / raw)
To: Fred Isaman; +Cc: linux-nfs
sure.
On Mon, Feb 14, 2011 at 12:27 PM, Fred Isaman <iisaman@netapp.com> wrote:
>
> On Feb 14, 2011, at 12:22 PM, Andy Adamson wrote:
>
>> Since I'm the author for the read version of this code, I'll raise my hand to be the author for this version.
>>
>> -->Andy
>>
>
> OK, do you wan't patch 5 as well?...they were both split from the same "pNFS Team" patch.
>
> Fred
>
>> On Feb 11, 2011, at 11:54 AM, Fred Isaman wrote:
>>
>>> From: The pNFS Team <linux-nfs@vger.kernel.org>
>>>
>>> Reorder nfs_write_rpcsetup, preparing for a pnfs switch.
>>>
>>> FIXME: these signoffs are ridiculous for this simple patch.
>>> Lets just pick an author and be done with it.
>>>
>>> Reported-by: Alexandros Batsakis <batsakis@netapp.com>
>>> Signed-off-by: Andy Adamson <andros@citi.umich.edu>
>>> Signed-off-by: Boaz Harrosh <bharrosh@panasas.com>
>>> Signed-off-by: Dean Hildebrand <dhildeb@us.ibm.com>
>>> Signed-off-by: Fred Isaman <iisaman@citi.umich.edu>
>>> Signed-off-by: J. Bruce Fields <bfields@fieldses.org>
>>> Signed-off-by: Mike Sager <sager@netapp.com>
>>> Signed-off-by: Ricardo Labiaga <Ricardo.Labiaga@netapp.com>
>>> Signed-off-by: Tao Guo <guotao@nrchpc.ac.cn>
>>> Signed-off-by: Andy Adamson <andros@netapp.com>
>>> Signed-off-by: Benny Halevy <bhalevy@panasas.com>
>>> Signed-off-by: Fred Isaman <iisaman@netapp.com>
>>> ---
>>> fs/nfs/write.c | 82 +++++++++++++++++++++++++++++++------------------------
>>> 1 files changed, 46 insertions(+), 36 deletions(-)
>>>
>>> diff --git a/fs/nfs/write.c b/fs/nfs/write.c
>>> index aca0268..5604854 100644
>>> --- a/fs/nfs/write.c
>>> +++ b/fs/nfs/write.c
>>> @@ -782,25 +782,21 @@ static int flush_task_priority(int how)
>>> return RPC_PRIORITY_NORMAL;
>>> }
>>>
>>> -/*
>>> - * Set up the argument/result storage required for the RPC call.
>>> - */
>>> -static int nfs_write_rpcsetup(struct nfs_page *req,
>>> - struct nfs_write_data *data,
>>> - const struct rpc_call_ops *call_ops,
>>> - unsigned int count, unsigned int offset,
>>> - int how)
>>> +static int nfs_initiate_write(struct nfs_write_data *data,
>>> + struct rpc_clnt *clnt,
>>> + const struct rpc_call_ops *call_ops,
>>> + int how)
>>> {
>>> - struct inode *inode = req->wb_context->path.dentry->d_inode;
>>> + struct inode *inode = data->inode;
>>> int priority = flush_task_priority(how);
>>> struct rpc_task *task;
>>> struct rpc_message msg = {
>>> .rpc_argp = &data->args,
>>> .rpc_resp = &data->res,
>>> - .rpc_cred = req->wb_context->cred,
>>> + .rpc_cred = data->cred,
>>> };
>>> struct rpc_task_setup task_setup_data = {
>>> - .rpc_client = NFS_CLIENT(inode),
>>> + .rpc_client = clnt,
>>> .task = &data->task,
>>> .rpc_message = &msg,
>>> .callback_ops = call_ops,
>>> @@ -811,12 +807,49 @@ static int nfs_write_rpcsetup(struct nfs_page *req,
>>> };
>>> int ret = 0;
>>>
>>> + /* Set up the initial task struct. */
>>> + NFS_PROTO(inode)->write_setup(data, &msg);
>>> +
>>> + dprintk("NFS: %5u initiated write call "
>>> + "(req %s/%lld, %u bytes @ offset %llu)\n",
>>> + data->task.tk_pid,
>>> + inode->i_sb->s_id,
>>> + (long long)NFS_FILEID(inode),
>>> + data->args.count,
>>> + (unsigned long long)data->args.offset);
>>> +
>>> + task = rpc_run_task(&task_setup_data);
>>> + if (IS_ERR(task)) {
>>> + ret = PTR_ERR(task);
>>> + goto out;
>>> + }
>>> + if (how & FLUSH_SYNC) {
>>> + ret = rpc_wait_for_completion_task(task);
>>> + if (ret == 0)
>>> + ret = task->tk_status;
>>> + }
>>> + rpc_put_task(task);
>>> +out:
>>> + return ret;
>>> +}
>>> +
>>> +/*
>>> + * Set up the argument/result storage required for the RPC call.
