linux-kernel.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Greg KH <gregkh@suse.de>
To: linux-kernel@vger.kernel.org, stable@kernel.org
Cc: stable-review@kernel.org, torvalds@linux-foundation.org,
	akpm@linux-foundation.org, alan@lxorguk.ukuu.org.uk,
	Greg Kroah-Hartman <gregkh@suse.de>
Subject: [03/74] mm: add new read_cache_page_gfp() helper function
Date: Thu, 04 Feb 2010 09:11:34 -0800	[thread overview]
Message-ID: <20100204171448.330762047@linux.site> (raw)
In-Reply-To: <20100204171850.GA16539@kroah.com>

2.6.32-stable review patch.  If anyone has any objections, please let us know.

------------------

From: Linus Torvalds <torvalds@linux-foundation.org>

commit 0531b2aac59c2296570ac52bfc032ef2ace7d5e1 upstream.

It's a simplified 'read_cache_page()' which takes a page allocation
flag, so that different paths can control how aggressive the memory
allocations are that populate a address space.

In particular, the intel GPU object mapping code wants to be able to do
a certain amount of own internal memory management by automatically
shrinking the address space when memory starts getting tight.  This
allows it to dynamically use different memory allocation policies on a
per-allocation basis, rather than depend on the (static) address space
gfp policy.

The actual new function is a one-liner, but re-organizing the helper
functions to the point where you can do this with a single line of code
is what most of the patch is all about.

Tested-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

---
 include/linux/pagemap.h |    2 
 mm/filemap.c            |  100 ++++++++++++++++++++++++++++++++----------------
 2 files changed, 70 insertions(+), 32 deletions(-)

--- a/include/linux/pagemap.h
+++ b/include/linux/pagemap.h
@@ -253,6 +253,8 @@ extern struct page * read_cache_page_asy
 extern struct page * read_cache_page(struct address_space *mapping,
 				pgoff_t index, filler_t *filler,
 				void *data);
+extern struct page * read_cache_page_gfp(struct address_space *mapping,
+				pgoff_t index, gfp_t gfp_mask);
 extern int read_cache_pages(struct address_space *mapping,
 		struct list_head *pages, filler_t *filler, void *data);
 
--- a/mm/filemap.c
+++ b/mm/filemap.c
@@ -1655,14 +1655,15 @@ EXPORT_SYMBOL(generic_file_readonly_mmap
 static struct page *__read_cache_page(struct address_space *mapping,
 				pgoff_t index,
 				int (*filler)(void *,struct page*),
-				void *data)
+				void *data,
+				gfp_t gfp)
 {
 	struct page *page;
 	int err;
 repeat:
 	page = find_get_page(mapping, index);
 	if (!page) {
-		page = page_cache_alloc_cold(mapping);
+		page = __page_cache_alloc(gfp | __GFP_COLD);
 		if (!page)
 			return ERR_PTR(-ENOMEM);
 		err = add_to_page_cache_lru(page, mapping, index, GFP_KERNEL);
@@ -1682,31 +1683,18 @@ repeat:
 	return page;
 }
 
-/**
- * read_cache_page_async - read into page cache, fill it if needed
- * @mapping:	the page's address_space
- * @index:	the page index
- * @filler:	function to perform the read
- * @data:	destination for read data
- *
- * Same as read_cache_page, but don't wait for page to become unlocked
- * after submitting it to the filler.
- *
- * Read into the page cache. If a page already exists, and PageUptodate() is
- * not set, try to fill the page but don't wait for it to become unlocked.
- *
- * If the page does not get brought uptodate, return -EIO.
- */
-struct page *read_cache_page_async(struct address_space *mapping,
+static struct page *do_read_cache_page(struct address_space *mapping,
 				pgoff_t index,
 				int (*filler)(void *,struct page*),
-				void *data)
+				void *data,
+				gfp_t gfp)
+
 {
 	struct page *page;
 	int err;
 
