mirror of https://lore.kernel.org/lkml/
 help / color / mirror / Atom feed
From: Octavian Purdila <octavian.purdila@intel.com>
To: xfs@oss.sgi.com
Cc: linux-fsdevel@vger.kernel.org, linux-kernel@vger.kernel.org,
	Octavian Purdila <octavian.purdila@intel.com>
Subject: [RFC PATCH] xfs: support for non-mmu architectures
Date: Wed, 18 Nov 2015 00:46:21 +0200	[thread overview]
Message-ID: <1447800381-20167-1-git-send-email-octavian.purdila@intel.com> (raw)

Naive implementation for non-mmu architectures: allocate physically
contiguous xfs buffers with alloc_pages. Terribly inefficient with
memory and fragmentation on high I/O loads but it may be good enough
for basic usage (which most non-mmu architectures will need).

This patch was tested with lklfuse [1] and basic operations seems to
work even with 16MB allocated for LKL.

[1] https://github.com/lkl/linux

Signed-off-by: Octavian Purdila <octavian.purdila@intel.com>
---
 fs/xfs/xfs_buf.c | 27 +++++++++++++++++++++++++++
 1 file changed, 27 insertions(+)

diff --git a/fs/xfs/xfs_buf.c b/fs/xfs/xfs_buf.c
index 8ecffb3..50b5246 100644
--- a/fs/xfs/xfs_buf.c
+++ b/fs/xfs/xfs_buf.c
@@ -261,6 +261,7 @@ xfs_buf_free(
 	ASSERT(list_empty(&bp->b_lru));
 
 	if (bp->b_flags & _XBF_PAGES) {
+#ifdef CONFIG_MMU
 		uint		i;
 
 		if (xfs_buf_is_vmapped(bp))
@@ -272,6 +273,10 @@ xfs_buf_free(
 
 			__free_page(page);
 		}
+#else
+		free_pages((unsigned long)page_to_virt(bp->b_pages[0]),
+			   order_base_2(bp->b_page_count));
+#endif
 	} else if (bp->b_flags & _XBF_KMEM)
 		kmem_free(bp->b_addr);
 	_xfs_buf_free_pages(bp);
@@ -338,7 +343,14 @@ use_alloc_page:
 		struct page	*page;
 		uint		retries = 0;
 retry:
+#ifdef CONFIG_MMU
 		page = alloc_page(gfp_mask);
+#else
+		if (i == 0)
+			page = alloc_pages(gfp_mask, order_base_2(page_count));
+		else
+			page = bp->b_pages[0] + i;
+#endif
 		if (unlikely(page == NULL)) {
 			if (flags & XBF_READ_AHEAD) {
 				bp->b_page_count = i;
@@ -372,8 +384,11 @@ retry:
 	return 0;
 
 out_free_pages:
+#ifdef CONFIG_MMU
 	for (i = 0; i < bp->b_page_count; i++)
 		__free_page(bp->b_pages[i]);
+#endif
+
 	return error;
 }
 
@@ -392,6 +407,7 @@ _xfs_buf_map_pages(
 	} else if (flags & XBF_UNMAPPED) {
 		bp->b_addr = NULL;
 	} else {
+#ifdef CONFIG_MMU
 		int retried = 0;
 		unsigned noio_flag;
 
@@ -412,6 +428,9 @@ _xfs_buf_map_pages(
 			vm_unmap_aliases();
 		} while (retried++ <= 1);
 		memalloc_noio_restore(noio_flag);
+#else
+		bp->b_addr = page_to_virt(bp->b_pages[0]);
+#endif
 
 		if (!bp->b_addr)
 			return -ENOMEM;
@@ -816,11 +835,19 @@ xfs_buf_get_uncached(
 	if (error)
 		goto fail_free_buf;
 
+#ifdef CONFIG_MMU
 	for (i = 0; i < page_count; i++) {
 		bp->b_pages[i] = alloc_page(xb_to_gfp(flags));
 		if (!bp->b_pages[i])
 			goto fail_free_mem;
 	}
+#else
+	bp->b_pages[0] = alloc_pages(flags, order_base_2(page_count));
+	if (!bp->b_pages[0])
+		goto fail_free_buf;
+	for (i = 1; i < page_count; i++)
+		bp->b_pages[i] = bp->b_pages[i-1] + 1;
+#endif
 	bp->b_flags |= _XBF_PAGES;
 
 	error = _xfs_buf_map_pages(bp, 0);
-- 
1.9.1


             reply	other threads:[~2015-11-17 22:46 UTC|newest]

Thread overview: 28+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2015-11-17 22:46 Octavian Purdila [this message]
2015-11-19 15:55 ` Brian Foster
2015-11-19 20:54   ` Octavian Purdila
2015-11-20 15:11     ` Brian Foster
2015-11-19 23:35   ` Dave Chinner
2015-11-20 14:09     ` Octavian Purdila
2015-11-20 15:11     ` Brian Foster
2015-11-20 15:35       ` Octavian Purdila
2015-11-20 15:40         ` Brian Foster
2015-11-20 20:36       ` Dave Chinner
2015-11-20 22:47         ` Brian Foster
2015-11-22 22:04           ` Dave Chinner
2015-11-23 12:50             ` Brian Foster
2015-11-23 21:00               ` Dave Chinner
2015-11-19 23:24 ` Dave Chinner
2015-11-19 23:54   ` Richard Weinberger
2015-11-20  0:58     ` Dave Chinner
2015-11-20 14:26       ` Octavian Purdila
2015-11-20 15:24         ` Brian Foster
2015-11-20 15:31           ` Octavian Purdila
2015-11-20 15:43             ` Brian Foster
2015-11-20 20:07         ` Theodore Ts'o
2015-11-20 13:43   ` Octavian Purdila
2015-11-20 21:08     ` Dave Chinner
2015-11-20 22:26       ` Octavian Purdila
2015-11-22 22:44         ` Dave Chinner
2015-11-23  1:41           ` Octavian Purdila
2015-11-23 21:46             ` Dave Chinner

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=1447800381-20167-1-git-send-email-octavian.purdila@intel.com \
    --to=octavian.purdila@intel.com \
    --cc=linux-fsdevel@vger.kernel.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=xfs@oss.sgi.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox

all inboxes | Powered by JetHome®