>>> + */
>>> +static int nfs_write_rpcsetup(struct nfs_page *req,
>>> + struct nfs_write_data *data,
>>> + const struct rpc_call_ops *call_ops,
>>> + unsigned int count, unsigned int offset,
>>> + int how)
>>> +{
>>> + struct inode *inode = req->wb_context->path.dentry->d_inode;
>>> +
>>> /* Set up the RPC argument and reply structs
>>> * NB: take care not to mess about with data->commit et al. */
>>>
>>> data->req = req;
>>> data->inode = inode = req->wb_context->path.dentry->d_inode;
>>> - data->cred = msg.rpc_cred;
>>> + data->cred = req->wb_context->cred;
>>>
>>> data->args.fh = NFS_FH(inode);
>>> data->args.offset = req_offset(req) + offset;
>>> @@ -837,30 +870,7 @@ static int nfs_write_rpcsetup(struct nfs_page *req,
>>> data->res.verf = &data->verf;
>>> nfs_fattr_init(&data->fattr);
>>>
>>> - /* Set up the initial task struct. */
>>> - NFS_PROTO(inode)->write_setup(data, &msg);
>>> -
>>> - dprintk("NFS: %5u initiated write call "
>>> - "(req %s/%lld, %u bytes @ offset %llu)\n",
>>> - data->task.tk_pid,
>>> - inode->i_sb->s_id,
>>> - (long long)NFS_FILEID(inode),
>>> - count,
>>> - (unsigned long long)data->args.offset);
>>> -
>>> - task = rpc_run_task(&task_setup_data);
>>> - if (IS_ERR(task)) {
>>> - ret = PTR_ERR(task);
>>> - goto out;
>>> - }
>>> - if (how & FLUSH_SYNC) {
>>> - ret = rpc_wait_for_completion_task(task);
>>> - if (ret == 0)
>>> - ret = task->tk_status;
>>> - }
>>> - rpc_put_task(task);
>>> -out:
>>> - return ret;
>>> + return nfs_initiate_write(data, NFS_CLIENT(inode), call_ops, how);
>>> }
>>>
>>> /* If a nfs_flush_* function fails, it should remove reqs from @head and
>>> --
>>> 1.7.2.1
>>>
>>> --
>>> To unsubscribe from this list: send the line "unsubscribe linux-nfs" in
>>> the body of a message to majordomo@vger.kernel.org
>>> More majordomo info at http://vger.kernel.org/majordomo-info.html
>>
>
> --
> To unsubscribe from this list: send the line "unsubscribe linux-nfs" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at http://vger.kernel.org/majordomo-info.html
>
^ permalink raw reply [flat|nested] 15+ messages in thread
* [PATCH 2/7] wave4: Send lseg down through nfs_write_rpcsetup into nfs_initiate_write
2011-02-11 16:54 [PATCH 0/7] pnfs wave 4 proposal Fred Isaman
2011-02-11 16:54 ` [PATCH 1/7] wave4: nfs: rearrange nfs_write_rpcsetup Fred Isaman
@ 2011-02-11 16:54 ` Fred Isaman
2011-02-11 16:54 ` [PATCH 3/7] wave4: remove GETATTR from ds write Fred Isaman
` (5 subsequent siblings)
7 siblings, 0 replies; 15+ messages in thread
From: Fred Isaman @ 2011-02-11 16:54 UTC (permalink / raw)
To: linux-nfs
We grab the lseg sent in to the doio function and attach it to
each struct nfs_write_data created. This is how the lseg will be
sent to the layout driver.
Signed-off-by: Fred Isaman <iisaman@netapp.com>
---
fs/nfs/write.c | 7 +++++--
include/linux/nfs_xdr.h | 1 +
2 files changed, 6 insertions(+), 2 deletions(-)
diff --git a/fs/nfs/write.c b/fs/nfs/write.c
index 5604854..fa92fad 100644
--- a/fs/nfs/write.c
+++ b/fs/nfs/write.c
@@ -98,6 +98,7 @@ void nfs_writedata_free(struct nfs_write_data *p)
static void nfs_writedata_release(struct nfs_write_data *wdata)
{
put_nfs_open_context(wdata->args.context);
+ put_lseg(wdata->lseg);
nfs_writedata_free(wdata);
}
@@ -840,6 +841,7 @@ static int nfs_write_rpcsetup(struct nfs_page *req,
struct nfs_write_data *data,
const struct rpc_call_ops *call_ops,
unsigned int count, unsigned int offset,
+ struct pnfs_layout_segment *lseg,
int how)
{
struct inode *inode = req->wb_context->path.dentry->d_inode;
@@ -850,6 +852,7 @@ static int nfs_write_rpcsetup(struct nfs_page *req,
data->req = req;
data->inode = inode = req->wb_context->path.dentry->d_inode;
data->cred = req->wb_context->cred;
+ data->lseg = get_lseg(lseg);
data->args.fh = NFS_FH(inode);
data->args.offset = req_offset(req) + offset;
@@ -930,7 +933,7 @@ static int nfs_flush_multi(struct inode *inode, struct list_head *head, unsigned
if (nbytes < wsize)
wsize = nbytes;
ret2 = nfs_write_rpcsetup(req, data, &nfs_write_partial_ops,
- wsize, offset, how);
+ wsize, offset, lseg, how);
if (ret == 0)
ret = ret2;
offset += wsize;
@@ -978,7 +981,7 @@ static int nfs_flush_one(struct inode *inode, struct list_head *head, unsigned i
req = nfs_list_entry(data->pages.next);
/* Set up the argument struct */
- return nfs_write_rpcsetup(req, data, &nfs_write_full_ops, count, 0, how);
+ return nfs_write_rpcsetup(req, data, &nfs_write_full_ops, count, 0, lseg, how);
out_bad:
while (!list_empty(head)) {
req = nfs_list_entry(head->next);
diff --git a/include/linux/nfs_xdr.h b/include/linux/nfs_xdr.h
index 4121c3e..d60ec32 100644
--- a/include/linux/nfs_xdr.h
+++ b/include/linux/nfs_xdr.h
@@ -1038,6 +1038,7 @@ struct nfs_write_data {
unsigned int npages; /* Max length of pagevec */
struct nfs_writeargs args; /* argument struct */
struct nfs_writeres res; /* result struct */
+ struct pnfs_layout_segment *lseg;
#ifdef CONFIG_NFS_V4
unsigned long timestamp; /* For lease renewal */
#endif
--
1.7.2.1
^ permalink raw reply related [flat|nested] 15+ messages in thread
* [PATCH 3/7] wave4: remove GETATTR from ds write
2011-02-11 16:54 [PATCH 0/7] pnfs wave 4 proposal Fred Isaman
2011-02-11 16:54 ` [PATCH 1/7] wave4: nfs: rearrange nfs_write_rpcsetup Fred Isaman
2011-02-11 16:54 ` [PATCH 2/7] wave4: Send lseg down through nfs_write_rpcsetup into nfs_initiate_write Fred Isaman
@ 2011-02-11 16:54 ` Fred Isaman
2011-02-11 16:54 ` [PATCH 4/7] wave4: nfs: rearrange nfs4_write_done Fred Isaman
` (4 subsequent siblings)
7 siblings, 0 replies; 15+ messages in thread
From: Fred Isaman @ 2011-02-11 16:54 UTC (permalink / raw)
To: linux-nfs
Any WRITE compounds directed to a data server need to have the
GETATTR calls suppressed.
