Re: [Xen-devel] Re: devel/merge.2.6.35-rc3 [WAS: [GIT PULL] pv/pcifront-2.6.32 + pv/pciback-2.6.32 = [INTx/MSI/MSI-X work now]]

[Sander Eikelenboom <linux@xxxxxxxxxxxxxx>]

Hmmm seems i got this one only once, next boots of same domU didn't give this 
warning.
It's still not working, but it seems xhci isoc still isn't in 2.6.35-rc3, so i 
guess i will have to wait for 2.6.36 for that.

I will see if i can try your tree with one of my other domU's with USB 2.0 
cards somewhere this week.
--

Sander


Sunday, June 27, 2010, 7:35:33 PM, you wrote:

> Hi Konrad,

> I have tried the 2.6.35-rc3, because i hoped the xhci (usb3) isoc support 
> would be in by now, but it seems nog fully functional yet :(.
> On boot i did get this, don't know if it's something with xen-swiotlb of with 
> the xhci since both are in the trace :-)

> [    1.329851] Linux video capture interface: v2.00
> [    1.379299] pvrusb2: Hardware description: WinTV PVR USB2 Model 24xxx
> [    1.379476] ------------[ cut here ]------------
> [    1.379489] WARNING: at kernel/smp.c:421 
> smp_call_function_many+0x76/0x1be()
> [    1.379496] Modules linked in: pvrusb2(+) cx2341x v4l2_common videodev 
> v4l1_compat v4l2_compat_ioctl32 tveeprom [last unloaded: scsi_wait_scan]
> [    1.379520] Pid: 1313, comm: modprobe Not tainted 2.6.35-rc3+ #2
> [    1.379526] Call Trace:
> [    1.379534]  [<ffffffff8109ad7b>] ? smp_call_function_many+0x76/0x1be
> [    1.379543]  [<ffffffff81070a5b>] warn_slowpath_common+0x80/0xae
> [    1.379551]  [<ffffffff8105fdf2>] ? do_flush_tlb_all+0x0/0x61
> [    1.379558]  [<ffffffff81070a9e>] warn_slowpath_null+0x15/0x17
> [    1.379566]  [<ffffffff8109ad7b>] smp_call_function_many+0x76/0x1be
> [    1.379573]  [<ffffffff8105fdf2>] ? do_flush_tlb_all+0x0/0x61
> [    1.379580]  [<ffffffff8105fdf2>] ? do_flush_tlb_all+0x0/0x61
> [    1.379587]  [<ffffffff8109aee3>] smp_call_function+0x20/0x24
> [    1.379594]  [<ffffffff81076774>] on_each_cpu+0x18/0x57
> [    1.379601]  [<ffffffff8105fc4b>] flush_tlb_all+0x17/0x19
> [    1.379608]  [<ffffffff8110030e>] __purge_vmap_area_lazy+0x128/0x183
> [    1.379616]  [<ffffffff811024cb>] vm_unmap_aliases+0x162/0x173
> [    1.379624]  [<ffffffff81031d4d>] xen_create_contiguous_region+0x44/0xea
> [    1.379633]  [<ffffffff81284d98>] xen_swiotlb_alloc_coherent+0x6e/0xe2
> [    1.379640]  [<ffffffff81106a23>] dma_pool_alloc+0x191/0x319
> [    1.379648]  [<ffffffff810337f2>] ? check_events+0x12/0x20
> [    1.379655]  [<ffffffff8110dc1b>] ? __kmalloc+0x170/0x182
> [    1.379663]  [<ffffffff8142de91>] ? kzalloc+0xf/0x11
> [    1.379670]  [<ffffffff8142de91>] ? kzalloc+0xf/0x11
> [    1.379677]  [<ffffffff8142df6f>] xhci_segment_alloc+0x38/0x76
> [    1.379684]  [<ffffffff8142e7fc>] xhci_ring_alloc+0x55/0x19d
> [    1.379691]  [<ffffffff8142ea79>] xhci_endpoint_init+0x74/0x4f0
> [    1.379699]  [<ffffffff81277c2a>] ? string+0x4c/0xb2
> [    1.379706]  [<ffffffff8142cfd8>] xhci_add_endpoint+0x1ce/0x21f
> [    1.379714]  [<ffffffff8141aa13>] usb_hcd_alloc_bandwidth+0x256/0x2a9
> [    1.379722]  [<ffffffff8141e125>] usb_set_interface+0xc7/0x253
> [    1.379730]  [<ffffffff8108607c>] ? __create_workqueue_key+0x12d/0x22d
> [    1.379747]  [<ffffffffa002d9ef>] pvr2_hdw_create+0x9af/0xacd [pvrusb2]
> [    1.379754]  [<ffffffff81271a4a>] ? ida_get_new_above+0xfd/0x1b4
> [    1.379762]  [<ffffffff8110d1ec>] ? kmem_cache_alloc_notrace+0xc0/0xe4
> [    1.379775]  [<ffffffffa0028ba2>] ? pvr_setup_attach+0x0/0x26 [pvrusb2]
> [    1.379791]  [<ffffffffa00333ac>] pvr2_context_create+0x10c/0x13f [pvrusb2]
> [    1.379805]  [<ffffffffa0028b4b>] pvr_probe+0x16/0x6d [pvrusb2]
> [    1.379813]  [<ffffffff8141fa0d>] usb_probe_interface+0xe9/0x134
> [    1.379821]  [<ffffffff81359ca3>] driver_probe_device+0xa3/0x151
> [    1.379828]  [<ffffffff81359da9>] __driver_attach+0x58/0x7b
> [    1.379835]  [<ffffffff81359d51>] ? __driver_attach+0x0/0x7b
> [    1.379842]  [<ffffffff81359528>] bus_for_each_dev+0x4e/0x84
> [    1.379849]  [<ffffffff81359b10>] driver_attach+0x1c/0x1e
> [    1.379855]  [<ffffffff81358e3f>] bus_add_driver+0xb8/0x20e
> [    1.379862]  [<ffffffff8135a086>] driver_register+0xb3/0x121
> [    1.379869]  [<ffffffff8141f79d>] usb_register_driver+0x80/0x139
> [    1.379882]  [<ffffffffa0047000>] ? pvr_init+0x0/0xb7 [pvrusb2]
> [    1.379894]  [<ffffffffa004706c>] pvr_init+0x6c/0xb7 [pvrusb2]
> [    1.379903]  [<ffffffff81002088>] do_one_initcall+0x5a/0x171
> [    1.379910]  [<ffffffff8109fa45>] sys_init_module+0x9a/0x1d8
> [    1.379917]  [<ffffffff81036c82>] system_call_fastpath+0x16/0x1b
> [    1.379924] ---[ end trace 7d0fc40935291479 ]---

