3.14.27-stable review

This is the start of the stable review cycle for the 3.14.27 release. There are 34 patches in this series, all will be posted as a response to this one. If anyone has any issues with these being applied, please let me know.

Responses should be made by Tue Dec 16 20:17:55 UTC 2014. Anything received after that time might be too late.

The whole patch series can be found in one patch at:

	kernel.org/pub/linux/kernel/v3.0/stable-review/patch-3.14.27-rc1.gz

and the diffstat can be found below.

thanks,

greg k-h

-------------
Pseudo-Shortlog of commits:

Greg Kroah-Hartman <[email protected]>

    Linux 3.14.27-rc1

Takashi Iwai <[email protected]>
ALSA: usb-audio: Don't resubmit pending URBs at MIDI error recovery

Takashi Iwai <[email protected]>
ALSA: hda - Fix built-in mic at resume on Lenovo Ideapad S210

Takashi Iwai <[email protected]>
ALSA: hda - Add EAPD fixup for ASUS Z99He laptop

Ronald Wahl <[email protected]>
mac80211: Fix regression that triggers a kernel BUG with CCMP

Anton Blanchard <[email protected]>
powerpc: 32 bit getcpu VDSO function uses 64 bit instructions

Todd Fujinaka <[email protected]>
igb: bring link up when PHY is powered up

Kan Liang <[email protected]>
perf/x86/intel: Protect LBR and extra_regs against KVM lying

Daniel Borkmann <[email protected]>
net: sctp: use MAX_HEADER for headroom reserve in output path

Eric Dumazet <[email protected]>
net: mvneta: fix race condition in mvneta_tx()

willy tarreau <[email protected]>
net: mvneta: fix Tx interrupt delay

Tom Herbert <[email protected]>
gre: Set inner mac header in gro complete

Nicolas Dichtel <[email protected]>
rtnetlink: release net refcnt on error in do_setlink()

Jack Morgenstein <[email protected]>
net/mlx4_core: Limit count field to 24 bits in qp_alloc_res

Thadeu Lima de Souza Cascardo <[email protected]>
tg3: fix ring init when there are more TX than RX channels

Marcelo Leitner <[email protected]>
Fix race condition between vxlan_sock_add and vxlan_sock_release

Yuri Chislov <[email protected]>

    ipv6: gre: fix wrong skb->protocol in WCCP

lucien <[email protected]>
ip_tunnel: the lack of vti_link_ops' dellink() cause kernel panic

Dmitry Torokhov <[email protected]>
sata_fsl: fix error handling of irq_of_parse_and_map

Tejun Heo <[email protected]>

    ahci: disable MSI on SAMSUNG 0xa800 SSD

Devin Ryles <[email protected]>
AHCI: Add DeviceIDs for Sunrise Point-LP SATA controller

Mathias Nyman <[email protected]>
USB: xhci: Reset a halted endpoint immediately when we encounter a stall.

Sakari Ailus <[email protected]>
media: smiapp: Only some selection targets are settable

Chris Clayton <[email protected]>

    x86: Use $(OBJDUMP) instead of plain objdump

Daniel Vetter <[email protected]>
drm/i915: Unlock panel even when LVDS is disabled

Daniel Vetter <[email protected]>
drm/i915: More cautious with pch fifo underruns

Petr Mladek <[email protected]>
drm/radeon: kernel panic in drm_calc_vbltimestamp_from_scanoutpos with 3.18.0-rc6

Grygorii Strashko <[email protected]>
i2c: davinci: generate STP always when NACK is received

Alexander Kochetkov <[email protected]>
i2c: omap: fix i207 errata handling

Alexander Kochetkov <[email protected]>
i2c: omap: fix NACK and Arbitration Lost irq handling

Seth Forshee <[email protected]>
xen-netfront: Remove BUGs on paged skb data which crosses a page boundary

Daniel Forrest <[email protected]>
mm: fix anon_vma_clone() error treatment

Hugh Dickins <[email protected]>
mm: fix swapoff hang after page migration and fork

Andrew Morton <[email protected]>
mm/vmpressure.c: fix race in vmpressure_work_fn()

Weijie Yang <[email protected]>
mm: frontswap: invalidate expired data on a dup-store failure

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

Diffstat:

 Makefile                                           |  4 +-
 arch/powerpc/kernel/vdso32/getcpu.S                |  4 +-
 arch/x86/boot/compressed/Makefile                  |  2 +-
 arch/x86/kernel/cpu/perf_event.c                   |  3 +
 arch/x86/kernel/cpu/perf_event.h                   | 12 ++--
 arch/x86/kernel/cpu/perf_event_intel.c             | 66 +++++++++++++++++++++-
 drivers/ata/ahci.c                                 |  4 ++
 drivers/ata/sata_fsl.c                             |  2 +-
 drivers/gpu/drm/i915/intel_display.c               |  2 -
 drivers/gpu/drm/i915/intel_lvds.c                  | 22 ++++----
 drivers/gpu/drm/radeon/radeon_kms.c                |  2 +
 drivers/i2c/busses/i2c-davinci.c                   |  8 +--
 drivers/i2c/busses/i2c-omap.c                      | 10 ++--
 drivers/media/i2c/smiapp/smiapp-core.c             |  2 +-
 drivers/net/ethernet/broadcom/tg3.c                |  3 +-
 drivers/net/ethernet/intel/igb/igb_main.c          |  2 +
 drivers/net/ethernet/marvell/mvneta.c              |  5 +-
 .../net/ethernet/mellanox/mlx4/resource_tracker.c  |  2 +-
 drivers/net/vxlan.c                                | 10 +---
 drivers/net/xen-netfront.c                         |  5 --
 drivers/usb/host/xhci-ring.c                       | 40 ++++---------
 drivers/usb/host/xhci.c                            | 60 +++++---------------
 mm/frontswap.c                                     |  4 +-
 mm/memory.c                                        | 26 ++++-----
 mm/mmap.c                                          | 10 +++-
 mm/rmap.c                                          |  6 +-
 mm/vmpressure.c                                    |  8 ++-
 net/core/rtnetlink.c                               |  1 +
 net/ipv4/gre_offload.c                             |  3 +
 net/ipv4/ip_vti.c                                  |  1 +
 net/ipv6/ip6_gre.c                                 |  4 +-
 net/ipv6/ip6_vti.c                                 | 11 ++++
 net/mac80211/aes_ccm.c                             |  3 +
 net/sctp/output.c                                  |  4 +-
 sound/pci/hda/patch_analog.c                       |  1 +
 sound/pci/hda/patch_realtek.c                      |  1 +
 sound/usb/midi.c                                   |  2 +
 37 files changed, 202 insertions(+), 153 deletions(-)

3.14-stable review patch. If anyone has any objections, please let me know.

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

From: Weijie Yang <[email protected]>

commit fb993fa1a2f669215fa03a09eed7848f2663e336 upstream.

If a frontswap dup-store failed, it should invalidate the expired page in the backend, or it could trigger some data corruption issue. Such as:
1. use zswap as the frontswap backend with writeback feature
2. store a swap page(version_1) to entry A, success
3. dup-store a newer page(version_2) to the same entry A, fail
4. use __swap_writepage() write version_2 page to swapfile, success
5. zswap do shrink, writeback version_1 page to swapfile
6. version_2 page is overwrited by version_1, data corrupt.

This patch fixes this issue by invalidating expired data immediately when meet a dup-store failure.

---
 mm/frontswap.c |    4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)

@@ -244,8 +244,10 @@ int __frontswap_store(struct page *page)
 		  the (older) page from frontswap
 		 */
 		inc_frontswap_failed_stores();
-		if (dup)
+		if (dup) {
 			__frontswap_clear(sis, offset);
+			frontswap_ops->invalidate_page(type, offset);
+		}
 	}
 	if (frontswap_writethrough_enabled)
 		/* report failure so swap also writes to swap device */

3.14-stable review patch. If anyone has any objections, please let me know.

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

From: Andrew Morton <[email protected]>

commit 91b57191cfd152c02ded0745250167d0263084f8 upstream.

In some android devices, there will be a "divide by zero" exception. vmpr->scanned could be zero before spin_lock(&vmpr->sr_lock).

Addresses https://bugzilla.kernel.org/show_bug.cgi?id=88051

---
 mm/vmpressure.c |    8 +++++---
 1 file changed, 5 insertions(+), 3 deletions(-)

@@ -165,6 +165,7 @@ static void vmpressure_work_fn(struct wo
 	unsigned long scanned;
 	unsigned long reclaimed;
 
+	spin_lock(&vmpr->sr_lock);
 	/*
 	 * Several contexts might be calling vmpressure(), so it is
 	 * possible that the work was rescheduled again before the old

@@ -173,11 +174,12 @@ static void vmpressure_work_fn(struct wo
 	 * here. No need for any locks here since we don't care if
 	 * vmpr->reclaimed is in sync.
 	 */
-	if (!vmpr->scanned)
+	scanned = vmpr->scanned;
+	if (!scanned) {
+		spin_unlock(&vmpr->sr_lock);
 		return;
+	}
 
-	spin_lock(&vmpr->sr_lock);
-	scanned = vmpr->scanned;
 	reclaimed = vmpr->reclaimed;
 	vmpr->scanned = 0;
 	vmpr->reclaimed = 0;

3.14-stable review patch. If anyone has any objections, please let me know.