Signed-off-by: Fred Isaman <iisaman@netapp.com>
---
fs/nfs/nfs4proc.c | 6 +++++-
fs/nfs/nfs4xdr.c | 8 +++++---
include/linux/nfs_xdr.h | 1 +
3 files changed, 11 insertions(+), 4 deletions(-)
diff --git a/fs/nfs/nfs4proc.c b/fs/nfs/nfs4proc.c
index f396853..a40a281 100644
--- a/fs/nfs/nfs4proc.c
+++ b/fs/nfs/nfs4proc.c
@@ -3144,7 +3144,11 @@ static void nfs4_proc_write_setup(struct nfs_write_data *data, struct rpc_messag
{
struct nfs_server *server = NFS_SERVER(data->inode);
- data->args.bitmask = server->cache_consistency_bitmask;
+ if (data->ds_clp) {
+ data->args.bitmask = NULL;
+ data->res.fattr = NULL;
+ } else
+ data->args.bitmask = server->cache_consistency_bitmask;
data->res.server = server;
data->timestamp = jiffies;
diff --git a/fs/nfs/nfs4xdr.c b/fs/nfs/nfs4xdr.c
index 2380c45..c35880c 100644
--- a/fs/nfs/nfs4xdr.c
+++ b/fs/nfs/nfs4xdr.c
@@ -2275,7 +2275,8 @@ static void nfs4_xdr_enc_write(struct rpc_rqst *req, struct xdr_stream *xdr,
encode_putfh(xdr, args->fh, &hdr);
encode_write(xdr, args, &hdr);
req->rq_snd_buf.flags |= XDRBUF_WRITE;
- encode_getfattr(xdr, args->bitmask, &hdr);
+ if (args->bitmask)
+ encode_getfattr(xdr, args->bitmask, &hdr);
encode_nops(&hdr);
}
@@ -5694,8 +5695,9 @@ static int nfs4_xdr_dec_write(struct rpc_rqst *rqstp, struct xdr_stream *xdr,
status = decode_write(xdr, res);
if (status)
goto out;
- decode_getfattr(xdr, res->fattr, res->server,
- !RPC_IS_ASYNC(rqstp->rq_task));
+ if (res->fattr)
+ decode_getfattr(xdr, res->fattr, res->server,
+ !RPC_IS_ASYNC(rqstp->rq_task));
if (!status)
status = res->count;
out:
diff --git a/include/linux/nfs_xdr.h b/include/linux/nfs_xdr.h
index d60ec32..726ec2a 100644
--- a/include/linux/nfs_xdr.h
+++ b/include/linux/nfs_xdr.h
@@ -1039,6 +1039,7 @@ struct nfs_write_data {
struct nfs_writeargs args; /* argument struct */
struct nfs_writeres res; /* result struct */
struct pnfs_layout_segment *lseg;
+ struct nfs_client *ds_clp; /* pNFS data server */
#ifdef CONFIG_NFS_V4
unsigned long timestamp; /* For lease renewal */
#endif
--
1.7.2.1
^ permalink raw reply related [flat|nested] 15+ messages in thread
* [PATCH 4/7] wave4: nfs: rearrange nfs4_write_done
2011-02-11 16:54 [PATCH 0/7] pnfs wave 4 proposal Fred Isaman
` (2 preceding siblings ...)
2011-02-11 16:54 ` [PATCH 3/7] wave4: remove GETATTR from ds write Fred Isaman
@ 2011-02-11 16:54 ` Fred Isaman
2011-02-11 16:54 ` [PATCH 5/7] wave4: implement pnfs_try_to_write_data Fred Isaman
` (3 subsequent siblings)
7 siblings, 0 replies; 15+ messages in thread
From: Fred Isaman @ 2011-02-11 16:54 UTC (permalink / raw)
To: linux-nfs
Add hook that layout driver can use to take control here.
