[net-next PATCH v2 0/5] Replace __skb_alloc_pages with simpler function

[net-next PATCH v2 0/5] Replace __skb_alloc_pages with simpler function

[Alexander Duyck]

This patch series replaces __skb_alloc_pages with a much simpler function,
__dev_alloc_pages.  The main difference between the two is that
__skb_alloc_pages had an sk_buff pointer that was being passed as NULL in
call places where it was called.  In a couple of cases the NULL was passed
by variable and this led to unnecessary code being run.

As such in order to simplify things the __dev_alloc_pages call only takes a
mask and the page order being requested.  In addition it takes advantage of
several behaviors already built into the page allocator so that it can just
set GFP flags unconditionally.

v2: Renamed functions to dev_alloc_page(s) instead of netdev_alloc_page(s)
    Removed __GFP_COLD flag from usb code as it was redundant

---

Alexander Duyck (5):
      net: Add device Rx page allocation function
      cxgb4/cxgb4vf: Replace __skb_alloc_page with __netdev_alloc_page
      phonet: Replace calls to __skb_alloc_page with __netdev_alloc_page
      fm10k/igb/ixgbe: Replace __skb_alloc_page with netdev_alloc_page
      net: Remove __skb_alloc_page and __skb_alloc_pages


 drivers/net/ethernet/chelsio/cxgb4/sge.c      |    6 +-
 drivers/net/ethernet/chelsio/cxgb4vf/sge.c    |    7 ++-
 drivers/net/ethernet/intel/fm10k/fm10k_main.c |    2 -
 drivers/net/ethernet/intel/igb/igb_main.c     |    2 -
 drivers/net/ethernet/intel/ixgbe/ixgbe_main.c |    3 -
 drivers/net/usb/cdc-phonet.c                  |    6 +-
 drivers/usb/gadget/function/f_phonet.c        |    6 +-
 include/linux/skbuff.h                        |   61 ++++++++++++++-----------
 8 files changed, 49 insertions(+), 44 deletions(-)

--
--
To unsubscribe from this list: send the line "unsubscribe linux-usb" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

[net-next PATCH v2 1/5] net: Add device Rx page allocation function

[Alexander Duyck]

This patch implements __dev_alloc_pages and __dev_alloc_page.  These are
meant to replace the __skb_alloc_pages and __skb_alloc_page functions.  The
reason for doing this is that it occurred to me that __skb_alloc_page is
supposed to be passed an sk_buff pointer, but it is NULL in all cases where
it is used.  Worse is that in the case of ixgbe it is passed NULL via the
sk_buff pointer in the rx_buffer info structure which means the compiler is
not correctly stripping it out.

The naming for these functions is based on dev_alloc_skb and __dev_alloc_skb.
There was originally a netdev_alloc_page, however that was passed a
net_device pointer and this function is not so I thought it best to follow
that naming scheme since that is the same difference between dev_alloc_skb
and netdev_alloc_skb.

In the case of anything greater than order 0 it is assumed that we want a
compound page so __GFP_COMP is set for all allocations as we expect a
compound page when assigning a page frag.

The other change in this patch is to exploit the behaviors of the page
allocator in how it handles flags.  So for example we can always set
__GFP_COMP and __GFP_MEMALLOC since they are ignored if they are not
applicable or are overridden by another flag.

Signed-off-by: Alexander Duyck <alexander.h.duyck@redhat.com>
---
 include/linux/skbuff.h |   48 ++++++++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 48 insertions(+)

diff --git a/include/linux/skbuff.h b/include/linux/skbuff.h
index 103fbe8..2e5221f 100644
--- a/include/linux/skbuff.h
+++ b/include/linux/skbuff.h
@@ -2185,6 +2185,54 @@ static inline struct sk_buff *netdev_alloc_skb_ip_align(struct net_device *dev,
 }
 
