Re: [Xen-devel] [linux-arm-xen test] 58849: regressions - FAIL

[Stefano Stabellini <stefano.stabellini@xxxxxxxxxxxxx>]

On Wed, 24 Jun 2015, Ian Campbell wrote:
> On Wed, 2015-06-24 at 06:03 +0000, osstest service user wrote:
> > flight 58849 linux-arm-xen real [real]
> > http://logs.test-lab.xenproject.org/osstest/logs/58849/
> > 
> > Regressions :-(
> > 
> > Tests which did not succeed and are blocking,
> > including tests which could not be run:
> >  test-armhf-armhf-xl-cubietruck 11 guest-start             fail REGR. vs. 
> > 58830
> 
> This was:
> http://logs.test-lab.xenproject.org/osstest/logs/58849/test-armhf-armhf-xl-cubietruck/cubietruck-braque---var-log-kern.log
> 
> Jun 24 04:09:13 cubietruck-braque kernel: [  807.637687] ------------[ cut 
> here ]------------
> Jun 24 04:09:13 cubietruck-braque kernel: [  807.637756] kernel BUG at 
> drivers/xen/grant-table.c:923!
> Jun 24 04:09:13 cubietruck-braque kernel: [  807.637784] Internal error: Oops 
> - BUG: 0 [#1] SMP ARM
> Jun 24 04:09:13 cubietruck-braque kernel: [  807.637810] Modules linked in: 
> xen_gntalloc bridge stp ipv6 llc brcmfmac brcmutil cfg80211
> Jun 24 04:09:13 cubietruck-braque kernel: [  807.637899] CPU: 0 PID: 16206 
> Comm: vif1.0-q0-guest Not tainted 3.16.7-ckt12+ #1
> Jun 24 04:09:13 cubietruck-braque kernel: [  807.637936] task: c12fc480 ti: 
> d2d3c000 task.ti: d2d3c000
> Jun 24 04:09:13 cubietruck-braque kernel: [  807.637977] PC is at 
> gnttab_batch_copy+0xd4/0xe0
> Jun 24 04:09:13 cubietruck-braque kernel: [  807.638004] LR is at 
> gnttab_batch_copy+0x1c/0xe0
> Jun 24 04:09:13 cubietruck-braque kernel: [  807.638030] pc : [<c04abf7c>]    
> lr : [<c04abec4>]    psr: a0000013
> Jun 24 04:09:13 cubietruck-braque kernel: [  807.638030] sp : d2d3deb0  ip : 
> deadbeef  fp : d2d3df3c
> Jun 24 04:09:13 cubietruck-braque kernel: [  807.638091] r10: 00000001  r9 : 
> 00000000  r8 : 00000008
> Jun 24 04:09:13 cubietruck-braque kernel: [  807.638124] r7 : 00000001  r6 : 
> 00000001  r5 : 00000000  r4 : e1e38d30
> Jun 24 04:09:13 cubietruck-braque kernel: [  807.638159] r3 : 00000001  r2 : 
> deadbeef  r1 : deadbeef  r0 : fffffff2
> Jun 24 04:09:13 cubietruck-braque kernel: [  807.638193] Flags: NzCv  IRQs on 
>  FIQs on  Mode SVC_32  ISA ARM  Segment kernel
> Jun 24 04:09:13 cubietruck-braque kernel: [  807.638227] Control: 10c5387d  
> Table: 7b50406a  DAC: 00000015
> Jun 24 04:09:13 cubietruck-braque kernel: [  807.638257] Process 
> vif1.0-q0-guest (pid: 16206, stack limit = 0xd2d3c248)
> Jun 24 04:09:13 cubietruck-braque kernel: [  807.638287] Stack: (0xd2d3deb0 
> to 0xd2d3e000)
> Jun 24 04:09:13 cubietruck-braque kernel: [  807.638316] dea0:                
>                      00000001 00000000 00000000 e1e30000
> Jun 24 04:09:13 cubietruck-braque kernel: [  807.638353] dec0: 00000001 
> c05d7c44 0000003e 00000ec2 d2d3df3c 00000000 00000000 00000001
> Jun 24 04:09:13 cubietruck-braque kernel: [  807.638391] dee0: dbbb7a80 
> 00000000 00000000 00000008 00000000 d2d3df20 e1e38cfc e1e38d30