 retry:
-	page = __read_cache_page(mapping, index, filler, data);
+	page = __read_cache_page(mapping, index, filler, data, gfp);
 	if (IS_ERR(page))
 		return page;
 	if (PageUptodate(page))
@@ -1731,8 +1719,67 @@ out:
 	mark_page_accessed(page);
 	return page;
 }
+
+/**
+ * read_cache_page_async - read into page cache, fill it if needed
+ * @mapping:	the page's address_space
+ * @index:	the page index
+ * @filler:	function to perform the read
+ * @data:	destination for read data
+ *
+ * Same as read_cache_page, but don't wait for page to become unlocked
+ * after submitting it to the filler.
+ *
+ * Read into the page cache. If a page already exists, and PageUptodate() is
+ * not set, try to fill the page but don't wait for it to become unlocked.
+ *
+ * If the page does not get brought uptodate, return -EIO.
+ */
+struct page *read_cache_page_async(struct address_space *mapping,
+				pgoff_t index,
+				int (*filler)(void *,struct page*),
+				void *data)
+{
+	return do_read_cache_page(mapping, index, filler, data, mapping_gfp_mask(mapping));
+}
 EXPORT_SYMBOL(read_cache_page_async);
 
+static struct page *wait_on_page_read(struct page *page)
+{
+	if (!IS_ERR(page)) {
+		wait_on_page_locked(page);
+		if (!PageUptodate(page)) {
+			page_cache_release(page);
+			page = ERR_PTR(-EIO);
+		}
+	}
+	return page;
+}
+
+/**
+ * read_cache_page_gfp - read into page cache, using specified page allocation flags.
+ * @mapping:	the page's address_space
+ * @index:	the page index
+ * @gfp:	the page allocator flags to use if allocating
+ *
+ * This is the same as "read_mapping_page(mapping, index, NULL)", but with
+ * any new page allocations done using the specified allocation flags. Note
+ * that the Radix tree operations will still use GFP_KERNEL, so you can't
+ * expect to do this atomically or anything like that - but you can pass in
+ * other page requirements.
+ *
+ * If the page does not get brought uptodate, return -EIO.
+ */
+struct page *read_cache_page_gfp(struct address_space *mapping,
+				pgoff_t index,
+				gfp_t gfp)
+{
+	filler_t *filler = (filler_t *)mapping->a_ops->readpage;
+
+	return wait_on_page_read(do_read_cache_page(mapping, index, filler, NULL, gfp));
+}
+EXPORT_SYMBOL(read_cache_page_gfp);
+
 /**
  * read_cache_page - read into page cache, fill it if needed
  * @mapping:	the page's address_space
@@ -1750,18 +1797,7 @@ struct page *read_cache_page(struct addr
 				int (*filler)(void *,struct page*),
 				void *data)
 {
-	struct page *page;
-
-	page = read_cache_page_async(mapping, index, filler, data);
-	if (IS_ERR(page))
-		goto out;
-	wait_on_page_locked(page);
-	if (!PageUptodate(page)) {
-		page_cache_release(page);
-		page = ERR_PTR(-EIO);
-	}
- out:
-	return page;
+	return wait_on_page_read(read_cache_page_async(mapping, index, filler, data));
 }
 EXPORT_SYMBOL(read_cache_page);
 



  parent reply	other threads:[~2010-02-04 17:38 UTC|newest]