 /**
+ * __dev_alloc_pages - allocate page for network Rx
+ * @gfp_mask: allocation priority. Set __GFP_NOMEMALLOC if not for network Rx
+ * @order: size of the allocation
+ *
+ * Allocate a new page.
+ *
+ * %NULL is returned if there is no free memory.
+*/
+static inline struct page *__dev_alloc_pages(gfp_t gfp_mask,
+					     unsigned int order)
+{
+	/* This piece of code contains several assumptions.
+	 * 1.  This is for device Rx, therefor a cold page is preferred.
+	 * 2.  The expectation is the user wants a compound page.
+	 * 3.  If requesting a order 0 page it will not be compound
+	 *     due to the check to see if order has a value in prep_new_page
+	 * 4.  __GFP_MEMALLOC is ignored if __GFP_NOMEMALLOC is set due to
+	 *     code in gfp_to_alloc_flags that should be enforcing this.
+	 */
+	gfp_mask |= __GFP_COLD | __GFP_COMP | __GFP_MEMALLOC;
+
+	return alloc_pages_node(NUMA_NO_NODE, gfp_mask, order);
+}
+
+static inline struct page *dev_alloc_pages(unsigned int order)
+{
+	return __dev_alloc_pages(GFP_ATOMIC, order);
+}
+
+/**
+ * __dev_alloc_page - allocate a page for network Rx
+ * @gfp_mask: allocation priority. Set __GFP_NOMEMALLOC if not for network Rx
+ *
+ * Allocate a new page.
+ *
+ * %NULL is returned if there is no free memory.
+ */
+static inline struct page *__dev_alloc_page(gfp_t gfp_mask)
+{
+	return __dev_alloc_pages(gfp_mask, 0);
+}
+
+static inline struct page *dev_alloc_page(void)
+{
+	return __dev_alloc_page(GFP_ATOMIC);
+}
+
+/**
  *	__skb_alloc_pages - allocate pages for ps-rx on a skb and preserve pfmemalloc data
  *	@gfp_mask: alloc_pages_node mask. Set __GFP_NOMEMALLOC if not for network packet RX
  *	@skb: skb to set pfmemalloc on if __GFP_MEMALLOC is used

[net-next PATCH v2 2/5] cxgb4/cxgb4vf: Replace __skb_alloc_page with __netdev_alloc_page

[Alexander Duyck]

Drop the bloated use of __skb_alloc_page and replace it with
__netdev_alloc_page.  In addition update the one other spot that is
allocating a page so that it allocates with the correct flags.

Cc: Hariprasad S <hariprasad@chelsio.com>
Cc: Casey Leedom <leedom@chelsio.com>
Signed-off-by: Alexander Duyck <alexander.h.duyck@redhat.com>
---
 drivers/net/ethernet/chelsio/cxgb4/sge.c   |    6 +++---
 drivers/net/ethernet/chelsio/cxgb4vf/sge.c |    7 ++++---
 2 files changed, 7 insertions(+), 6 deletions(-)

diff --git a/drivers/net/ethernet/chelsio/cxgb4/sge.c b/drivers/net/ethernet/chelsio/cxgb4/sge.c
index 5e1b314..20ee002 100644
--- a/drivers/net/ethernet/chelsio/cxgb4/sge.c
+++ b/drivers/net/ethernet/chelsio/cxgb4/sge.c
@@ -576,7 +576,7 @@ static unsigned int refill_fl(struct adapter *adap, struct sge_fl *q, int n,
 	__be64 *d = &q->desc[q->pidx];
 	struct rx_sw_desc *sd = &q->sdesc[q->pidx];
 
-	gfp |= __GFP_NOWARN | __GFP_COLD;
+	gfp |= __GFP_NOWARN;
 
 	if (s->fl_pg_order == 0)
 		goto alloc_small_pages;
@@ -585,7 +585,7 @@ static unsigned int refill_fl(struct adapter *adap, struct sge_fl *q, int n,
 	 * Prefer large buffers
 	 */
 	while (n) {
-		pg = alloc_pages(gfp | __GFP_COMP, s->fl_pg_order);
+		pg = __dev_alloc_pages(gfp, s->fl_pg_order);
 		if (unlikely(!pg)) {
 			q->large_alloc_failed++;
 			break;       /* fall back to single pages */
@@ -615,7 +615,7 @@ static unsigned int refill_fl(struct adapter *adap, struct sge_fl *q, int n,
 
 alloc_small_pages:
 	while (n--) {
-		pg = __skb_alloc_page(gfp, NULL);
+		pg = __dev_alloc_page(gfp);
 		if (unlikely(!pg)) {
 			q->alloc_failed++;
 			break;
diff --git a/drivers/net/ethernet/chelsio/cxgb4vf/sge.c b/drivers/net/ethernet/chelsio/cxgb4vf/sge.c
index 85036e6..9df40df 100644
--- a/drivers/net/ethernet/chelsio/cxgb4vf/sge.c
+++ b/drivers/net/ethernet/chelsio/cxgb4vf/sge.c
@@ -602,6 +602,8 @@ static unsigned int refill_fl(struct adapter *adapter, struct sge_fl *fl,
 	 */
 	BUG_ON(fl->avail + n > fl->size - FL_PER_EQ_UNIT);
 