Signed-off-by: Fred Isaman <iisaman@netapp.com>
---
fs/nfs/nfs4proc.c | 14 ++++++++++----
include/linux/nfs_xdr.h | 1 +
2 files changed, 11 insertions(+), 4 deletions(-)
diff --git a/fs/nfs/nfs4proc.c b/fs/nfs/nfs4proc.c
index a40a281..3952f7c 100644
--- a/fs/nfs/nfs4proc.c
+++ b/fs/nfs/nfs4proc.c
@@ -3122,13 +3122,10 @@ void nfs4_reset_read(struct rpc_task *task, struct nfs_read_data *data)
}
EXPORT_SYMBOL_GPL(nfs4_reset_read);
-static int nfs4_write_done(struct rpc_task *task, struct nfs_write_data *data)
+static int nfs4_write_done_cb(struct rpc_task *task, struct nfs_write_data *data)
{
struct inode *inode = data->inode;
- if (!nfs4_sequence_done(task, &data->res.seq_res))
- return -EAGAIN;
-
if (nfs4_async_handle_error(task, NFS_SERVER(inode), data->args.context->state) == -EAGAIN) {
nfs_restart_rpc(task, NFS_SERVER(inode)->nfs_client);
return -EAGAIN;
@@ -3140,6 +3137,13 @@ static int nfs4_write_done(struct rpc_task *task, struct nfs_write_data *data)
return 0;
}
+static int nfs4_write_done(struct rpc_task *task, struct nfs_write_data *data)
+{
+ if (!nfs4_sequence_done(task, &data->res.seq_res))
+ return -EAGAIN;
+ return data->write_done_cb(task, data);
+}
+
static void nfs4_proc_write_setup(struct nfs_write_data *data, struct rpc_message *msg)
{
struct nfs_server *server = NFS_SERVER(data->inode);
@@ -3149,6 +3153,8 @@ static void nfs4_proc_write_setup(struct nfs_write_data *data, struct rpc_messag
data->res.fattr = NULL;
} else
data->args.bitmask = server->cache_consistency_bitmask;
+ if (!data->write_done_cb)
+ data->write_done_cb = nfs4_write_done_cb;
data->res.server = server;
data->timestamp = jiffies;
diff --git a/include/linux/nfs_xdr.h b/include/linux/nfs_xdr.h
index 726ec2a..23feaf6 100644
--- a/include/linux/nfs_xdr.h
+++ b/include/linux/nfs_xdr.h
@@ -1040,6 +1040,7 @@ struct nfs_write_data {
struct nfs_writeres res; /* result struct */
struct pnfs_layout_segment *lseg;
struct nfs_client *ds_clp; /* pNFS data server */
+ int (*write_done_cb) (struct rpc_task *task, struct nfs_write_data *data);
#ifdef CONFIG_NFS_V4
unsigned long timestamp; /* For lease renewal */
#endif
--
1.7.2.1
^ permalink raw reply related [flat|nested] 15+ messages in thread
* [PATCH 5/7] wave4: implement pnfs_try_to_write_data
2011-02-11 16:54 [PATCH 0/7] pnfs wave 4 proposal Fred Isaman
` (3 preceding siblings ...)
2011-02-11 16:54 ` [PATCH 4/7] wave4: nfs: rearrange nfs4_write_done Fred Isaman
@ 2011-02-11 16:54 ` Fred Isaman
2011-02-11 16:54 ` [PATCH 6/7] wave4: filelayout write Fred Isaman
` (2 subsequent siblings)
7 siblings, 0 replies; 15+ messages in thread
From: Fred Isaman @ 2011-02-11 16:54 UTC (permalink / raw)
To: linux-nfs
From: The pNFS Team <linux-nfs@vger.kernel.org>
split from "pnfs_submit: generic write" and modified
Signed-off-by: Andy Adamson <andros@citi.umich.edu>
Signed-off-by: Boaz Harrosh <bharrosh@panasas.com>
Signed-off-by: Dean Hildebrand <dhildeb@us.ibm.com>
Signed-off-by: Fred Isaman <iisaman@citi.umich.edu>
Signed-off-by: J. Bruce Fields <bfields@fieldses.org>
Signed-off-by: Mike Sager <sager@netapp.com>
Signed-off-by: Ricardo Labiaga <Ricardo.Labiaga@netapp.com>
Signed-off-by: Tao Guo <guotao@nrchpc.ac.cn>
Signed-off-by: Andy Adamson <andros@netapp.com>
Signed-off-by: Benny Halevy <bhalevy@panasas.com>
Signed-off-by: Fred Isaman <iisaman@netapp.com>
---
fs/nfs/pnfs.c | 24 ++++++++++++++++++++++++
fs/nfs/pnfs.h | 11 +++++++++++
fs/nfs/write.c | 4 ++++
include/linux/nfs_iostat.h | 1 +
include/linux/nfs_xdr.h | 1 +
5 files changed, 41 insertions(+), 0 deletions(-)
diff --git a/fs/nfs/pnfs.c b/fs/nfs/pnfs.c
index 912b1ff..9d2351a 100644
--- a/fs/nfs/pnfs.c
+++ b/fs/nfs/pnfs.c
@@ -892,6 +892,30 @@ pnfs_pageio_init_read(struct nfs_pageio_descriptor *pgio,
pnfs_set_pg_test(inode, pgio);
}
+enum pnfs_try_status
+pnfs_try_to_write_data(struct nfs_write_data *wdata,
+ const struct rpc_call_ops *call_ops, int how)
+{
+ struct inode *inode = wdata->inode;
+ enum pnfs_try_status trypnfs;
+ struct nfs_server *nfss = NFS_SERVER(inode);
+
+ wdata->mds_ops = call_ops;
+
+ dprintk("%s: Writing ino:%lu %u@%llu (how %d)\n", __func__,
+ inode->i_ino, wdata->args.count, wdata->args.offset, how);
+
+ trypnfs = nfss->pnfs_curr_ld->write_pagelist(wdata, how);
+ if (trypnfs == PNFS_NOT_ATTEMPTED) {
+ wdata->lseg = NULL;
+ put_lseg(wdata->lseg);
+ } else
+ nfs_inc_stats(inode, NFSIOS_PNFS_WRITE);
+
+ dprintk("%s End (trypnfs:%d)\n", __func__, trypnfs);
+ return trypnfs;
+}
+
/*
* Call the appropriate parallel I/O subsystem read function.