> Jun 24 04:09:13 cubietruck-braque kernel: [  807.638430] df00: 00000001 
> 00000000 00000001 00000000 e1e38d30 e1e63530 0000003e 00000208
> Jun 24 04:09:13 cubietruck-braque kernel: [  807.638468] df20: d9f0f480 
> d9f0f480 00000001 00000000 d2d3df2c d2d3df34 d2d3df34 00000000
> Jun 24 04:09:13 cubietruck-braque kernel: [  807.638505] df40: 00000000 
> db34e380 00000000 e1e30000 c05d776c 00000000 00000000 00000000
> Jun 24 04:09:13 cubietruck-braque kernel: [  807.638543] df60: 00000000 
> c0264138 00000000 00000000 00000000 e1e30000 00000000 00000000
> Jun 24 04:09:13 cubietruck-braque kernel: [  807.638581] df80: d2d3df80 
> d2d3df80 00000000 00000000 d2d3df90 d2d3df90 d2d3dfac db34e380
> Jun 24 04:09:13 cubietruck-braque kernel: [  807.638638] dfa0: c026406c 
> 00000000 00000000 c020f038 00000000 00000000 00000000 00000000
> Jun 24 04:09:13 cubietruck-braque kernel: [  807.638686] dfc0: 00000000 
> 00000000 00000000 00000000 00000000 00000000 00000000 00000000
> Jun 24 04:09:13 cubietruck-braque kernel: [  807.638723] dfe0: 00000000 
> 00000000 00000000 00000000 00000013 00000000 00000000 00000000
> Jun 24 04:09:13 cubietruck-braque kernel: [  807.638785] [<c04abf7c>] 
> (gnttab_batch_copy) from [<c05d7c44>] (xenvif_kthread_guest_rx+0x4d8/0xbc0)
> Jun 24 04:09:13 cubietruck-braque kernel: [  807.638841] [<c05d7c44>] 
> (xenvif_kthread_guest_rx) from [<c0264138>] (kthread+0xcc/0xe8)
> Jun 24 04:09:13 cubietruck-braque kernel: [  807.638887] [<c0264138>] 
> (kthread) from [<c020f038>] (ret_from_fork+0x14/0x3c)
> Jun 24 04:09:13 cubietruck-braque kernel: [  807.638929] Code: 0affffe5 
> eaffffed e8bd80f8 e7f001f2 (e7f001f2) 
> Jun 24 04:09:13 cubietruck-braque kernel: [  807.638978] ---[ end trace 
> 98c74482d9a5771d ]---
> 
> Which looks familiar, although I can't seem to find it, does anyone
> remember it? Are we missing a backport perhaps?
> 
> This is the 3.16.y based linux-arm-xen tree, which was recently updated
> from a baseline of v3.16.4-ckt7 to v3.16.7-ckt12 (flight 58830) in both
> cases plus "xen_arch_need_swiotlb" for swiotlb stuff.
> 
> This here was the next flight which only added the "xen: netback: read
> hotplug script once at start of day.", which I don't think is related to
> the failure, which I suspect is intermittent.

The only explanation that I can come up with is that GNTTABOP_query_size
returns different max_nr_frames values at the time
gnttab_setup_auto_xlat_frames is called and when
boot_max_nr_grant_frames is initialized.

I wonder what would happen if:


diff --git a/drivers/xen/grant-table.c b/drivers/xen/grant-table.c
index b1c7170..77eae32 100644
--- a/drivers/xen/grant-table.c
+++ b/drivers/xen/grant-table.c
@@ -636,7 +636,7 @@ EXPORT_SYMBOL_GPL(gnttab_max_grant_frames);
 int gnttab_setup_auto_xlat_frames(phys_addr_t addr)
 {
        xen_pfn_t *pfn;
-       unsigned int max_nr_gframes = __max_nr_grant_frames();
+       unsigned int max_nr_gframes = gnttab_max_grant_frames();
        unsigned int i;
        void *vaddr;
 

_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxx
http://lists.xen.org/xen-devel