+	gfp |= __GFP_NOWARN;
+
 	/*
 	 * If we support large pages, prefer large buffers and fail over to
 	 * small pages if we can't allocate large pages to satisfy the refill.
@@ -612,8 +614,7 @@ static unsigned int refill_fl(struct adapter *adapter, struct sge_fl *fl,
 		goto alloc_small_pages;
 
 	while (n) {
-		page = alloc_pages(gfp | __GFP_COMP | __GFP_NOWARN,
-				   FL_PG_ORDER);
+		page = __dev_alloc_pages(gfp, FL_PG_ORDER);
 		if (unlikely(!page)) {
 			/*
 			 * We've failed inour attempt to allocate a "large
@@ -657,7 +658,7 @@ static unsigned int refill_fl(struct adapter *adapter, struct sge_fl *fl,
 
 alloc_small_pages:
 	while (n--) {
-		page = __skb_alloc_page(gfp | __GFP_NOWARN, NULL);
+		page = __dev_alloc_page(gfp);
 		if (unlikely(!page)) {
 			fl->alloc_failed++;
 			break;

[net-next PATCH v2 3/5] phonet: Replace calls to __skb_alloc_page with __netdev_alloc_page

[Alexander Duyck]

Replace the calls to __skb_alloc_page that are passed NULL with calls to
__netdev_alloc_page.

In addition remove __GFP_COLD flag from allocations as we only want it for
the Rx buffer which is taken care of by __dev_alloc_skb, not for any
secondary allocations such as the queue element transmit descriptors.

Cc: Oliver Neukum <oliver@neukum.org>
Cc: Felipe Balbi <balbi@ti.com>
Signed-off-by: Alexander Duyck <alexander.h.duyck@redhat.com>
---
 drivers/net/usb/cdc-phonet.c           |    6 +++---
 drivers/usb/gadget/function/f_phonet.c |    6 +++---
 2 files changed, 6 insertions(+), 6 deletions(-)

diff --git a/drivers/net/usb/cdc-phonet.c b/drivers/net/usb/cdc-phonet.c
index 2ec1500..415ce8b 100644
--- a/drivers/net/usb/cdc-phonet.c
+++ b/drivers/net/usb/cdc-phonet.c
@@ -130,7 +130,7 @@ static int rx_submit(struct usbpn_dev *pnd, struct urb *req, gfp_t gfp_flags)
 	struct page *page;
 	int err;
 
-	page = __skb_alloc_page(gfp_flags | __GFP_NOMEMALLOC, NULL);
+	page = __dev_alloc_page(gfp_flags | __GFP_NOMEMALLOC);
 	if (!page)
 		return -ENOMEM;
 
@@ -212,7 +212,7 @@ resubmit:
 	if (page)
 		put_page(page);
 	if (req)
-		rx_submit(pnd, req, GFP_ATOMIC | __GFP_COLD);
+		rx_submit(pnd, req, GFP_ATOMIC);
 }
 
 static int usbpn_close(struct net_device *dev);
@@ -231,7 +231,7 @@ static int usbpn_open(struct net_device *dev)
 	for (i = 0; i < rxq_size; i++) {
 		struct urb *req = usb_alloc_urb(0, GFP_KERNEL);
 
-		if (!req || rx_submit(pnd, req, GFP_KERNEL | __GFP_COLD)) {
+		if (!req || rx_submit(pnd, req, GFP_KERNEL)) {
 			usb_free_urb(req);
 			usbpn_close(dev);
 			return -ENOMEM;
diff --git a/drivers/usb/gadget/function/f_phonet.c b/drivers/usb/gadget/function/f_phonet.c
index b9cfc15..cde7397 100644
--- a/drivers/usb/gadget/function/f_phonet.c
+++ b/drivers/usb/gadget/function/f_phonet.c
@@ -303,7 +303,7 @@ pn_rx_submit(struct f_phonet *fp, struct usb_request *req, gfp_t gfp_flags)
 	struct page *page;
 	int err;
 
-	page = __skb_alloc_page(gfp_flags | __GFP_NOMEMALLOC, NULL);
+	page = __dev_alloc_page(gfp_flags | __GFP_NOMEMALLOC);
 	if (!page)
 		return -ENOMEM;
 
@@ -377,7 +377,7 @@ static void pn_rx_complete(struct usb_ep *ep, struct usb_request *req)
 	if (page)
 		put_page(page);
 	if (req)
-		pn_rx_submit(fp, req, GFP_ATOMIC | __GFP_COLD);
+		pn_rx_submit(fp, req, GFP_ATOMIC);
 }
 