*/
diff --git a/fs/nfs/pnfs.h b/fs/nfs/pnfs.h
index d2d708a..50bcb1f 100644
--- a/fs/nfs/pnfs.h
+++ b/fs/nfs/pnfs.h
@@ -81,6 +81,8 @@ struct pnfs_layoutdriver_type {
* I/O, else return PNFS_NOT_ATTEMPTED to fall back to normal NFS
*/
enum pnfs_try_status (*read_pagelist) (struct nfs_read_data *nfs_data);
+ enum pnfs_try_status
+ (*write_pagelist) (struct nfs_write_data *nfs_data, int how);
};
struct pnfs_layout_hdr {
@@ -172,6 +174,8 @@ pnfs_update_layout(struct inode *ino, struct nfs_open_context *ctx,
enum pnfs_iomode access_type);
void set_pnfs_layoutdriver(struct nfs_server *, u32 id);
void unset_pnfs_layoutdriver(struct nfs_server *);
+enum pnfs_try_status pnfs_try_to_write_data(struct nfs_write_data *,
+ const struct rpc_call_ops *, int);
enum pnfs_try_status pnfs_try_to_read_data(struct nfs_read_data *,
const struct rpc_call_ops *);
void pnfs_pageio_init_read(struct nfs_pageio_descriptor *, struct inode *);
@@ -250,6 +254,13 @@ pnfs_try_to_read_data(struct nfs_read_data *data,
return PNFS_NOT_ATTEMPTED;
}
+static inline enum pnfs_try_status
+pnfs_try_to_write_data(struct nfs_write_data *data,
+ const struct rpc_call_ops *call_ops, int how)
+{
+ return PNFS_NOT_ATTEMPTED;
+}
+
static inline bool
pnfs_roc(struct inode *ino)
{
diff --git a/fs/nfs/write.c b/fs/nfs/write.c
index fa92fad..24cc34d 100644
--- a/fs/nfs/write.c
+++ b/fs/nfs/write.c
@@ -873,6 +873,10 @@ static int nfs_write_rpcsetup(struct nfs_page *req,
data->res.verf = &data->verf;
nfs_fattr_init(&data->fattr);
+ if (data->lseg &&
+ (pnfs_try_to_write_data(data, call_ops, how) == PNFS_ATTEMPTED))
+ return 0;
+
return nfs_initiate_write(data, NFS_CLIENT(inode), call_ops, how);
}
diff --git a/include/linux/nfs_iostat.h b/include/linux/nfs_iostat.h
index 37a1437..8866bb3 100644
--- a/include/linux/nfs_iostat.h
+++ b/include/linux/nfs_iostat.h
@@ -114,6 +114,7 @@ enum nfs_stat_eventcounters {
NFSIOS_SHORTWRITE,
NFSIOS_DELAY,
NFSIOS_PNFS_READ,
+ NFSIOS_PNFS_WRITE,
__NFSIOS_COUNTSMAX,
};
diff --git a/include/linux/nfs_xdr.h b/include/linux/nfs_xdr.h
index 23feaf6..9ff26f9 100644
--- a/include/linux/nfs_xdr.h
+++ b/include/linux/nfs_xdr.h
@@ -1040,6 +1040,7 @@ struct nfs_write_data {
struct nfs_writeres res; /* result struct */
struct pnfs_layout_segment *lseg;
struct nfs_client *ds_clp; /* pNFS data server */
+ const struct rpc_call_ops *mds_ops;
int (*write_done_cb) (struct rpc_task *task, struct nfs_write_data *data);
#ifdef CONFIG_NFS_V4
unsigned long timestamp; /* For lease renewal */
--
1.7.2.1
^ permalink raw reply related [flat|nested] 15+ messages in thread
* [PATCH 6/7] wave4: filelayout write
2011-02-11 16:54 [PATCH 0/7] pnfs wave 4 proposal Fred Isaman
` (4 preceding siblings ...)
2011-02-11 16:54 ` [PATCH 5/7] wave4: implement pnfs_try_to_write_data Fred Isaman
@ 2011-02-11 16:54 ` Fred Isaman
2011-02-11 16:54 ` [PATCH 7/7] wave4: nfs_pageio_do_add_request needs to distinguish between read and write Fred Isaman
2011-02-11 20:19 ` [PATCH 0/7] pnfs wave 4 proposal Andy Adamson
7 siblings, 0 replies; 15+ messages in thread
From: Fred Isaman @ 2011-02-11 16:54 UTC (permalink / raw)
To: linux-nfs
From: The pNFS Team <linux-nfs@vger.kernel.org>
Signed-off-by: Andy Adamson <andros@citi.umich.edu>
Signed-off-by: Dean Hildebrand <dhildeb@us.ibm.com>
Signed-off-by: Fred Isaman <iisaman@citi.umich.edu>
Signed-off-by: Mingyang Guo <guomingyang@nrchpc.ac.cn>
Signed-off-by: Oleg Drokin <green@linuxhacker.ru>
Signed-off-by: Ricardo Labiaga <Ricardo.Labiaga@netapp.com>
Tested-by: Guo Mingyang <guomingyang@nrchpc.ac.cn>
Signed-off-by: Andy Adamson <andros@netapp.com>
Signed-off-by: Benny Halevy <bhalevy@panasas.com>
Signed-off-by: Fred Isaman <iisaman@netapp.com>
---
fs/nfs/internal.h | 5 ++
fs/nfs/nfs4filelayout.c | 103 +++++++++++++++++++++++++++++++++++++++++++++++
fs/nfs/nfs4proc.c | 17 ++++++++
fs/nfs/write.c | 5 ++-
include/linux/nfs_xdr.h | 1 +
5 files changed, 130 insertions(+), 1 deletions(-)
diff --git a/fs/nfs/internal.h b/fs/nfs/internal.h
index f69a322..34096a1 100644
--- a/fs/nfs/internal.h