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

From: Hugh Dickins <[email protected]>

commit 2022b4d18a491a578218ce7a4eca8666db895a73 upstream.

I've been seeing swapoff hangs in recent testing: it's cycling around trying unsuccessfully to find an mm for some remaining pages of swap.

I have been exercising swap and page migration more heavily recently, and now notice a long-standing error in copy_one_pte(): it's trying to add dst_mm to swapoff's mmlist when it finds a swap entry, but is doing so even when it's a migration entry or an hwpoison entry.

Which wouldn't matter much, except it adds dst_mm next to src_mm, assuming src_mm is already on the mmlist: which may not be so. Then if pages are later swapped out from dst_mm, swapoff won't be able to find where to replace them.

There's already a !non_swap_entry() test for stats: move that up before the swap_duplicate() and the addition to mmlist.

---
 mm/memory.c |   24 ++++++++++++------------
 1 file changed, 12 insertions(+), 12 deletions(-)

@@ -808,20 +808,20 @@ copy_one_pte(struct mm_struct *dst_mm, s
 		if (!pte_file(pte)) {
 			swp_entry_t entry = pte_to_swp_entry(pte);
 
-			if (swap_duplicate(entry) < 0)
-				return entry.val;
+			if (likely(!non_swap_entry(entry))) {
+				if (swap_duplicate(entry) < 0)
+					return entry.val;
 
-			/* make sure dst_mm is on swapoff's mmlist. */
-			if (unlikely(list_empty(&dst_mm->mmlist))) {
-				spin_lock(&mmlist_lock);
-				if (list_empty(&dst_mm->mmlist))
-					list_add(&dst_mm->mmlist,
-						 &src_mm->mmlist);
-				spin_unlock(&mmlist_lock);
-			}
-			if (likely(!non_swap_entry(entry)))
+				/* make sure dst_mm is on swapoff's mmlist. */
+				if (unlikely(list_empty(&dst_mm->mmlist))) {
+					spin_lock(&mmlist_lock);
+					if (list_empty(&dst_mm->mmlist))
+						list_add(&dst_mm->mmlist,
+							 &src_mm->mmlist);
+					spin_unlock(&mmlist_lock);
+				}
 				rss[MM_SWAPENTS]++;
-			else if (is_migration_entry(entry)) {
+			} else if (is_migration_entry(entry)) {
 				page = migration_entry_to_page(entry);
 
 				if (PageAnon(page))

3.14-stable review patch. If anyone has any objections, please let me know.

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

From: Daniel Forrest <[email protected]>

commit c4ea95d7cd08d9ffd7fa75e6c5e0332d596dd11e upstream.

Andrew Morton noticed that the error return from anon_vma_clone() was being dropped and replaced with -ENOMEM (which is not itself a bug because the only error return value from anon_vma_clone() is -ENOMEM).

I did an audit of callers of anon_vma_clone() and discovered an actual bug where the error return was being lost. In __split_vma(), between Linux 3.11 and 3.12 the code was changed so the err variable is used before the call to anon_vma_clone() and the default initial value of -ENOMEM is overwritten. So a failure of anon_vma_clone() will return success since err at this point is now zero.

Below is a patch which fixes this bug and also propagates the error return value from anon_vma_clone() in all cases.

---
 mm/mmap.c |   10 +++++++---
 mm/rmap.c |    6 ++++--
 2 files changed, 11 insertions(+), 5 deletions(-)

@@ -745,8 +745,11 @@ again:			remove_next = 1 + (end > next->
 		 * shrinking vma had, to cover any anon pages imported.
 		 */
 		if (exporter && exporter->anon_vma && !importer->anon_vma) {
-			if (anon_vma_clone(importer, exporter))
-				return -ENOMEM;
+			int error;
+
+			error = anon_vma_clone(importer, exporter);
+			if (error)
+				return error;
 			importer->anon_vma = exporter->anon_vma;
 		}
 	}

@@ -2428,7 +2431,8 @@ static int __split_vma(struct mm_struct
 	if (err)
 		goto out_free_vma;
 
-	if (anon_vma_clone(new, vma))
+	err = anon_vma_clone(new, vma);
+	if (err)
 		goto out_free_mpol;
 
 	if (new->vm_file)

@@ -274,6 +274,7 @@ int anon_vma_fork(struct vm_area_struct
 {
 	struct anon_vma_chain *avc;
 	struct anon_vma *anon_vma;
+	int error;
 
 	/* Don't bother if the parent process has no anon_vma here. */
 	if (!pvma->anon_vma)

@@ -283,8 +284,9 @@ int anon_vma_fork(struct vm_area_struct
 	 * First, attach the new VMA to the parent VMA's anon_vmas,
 	 * so rmap can find non-COWed pages in child processes.
 	 */
-	if (anon_vma_clone(vma, pvma))
-		return -ENOMEM;
+	error = anon_vma_clone(vma, pvma);
+	if (error)
+		return error;
 
 	/* Then add our own anon_vma. */
 	anon_vma = anon_vma_alloc();

3.14-stable review patch. If anyone has any objections, please let me know.

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

From: Seth Forshee <[email protected]>

commit 8d609725d4357f499e2103e46011308b32f53513 upstream.

These BUGs can be erroneously triggered by frags which refer to tail pages within a compound page. The data in these pages may overrun the hardware page while still being contained within the compound page, but since compound_order() evaluates to 0 for tail pages the assertion fails. The code already iterates through subsequent pages correctly in this scenario, so the BUGs are unnecessary and can be removed.

---
 drivers/net/xen-netfront.c |    5 -----
 1 file changed, 5 deletions(-)

@@ -469,9 +469,6 @@ static void xennet_make_frags(struct sk_
 		len = skb_frag_size(frag);
 		offset = frag->page_offset;
 
-		/* Data must not cross a page boundary. */
-		BUG_ON(len + offset > PAGE_SIZE<<compound_order(page));
-
 		/* Skip unused frames from start of page */
 		page += offset >> PAGE_SHIFT;
 		offset &= ~PAGE_MASK;

@@ -479,8 +476,6 @@ static void xennet_make_frags(struct sk_
 		while (len > 0) {
 			unsigned long bytes;
 
-			BUG_ON(offset >= PAGE_SIZE);
-
 			bytes = PAGE_SIZE - offset;
 			if (bytes > len)
 				bytes = len;

3.14-stable review patch. If anyone has any objections, please let me know.

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

From: Alexander Kochetkov <[email protected]>

commit 27caca9d2e01c92b26d0690f065aad093fea01c7 upstream.

commit 1d7afc95946487945cc7f5019b41255b72224b70 (i2c: omap: ack IRQ in parts) changed the interrupt handler to complete transfers without clearing XRDY (AL case) and ARDY (NACK case) flags. XRDY or ARDY interrupts will be fired again. As a result, ISR keep processing transfer after it was already complete (from the driver code point of view).

A didn't see real impacts of the 1d7afc9, but it is really bad idea to have ISR running on user data after transfer was complete.

It looks, what 1d7afc9 violate TI specs in what how AL and NACK should be handled (see Note 1, sprugn4r, Figure 17-31 and Figure 17-32).

According to specs (if I understood correctly), in case of NACK and AL driver must reset NACK, AL, ARDY, RDR, and RRDY (Master Receive Mode), and NACK, AL, ARDY, and XDR (Master Transmitter Mode).

All that is done down the code under the if condition:
if (stat & (OMAP_I2C_STAT_ARDY | OMAP_I2C_STAT_NACK | OMAP_I2C_STAT_AL)) ...

The patch restore pre 1d7afc9 logic of handling NACK and AL interrupts, so no interrupts is fired after ISR informs the rest of driver what transfer complete.

Note: instead of removing break under NACK case, we could just replace 'break' with 'continue' and allow NACK transfer to finish using ARDY event. I found that NACK and ARDY bits usually set together. That case confirm TI wiki:
http://processors.wiki.ti.com/index.php/I2C_Tips#Detecting_and_handling_NACK

In order if someone interested in the event traces for NACK and AL cases, I sent them to mailing list.

Tested on Beagleboard XM C.

---
 drivers/i2c/busses/i2c-omap.c |    2 --
 1 file changed, 2 deletions(-)

@@ -926,14 +926,12 @@ omap_i2c_isr_thread(int this_irq, void *
 		if (stat & OMAP_I2C_STAT_NACK) {
 			err |= OMAP_I2C_STAT_NACK;
 			omap_i2c_ack_stat(dev, OMAP_I2C_STAT_NACK);
-			break;
 		}
 
 		if (stat & OMAP_I2C_STAT_AL) {
 			dev_err(dev->dev, "Arbitration lost\n");
 			err |= OMAP_I2C_STAT_AL;
 			omap_i2c_ack_stat(dev, OMAP_I2C_STAT_AL);
-			break;
 		}
 
 		/*

3.14-stable review patch. If anyone has any objections, please let me know.

