Boaz Harrosh wrote:
On 07/27/2011 01:58 PM, Jim Rees wrote:
> But I thought for simplicity since this patchset was reviewed then, I've
> tested with this patch and it works good as well. If you are going to
> break things up, I can test with the minimal patches and report what is
> need in Stable. (Is anyone working on it?)
>
> We may have over-squashed that one for block layout. I can dig up the
> original patches if necessary, but I'm guessing you probably have them.
Please do. Because by now I'm confused. the first I discovered it was
by git diff of a branch that worked. I lost tracked of where the code
was originated from.
Here is just the original multiple segments patch. This is from the
88-patch set I sent out before Bakeathon. I don't know if this would still
apply separately.
>From fc3381fa66adfcee4203b5a55ca6605bcfc8dafb Mon Sep 17 00:00:00 2001
From: Peng Tao <bergwolf@xxxxxxxxx>
Date: Tue, 31 May 2011 00:42:56 -0400
Subject: [PATCH 03/89] pnfs: let layoutcommit code handle multiple segments
Some layout driver like block will have multiple segments.
Generic code should be able to handle it.
---
fs/nfs/pnfs.c | 15 ++++++++++++---
fs/nfs/pnfs.h | 1 +
2 files changed, 13 insertions(+), 3 deletions(-)
diff --git a/fs/nfs/pnfs.c b/fs/nfs/pnfs.c
index e3d618b..c18cda0 100644
--- a/fs/nfs/pnfs.c
+++ b/fs/nfs/pnfs.c
@@ -1193,10 +1193,18 @@ pnfs_try_to_read_data(struct nfs_read_data *rdata,
static struct pnfs_layout_segment *pnfs_list_write_lseg(struct inode *inode)
{
struct pnfs_layout_segment *lseg, *rv = NULL;
+ loff_t max_pos = 0;
+
+ list_for_each_entry(lseg, &NFS_I(inode)->layout->plh_segs, pls_list) {
+ if (lseg->pls_range.iomode == IOMODE_RW) {
+ if (max_pos < lseg->pls_end_pos)
+ max_pos = lseg->pls_end_pos;
+ if (test_and_clear_bit(NFS_LSEG_LAYOUTCOMMIT, &lseg->pls_flags))
+ rv = lseg;
+ }
+ }
+ rv->pls_end_pos = max_pos;
- list_for_each_entry(lseg, &NFS_I(inode)->layout->plh_segs, pls_list)
- if (lseg->pls_range.iomode == IOMODE_RW)
- rv = lseg;
return rv;
}
@@ -1211,6 +1219,7 @@ pnfs_set_layoutcommit(struct nfs_write_data *wdata)
if (!test_and_set_bit(NFS_INO_LAYOUTCOMMIT, &nfsi->flags)) {
/* references matched in nfs4_layoutcommit_release */
get_lseg(wdata->lseg);
+ set_bit(NFS_LSEG_LAYOUTCOMMIT, &wdata->lseg->pls_flags);
wdata->lseg->pls_lc_cred =
get_rpccred(wdata->args.context->state->owner->so_cred);
mark_as_dirty = true;
diff --git a/fs/nfs/pnfs.h b/fs/nfs/pnfs.h
index 53819d0..c6a9af2 100644
--- a/fs/nfs/pnfs.h
+++ b/fs/nfs/pnfs.h
@@ -36,6 +36,7 @@
enum {
NFS_LSEG_VALID = 0, /* cleared when lseg is recalled/returned */
NFS_LSEG_ROC, /* roc bit received from server */
+ NFS_LSEG_LAYOUTCOMMIT, /* layoutcommit bit set for layoutcommit */
};
struct pnfs_layout_segment {
--
1.7.4.2
--
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