+++ b/fs/nfs/internal.h
@@ -272,6 +272,10 @@ extern int nfs_initiate_read(struct nfs_read_data *data, struct rpc_clnt *clnt,
extern void nfs_read_prepare(struct rpc_task *task, void *calldata);
/* write.c */
+extern int nfs_initiate_write(struct nfs_write_data *data,
+ struct rpc_clnt *clnt,
+ const struct rpc_call_ops *call_ops,
+ int how);
extern void nfs_write_prepare(struct rpc_task *task, void *calldata);
#ifdef CONFIG_MIGRATION
extern int nfs_migrate_page(struct address_space *,
@@ -282,6 +286,7 @@ extern int nfs_migrate_page(struct address_space *,
/* nfs4proc.c */
extern void nfs4_reset_read(struct rpc_task *task, struct nfs_read_data *data);
+extern void nfs4_reset_write(struct rpc_task *task, struct nfs_write_data *data);
extern int _nfs4_call_sync(struct nfs_server *server,
struct rpc_message *msg,
struct nfs4_sequence_args *args,
diff --git a/fs/nfs/nfs4filelayout.c b/fs/nfs/nfs4filelayout.c
index 80e7a66..be4af22 100644
--- a/fs/nfs/nfs4filelayout.c
+++ b/fs/nfs/nfs4filelayout.c
@@ -208,12 +208,69 @@ static void filelayout_read_release(void *data)
rdata->mds_ops->rpc_release(data);
}
+static int filelayout_write_done_cb(struct rpc_task *task,
+ struct nfs_write_data *data)
+{
+ int reset = 0;
+
+ if (filelayout_async_handle_error(task, data->args.context->state,
+ data->ds_clp, &reset) == -EAGAIN) {
+ struct nfs_client *clp;
+
+ dprintk("%s calling restart ds_clp %p ds_clp->cl_session %p\n",
+ __func__, data->ds_clp, data->ds_clp->cl_session);
+ if (reset) {
+ nfs4_reset_write(task, data);
+ filelayout_set_lo_fail(data->lseg);
+ clp = NFS_SERVER(data->inode)->nfs_client;
+ } else
+ clp = data->ds_clp;
+ nfs_restart_rpc(task, clp);
+ return -EAGAIN;
+ }
+
+ return 0;
+}
+
+static void filelayout_write_prepare(struct rpc_task *task, void *data)
+{
+ struct nfs_write_data *wdata = (struct nfs_write_data *)data;
+
+ if (nfs41_setup_sequence(wdata->ds_clp->cl_session,
+ &wdata->args.seq_args, &wdata->res.seq_res,
+ 0, task))
+ return;
+
+ rpc_call_start(task);
+}
+
+static void filelayout_write_call_done(struct rpc_task *task, void *data)
+{
+ struct nfs_write_data *wdata = (struct nfs_write_data *)data;
+
+ /* Note this may cause RPC to be resent */
+ wdata->mds_ops->rpc_call_done(task, data);
+}
+
+static void filelayout_write_release(void *data)
+{
+ struct nfs_write_data *wdata = (struct nfs_write_data *)data;
+
+ wdata->mds_ops->rpc_release(data);
+}
+
struct rpc_call_ops filelayout_read_call_ops = {
.rpc_call_prepare = filelayout_read_prepare,
.rpc_call_done = filelayout_read_call_done,
.rpc_release = filelayout_read_release,
};
+struct rpc_call_ops filelayout_write_call_ops = {
+ .rpc_call_prepare = filelayout_write_prepare,
+ .rpc_call_done = filelayout_write_call_done,
+ .rpc_release = filelayout_write_release,
+};
+
static enum pnfs_try_status
filelayout_read_pagelist(struct nfs_read_data *data)
{
@@ -254,6 +311,51 @@ filelayout_read_pagelist(struct nfs_read_data *data)
return PNFS_ATTEMPTED;
}
+/* Perform async writes. */
+static enum pnfs_try_status
+filelayout_write_pagelist(struct nfs_write_data *data, int sync)
+{
+ struct pnfs_layout_segment *lseg = data->lseg;
+ struct nfs4_pnfs_ds *ds;
+ loff_t offset = data->args.offset;
+ u32 idx;
+ struct nfs_fh *fh;
+
+ /* 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__);
+ set_bit(lo_fail_bit(IOMODE_RW), &lseg->pls_layout->plh_flags);
+ set_bit(lo_fail_bit(IOMODE_READ), &lseg->pls_layout->plh_flags);
+ return PNFS_NOT_ATTEMPTED;
+ }
+ dprintk("%s ino %lu sync %d req %Zu@%llu DS:%x:%hu\n", __func__,
+ data->inode->i_ino, sync, (size_t) data->args.count, offset,
+ ntohl(ds->ds_ip_addr), ntohs(ds->ds_port));
+
+ /* We can't handle commit to ds yet */
+ if (!FILELAYOUT_LSEG(lseg)->commit_through_mds)
+ data->args.stable = NFS_FILE_SYNC;
+
+ data->write_done_cb = filelayout_write_done_cb;
+ data->ds_clp = ds->ds_clp;
+ fh = nfs4_fl_select_ds_fh(lseg, offset);
+ if (fh)
+ data->args.fh = fh;
+ /*
+ * Get the file offset on the dserver. Set the write offset to
+ * this offset and save the original offset.