> --
> Sander



> Tuesday, June 22, 2010, 8:45:21 PM, you wrote:

>> On Wed, Apr 21, 2010 at 10:43:36AM +0200, Sander Eikelenboom wrote:
>>> Is your xen-swiotlb  work also going to be merged with this?

>> So, I never got to answer this. The answer is: eventually. When
>> we get to stabilizing 2.6.34 to work as Dom0, we will be definitely using
>> that.


>>> Guess i will wait for this one to get merged by Jeremy to test and replace 
>>> my dom0 and domU kernels.

>> For the DomU, if you feel adventurous, I've just updated two branches
>> with the latest Xen PCI frontend driver:

>>  pv/merge.2.6.34
>>  devel/merge.2.6.35-rc3

>> The 2.6.35-rc-3 contains the latest of the swiotlb, xen-swiotlb, xen
>> pcifront, etc. Lots of cleanup, merging of patches, etc. to make it
>> nicer for Linux Kernel Mailing List (LKML).

>> The outstanding bugs are:
>>  - pvgrub isn't working well. If you specify 'iommu=soft' pvgrub breaks.
>>  - the xen tools nor the xen pci front specify an PCI region, hence your
>>    guess is limited to 3GB so that you have a memory region below 4GB
>>    for 32bit guest devices.

>>> 
>>> Is there anything you need to be tested upfront ?

>> Oh man, I like that. If you feel you have the time, booting the DomU
>> kernel with any hardware would be appreciated. And try to do funky
>> stuff. I don't think I've completely hardened the driver.







-- 
Best regards,
 Sander                            mailto:linux@xxxxxxxxxxxxxx


_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/xen-devel