 /*-------------------------------------------------------------------------*/
@@ -437,7 +437,7 @@ static int pn_set_alt(struct usb_function *f, unsigned intf, unsigned alt)
 
 			netif_carrier_on(dev);
 			for (i = 0; i < phonet_rxq_size; i++)
-				pn_rx_submit(fp, fp->out_reqv[i], GFP_ATOMIC | __GFP_COLD);
+				pn_rx_submit(fp, fp->out_reqv[i], GFP_ATOMIC);
 		}
 		spin_unlock(&port->lock);
 		return 0;

[net-next PATCH v2 4/5] fm10k/igb/ixgbe: Replace __skb_alloc_page with netdev_alloc_page

[Alexander Duyck]

The Intel drivers were pretty much just using the plain vanilla GFP flags
in their calls to __skb_alloc_page so this change makes it so that they use
netdev_alloc_page which just uses GFP_ATOMIC for the gfp_flags value.

Cc: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Cc: Matthew Vick <matthew.vick@intel.com>
Cc: Don Skidmore <donald.c.skidmore@intel.com>
Signed-off-by: Alexander Duyck <alexander.h.duyck@redhat.com>
---
 drivers/net/ethernet/intel/fm10k/fm10k_main.c |    2 +-
 drivers/net/ethernet/intel/igb/igb_main.c     |    2 +-
 drivers/net/ethernet/intel/ixgbe/ixgbe_main.c |    3 +--
 3 files changed, 3 insertions(+), 4 deletions(-)

diff --git a/drivers/net/ethernet/intel/fm10k/fm10k_main.c b/drivers/net/ethernet/intel/fm10k/fm10k_main.c
index e645af4..73457ed 100644
--- a/drivers/net/ethernet/intel/fm10k/fm10k_main.c
+++ b/drivers/net/ethernet/intel/fm10k/fm10k_main.c
@@ -83,7 +83,7 @@ static bool fm10k_alloc_mapped_page(struct fm10k_ring *rx_ring,
 		return true;
 
 	/* alloc new page for storage */
-	page = alloc_page(GFP_ATOMIC | __GFP_COLD);
+	page = dev_alloc_page();
 	if (unlikely(!page)) {
 		rx_ring->rx_stats.alloc_failed++;
 		return false;
diff --git a/drivers/net/ethernet/intel/igb/igb_main.c b/drivers/net/ethernet/intel/igb/igb_main.c
index a2d72a8..1e35fae 100644
--- a/drivers/net/ethernet/intel/igb/igb_main.c
+++ b/drivers/net/ethernet/intel/igb/igb_main.c
@@ -6988,7 +6988,7 @@ static bool igb_alloc_mapped_page(struct igb_ring *rx_ring,
 		return true;
 
 	/* alloc new page for storage */
-	page = __skb_alloc_page(GFP_ATOMIC | __GFP_COLD, NULL);
+	page = dev_alloc_page();
 	if (unlikely(!page)) {
 		rx_ring->rx_stats.alloc_failed++;
 		return false;
diff --git a/drivers/net/ethernet/intel/ixgbe/ixgbe_main.c b/drivers/net/ethernet/intel/ixgbe/ixgbe_main.c
index d2df4e3..7405478 100644
--- a/drivers/net/ethernet/intel/ixgbe/ixgbe_main.c
+++ b/drivers/net/ethernet/intel/ixgbe/ixgbe_main.c
@@ -1440,8 +1440,7 @@ static bool ixgbe_alloc_mapped_page(struct ixgbe_ring *rx_ring,
 
 	/* alloc new page for storage */
 	if (likely(!page)) {
-		page = __skb_alloc_pages(GFP_ATOMIC | __GFP_COLD | __GFP_COMP,
-					 bi->skb, ixgbe_rx_pg_order(rx_ring));
+		page = dev_alloc_pages(ixgbe_rx_pg_order(rx_ring));
 		if (unlikely(!page)) {
 			rx_ring->rx_stats.alloc_rx_page_failed++;
 			return false;

[net-next PATCH v2 5/5] net: Remove __skb_alloc_page and __skb_alloc_pages

[Alexander Duyck]

Remove the two functions which are now dead code.