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

From: Alexander Kochetkov <[email protected]>

commit ccfc866356674cb3a61829d239c685af6e85f197 upstream.

commit 6d9939f651419a63e091105663821f9c7d3fec37 (i2c: omap: split out [XR]DR and [XR]RDY) changed the way how errata i207 (I2C: RDR Flag May Be Incorrectly Set) get handled. 6d9939f6514 code doesn't correspond to workaround provided by errata.

According to errata ISR must filter out spurious RDR before data read not after. ISR must read RXSTAT to get number of bytes available to read. Because RDR could be set while there could no data in the receive FIFO.

Restored pre 6d9939f6514 way of handling errata.

Found by code review. Real impact haven't seen.
Tested on Beagleboard XM C.

---
 drivers/i2c/busses/i2c-omap.c |    8 +++++---
 1 file changed, 5 insertions(+), 3 deletions(-)

@@ -956,11 +956,13 @@ omap_i2c_isr_thread(int this_irq, void *
 			if (dev->fifo_size)
 				num_bytes = dev->buf_len;
 
-			omap_i2c_receive_data(dev, num_bytes, true);
-
-			if (dev->errata & I2C_OMAP_ERRATA_I207)
+			if (dev->errata & I2C_OMAP_ERRATA_I207) {
 				i2c_omap_errata_i207(dev, stat);
+				num_bytes = (omap_i2c_read_reg(dev,
+					OMAP_I2C_BUFSTAT_REG) >> 8) & 0x3F;
+			}
 
+			omap_i2c_receive_data(dev, num_bytes, true);
 			omap_i2c_ack_stat(dev, OMAP_I2C_STAT_RDR);
 			continue;
 		}

3.14-stable review patch. If anyone has any objections, please let me know.

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

From: Petr Mladek <[email protected]>

commit f5475cc43c899e33098d4db44b7c5e710f16589d upstream.

I was unable too boot 3.18.0-rc6 because of the following kernel panic in drm_calc_vbltimestamp_from_scanoutpos():

    [drm] Initialized drm 1.1.0 20060810
    [drm] radeon kernel modesetting enabled.
    [drm] initializing kernel modesetting (RV100 0x1002:0x515E 0x15D9:0x8080).
    [drm] register mmio base: 0xC8400000
    [drm] register mmio size: 65536
    radeon 0000:0b:01.0: VRAM: 128M 0x00000000D0000000 - 0x00000000D7FFFFFF (16M used)
    radeon 0000:0b:01.0: GTT: 512M 0x00000000B0000000 - 0x00000000CFFFFFFF
    [drm] Detected VRAM RAM=128M, BAR=128M
    [drm] RAM width 16bits DDR
    [TTM] Zone  kernel: Available graphics memory: 3829346 kiB
    [TTM] Zone   dma32: Available graphics memory: 2097152 kiB
    [TTM] Initializing pool allocator
    [TTM] Initializing DMA pool allocator
    [drm] radeon: 16M of VRAM memory ready
    [drm] radeon: 512M of GTT memory ready.
    [drm] GART: num cpu pages 131072, num gpu pages 131072
    [drm] PCI GART of 512M enabled (table at 0x0000000037880000).
    radeon 0000:0b:01.0: WB disabled
    radeon 0000:0b:01.0: fence driver on ring 0 use gpu addr 0x00000000b0000000 and cpu addr 0xffff8800bbbfa000
    [drm] Supports vblank timestamp caching Rev 2 (21.10.2013).
    [drm] Driver supports precise vblank timestamp query.
    [drm] radeon: irq initialized.
    [drm] Loading R100 Microcode
    radeon 0000:0b:01.0: Direct firmware load for radeon/R100_cp.bin failed with error -2
    radeon_cp: Failed to load firmware "radeon/R100_cp.bin"
    [drm:r100_cp_init] *ERROR* Failed to load firmware!
    radeon 0000:0b:01.0: failed initializing CP (-2).
    radeon 0000:0b:01.0: Disabling GPU acceleration
    [drm] radeon: cp finalized
    BUG: unable to handle kernel NULL pointer dereference at 000000000000025c
    IP: [<ffffffff8150423b>] drm_calc_vbltimestamp_from_scanoutpos+0x4b/0x320
    PGD 0
    Oops: 0000 [#1] SMP
    Modules linked in:
    CPU: 1 PID: 1 Comm: swapper/0 Not tainted 3.18.0-rc6-4-default #2649
    Hardware name: Supermicro X7DB8/X7DB8, BIOS 6.00 07/26/2006
    task: ffff880234da2010 ti: ffff880234da4000 task.ti: ffff880234da4000
    RIP: 0010:[<ffffffff8150423b>]  [<ffffffff8150423b>] drm_calc_vbltimestamp_from_scanoutpos+0x4b/0x320
    RSP: 0000:ffff880234da7918  EFLAGS: 00010086
    RAX: ffffffff81557890 RBX: 0000000000000000 RCX: ffff880234da7a48
    RDX: ffff880234da79f4 RSI: 0000000000000000 RDI: ffff880232e15000
    RBP: ffff880234da79b8 R08: 0000000000000000 R09: 0000000000000000
    R10: 000000000000000a R11: 0000000000000001 R12: ffff880232dda1c0
    R13: ffff880232e1518c R14: 0000000000000292 R15: ffff880232e15000
    FS:  0000000000000000(0000) GS:ffff88023fc40000(0000) knlGS:0000000000000000
    CS:  0010 DS: 0000 ES: 0000 CR0: 000000008005003b
    CR2: 000000000000025c CR3: 0000000002014000 CR4: 00000000000007e0
    Stack:
     ffff880234da79d8 0000000000000286 ffff880232dcbc00 0000000000002480
     ffff880234da7958 0000000000000296 ffff880234da7998 ffffffff8151b51d
     ffff880234da7a48 0000000032dcbeb0 ffff880232dcbc00 ffff880232dcbc58
    Call Trace:
     [<ffffffff8151b51d>] ? drm_vma_offset_remove+0x1d/0x110
     [<ffffffff8152dc98>] radeon_get_vblank_timestamp_kms+0x38/0x60
     [<ffffffff8152076a>] ? ttm_bo_release_list+0xba/0x180
     [<ffffffff81503751>] drm_get_last_vbltimestamp+0x41/0x70
     [<ffffffff81503933>] vblank_disable_and_save+0x73/0x1d0
     [<ffffffff81106b2f>] ? try_to_del_timer_sync+0x4f/0x70
     [<ffffffff81505245>] drm_vblank_cleanup+0x65/0xa0
     [<ffffffff815604fa>] radeon_irq_kms_fini+0x1a/0x70
     [<ffffffff8156c07e>] r100_init+0x26e/0x410
     [<ffffffff8152ae3e>] radeon_device_init+0x7ae/0xb50
     [<ffffffff8152d57f>] radeon_driver_load_kms+0x8f/0x210
     [<ffffffff81506965>] drm_dev_register+0xb5/0x110
     [<ffffffff8150998f>] drm_get_pci_dev+0x8f/0x200
     [<ffffffff815291cd>] radeon_pci_probe+0xad/0xe0
     [<ffffffff8141a365>] local_pci_probe+0x45/0xa0
     [<ffffffff8141b741>] pci_device_probe+0xd1/0x130
     [<ffffffff81633dad>] driver_probe_device+0x12d/0x3e0
     [<ffffffff8163413b>] __driver_attach+0x9b/0xa0
     [<ffffffff816340a0>] ? __device_attach+0x40/0x40
     [<ffffffff81631cd3>] bus_for_each_dev+0x63/0xa0
     [<ffffffff8163378e>] driver_attach+0x1e/0x20
     [<ffffffff81633390>] bus_add_driver+0x180/0x240
     [<ffffffff81634914>] driver_register+0x64/0xf0
     [<ffffffff81419cac>] __pci_register_driver+0x4c/0x50
     [<ffffffff81509bf5>] drm_pci_init+0xf5/0x120
     [<ffffffff821dc871>] ? ttm_init+0x6a/0x6a
     [<ffffffff821dc908>] radeon_init+0x97/0xb5
     [<ffffffff810002fc>] do_one_initcall+0xbc/0x1f0
     [<ffffffff810e3278>] ? __wake_up+0x48/0x60
     [<ffffffff8218e256>] kernel_init_freeable+0x18a/0x215
     [<ffffffff8218d983>] ? initcall_blacklist+0xc0/0xc0
     [<ffffffff818a78f0>] ? rest_init+0x80/0x80
     [<ffffffff818a78fe>] kernel_init+0xe/0xf0
     [<ffffffff818c0c3c>] ret_from_fork+0x7c/0xb0
     [<ffffffff818a78f0>] ? rest_init+0x80/0x80
    Code: 45 ac 0f 88 a8 01 00 00 3b b7 d0 01 00 00 49 89 ff 0f 83 99 01 00 00 48 8b 47 20 48 8b 80 88 00 00 00 48 85 c0 0f 84 cd 01 00 00 <41> 8b b1 5c 02 00 00 41 8b 89 58 02 00 00 89 75 98 41 8b b1 60
    RIP  [<ffffffff8150423b>] drm_calc_vbltimestamp_from_scanoutpos+0x4b/0x320
     RSP <ffff880234da7918>
    CR2: 000000000000025c
    ---[ end trace ad2c0aadf48e2032 ]---
    Kernel panic - not syncing: Attempted to kill init! exitcode=0x00000009

It has helped me to add a NULL pointer check that was suggested at
http://lists.freedesktop.org/archives/dri-devel/2014-October/070663.html

I am not familiar with the code. But the change looks sane and we need something fast at this stage of 3.18 development.

