Menu
▾
▴
Re: [E1000-devel] WARNING triggered by e100 driver
|
From: Ronciak, J. <joh...@in...> - 2013-03-19 17:20:52
|
Hi Alan, You found the right people for looking into this. Indeed this looks to be an issue for one of the two calls. Since we are going to have to recode the area around the call a bit please give a bit to get a test version ready. Once we versify that it works we'll push upstream and give you credit for find the bug. Since I'm travelling on business this week I'll have some ready for you next week. Sorry for the delay. If I get a chance while I'm traveling I'll do so. Thanks. Cheers, John > -----Original Message----- > From: Alan Stern [mailto:st...@ro...] > Sent: Friday, March 15, 2013 9:08 AM > To: Kirsher, Jeffrey T; Brandeburg, Jesse; Allan, Bruce W > Cc: e1000-devel Mailing List > Subject: [E1000-devel] WARNING triggered by e100 driver > > [The Intel ethernet drivers have lots of maintainers listed, so I > arbitrarily addressed this to the first three names on the list.] > > Ever since this check was added to the DMA API, I have been getting the > following WARNING at boot: > > [ 45.843379] ------------[ cut here ]------------ > [ 45.844008] WARNING: at lib/dma-debug.c:933 > check_unmap+0x5de/0x64d() > [ 45.844008] Hardware name: HP dx2000 MT (EE004AA) > [ 45.844008] e100 0000:01:08.0: DMA-API: device driver failed to > check map error[device address=0x000000002c6ecf32] [size=42 bytes] > [mapped as single] > [ 45.844008] Modules linked in: snd_intel8x0 snd_ac97_codec ac97_bus > evdev snd_pcm pcspkr snd_timer snd soundcore e100 snd_page_alloc mii > i915 ohci_hcd(O) fbcon font bitblit softcursor drm_kms_helper drm fb > ehci_pci(O) fbdev ehci_hcd(O) cfbcopyarea i2c_algo_bit fan i2c_core > button uhci_hcd(O) video processor backlight thermal_sys usbcore > usb_common cfbimgblt cfbfillrect > [ 45.844008] Pid: 526, comm: nash-hotplug Tainted: G W O > 3.9.0-rc1AS-00031-gc7b496b-dirty #22 > [ 45.844008] Call Trace: > [ 45.844008] [<c1022aca>] warn_slowpath_common+0x63/0x7a > [ 45.844008] [<c11473c4>] ? check_unmap+0x5de/0x64d > [ 45.844008] [<c1022b45>] warn_slowpath_fmt+0x26/0x2a > [ 45.844008] [<c11473c4>] check_unmap+0x5de/0x64d > [ 45.844008] [<c10573d8>] ? __lock_acquire+0xc7c/0xc8b > [ 45.844008] [<c11475a4>] debug_dma_unmap_page+0x61/0x69 > [ 45.844008] [<f00df845>] pci_unmap_single+0x4a/0x54 [e100] > [ 45.844008] [<f00e137b>] e100_poll+0x35a/0x3e0 [e100] > [ 45.844008] [<c106994c>] ? rcu_process_callbacks+0x145/0x3c5 > [ 45.844008] [<c11e340d>] net_rx_action+0x65/0x133 > [ 45.844008] [<c10280ad>] __do_softirq+0x96/0x140 > [ 45.844008] [<c1028017>] ? local_bh_enable+0xd/0xd > [ 45.844008] <IRQ> [<c102823a>] ? irq_exit+0x41/0x9a > [ 45.844008] [<c1003447>] ? do_IRQ+0x6b/0x7f > [ 45.844008] [<c1245ff1>] ? common_interrupt+0x31/0x36 > [ 45.844008] ---[ end trace e9161d67ae418510 ]--- > [ 45.844008] Mapped at: > [ 45.844008] [<c114781b>] debug_dma_map_page+0x3a/0xc3 > [ 45.844008] [<f00de435>] pci_map_single+0x6a/0x74 [e100] > [ 45.844008] [<f00e0113>] e100_xmit_prepare+0x6c/0x99 [e100] > [ 45.844008] [<f00df343>] e100_exec_cb+0x6c/0xee [e100] > [ 45.844008] [<f00dfbea>] e100_xmit_frame+0x61/0xd0 [e100] > > The reason for the warning is very simple: e100_xmit_prepare() calls > pci_map_single() but doesn't check the result. Since DMA mapping > requests can fail, this is clearly a bug. There's even a comment about > it in the source code. > > I don't know enough about networking to fix this. Can someone please > take a look at it? > > Thanks, > > Alan Stern > > > ----------------------------------------------------------------------- > ------- > Everyone hates slow websites. So do we. > Make your web apps faster with AppDynamics Download AppDynamics Lite > for free today: > http://p.sf.net/sfu/appdyn_d2d_mar > _______________________________________________ > E1000-devel mailing list > E10...@li... > https://lists.sourceforge.net/lists/listinfo/e1000-devel > To learn more about Intel® Ethernet, visit > http://communities.intel.com/community/wired |