Signed-off-by: Alexander Duyck <alexander.h.duyck@redhat.com>
---
 include/linux/skbuff.h |   43 -------------------------------------------
 1 file changed, 43 deletions(-)

diff --git a/include/linux/skbuff.h b/include/linux/skbuff.h
index 2e5221f..73c370e 100644
--- a/include/linux/skbuff.h
+++ b/include/linux/skbuff.h
@@ -2233,49 +2233,6 @@ static inline struct page *dev_alloc_page(void)
 }
 
 /**
- *	__skb_alloc_pages - allocate pages for ps-rx on a skb and preserve pfmemalloc data
- *	@gfp_mask: alloc_pages_node mask. Set __GFP_NOMEMALLOC if not for network packet RX
- *	@skb: skb to set pfmemalloc on if __GFP_MEMALLOC is used
- *	@order: size of the allocation
- *
- * 	Allocate a new page.
- *
- * 	%NULL is returned if there is no free memory.
-*/
-static inline struct page *__skb_alloc_pages(gfp_t gfp_mask,
-					      struct sk_buff *skb,
-					      unsigned int order)
-{
-	struct page *page;
-
-	gfp_mask |= __GFP_COLD;
-
-	if (!(gfp_mask & __GFP_NOMEMALLOC))
-		gfp_mask |= __GFP_MEMALLOC;
-
-	page = alloc_pages_node(NUMA_NO_NODE, gfp_mask, order);
-	if (skb && page && page->pfmemalloc)
-		skb->pfmemalloc = true;
-
-	return page;
-}
-
-/**
- *	__skb_alloc_page - allocate a page for ps-rx for a given skb and preserve pfmemalloc data
- *	@gfp_mask: alloc_pages_node mask. Set __GFP_NOMEMALLOC if not for network packet RX
- *	@skb: skb to set pfmemalloc on if __GFP_MEMALLOC is used
- *
- * 	Allocate a new page.
- *
- * 	%NULL is returned if there is no free memory.
- */
-static inline struct page *__skb_alloc_page(gfp_t gfp_mask,
-					     struct sk_buff *skb)
-{
-	return __skb_alloc_pages(gfp_mask, skb, 0);
-}
-
-/**
  *	skb_propagate_pfmemalloc - Propagate pfmemalloc if skb is allocated after RX page
  *	@page: The page that was allocated from skb_alloc_page
  *	@skb: The skb that may need pfmemalloc set

Re: [net-next PATCH v2 4/5] fm10k/igb/ixgbe: Replace __skb_alloc_page with netdev_alloc_page

[Cong Wang]

On Tue, Nov 11, 2014 at 9:11 AM, Alexander Duyck
<alexander.h.duyck@redhat.com> wrote:
> diff --git a/drivers/net/ethernet/intel/fm10k/fm10k_main.c b/drivers/net/ethernet/intel/fm10k/fm10k_main.c
> index e645af4..73457ed 100644
> --- a/drivers/net/ethernet/intel/fm10k/fm10k_main.c
> +++ b/drivers/net/ethernet/intel/fm10k/fm10k_main.c
> @@ -83,7 +83,7 @@ static bool fm10k_alloc_mapped_page(struct fm10k_ring *rx_ring,
>                 return true;
>
>         /* alloc new page for storage */
> -       page = alloc_page(GFP_ATOMIC | __GFP_COLD);
> +       page = dev_alloc_page();

Doesn't match $subject.

Re: [net-next PATCH v2 4/5] fm10k/igb/ixgbe: Replace __skb_alloc_page with netdev_alloc_page

[Alexander Duyck]

On 11/11/2014 09:15 AM, Cong Wang wrote:
> On Tue, Nov 11, 2014 at 9:11 AM, Alexander Duyck
> <alexander.h.duyck-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org> wrote:
>> diff --git a/drivers/net/ethernet/intel/fm10k/fm10k_main.c b/drivers/net/ethernet/intel/fm10k/fm10k_main.c
>> index e645af4..73457ed 100644
>> --- a/drivers/net/ethernet/intel/fm10k/fm10k_main.c
>> +++ b/drivers/net/ethernet/intel/fm10k/fm10k_main.c
>> @@ -83,7 +83,7 @@ static bool fm10k_alloc_mapped_page(struct fm10k_ring *rx_ring,
>>                  return true;
>>
>>          /* alloc new page for storage */
>> -       page = alloc_page(GFP_ATOMIC | __GFP_COLD);
>> +       page = dev_alloc_page();
> Doesn't match $subject.

Yeah, I just noticed that.  I missed the patch title and comments when I 
was doing the replacement.

v3 on the way.

Thanks,

Alex
--
To unsubscribe from this list: send the line "unsubscribe linux-usb" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html