---
 drivers/gpu/drm/radeon/radeon_kms.c |    2 ++
 1 file changed, 2 insertions(+)

@@ -740,6 +740,8 @@ int radeon_get_vblank_timestamp_kms(stru
 
 	/* Get associated drm_crtc: */
 	drmcrtc = &rdev->mode_info.crtcs[crtc]->base;
+	if (!drmcrtc)
+		return -EINVAL;
 
 	/* Helper routine in DRM core does all the work: */
 	return drm_calc_vbltimestamp_from_scanoutpos(dev, crtc, max_error,

3.14-stable review patch. If anyone has any objections, please let me know.

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

From: Daniel Vetter <[email protected]>

commit b68362278af94e1171f5be9d4e44988601fb0439 upstream.

Apparently PCH fifo underruns are tricky, we have plenty reports that we see the occasional underrun (especially at boot-up).

So for a change let's see what happens when we don't re-enable pch fifo underrun reporting when the pipe is disabled. This means that the kernel can't catch pch fifo underruns when they happen (except when all pipes are on on the pch). But we'll still catch underruns when disabling the pipe again. So not a terrible reduction in test coverage.

Since the DRM_ERROR is new and hence a regression plan B would be to revert it back to a debug output. Which would be a lot worse than this hack for underrun test coverage in the wild. See the referenced discussions for more.

References: http://mid.gmane.org/CA+gsUGRfGe3t4NcjdeA=qXysrhLY3r4CEu7z4bjTwxi1uOfy+g@mail.gmail.com
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=85898
References: https://bugs.freedesktop.org/show_bug.cgi?id=85898
References: https://bugs.freedesktop.org/show_bug.cgi?id=86233
References: https://bugs.freedesktop.org/show_bug.cgi?id=86478

---
 drivers/gpu/drm/i915/intel_display.c |    2 --
 1 file changed, 2 deletions(-)

@@ -3817,7 +3817,6 @@ static void ironlake_crtc_disable(struct
 		ironlake_fdi_disable(crtc);
 
 		ironlake_disable_pch_transcoder(dev_priv, pipe);
-		intel_set_pch_fifo_underrun_reporting(dev, pipe, true);
 
 		if (HAS_PCH_CPT(dev)) {
 			/* disable TRANS_DP_CTL */

@@ -3883,7 +3882,6 @@ static void haswell_crtc_disable(struct
 
 	if (intel_crtc->config.has_pch_encoder) {
 		lpt_disable_pch_transcoder(dev_priv);
-		intel_set_pch_fifo_underrun_reporting(dev, TRANSCODER_A, true);
 		intel_ddi_fdi_disable(crtc);
 	}
 

3.14-stable review patch. If anyone has any objections, please let me know.

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

From: Daniel Vetter <[email protected]>

commit b0616c5306b342ceca07044dbc4f917d95c4f825 upstream.

Otherwise we'll have backtraces in assert_panel_unlocked because the BIOS locks the register. In the reporter's case this regression was introduced in

commit c31407a3672aaebb4acddf90944a114fa5c8af7b

Author: Chris Wilson <[email protected]>
Date:   Thu Oct 18 21:07:01 2012 +0100

drm/i915: Add no-lvds quirk for Supermicro X7SPA-H

---
 drivers/gpu/drm/i915/intel_lvds.c |   22 +++++++++++-----------
 1 file changed, 11 insertions(+), 11 deletions(-)

@@ -905,6 +905,17 @@ void intel_lvds_init(struct drm_device *
 	int pipe;
 	u8 pin;
 
+	/*
+	 * Unlock registers and just leave them unlocked. Do this before
+	 * checking quirk lists to avoid bogus WARNINGs.
+	 */
+	if (HAS_PCH_SPLIT(dev)) {
+		I915_WRITE(PCH_PP_CONTROL,
+			   I915_READ(PCH_PP_CONTROL) | PANEL_UNLOCK_REGS);
+	} else {
+		I915_WRITE(PP_CONTROL,
+			   I915_READ(PP_CONTROL) | PANEL_UNLOCK_REGS);
+	}
 	if (!intel_lvds_supported(dev))
 		return;
 

@@ -1099,17 +1110,6 @@ out:
 	DRM_DEBUG_KMS("detected %s-link lvds configuration\n",
 		      lvds_encoder->is_dual_link ? "dual" : "single");
 
-	/*
-	 * Unlock registers and just
-	 * leave them unlocked
-	 */
-	if (HAS_PCH_SPLIT(dev)) {
-		I915_WRITE(PCH_PP_CONTROL,
-			   I915_READ(PCH_PP_CONTROL) | PANEL_UNLOCK_REGS);
-	} else {
-		I915_WRITE(PP_CONTROL,
-			   I915_READ(PP_CONTROL) | PANEL_UNLOCK_REGS);
-	}
 	lvds_connector->lid_notifier.notifier_call = intel_lid_notify;
 	if (acpi_lid_notifier_register(&lvds_connector->lid_notifier)) {
 		DRM_DEBUG_KMS("lid notifier registration failed\n");

3.14-stable review patch. If anyone has any objections, please let me know.

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

From: Chris Clayton <[email protected]>

commit e2e68ae688b0a3766cd75aedf4ed4e39be402009 upstream.

commit e6023367d779 'x86, kaslr: Prevent .bss from overlaping initrd' broke the cross compile of x86. It added a objdump invocation, which invokes the host native objdump and ignores an active cross tool chain.

Use $(OBJDUMP) instead which takes the CROSS_COMPILE prefix into account.

[ tglx: Massage changelog and use $(OBJDUMP) ]

---
 arch/x86/boot/compressed/Makefile |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

@@ -75,7 +75,7 @@ suffix-$(CONFIG_KERNEL_XZ)	:= xz
 suffix-$(CONFIG_KERNEL_LZO) 	:= lzo
 suffix-$(CONFIG_KERNEL_LZ4) 	:= lz4
 
-RUN_SIZE = $(shell objdump -h vmlinux | \
+RUN_SIZE = $(shell $(OBJDUMP) -h vmlinux | \
 	     perl $(srctree)/arch/x86/tools/calc_run_size.pl)
 quiet_cmd_mkpiggy = MKPIGGY $@
       cmd_mkpiggy = $(obj)/mkpiggy $< $(RUN_SIZE) > $@ || ( rm -f $@ ; false )

3.14-stable review patch. If anyone has any objections, please let me know.

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

From: Sakari Ailus <[email protected]>

commit b31eb901c4e5eeef4c83c43dfbc7fe0d4348cb21 upstream.

Setting a non-settable selection target caused BUG() to be called. The check for valid selections only takes the selection target into account, but does not tell whether it may be set, or only get. Fix the issue by simply returning an error to the user.

---
 drivers/media/i2c/smiapp/smiapp-core.c |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

@@ -2138,7 +2138,7 @@ static int smiapp_set_selection(struct v
 		ret = smiapp_set_compose(subdev, fh, sel);
 		break;
 	default:
-		BUG();
+		ret = -EINVAL;
 	}
 
 	mutex_unlock(&sensor->mutex);

3.14-stable review patch. If anyone has any objections, please let me know.

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

From: Mathias Nyman <[email protected]>

commit 8e71a322fdb127814bcba423a512914ca5bc6cf5 upstream.

If a device is halted and reuturns a STALL, then the halted endpoint needs to be cleared both on the host and device side. The host side halt is cleared by issueing a xhci reset endpoint command. The device side is cleared with a ClearFeature(ENDPOINT_HALT) request, which should be issued by the device driver if a URB reruen -EPIPE.

Previously we cleared the host side halt after the device side was cleared. To make sure the host side halt is cleared in time we want to issue the reset endpoint command immedialtely when a STALL status is encountered.

Otherwise we end up not following the specs and not returning -EPIPE several times in a row when trying to transfer data to a halted endpoint.

