Patchwork 8139cp dma-debug warning.

login
register
mail settings
Submitter Dave Jones
Date Aug. 12, 2009, 5:13 p.m.
Message ID <20090812171333.GD3116@redhat.com>
Download mbox | patch
Permalink /patch/31223/
State Not Applicable
Delegated to: David Miller
Headers show

Comments

Dave Jones - Aug. 12, 2009, 5:13 p.m.
On Thu, Aug 06, 2009 at 05:57:02PM -0400, Dave Jones wrote:
 > I'm chasing yet another dma-debug warning where we're unmapping a different
 > size to what we mapped.
 > 
 >  > WARNING: at lib/dma-debug.c:803 check_unmap+0x1f5/0x509() (Not tainted)
 >  > Hardware name: 
 >  > 8139cp 0000:00:03.0: DMA-API: device driver frees DMA memory with different
 >  > size [device address=0x000000001e9f8852] [map size=1536 bytes] [unmap size=1538
 >  > bytes]
 >  > Modules linked in: ipv6 dm_multipath uinput joydev 8139too virtio_balloon
 >  > 8139cp mii i2c_piix4 virtio_pci i2c_core floppy squashfs pata_acpi ata_generic
 >  > [last unloaded: scsi_wait_scan]
 >  > Pid: 0, comm: swapper Not tainted 2.6.31-0.125.rc5.git2.fc12.i686 #1
 >  > Call Trace:
 >  >  [<c0444408>] warn_slowpath_common+0x7b/0xa3
 >  >  [<c0606746>] ? check_unmap+0x1f5/0x509
 >  >  [<c0444499>] warn_slowpath_fmt+0x34/0x48
 >  >  [<c0606746>] check_unmap+0x1f5/0x509
 >  >  [<c04ecb43>] ? check_valid_pointer+0x2c/0x6c
 >  >  [<c0606c00>] debug_dma_unmap_page+0x62/0x7b
 >  >  [<e0dc80be>] dma_unmap_single_attrs.clone.2+0x5a/0x75 [8139cp]
 >  >  [<e0dc8220>] cp_rx_poll+0x147/0x301 [8139cp]
 >  >  [<c077e5d4>] net_rx_action+0xa7/0x1d3
 >  >  [<c044accf>] __do_softirq+0xc8/0x192
 >  >  [<c044ade2>] do_softirq+0x49/0x7f
 >  >  [<c044af36>] irq_exit+0x48/0x8c
 >  >  [<c041c302>] smp_apic_timer_interrupt+0x7a/0x99
 >  >  [<c0404416>] apic_timer_interrupt+0x36/0x3c
 >  >  [<c0425e98>] ? native_safe_halt+0xa/0xc
 >  >  [<c040ad6f>] default_idle+0x55/0x98
 >  >  [<c046e44d>] ? trace_hardirqs_off+0x19/0x2c
 >  >  [<c04029e3>] cpu_idle+0xac/0xcd
 >  >  [<c08080ee>] rest_init+0x66/0x79
 >  >  [<c0a72ae0>] start_kernel+0x36f/0x385
 >  >  [<c0a7207e>] __init_begin+0x7e/0x96
 >  > ---[ end trace f3c3298e5df24f15 ]---
 >  > Mapped at:
 >  >  [<c0606f6b>] debug_dma_map_page+0x6b/0x13b
 >  >  [<e0dc7c22>] dma_map_single_attrs.clone.1+0x78/0x93 [8139cp]
 >  >  [<e0dc86d0>] cp_init_rings+0xaa/0x12c [8139cp]
 >  >  [<e0dc87d6>] cp_open+0x84/0x154 [8139cp]
 >  >  [<c077ff2b>] dev_open+0x99/0xe4
 >  
 > Looking at 8139cp.c I see this code in cp_rx_poll ..
 > 
 >  552                 buflen = cp->rx_buf_sz + NET_IP_ALIGN;
 >  553                 new_skb = netdev_alloc_skb(dev, buflen);
 >  554                 if (!new_skb) {
 >  555                         dev->stats.rx_dropped++;
 >  556                         goto rx_next;
 >  557                 }
 >  558 
 >  559                 skb_reserve(new_skb, NET_IP_ALIGN);
 > 
 > 
 > Aren't we padding the alignment twice here? or am I missing something?
 > The warning was a difference of two bytes, which is NET_IP_ALIGN,
 > so I'm wondering if that skb_reserve line needs to be nuked ?

There's another instance of the same further in the file.

Does this look right?

	Dave

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

Patch

diff --git a/drivers/net/8139cp.c b/drivers/net/8139cp.c
index 07919d0..065c9c3 100644
--- a/drivers/net/8139cp.c
+++ b/drivers/net/8139cp.c
@@ -556,8 +556,6 @@  rx_status_loop:
 			goto rx_next;
 		}
 
-		skb_reserve(new_skb, NET_IP_ALIGN);
-
 		dma_unmap_single(&cp->pdev->dev, mapping,
 				 buflen, PCI_DMA_FROMDEVICE);
 
@@ -1061,8 +1059,6 @@  static int cp_refill_rx(struct cp_private *cp)
 		if (!skb)
 			goto err_out;
 
-		skb_reserve(skb, NET_IP_ALIGN);
-
 		mapping = dma_map_single(&cp->pdev->dev, skb->data,
 					 cp->rx_buf_sz, PCI_DMA_FROMDEVICE);
 		cp->rx_skb[i] = skb;