Re: [PATCH 13/24] pnfs_submit: read path changeover

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



On Wed, Jun 9, 2010 at 3:19 PM, Boaz Harrosh <bharrosh@xxxxxxxxxxx> wrote:
> On 06/08/2010 07:19 AM, Fred Isaman wrote:
>> Change readpages path to only call LAYOUTGET once.
>>
>> Signed-off-by: Fred Isaman <iisaman@xxxxxxxxxx>
>> ---
>>  fs/nfs/pagelist.c |    2 ++
>>  fs/nfs/pnfs.c     |   37 +++++++------------------------------
>>  fs/nfs/pnfs.h     |   25 ++++++++++++++++---------
>>  3 files changed, 25 insertions(+), 39 deletions(-)
>>
>> diff --git a/fs/nfs/pagelist.c b/fs/nfs/pagelist.c
>> index ed647b9..c3e5a1f 100644
>> --- a/fs/nfs/pagelist.c
>> +++ b/fs/nfs/pagelist.c
>> @@ -253,6 +253,8 @@ static int nfs_can_coalesce_requests(struct nfs_page *prev,
>>               return 0;
>>       if (prev->wb_pgbase + prev->wb_bytes != PAGE_CACHE_SIZE)
>>               return 0;
>> +     if (req->wb_lseg != prev->wb_lseg)
>> +             return 0;
>>  #ifdef CONFIG_NFS_V4_1
>>       if (pgio->pg_test && !pgio->pg_test(pgio, prev, req))
>>               return 0;
>> diff --git a/fs/nfs/pnfs.c b/fs/nfs/pnfs.c
>> index 2b5f6fc..692a18e 100644
>> --- a/fs/nfs/pnfs.c
>> +++ b/fs/nfs/pnfs.c
>> @@ -1689,7 +1689,7 @@ pnfs_readpages(struct nfs_read_data *rdata)
>>  {
>>       struct nfs_readargs *args = &rdata->args;
>>       struct inode *inode = rdata->inode;
>> -     int numpages, status, pgcount, temp;
>> +     int numpages, pgcount, temp;
>>       struct nfs_server *nfss = NFS_SERVER(inode);
>>       struct nfs_inode *nfsi = NFS_I(inode);
>>       struct pnfs_layout_segment *lseg;
>> @@ -1701,19 +1701,8 @@ pnfs_readpages(struct nfs_read_data *rdata)
>>               args->count,
>>               args->offset);
>>
>> -     /* Retrieve and set layout if not allready cached */
>> -     status = _pnfs_update_layout(inode,
>> -                                 args->context,
>> -                                 args->count,
>> -                                 args->offset,
>> -                                 IOMODE_READ,
>> -                                 &lseg);
>> -     if (status) {
>> -             dprintk("%s: Updating layout failed (%d), retry with NFS \n",
>> -                     __func__, status);
>> -             trypnfs = PNFS_NOT_ATTEMPTED;
>> -             goto out;
>> -     }
>> +     lseg = rdata->req->wb_lseg;
>> +     get_lseg(lseg);
>>
>>       /* Determine number of pages. */
>>       pgcount = args->pgbase + args->count;
>> @@ -1740,7 +1729,6 @@ pnfs_readpages(struct nfs_read_data *rdata)
>>               rdata->pdata.lseg = NULL;
>>               put_lseg(lseg);
>>       }
>> - out:
>>       dprintk("%s End (trypnfs:%d)\n", __func__, trypnfs);
>>       return trypnfs;
>>  }
>> @@ -1749,21 +1737,10 @@ enum pnfs_try_status
>>  _pnfs_try_to_read_data(struct nfs_read_data *data,
>>                      const struct rpc_call_ops *call_ops)
>>  {
>> -     struct inode *ino = data->inode;
>> -     struct nfs_server *nfss = NFS_SERVER(ino);
>> -
>> -     dprintk("--> %s\n", __func__);
>> -     /* Only create an rpc request if utilizing NFSv4 I/O */
>> -     if (!pnfs_enabled_sb(nfss) ||
>> -         !nfss->pnfs_curr_ld->ld_io_ops->read_pagelist) {
>> -             dprintk("<-- %s: not using pnfs\n", __func__);
>> -             return PNFS_NOT_ATTEMPTED;
>> -     } else {
>> -             dprintk("%s: Utilizing pNFS I/O\n", __func__);
>> -             data->pdata.call_ops = call_ops;
>> -             data->pdata.pnfs_error = 0;
>> -             return pnfs_readpages(data);
>> -     }
>
> Wahoo, nice stuff Ha
>
> By now this can go into it's only caller.
> (And caller de-inlined, what was that all about)
>
>> +     dprintk("%s: Utilizing pNFS I/O\n", __func__);
>> +     data->pdata.call_ops = call_ops;
>> +     data->pdata.pnfs_error = 0;
>> +     return pnfs_readpages(data);
>>  }
>>
>>  enum pnfs_try_status
>> diff --git a/fs/nfs/pnfs.h b/fs/nfs/pnfs.h
>> index 6326ed5..816ebe1 100644
>> --- a/fs/nfs/pnfs.h
>> +++ b/fs/nfs/pnfs.h
>> @@ -94,22 +94,29 @@ static inline int pnfs_enabled_sb(struct nfs_server *nfss)
>>       return nfss->pnfs_curr_ld != NULL;
>>  }
>>
>> +static inline void _pnfs_clear_lseg_from_pages(struct list_head *head)
>> +{
>> +     struct nfs_page *req;
>> +
>> +     list_for_each_entry(req, head, wb_list) {
>> +             put_lseg(req->wb_lseg);
>> +             req->wb_lseg = NULL;
>> +     }
>> +}
>> +
>>  static inline enum pnfs_try_status
>>  pnfs_try_to_read_data(struct nfs_read_data *data,
>>                     const struct rpc_call_ops *call_ops)
>
> Don't think this needs to be inline, whats the point?
>

The point is that it is in the header file, not a c file.

Fred

>>  {
>> -     struct inode *inode = data->inode;
>> -     struct nfs_server *nfss = NFS_SERVER(inode);
>>       enum pnfs_try_status ret;
>>
>> -     /* FIXME: read_pagelist should probably be mandated */
>> -     if (PNFS_EXISTS_LDIO_OP(nfss, read_pagelist))
>> -             ret = _pnfs_try_to_read_data(data, call_ops);
>> -     else
>> -             ret = PNFS_NOT_ATTEMPTED;
>> -
>> +     if (!data->req->wb_lseg)
>> +             return PNFS_NOT_ATTEMPTED;
>> +     ret = _pnfs_try_to_read_data(data, call_ops);
>>       if (ret == PNFS_ATTEMPTED)
>> -             nfs_inc_stats(inode, NFSIOS_PNFS_READ);
>> +             nfs_inc_stats(data->inode, NFSIOS_PNFS_READ);
>> +     else
>> +             _pnfs_clear_lseg_from_pages(&data->pages);
>>       return ret;
>>  }
>>
>
> Boaz
> --
> To unsubscribe from this list: send the line "unsubscribe linux-nfs" in
> the body of a message to majordomo@xxxxxxxxxxxxxxx
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
>
--
To unsubscribe from this list: send the line "unsubscribe linux-nfs" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html


[Index of Archives]     [Linux Filesystem Development]     [Linux USB Development]     [Linux Media Development]     [Video for Linux]     [Linux NILFS]     [Linux Audio Users]     [Yosemite Info]     [Linux SCSI]

  Powered by Linux