---
 drivers/usb/host/xhci-ring.c |   40 +++++++---------------------
 drivers/usb/host/xhci.c      |   60 ++++++++++---------------------------------
 2 files changed, 25 insertions(+), 75 deletions(-)

@@ -1998,22 +1998,13 @@ static int finish_td(struct xhci_hcd *xh
 		ep->stopped_td = td;
 		return 0;
 	} else {
-		if (trb_comp_code == COMP_STALL) {
-			/* The transfer is completed from the driver's
-			 * perspective, but we need to issue a set dequeue
-			 * command for this stalled endpoint to move the dequeue
-			 * pointer past the TD.  We can't do that here because
-			 * the halt condition must be cleared first.  Let the
-			 * USB class driver clear the stall later.
-			 */
-			ep->stopped_td = td;
-			ep->stopped_stream = ep_ring->stream_id;
-		} else if (xhci_requires_manual_halt_cleanup(xhci,
-					ep_ctx, trb_comp_code)) {
-			/* Other types of errors halt the endpoint, but the
-			 * class driver doesn't call usb_reset_endpoint() unless
-			 * the error is -EPIPE.  Clear the halted status in the
-			 * xHCI hardware manually.
+		if (trb_comp_code == COMP_STALL ||
+		    xhci_requires_manual_halt_cleanup(xhci, ep_ctx,
+						      trb_comp_code)) {
+			/* Issue a reset endpoint command to clear the host side
+			 * halt, followed by a set dequeue command to move the
+			 * dequeue pointer past the TD.
+			 * The class driver clears the device side halt later.
 			 */
 			xhci_cleanup_halted_endpoint(xhci,
 					slot_id, ep_index, ep_ring->stream_id,

@@ -2133,9 +2124,7 @@ static int process_ctrl_td(struct xhci_h
 		else
 			td->urb->actual_length = 0;
 
-		xhci_cleanup_halted_endpoint(xhci,
-			slot_id, ep_index, 0, td, event_trb);
-		return finish_td(xhci, td, event_trb, event, ep, status, true);
+		return finish_td(xhci, td, event_trb, event, ep, status, false);
 	}
 	/*
 	 * Did we transfer any data, despite the errors that might have

@@ -2689,17 +2678,8 @@ cleanup:
 		if (ret) {
 			urb = td->urb;
 			urb_priv = urb->hcpriv;
-			/* Leave the TD around for the reset endpoint function
-			 * to use(but only if it's not a control endpoint,
-			 * since we already queued the Set TR dequeue pointer
-			 * command for stalled control endpoints).
-			 */
-			if (usb_endpoint_xfer_control(&urb->ep->desc) ||
-				(trb_comp_code != COMP_STALL &&
-					trb_comp_code != COMP_BABBLE))
-				xhci_urb_free_priv(xhci, urb_priv);
-			else
-				kfree(urb_priv);
+
+			xhci_urb_free_priv(xhci, urb_priv);
 
 			usb_hcd_unlink_urb_from_ep(bus_to_hcd(urb->dev->bus), urb);
 			if ((urb->actual_length != urb->transfer_buffer_length &&

3.14-stable review patch. If anyone has any objections, please let me know.

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

From: Devin Ryles <[email protected]>

commit 249cd0a187ed4ef1d0af7f74362cc2791ec5581b upstream.

This patch adds DeviceIDs for Sunrise Point-LP.

---
 drivers/ata/ahci.c |    3 +++
 1 file changed, 3 insertions(+)

@@ -322,6 +322,9 @@ static const struct pci_device_id ahci_p
 	{ PCI_VDEVICE(INTEL, 0x8c87), board_ahci }, /* 9 Series RAID */
 	{ PCI_VDEVICE(INTEL, 0x8c8e), board_ahci }, /* 9 Series RAID */
 	{ PCI_VDEVICE(INTEL, 0x8c8f), board_ahci }, /* 9 Series RAID */
+	{ PCI_VDEVICE(INTEL, 0x9d03), board_ahci }, /* Sunrise Point-LP AHCI */
+	{ PCI_VDEVICE(INTEL, 0x9d05), board_ahci }, /* Sunrise Point-LP RAID */
+	{ PCI_VDEVICE(INTEL, 0x9d07), board_ahci }, /* Sunrise Point-LP RAID */
 	{ PCI_VDEVICE(INTEL, 0xa103), board_ahci }, /* Sunrise Point-H AHCI */
 	{ PCI_VDEVICE(INTEL, 0xa103), board_ahci }, /* Sunrise Point-H RAID */
 	{ PCI_VDEVICE(INTEL, 0xa105), board_ahci }, /* Sunrise Point-H RAID */

3.14-stable review patch. If anyone has any objections, please let me know.

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

From: Tejun Heo <[email protected]>

commit 2b21ef0aae65f22f5ba86b13c4588f6f0c2dbefb upstream.

Just like 0x1600 which got blacklisted by 66a7cbc303f4 ("ahci: disable MSI instead of NCQ on Samsung pci-e SSDs on macbooks"), 0xa800 chokes on NCQ commands if MSI is enabled. Disable MSI.

---
 drivers/ata/ahci.c |    1 +
 1 file changed, 1 insertion(+)

@@ -496,6 +496,7 @@ static const struct pci_device_id ahci_p
 	 * enabled.  https://bugzilla.kernel.org/show_bug.cgi?id=60731
 	 */
 	{ PCI_VDEVICE(SAMSUNG, 0x1600), board_ahci_nomsi },
+	{ PCI_VDEVICE(SAMSUNG, 0xa800), board_ahci_nomsi },
 
 	/* Enmotus */
 	{ PCI_DEVICE(0x1c44, 0x8000), board_ahci },

3.14-stable review patch. If anyone has any objections, please let me know.

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

From: Dmitry Torokhov <[email protected]>

commit aad0b624129709c94c2e19e583b6053520353fa8 upstream.

irq_of_parse_and_map() returns 0 on error (the result is unsigned int), so testing for negative result never works.

---
 drivers/ata/sata_fsl.c |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

@@ -1503,7 +1503,7 @@ static int sata_fsl_probe(struct platfor
 	host_priv->csr_base = csr_base;
 
 	irq = irq_of_parse_and_map(ofdev->dev.of_node, 0);
-	if (irq < 0) {
+	if (!irq) {
 		dev_err(&ofdev->dev, "invalid irq from platform\n");
 		goto error_exit_with_cleanup;
 	}

3.14-stable review patch. If anyone has any objections, please let me know.

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

From: lucien <[email protected]>

[ Upstream commit 20ea60ca9952bd19d4b0d74719daba305aef5178 ]

Now the vti_link_ops do not point the .dellink, for fb tunnel device (ip_vti0), the net_device will be removed as the default .dellink is unregister_netdevice_queue,but the tunnel still in the tunnel list, then if we add a new vti tunnel, in ip_tunnel_find():

        hlist_for_each_entry_rcu(t, head, hash_node) {
                if (local == t->parms.iph.saddr &&
                    remote == t->parms.iph.daddr &&
                    link == t->parms.link &&
==>                 type == t->dev->type &&
                    ip_tunnel_key_match(&t->parms, flags, key))
                        break;
        }

the panic will happen, cause dev of ip_tunnel *t is null:

[ 3835.072977] IP: [<ffffffffa04103fd>] ip_tunnel_find+0x9d/0xc0 [ip_tunnel]
[ 3835.073008] PGD b2c21067 PUD b7277067 PMD 0
[ 3835.073008] Oops: 0000 [#1] SMP
.....
[ 3835.073008] Stack:
[ 3835.073008]  ffff8800b72d77f0 ffffffffa0411924 ffff8800bb956000 ffff8800b72d78e0
[ 3835.073008]  ffff8800b72d78a0 0000000000000000 ffffffffa040d100 ffff8800b72d7858
[ 3835.073008]  ffffffffa040b2e3 0000000000000000 0000000000000000 0000000000000000
[ 3835.073008] Call Trace:
[ 3835.073008]  [<ffffffffa0411924>] ip_tunnel_newlink+0x64/0x160 [ip_tunnel]
[ 3835.073008]  [<ffffffffa040b2e3>] vti_newlink+0x43/0x70 [ip_vti]
[ 3835.073008]  [<ffffffff8150d4da>] rtnl_newlink+0x4fa/0x5f0
[ 3835.073008]  [<ffffffff812f68bb>] ? nla_strlcpy+0x5b/0x70
[ 3835.073008]  [<ffffffff81508fb0>] ? rtnl_link_ops_get+0x40/0x60
[ 3835.073008]  [<ffffffff8150d11f>] ? rtnl_newlink+0x13f/0x5f0
[ 3835.073008]  [<ffffffff81509cf4>] rtnetlink_rcv_msg+0xa4/0x270
[ 3835.073008]  [<ffffffff8126adf5>] ? sock_has_perm+0x75/0x90
[ 3835.073008]  [<ffffffff81509c50>] ? rtnetlink_rcv+0x30/0x30
[ 3835.073008]  [<ffffffff81529e39>] netlink_rcv_skb+0xa9/0xc0
[ 3835.073008]  [<ffffffff81509c48>] rtnetlink_rcv+0x28/0x30

....

modprobe ip_vti
ip link del ip_vti0 type vti
ip link add ip_vti0 type vti
rmmod ip_vti

do that one or more times, kernel will panic.

fix it by assigning ip_tunnel_dellink to vti_link_ops' dellink, in which we skip the unregister of fb tunnel device. do the same on ip6_vti.