Thread overview: 82+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2010-02-04 17:18 [00/74] 2.6.32.8-stable review Greg KH
2010-02-04 17:11 ` [01/74] [SCSI] scsi_lib: Fix bug in completion of bidi commands Greg KH
2010-02-04 17:11 ` [02/74] [SCSI] mptsas: Fix issue with chain pools allocation on katmai Greg KH
2010-02-04 17:11 ` Greg KH [this message]
2010-02-04 17:11 ` [04/74] drm/i915: Selectively enable self-reclaim Greg KH
2010-02-04 17:11 ` [05/74] firewire: ohci: fix crashes with TSB43AB23 on 64bit systems Greg KH
2010-02-04 17:11 ` [06/74] S390: fix single stepped svcs with TRACE_IRQFLAGS=y Greg KH
2010-02-04 17:11 ` [07/74] x86: Set hotpluggable nodes in nodes_possible_map Greg KH
2010-02-04 17:11 ` [08/74] x86: Remove "x86 CPU features in debugfs" (CONFIG_X86_CPU_DEBUG) Greg KH
2010-02-04 17:11 ` [09/74] libata: retry FS IOs even if it has failed with AC_ERR_INVALID Greg KH
2010-02-04 17:11 ` [10/74] [S390] zcrypt: Do not remove coprocessor for error 8/72 Greg KH
2010-02-04 17:11 ` [11/74] [S390] dasd: fix possible NULL pointer errors Greg KH
2010-02-11 23:15   ` Bastian Blank
2010-02-11 23:38     ` Greg KH
2010-02-04 17:11 ` [12/74] ACPI: Add a generic API for _OSC -v2 Greg KH
2010-02-04 17:11 ` [13/74] ACPI: Add platform-wide _OSC support Greg KH
2010-02-04 17:11 ` [14/74] ACPI: fix OSC regression that caused aer and pciehp not to load Greg KH
2010-02-04 17:11 ` [15/74] ACPI: Advertise to BIOS in _OSC: _OST on _PPC changes Greg KH
2010-02-04 17:11 ` [16/74] UBI: fix volume creation input checking Greg KH
2010-02-04 17:11 ` [17/74] e1000: enhance frame fragment detection Greg KH
2010-02-04 17:11 ` [18/74] e1000e: " Greg KH
2010-02-04 17:11 ` [19/74] e1000/e1000e: dont use small hardware rx buffers Greg KH
2010-02-04 17:11 ` [20/74] drm/i915: Reload hangcheck timer too for Ironlake Greg KH
2010-02-04 17:11 ` [21/74] Fix a leak in affs_fill_super() Greg KH
2010-02-04 17:11 ` [22/74] Fix failure exits in bfs_fill_super() Greg KH
2010-02-04 17:11 ` [23/74] fix oops in fs/9p late mount failure Greg KH
2010-02-04 17:11 ` [24/74] fix leak in romfs_fill_super() Greg KH
2010-02-04 17:11 ` [25/74] Fix remount races with symlink handling in affs Greg KH
2010-02-04 17:11 ` [26/74] fix affs parse_options() Greg KH
2010-02-04 17:11 ` [27/74] Fix failure exit in ipathfs Greg KH
2010-02-04 17:11 ` [28/74] mm: fix migratetype bug which slowed swapping Greg KH
2010-02-04 17:12 ` [29/74] FDPIC: Respect PT_GNU_STACK exec protection markings when creating NOMMU stack Greg KH
2010-02-04 17:12 ` [30/74] Split flush_old_exec into two functions Greg KH
2010-02-04 17:12 ` [31/74] sparc: TIF_ABI_PENDING bit removal Greg KH
2010-02-04 17:12 ` [32/74] x86: get rid of the insane TIF_ABI_PENDING bit Greg KH
2010-02-04 17:12 ` [33/74] Input: winbond-cir - remove dmesg spam Greg KH
2010-02-04 17:12 ` [34/74] x86: Disable HPET MSI on ATI SB700/SB800 Greg KH
2010-02-04 17:12 ` [35/74] iwlwifi: set default aggregation frame count limit to 31 Greg KH
2010-02-04 17:12 ` [36/74] drm/i915: only enable hotplug for detected outputs Greg KH
2010-02-04 17:12 ` [37/74] firewire: core: add_descriptor size check Greg KH
2010-02-04 17:12 ` [38/74] SECURITY: selinux, fix update_rlimit_cpu parameter Greg KH
2010-02-04 17:12 ` [39/74] regulator: Specify REGULATOR_CHANGE_STATUS for WM835x LED constraints Greg KH
2010-02-04 17:12 ` [40/74] x86: Add Dell OptiPlex 760 reboot quirk Greg KH