+ */
+ data->args.offset = filelayout_get_dserver_offset(lseg, offset);
+ data->mds_offset = offset;
+
+ /* Perform an asynchronous write */
+ nfs_initiate_write(data, ds->ds_clp->cl_rpcclient,
+ &filelayout_write_call_ops, sync);
+ return PNFS_ATTEMPTED;
+}
+
/*
* filelayout_check_layout()
*
@@ -475,6 +577,7 @@ static struct pnfs_layoutdriver_type filelayout_type = {
.free_lseg = filelayout_free_lseg,
.pg_test = filelayout_pg_test,
.read_pagelist = filelayout_read_pagelist,
+ .write_pagelist = filelayout_write_pagelist,
};
static int __init nfs4filelayout_init(void)
diff --git a/fs/nfs/nfs4proc.c b/fs/nfs/nfs4proc.c
index 3952f7c..6cce678 100644
--- a/fs/nfs/nfs4proc.c
+++ b/fs/nfs/nfs4proc.c
@@ -3144,6 +3144,23 @@ static int nfs4_write_done(struct rpc_task *task, struct nfs_write_data *data)
return data->write_done_cb(task, data);
}
+/* Reset the the nfs_write_data to send the write to the MDS. */
+void nfs4_reset_write(struct rpc_task *task, struct nfs_write_data *data)
+{
+ dprintk("%s Reset task for i/o through \n", __func__);
+ put_lseg(data->lseg);
+ data->lseg = NULL;
+ data->ds_clp = NULL;
+ data->write_done_cb = nfs4_write_done_cb;
+ data->args.fh = NFS_FH(data->inode);
+ data->args.bitmask = data->res.server->cache_consistency_bitmask;
+ data->args.offset = data->mds_offset;
+ data->res.fattr = &data->fattr;
+ task->tk_ops = data->mds_ops;
+ rpc_task_reset_client(task, NFS_CLIENT(data->inode));
+}
+EXPORT_SYMBOL_GPL(nfs4_reset_write);
+
static void nfs4_proc_write_setup(struct nfs_write_data *data, struct rpc_message *msg)
{
struct nfs_server *server = NFS_SERVER(data->inode);
diff --git a/fs/nfs/write.c b/fs/nfs/write.c
index 24cc34d..5757b06 100644
--- a/fs/nfs/write.c
+++ b/fs/nfs/write.c
@@ -783,7 +783,7 @@ static int flush_task_priority(int how)
return RPC_PRIORITY_NORMAL;
}
-static int nfs_initiate_write(struct nfs_write_data *data,
+int nfs_initiate_write(struct nfs_write_data *data,
struct rpc_clnt *clnt,
const struct rpc_call_ops *call_ops,
int how)
@@ -833,6 +833,7 @@ static int nfs_initiate_write(struct nfs_write_data *data,
out:
return ret;
}
+EXPORT_SYMBOL(nfs_initiate_write);
/*
* Set up the argument/result storage required for the RPC call.
@@ -1186,6 +1187,7 @@ int nfs_writeback_done(struct rpc_task *task, struct nfs_write_data *data)
*/
static unsigned long complain;
+ /* Note this will print the MDS for a DS write */
if (time_before(complain, jiffies)) {
dprintk("NFS: faulty NFS server %s:"
" (committed = %d) != (stable = %d)\n",
@@ -1206,6 +1208,7 @@ int nfs_writeback_done(struct rpc_task *task, struct nfs_write_data *data)
/* Was this an NFSv2 write or an NFSv3 stable write? */
if (resp->verf->committed != NFS_UNSTABLE) {
/* Resend from where the server left off */
+ data->mds_offset += resp->count;
argp->offset += resp->count;
argp->pgbase += resp->count;
argp->count -= resp->count;
diff --git a/include/linux/nfs_xdr.h b/include/linux/nfs_xdr.h
index 9ff26f9..a045d12 100644
--- a/include/linux/nfs_xdr.h
+++ b/include/linux/nfs_xdr.h
@@ -1045,6 +1045,7 @@ struct nfs_write_data {
#ifdef CONFIG_NFS_V4
unsigned long timestamp; /* For lease renewal */
#endif
+ __u64 mds_offset; /* Filelayout dense stripe */
struct page *page_array[NFS_PAGEVEC_SIZE];
};
--
1.7.2.1
^ permalink raw reply related [flat|nested] 15+ messages in thread
* [PATCH 7/7] wave4: nfs_pageio_do_add_request needs to distinguish between read and write
2011-02-11 16:54 [PATCH 0/7] pnfs wave 4 proposal Fred Isaman
` (5 preceding siblings ...)
2011-02-11 16:54 ` [PATCH 6/7] wave4: filelayout write Fred Isaman
@ 2011-02-11 16:54 ` Fred Isaman
2011-02-11 22:15 ` pNFS files server? Daniel.Muntz
2011-02-11 20:19 ` [PATCH 0/7] pnfs wave 4 proposal Andy Adamson
7 siblings, 1 reply; 15+ messages in thread
From: Fred Isaman @ 2011-02-11 16:54 UTC (permalink / raw)
To: linux-nfs
This is so that it can send the appropriate mode in LAYOUTGET.
Signed-off-by: Fred Isaman <iisaman@netapp.com>
---
fs/nfs/pagelist.c | 7 +++----
fs/nfs/pnfs.c | 8 ++++++++
fs/nfs/pnfs.h | 7 +++++++
fs/nfs/write.c | 2 +-
include/linux/nfs_page.h | 1 +
5 files changed, 20 insertions(+), 5 deletions(-)
diff --git a/fs/nfs/pagelist.c b/fs/nfs/pagelist.c
index f0aa20b..0518d62a 100644
--- a/fs/nfs/pagelist.c
+++ b/fs/nfs/pagelist.c
@@ -299,10 +299,9 @@ static int nfs_pageio_do_add_request(struct nfs_pageio_descriptor *desc,
} else {
put_lseg(desc->pg_lseg);
desc->pg_base = req->wb_pgbase;
- if (desc->pg_test)
- desc->pg_lseg = pnfs_update_layout(desc->pg_inode,
- req->wb_context,
- IOMODE_READ);
+ desc->pg_lseg = pnfs_update_layout(desc->pg_inode,
+ req->wb_context,
+ desc->pg_iomode);
}
nfs_list_remove_request(req);
nfs_list_add_request(req, &desc->pg_list);
diff --git a/fs/nfs/pnfs.c b/fs/nfs/pnfs.c
index 9d2351a..a2ee472 100644
--- a/fs/nfs/pnfs.c
+++ b/fs/nfs/pnfs.c
@@ -889,6 +889,14 @@ void
pnfs_pageio_init_read(struct nfs_pageio_descriptor *pgio,
struct inode *inode)
{
+ pgio->pg_iomode = IOMODE_READ;
+ pnfs_set_pg_test(inode, pgio);
+}
+
+void