---
 net/ipv4/ip_vti.c  |    1 +
 net/ipv6/ip6_vti.c |   11 +++++++++++
 2 files changed, 12 insertions(+)

@@ -369,6 +369,7 @@ static struct rtnl_link_ops vti_link_ops
 	.validate	= vti_tunnel_validate,
 	.newlink	= vti_newlink,
 	.changelink	= vti_changelink,
+	.dellink        = ip_tunnel_dellink,
 	.get_size	= vti_get_size,
 	.fill_info	= vti_fill_info,
 };

@@ -825,6 +825,15 @@ static int vti6_newlink(struct net *src_
 	return vti6_tnl_create2(dev);
 }
 
+static void vti6_dellink(struct net_device *dev, struct list_head *head)
+{
+	struct net *net = dev_net(dev);
+	struct vti6_net *ip6n = net_generic(net, vti6_net_id);
+
+	if (dev != ip6n->fb_tnl_dev)
+		unregister_netdevice_queue(dev, head);
+}
+
 static int vti6_changelink(struct net_device *dev, struct nlattr *tb[],
 			   struct nlattr *data[])
 {

@@ -900,6 +909,7 @@ static struct rtnl_link_ops vti6_link_op
 	.setup		= vti6_dev_setup,
 	.validate	= vti6_validate,
 	.newlink	= vti6_newlink,
+	.dellink	= vti6_dellink,
 	.changelink	= vti6_changelink,
 	.get_size	= vti6_get_size,
 	.fill_info	= vti6_fill_info,

@@ -945,6 +955,7 @@ static int __net_init vti6_init_net(stru
 	if (!ip6n->fb_tnl_dev)
 		goto err_alloc_dev;
 	dev_net_set(ip6n->fb_tnl_dev, net);
+	ip6n->fb_tnl_dev->rtnl_link_ops = &vti6_link_ops;
 
 	err = vti6_fb_tnl_dev_init(ip6n->fb_tnl_dev);
 	if (err < 0)

3.14-stable review patch. If anyone has any objections, please let me know.

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

From: Yuri Chislov <[email protected]>

[ Upstream commit be6572fdb1bfbe23b2624d477de50af50b02f5d6 ]

When using GRE redirection in WCCP, it sets the wrong skb->protocol, that is, ETH_P_IP instead of ETH_P_IPV6 for the encapuslated traffic.

---
 net/ipv6/ip6_gre.c |    4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

@@ -508,11 +508,11 @@ static int ip6gre_rcv(struct sk_buff *sk
 
 		skb->protocol = gre_proto;
 		/* WCCP version 1 and 2 protocol decoding.
-		 * - Change protocol to IP
+		 * - Change protocol to IPv6
 		 * - When dealing with WCCPv2, Skip extra 4 bytes in GRE header
 		 */
 		if (flags == 0 && gre_proto == htons(ETH_P_WCCP)) {
-			skb->protocol = htons(ETH_P_IP);
+			skb->protocol = htons(ETH_P_IPV6);
 			if ((*(h + offset) & 0xF0) != 0x40)
 				offset += 4;
 		}

3.14-stable review patch. If anyone has any objections, please let me know.

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

From: Marcelo Leitner <[email protected]>

[ Upstream commit 00c83b01d58068dfeb2e1351cca6fccf2a83fa8f ]

Currently, when trying to reuse a socket, vxlan_sock_add will grab vn->sock_lock, locate a reusable socket, inc refcount and release vn->sock_lock.

But vxlan_sock_release() will first decrement refcount, and then grab that lock. refcnt operations are atomic but as currently we have deferred works which hold vs->refcnt each, this might happen, leading to a use after free (specially after vxlan_igmp_leave):

CPU 1 CPU 2

deferred work vxlan_sock_add
... ...

                                   spin_lock(&vn->sock_lock)
                                   vs = vxlan_find_sock();
  vxlan_sock_release
    dec vs->refcnt, reaches 0
    spin_lock(&vn->sock_lock)
                                   vxlan_sock_hold(vs), refcnt=1
                                   spin_unlock(&vn->sock_lock)
    hlist_del_rcu(&vs->hlist);
    vxlan_notify_del_rx_port(vs)
    spin_unlock(&vn->sock_lock)

So when we look for a reusable socket, we check if it wasn't freed already before reusing it.

---
 drivers/net/vxlan.c |   10 +++-------
 1 file changed, 3 insertions(+), 7 deletions(-)

@@ -2106,9 +2106,8 @@ static int vxlan_init(struct net_device
 	spin_lock(&vn->sock_lock);
 	vs = vxlan_find_sock(dev_net(dev), ipv6 ? AF_INET6 : AF_INET,
 			     vxlan->dst_port);
-	if (vs) {
+	if (vs && atomic_add_unless(&vs->refcnt, 1, 0)) {
 		/* If we have a socket with same port already, reuse it */
-		atomic_inc(&vs->refcnt);
 		vxlan_vs_add_dev(vs, vxlan);
 	} else {
 		/* otherwise make new socket outside of RTNL */

@@ -2574,12 +2573,9 @@ struct vxlan_sock *vxlan_sock_add(struct
 
 	spin_lock(&vn->sock_lock);
 	vs = vxlan_find_sock(net, ipv6 ? AF_INET6 : AF_INET, port);
-	if (vs) {
-		if (vs->rcv == rcv)
-			atomic_inc(&vs->refcnt);
-		else
+	if (vs && ((vs->rcv != rcv) ||
+		   !atomic_add_unless(&vs->refcnt, 1, 0)))
 			vs = ERR_PTR(-EBUSY);
-	}
 	spin_unlock(&vn->sock_lock);
 
 	if (!vs)

3.14-stable review patch. If anyone has any objections, please let me know.

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

From: Thadeu Lima de Souza Cascardo <[email protected]>

[ Upstream commit a620a6bc1c94c22d6c312892be1e0ae171523125 ]

If TX channels are set to 4 and RX channels are set to less than 4, using ethtool -L, the driver will try to initialize more RX channels than it has allocated, causing an oops.

This fix only initializes the RX ring if it has been allocated.

---
 drivers/net/ethernet/broadcom/tg3.c |    3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

@@ -8548,7 +8548,8 @@ static int tg3_init_rings(struct tg3 *tp
 		if (tnapi->rx_rcb)
 			memset(tnapi->rx_rcb, 0, TG3_RX_RCB_RING_BYTES(tp));
 
-		if (tg3_rx_prodring_alloc(tp, &tnapi->prodring)) {
+		if (tnapi->prodring.rx_std &&
+		    tg3_rx_prodring_alloc(tp, &tnapi->prodring)) {
 			tg3_free_rings(tp);
 			return -ENOMEM;
 		}

3.14-stable review patch. If anyone has any objections, please let me know.

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

From: Jack Morgenstein <[email protected]>

[ Upstream commit 2d5c57d7fbfaa642fb7f0673df24f32b83d9066c ]

Some VF drivers use the upper byte of "param1" (the qp count field) in mlx4_qp_reserve_range() to pass flags which are used to optimize the range allocation.

Under the current code, if any of these flags are set, the 32-bit count field yields a count greater than 2^24, which is out of range, and this VF fails.

As these flags represent a "best-effort" allocation hint anyway, they may safely be ignored. Therefore, the PF driver may simply mask out the bits.

---
 drivers/net/ethernet/mellanox/mlx4/resource_tracker.c |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

@@ -1456,7 +1456,7 @@ static int qp_alloc_res(struct mlx4_dev
 
 	switch (op) {
 	case RES_OP_RESERVE:
-		count = get_param_l(&in_param);
+		count = get_param_l(&in_param) & 0xffffff;
 		align = get_param_h(&in_param);
 		err = mlx4_grant_resource(dev, slave, RES_QP, count, 0);
 		if (err)

3.14-stable review patch. If anyone has any objections, please let me know.

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

From: Nicolas Dichtel <[email protected]>

[ Upstream commit e0ebde0e131b529fd721b24f62872def5ec3718c ]

rtnl_link_get_net() holds a reference on the 'struct net', we need to release it in case of error.

---
 net/core/rtnetlink.c |    1 +
 1 file changed, 1 insertion(+)

@@ -1453,6 +1453,7 @@ static int do_setlink(const struct sk_bu
 			goto errout;
 		}
 		if (!netlink_ns_capable(skb, net->user_ns, CAP_NET_ADMIN)) {
+			put_net(net);
 			err = -EPERM;
 			goto errout;
 		}

3.14-stable review patch. If anyone has any objections, please let me know.