2010-02-04 17:12 ` [41/74] x86: Add quirk for Intel DG45FC board to avoid low memory corruption Greg KH
2010-02-04 17:12 ` [42/74] x86/amd-iommu: Fix possible integer overflow Greg KH
2010-02-04 17:12 ` [43/74] clocksource: fix compilation if no GENERIC_TIME Greg KH
2010-02-04 17:12 ` [44/74] tcp: update the netstamp_needed counter when cloning sockets Greg KH
2010-02-04 17:12 ` [45/74] sky2: Fix oops in sky2_xmit_frame() after TX timeout Greg KH
2010-02-04 17:12 ` [46/74] net: restore ip source validation Greg KH
2010-02-05 10:16   ` Sven Joachim
2010-02-04 17:12 ` [47/74] af_packet: Dont use skb after dev_queue_xmit() Greg KH
2010-02-04 17:12 ` [48/74] ax25: netrom: rose: Fix timer oopses Greg KH
2010-02-04 17:12 ` [49/74] KVM: allow userspace to adjust kvmclock offset Greg KH
2010-02-04 17:12 ` [50/74] oprofile/x86: add Xeon 7500 series support Greg KH
2010-02-04 17:12 ` [51/74] oprofile/x86: fix crash when profiling more than 28 events Greg KH
2010-02-04 17:12 ` [52/74] libata: retry link resume if necessary Greg KH
2010-02-04 17:12 ` [53/74] mm: percpu-vmap fix RCU list walking Greg KH
2010-02-04 17:12 ` [54/74] mm: purge fragmented percpu vmap blocks Greg KH
2010-02-04 17:12 ` [55/74] block: fix bio_add_page for non trivial merge_bvec_fn case Greg KH
2010-02-04 17:12 ` [56/74] Fix flush_old_exec()/setup_new_exec() split Greg KH
2010-02-04 17:12 ` [57/74] random: drop weird m_time/a_time manipulation Greg KH
2010-02-04 17:12 ` [58/74] random: Remove unused inode variable Greg KH
2010-02-04 17:12 ` [59/74] block: fix bugs in bio-integrity mempool usage Greg KH
2010-02-04 17:12 ` [60/74] usb: r8a66597-hdc disable interrupts fix Greg KH
2010-02-04 17:12 ` [61/74] connector: Delete buggy notification code Greg KH
2010-02-04 17:12 ` [62/74] be2net: Bug fix to support newer generation of BE ASIC Greg KH
2010-02-04 17:12 ` [63/74] be2net: Fix memset() arg ordering Greg KH
2010-02-04 17:12 ` [64/74] mm: flush dcache before writing into page to avoid alias Greg KH
2010-02-04 17:12 ` [65/74] mac80211: fix NULL pointer dereference when ftrace is enabled Greg KH
2010-02-04 17:12 ` [66/74] imxfb: correct location of callbacks in suspend and resume Greg KH
2010-02-04 17:12 ` [67/74] mx3fb: some debug and initialisation fixes Greg KH
2010-02-04 17:12 ` [68/74] starfire: clean up properly if firmware loading fails Greg KH
2010-02-04 17:12 ` [69/74] kernel/cred.c: use kmem_cache_free Greg KH
2010-02-04 17:12 ` [70/74] uartlite: fix crash when using as console Greg KH
2010-02-04 17:12 ` [71/74] pktcdvd: removing device does not remove its sysfs dir Greg KH
2010-02-04 17:12 ` [72/74] ath9k: fix eeprom INI values override for 2GHz-only cards Greg KH
2010-02-04 17:12 ` [73/74] ath9k: fix beacon slot/buffer leak Greg KH
2010-02-04 17:12 ` [74/74] powerpc: TIF_ABI_PENDING bit removal Greg KH
2010-02-05  7:36 ` [Stable-review] [00/74] 2.6.32.8-stable review Nikola Ciprich
2010-02-05 17:12   ` Greg KH
2010-02-07 10:26     ` Nikola Ciprich
2010-02-05 16:53 ` Greg KH

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=20100204171448.330762047@linux.site \
    --to=gregkh@suse.de \
    --cc=akpm@linux-foundation.org \
    --cc=alan@lxorguk.ukuu.org.uk \
    --cc=linux-kernel@vger.kernel.org \
    --cc=stable-review@kernel.org \
    --cc=stable@kernel.org \
    --cc=torvalds@linux-foundation.org \
    /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;
as well as URLs for NNTP newsgroup(s).