+pnfs_pageio_init_write(struct nfs_pageio_descriptor *pgio, struct inode *inode)
+{
+ pgio->pg_iomode = IOMODE_RW;
pnfs_set_pg_test(inode, pgio);
}
diff --git a/fs/nfs/pnfs.h b/fs/nfs/pnfs.h
index 50bcb1f..16f9fa5 100644
--- a/fs/nfs/pnfs.h
+++ b/fs/nfs/pnfs.h
@@ -179,6 +179,7 @@ enum pnfs_try_status pnfs_try_to_write_data(struct nfs_write_data *,
enum pnfs_try_status pnfs_try_to_read_data(struct nfs_read_data *,
const struct rpc_call_ops *);
void pnfs_pageio_init_read(struct nfs_pageio_descriptor *, struct inode *);
+void pnfs_pageio_init_write(struct nfs_pageio_descriptor *, struct inode *);
int pnfs_layout_process(struct nfs4_layoutget *lgp);
void pnfs_free_lseg_list(struct list_head *tmp_list);
void pnfs_destroy_layout(struct nfs_inode *);
@@ -297,6 +298,12 @@ pnfs_pageio_init_read(struct nfs_pageio_descriptor *pgio, struct inode *ino)
pgio->pg_test = NULL;
}
+static inline void
+pnfs_pageio_init_write(struct nfs_pageio_descriptor *pgio, struct inode *ino)
+{
+ pgio->pg_test = NULL;
+}
+
#endif /* CONFIG_NFS_V4_1 */
#endif /* FS_NFS_PNFS_H */
diff --git a/fs/nfs/write.c b/fs/nfs/write.c
index 5757b06..4c21045 100644
--- a/fs/nfs/write.c
+++ b/fs/nfs/write.c
@@ -1001,7 +1001,7 @@ static void nfs_pageio_init_write(struct nfs_pageio_descriptor *pgio,
{
size_t wsize = NFS_SERVER(inode)->wsize;
- pgio->pg_test = NULL;
+ pnfs_pageio_init_write(pgio, inode);
if (wsize < PAGE_CACHE_SIZE)
nfs_pageio_init(pgio, inode, nfs_flush_multi, wsize, ioflags);
diff --git a/include/linux/nfs_page.h b/include/linux/nfs_page.h
index ba88ff4..fd57be2 100644
--- a/include/linux/nfs_page.h
+++ b/include/linux/nfs_page.h
@@ -63,6 +63,7 @@ struct nfs_pageio_descriptor {
int pg_ioflags;
int pg_error;
struct pnfs_layout_segment *pg_lseg;
+ int pg_iomode;
int (*pg_test)(struct nfs_pageio_descriptor *, struct nfs_page *, struct nfs_page *);
};
--
1.7.2.1
^ permalink raw reply related [flat|nested] 15+ messages in thread
* pNFS files server?
2011-02-11 16:54 ` [PATCH 7/7] wave4: nfs_pageio_do_add_request needs to distinguish between read and write Fred Isaman
@ 2011-02-11 22:15 ` Daniel.Muntz
2011-02-14 17:20 ` Andy Adamson
0 siblings, 1 reply; 15+ messages in thread
From: Daniel.Muntz @ 2011-02-11 22:15 UTC (permalink / raw)
To: linux-nfs
Is anyone still working on a freely-available pNFS files server for Linux? I've had a couple of inquiries recently about spNFS, and would like to be able to point people to recent work. Last I knew, Andy's GFS2-based pnfs files server work was the POR for delivering a server. Is that work still going on, and if so, is there usable code I can point to?
Thanks,
Dan
^ permalink raw reply [flat|nested] 15+ messages in thread
* Re: pNFS files server?
2011-02-11 22:15 ` pNFS files server? Daniel.Muntz
@ 2011-02-14 17:20 ` Andy Adamson
0 siblings, 0 replies; 15+ messages in thread
From: Andy Adamson @ 2011-02-14 17:20 UTC (permalink / raw)
To: Daniel.Muntz; +Cc: linux-nfs
The GFS2 pnfs files server code has not moved forward for a while as we have all spent our time on the pnfs files client. The pnfsd-files branch of benny's tree git://linux-nfs.org/~bhalevy/linux-pnfs.git has the latest code.
-->Andy
On Feb 11, 2011, at 5:15 PM, <Daniel.Muntz@emc.com> wrote:
> Is anyone still working on a freely-available pNFS files server for Linux? I've had a couple of inquiries recently about spNFS, and would like to be able to point people to recent work. Last I knew, Andy's GFS2-based pnfs files server work was the POR for delivering a server. Is that work still going on, and if so, is there usable code I can point to?
>
> Thanks,
> Dan--
> To unsubscribe from this list: send the line "unsubscribe linux-nfs" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at http://vger.kernel.org/majordomo-info.html
^ permalink raw reply [flat|nested] 15+ messages in thread
* Re: [PATCH 0/7] pnfs wave 4 proposal
2011-02-11 16:54 [PATCH 0/7] pnfs wave 4 proposal Fred Isaman
` (6 preceding siblings ...)
2011-02-11 16:54 ` [PATCH 7/7] wave4: nfs_pageio_do_add_request needs to distinguish between read and write Fred Isaman
@ 2011-02-11 20:19 ` Andy Adamson
2011-02-11 20:32 ` Fred Isaman
7 siblings, 1 reply; 15+ messages in thread
From: Andy Adamson @ 2011-02-11 20:19 UTC (permalink / raw)
To: Fred Isaman; +Cc: linux-nfs
I don't see pnfs_initiate_write
-->Andy
On Feb 11, 2011, at 11:54 AM, Fred Isaman wrote:
> These apply on top of Andy's wave3-on-nfs-for-next-rev4.
>
> These patches give write functionality to the pnfs file layout driver.
> The primary limitation is that we have not yet modified the commit
> code, so we can not yet handle COMMITs to the data server. We get
> around this for the moment by sending any WRITE to a data server which
> does not have commit-to-mds set as a SYNCH write.
>
> Fred
>
> --
> To unsubscribe from this list: send the line "unsubscribe linux-nfs" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at http://vger.kernel.org/majordomo-info.html
^ permalink raw reply [flat|nested] 15+ messages in thread