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

From: Tom Herbert <[email protected]>

[ Upstream commit 6fb2a756739aa507c1fd5b8126f0bfc2f070dc46 ]

Set the inner mac header to point to the GRE payload when doing GRO. This is needed if we proceed to send the packet through GRE GSO which now uses the inner mac header instead of inner network header to determine the length of encapsulation headers.

---
 net/ipv4/gre_offload.c |    3 +++
 1 file changed, 3 insertions(+)

@@ -271,6 +271,9 @@ static int gre_gro_complete(struct sk_bu
 		err = ptype->callbacks.gro_complete(skb, nhoff + grehlen);
 
 	rcu_read_unlock();
+
+	skb_set_inner_mac_header(skb, nhoff + grehlen);
+
 	return err;
 }
 

3.14-stable review patch. If anyone has any objections, please let me know.

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

From: willy tarreau <[email protected]>

[ Upstream commit aebea2ba0f7495e1a1c9ea5e753d146cb2f6b845 ]

The mvneta driver sets the amount of Tx coalesce packets to 16 by default. Normally that does not cause any trouble since the driver uses a much larger Tx ring size (532 packets). But some sockets might run with very small buffers, much smaller than the equivalent of 16 packets. This is what ping is doing for example, by setting SNDBUF to 324 bytes rounded up to 2kB by the kernel.

The problem is that there is no documented method to force a specific packet to emit an interrupt (eg: the last of the ring) nor is it possible to make the NIC emit an interrupt after a given delay.

In this case, it causes trouble, because when ping sends packets over its raw socket, the few first packets leave the system, and the first 15 packets will be emitted without an IRQ being generated, so without the skbs being freed. And since the socket's buffer is small, there's no way to reach that amount of packets, and the ping ends up with "send: no buffer available" after sending 6 packets. Running with 3 instances of ping in parallel is enough to hide the problem, because with 6 packets per instance, that's 18 packets total, which is enough to grant a Tx interrupt before all are sent.

The original driver in the LSP kernel worked around this design flaw by using a software timer to clean up the Tx descriptors. This timer was slow and caused terrible network performance on some Tx-bound workloads (such as routing) but was enough to make tools like ping work correctly.

Instead here, we simply set the packet counts before interrupt to 1. This ensures that each packet sent will produce an interrupt. NAPI takes care of coalescing interrupts since the interrupt is disabled once generated.

No measurable performance impact nor CPU usage were observed on small nor large packets, including when saturating the link on Tx, and this fixes tools like ping which rely on too small a send buffer. If one wants to increase this value for certain workloads where it is safe to do so, "ethtool -C $dev tx-frames" will override this default setting.

This fix needs to be applied to stable kernels starting with 3.10.

---
 drivers/net/ethernet/marvell/mvneta.c |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

@@ -213,7 +213,7 @@
 /* Various constants */
 
 /* Coalescing */
-#define MVNETA_TXDONE_COAL_PKTS		16
+#define MVNETA_TXDONE_COAL_PKTS		1
 #define MVNETA_RX_COAL_PKTS		32
 #define MVNETA_RX_COAL_USEC		100
 

3.14-stable review patch. If anyone has any objections, please let me know.

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

From: Eric Dumazet <[email protected]>

[ Upstream commit 5f478b41033606d325e420df693162e2524c2b94 ]

mvneta_tx() dereferences skb to get skb->len too late, as hardware might have completed the transmit and TX completion could have freed the skb from another cpu.

---
 drivers/net/ethernet/marvell/mvneta.c |    3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

@@ -1612,6 +1612,7 @@ static int mvneta_tx(struct sk_buff *skb
 	u16 txq_id = skb_get_queue_mapping(skb);
 	struct mvneta_tx_queue *txq = &pp->txqs[txq_id];
 	struct mvneta_tx_desc *tx_desc;
+	int len = skb->len;
 	struct netdev_queue *nq;
 	int frags = 0;
 	u32 tx_cmd;

@@ -1675,7 +1676,7 @@ out:
 
 		u64_stats_update_begin(&stats->syncp);
 		stats->tx_packets++;
-		stats->tx_bytes  += skb->len;
+		stats->tx_bytes  += len;
 		u64_stats_update_end(&stats->syncp);
 	} else {
 		dev->stats.tx_dropped++;

3.14-stable review patch. If anyone has any objections, please let me know.

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

From: Kan Liang <[email protected]>

commit 338b522ca43cfd32d11a370f4203bcd089c6c877 upstream.

With -cpu host, KVM reports LBR and extra_regs support, if the host has support.

When the guest perf driver tries to access LBR or extra_regs MSR, it #GPs all MSR accesses,since KVM doesn't handle LBR and extra_regs support. So check the related MSRs access right once at initialization time to avoid the error access at runtime.

For reproducing the issue, please build the kernel with CONFIG_KVM_INTEL = y (for host kernel).
And CONFIG_PARAVIRT = n and CONFIG_KVM_GUEST = n (for guest kernel). Start the guest with -cpu host.
Run perf record with --branch-any or --branch-filter in guest to trigger LBR Run perf stat offcore events (E.g. LLC-loads/LLC-load-misses ...) in guest to trigger offcore_rsp #GP

---
 arch/x86/kernel/cpu/perf_event.c       |    3 +
 arch/x86/kernel/cpu/perf_event.h       |   12 +++---
 arch/x86/kernel/cpu/perf_event_intel.c |   66 ++++++++++++++++++++++++++++++++-
 3 files changed, 75 insertions(+), 6 deletions(-)

@@ -118,6 +118,9 @@ static int x86_pmu_extra_regs(u64 config
 			continue;
 		if (event->attr.config1 & ~er->valid_mask)
 			return -EINVAL;
+		/* Check if the extra msrs can be safely accessed*/
+		if (!er->extra_msr_access)
+			return -ENXIO;
 
 		reg->idx = er->idx;
 		reg->config = event->attr.config1;

@@ -293,14 +293,16 @@ struct extra_reg {
 	u64			config_mask;
 	u64			valid_mask;
 	int			idx;  /* per_xxx->regs[] reg index */
+	bool			extra_msr_access;
 };
 
 #define EVENT_EXTRA_REG(e, ms, m, vm, i) {	\
-	.event = (e),		\
-	.msr = (ms),		\
-	.config_mask = (m),	\
-	.valid_mask = (vm),	\
-	.idx = EXTRA_REG_##i,	\
+	.event = (e),			\
+	.msr = (ms),			\
+	.config_mask = (m),		\
+	.valid_mask = (vm),		\
+	.idx = EXTRA_REG_##i,		\
+	.extra_msr_access = true,	\
 	}
 
 #define INTEL_EVENT_EXTRA_REG(event, msr, vm, idx)	\

3.14-stable review patch. If anyone has any objections, please let me know.

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

From: Todd Fujinaka <[email protected]>

commit aec653c43b0c55667355e26d7de1236bda9fb4e3 upstream.

Call igb_setup_link() when the PHY is powered up.

---
 drivers/net/ethernet/intel/igb/igb_main.c |    2 ++
 1 file changed, 2 insertions(+)

@@ -1613,6 +1613,8 @@ void igb_power_up_link(struct igb_adapte
 		igb_power_up_phy_copper(&adapter->hw);
 	else
 		igb_power_up_serdes_link_82575(&adapter->hw);
+
+	igb_setup_link(&adapter->hw);
 }
 
 /**

3.14-stable review patch. If anyone has any objections, please let me know.

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

From: Anton Blanchard <[email protected]>

commit 152d44a853e42952f6c8a504fb1f8eefd21fd5fd upstream.

I used some 64 bit instructions when adding the 32 bit getcpu VDSO function. Fix it.

---
 arch/powerpc/kernel/vdso32/getcpu.S |    4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

@@ -30,8 +30,8 @@
 V_FUNCTION_BEGIN(__kernel_getcpu)
   .cfi_startproc
 	mfspr	r5,SPRN_USPRG3
-	cmpdi	cr0,r3,0
-	cmpdi	cr1,r4,0
+	cmpwi	cr0,r3,0
+	cmpwi	cr1,r4,0
 	clrlwi  r6,r5,16
 	rlwinm  r7,r5,16,31-15,31-0
 	beq	cr0,1f

3.14-stable review patch. If anyone has any objections, please let me know.

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

From: Ronald Wahl <[email protected]>

commit 4f031fa9f188b2b0641ac20087d9e16bcfb4e49d upstream.

Commit 7ec7c4a9a686c608315739ab6a2b0527a240883c (mac80211: port CCMP to cryptoapi's CCM driver) introduced a regression when decrypting empty packets (data_len == 0). This will lead to backtraces like:

(scatterwalk_start) from [<c01312f4>] (scatterwalk_map_and_copy+0x2c/0xa8) (scatterwalk_map_and_copy) from [<c013a5a0>] (crypto_ccm_decrypt+0x7c/0x25c) (crypto_ccm_decrypt) from [<c032886c>] (ieee80211_aes_ccm_decrypt+0x160/0x170) (ieee80211_aes_ccm_decrypt) from [<c031c628>] (ieee80211_crypto_ccmp_decrypt+0x1ac/0x238) (ieee80211_crypto_ccmp_decrypt) from [<c032ef28>] (ieee80211_rx_handlers+0x870/0x1d24) (ieee80211_rx_handlers) from [<c0330c7c>] (ieee80211_prepare_and_rx_handle+0x8a0/0x91c) (ieee80211_prepare_and_rx_handle) from [<c0331260>] (ieee80211_rx+0x568/0x730) (ieee80211_rx) from [<c01d3054>] (__carl9170_rx+0x94c/0xa20) (__carl9170_rx) from [<c01d3324>] (carl9170_rx_stream+0x1fc/0x320) (carl9170_rx_stream) from [<c01cbccc>] (carl9170_usb_tasklet+0x80/0xc8) (carl9170_usb_tasklet) from [<c00199dc>] (tasklet_hi_action+0x88/0xcc) (tasklet_hi_action) from [<c00193c8>] (__do_softirq+0xcc/0x200) (__do_softirq) from [<c0019734>] (irq_exit+0x80/0xe0)
(irq_exit) from [<c0009c10>] (handle_IRQ+0x64/0x80)
(handle_IRQ) from [<c000c3a0>] (__irq_svc+0x40/0x4c)
(__irq_svc) from [<c0009d44>] (arch_cpu_idle+0x2c/0x34)

Such packets can appear for example when using the carl9170 wireless driver because hardware sometimes generates garbage when the internal FIFO overruns.

This patch adds an additional length check.

---
 net/mac80211/aes_ccm.c |    3 +++
 1 file changed, 3 insertions(+)

@@ -54,6 +54,9 @@ int ieee80211_aes_ccm_decrypt(struct cry
 
 	memset(&aead_req, 0, sizeof(aead_req));
 
+	if (data_len == 0)
+		return -EINVAL;
+
 	sg_init_one(&pt, data, data_len);
 	sg_init_one(&assoc, &aad[2], be16_to_cpup((__be16 *)aad));
 	sg_init_table(ct, 2);

3.14-stable review patch. If anyone has any objections, please let me know.

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

From: Takashi Iwai <[email protected]>

commit f62f5eff3d40a56ad1cf0d81a6cac8dd8743e8a1 upstream.

The same fixup to enable EAPD is needed for ASUS Z99He with AD1986A codec like another ASUS machine.

---
 sound/pci/hda/patch_analog.c |    1 +
 1 file changed, 1 insertion(+)

@@ -333,6 +333,7 @@ static const struct hda_fixup ad1986a_fi
 
 static const struct snd_pci_quirk ad1986a_fixup_tbl[] = {
 	SND_PCI_QUIRK(0x103c, 0x30af, "HP B2800", AD1986A_FIXUP_LAPTOP_IMIC),
+	SND_PCI_QUIRK(0x1043, 0x1443, "ASUS Z99He", AD1986A_FIXUP_EAPD),
 	SND_PCI_QUIRK(0x1043, 0x1447, "ASUS A8JN", AD1986A_FIXUP_EAPD),
 	SND_PCI_QUIRK_MASK(0x1043, 0xff00, 0x8100, "ASUS P5", AD1986A_FIXUP_3STACK),
 	SND_PCI_QUIRK_MASK(0x1043, 0xff00, 0x8200, "ASUS M2", AD1986A_FIXUP_3STACK),

3.14-stable review patch. If anyone has any objections, please let me know.

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

From: Takashi Iwai <[email protected]>

commit fedb2245cbb8d823e449ebdd48ba9bb35c071ce0 upstream.

The built-in mic boost volume gets almost muted after suspend/resume on Lenovo Ideapad S210.

Bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=88121

---
 sound/pci/hda/patch_realtek.c |    1 +
 1 file changed, 1 insertion(+)

@@ -4554,6 +4554,7 @@ static const struct snd_pci_quirk alc269
 	SND_PCI_QUIRK(0x17aa, 0x2212, "Thinkpad", ALC269_FIXUP_LIMIT_INT_MIC_BOOST),
 	SND_PCI_QUIRK(0x17aa, 0x2214, "Thinkpad", ALC269_FIXUP_LIMIT_INT_MIC_BOOST),
 	SND_PCI_QUIRK(0x17aa, 0x2215, "Thinkpad", ALC269_FIXUP_LIMIT_INT_MIC_BOOST),
+	SND_PCI_QUIRK(0x17aa, 0x3977, "IdeaPad S210", ALC283_FIXUP_INT_MIC),
 	SND_PCI_QUIRK(0x17aa, 0x3978, "IdeaPad Y410P", ALC269_FIXUP_NO_SHUTUP),
 	SND_PCI_QUIRK(0x17aa, 0x5013, "Thinkpad", ALC269_FIXUP_LIMIT_INT_MIC_BOOST),
 	SND_PCI_QUIRK(0x17aa, 0x501a, "Thinkpad", ALC283_FIXUP_INT_MIC),

3.14-stable review patch. If anyone has any objections, please let me know.

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

From: Takashi Iwai <[email protected]>

commit 66139a48cee1530c91f37c145384b4ee7043f0b7 upstream.

In snd_usbmidi_error_timer(), the driver tries to resubmit MIDI input URBs to reactivate the MIDI stream, but this causes the error when some of URBs are still pending like:

 WARNING: CPU: 0 PID: 0 at ../drivers/usb/core/urb.c:339 usb_submit_urb+0x5f/0x70()
 URB ef705c40 submitted while active
 CPU: 0 PID: 0 Comm: swapper/0 Not tainted 3.16.6-2-desktop #1
 Hardware name: FOXCONN TPS01/TPS01, BIOS 080015  03/23/2010
  c0984bfa f4009ed4 c078deaf f4009ee4 c024c884 c09a135c f4009f00 00000000
  c0984bfa 00000153 c061ac4f c061ac4f 00000009 00000001 ef705c40 e854d1c0
  f4009eec c024c8d3 00000009 f4009ee4 c09a135c f4009f00 f4009f04 c061ac4f
 Call Trace:
  [<c0205df6>] try_stack_unwind+0x156/0x170
  [<c020482a>] dump_trace+0x5a/0x1b0
  [<c0205e56>] show_trace_log_lvl+0x46/0x50
  [<c02049d1>] show_stack_log_lvl+0x51/0xe0
  [<c0205eb7>] show_stack+0x27/0x50
  [<c078deaf>] dump_stack+0x45/0x65
  [<c024c884>] warn_slowpath_common+0x84/0xa0
  [<c024c8d3>] warn_slowpath_fmt+0x33/0x40
  [<c061ac4f>] usb_submit_urb+0x5f/0x70
  [<f7974104>] snd_usbmidi_submit_urb+0x14/0x60 [snd_usbmidi_lib]
  [<f797483a>] snd_usbmidi_error_timer+0x6a/0xa0 [snd_usbmidi_lib]
  [<c02570c0>] call_timer_fn+0x30/0x130
  [<c0257442>] run_timer_softirq+0x1c2/0x260
  [<c0251493>] __do_softirq+0xc3/0x270
  [<c0204732>] do_softirq_own_stack+0x22/0x30
  [<c025186d>] irq_exit+0x8d/0xa0
  [<c0795228>] smp_apic_timer_interrupt+0x38/0x50
  [<c0794a3c>] apic_timer_interrupt+0x34/0x3c
  [<c0673d9e>] cpuidle_enter_state+0x3e/0xd0
  [<c028bb8d>] cpu_idle_loop+0x29d/0x3e0
  [<c028bd23>] cpu_startup_entry+0x53/0x60
  [<c0bfac1e>] start_kernel+0x415/0x41a

For avoiding these errors, check the pending URBs and skip resubmitting such ones.

---
 sound/usb/midi.c |    2 ++
 1 file changed, 2 insertions(+)

@@ -364,6 +364,8 @@ static void snd_usbmidi_error_timer(unsi
 		if (in && in->error_resubmit) {
 			in->error_resubmit = 0;
 			for (j = 0; j < INPUT_URBS; ++j) {
+				if (atomic_read(&in->urbs[j]->use_count))
+					continue;
 				in->urbs[j]->dev = umidi->dev;
 				snd_usbmidi_submit_urb(in->urbs[j], GFP_ATOMIC);
 			}

Build results:

	total: 137 pass: 137 fail: 0

Qemu tests:

	total: 30 pass: 30 fail: 0

Details are available at http://server.roeck-us.net:8010/builders.

Guenter

Compiled and booted on my test system. No dmesg regressions.

-- Shuah

Shuah Khan
Sr. Linux Kernel Developer
Samsung Open Source Group
Samsung Research America (Silicon Valley)
[email protected] | (970) 217-8978