Jan,

i updated to 3.9.3 from stable branch

*Dom0:*
#hv04: xl info
host                   : hv04.edss.local
release                : 3.9.3-1.g0b5d8f5-xen
version                : #1 SMP Mon May 20 09:43:24 UTC 2013 (0b5d8f5)
machine                : x86_64
nr_cpus                : 24
max_cpu_id             : 23
nr_nodes               : 2
cores_per_socket       : 6
threads_per_core       : 2
cpu_mhz                : 2666
hw_caps                : 
bfebfbff:2c100800:00000000:00003f40:029ee3ff:00000000:00000001:000
00000
virt_caps              : hvm hvm_directio
total_memory           : 98294
free_memory            : 92985
sharing_freed_memory   : 0
sharing_used_memory    : 0
free_cpus              : 0
xen_major              : 4
xen_minor              : 2
xen_extra              : .2_01-240.2
xen_caps               : xen-3.0-x86_64 xen-3.0-x86_32p hvm-3.0-x86_32 
hvm-3.0-x86_32p hvm-3.0-x86_64 
xen_scheduler          : credit
xen_pagesize           : 4096
platform_params        : virt_start=0xffff800000000000
xen_changeset          : 26064
xen_commandline        : vga=mode-0x314 dom0_max_vcpus=8 
dom0_mem=4G,max:4G msi=on iommu=on console=ttyS0,115200 xen-
netback.netback_max_groups=6
cc_compiler            : gcc (SUSE Linux) 4.7.2 20130108 [gcc-4_7-branch 
revision 195012
cc_compile_by          : abuild
cc_compile_domain      : 
cc_compile_date        : Fri May 10 15:05:06 UTC 2013
xend_config_format     : 4


*Both DomU:*
test:/home/local # uname -a
Linux test.edss.local 3.9.3-1.g0b5d8f5-xen #1 SMP Mon May 20 09:43:24 
UTC 2013 (0b5d8f5) x86_64 x86_64 x86_64 GNU/Linux


*From xl console*

[  4] local 10.251.2.201 port 5001 connected with 10.251.2.202 port 
26946
[ ID] Interval       Transfer     Bandwidth
[  4]  0.0- 2.0 sec  8.38 MBytes  35.2 Mbits/sec
[  4]  2.0- 4.0 sec  90.3 MBytes   379 Mbits/sec
[  4]  4.0- 6.0 sec  7.30 MBytes  30.6 Mbits/sec
[  105.391855] BUG: unable to handle kernel paging request at 
ffff880078c6f000
[  105.391918] IP: [<ffffffffa001a75c>] netif_poll+0x49c/0xe80 [xennet]
[  105.391970] PGD a85067 PUD a95067 PMD 7fc2b067 PTE 
8010000078c6f065
[  105.392029] Oops: 0003 [#1] SMP 
[  105.392058] Modules linked in: iptable_filter ip_tables x_tables af_packet 
hwmon domctl crc32_pclmul crc32c_intel ghash_clmulni_intel aesni_intel 
ablk_helper joydev cryptd lrw aes_x86_64 xts gf128mul autofs4 
scsi_dh_emc scsi_dh_alua scsi_dh_rdac scsi_dh_hp_sw scsi_dh xenblk 
cdrom xennet ata_generic ata_piix
[  105.392319] CPU 0 
[  105.392336] Pid: 0, comm: swapper/0 Not tainted 3.9.3-1.g0b5d8f5-xen 
#1  
[  105.392383] RIP: e030:[<ffffffffa001a75c>]  [<ffffffffa001a75c>] 
netif_poll+0x49c/0xe80 [xennet]
[  105.392450] RSP: e02b:ffff88007b403d18  EFLAGS: 00010286
[  105.392485] RAX: ffff88007da85638 RBX: ffff880078c6eec0 RCX: 
ffff880078c6f000
[  105.392531] RDX: ffff880078b54170 RSI: ffff880078c6eec0 RDI: 
ffff880078479280
[  105.392576] RBP: ffff88007869c6c0 R08: 0000000000000dc0 R09: 
0000000000000000
[  105.392622] R10: 0000000000000000 R11: 0000000000000001 R12: 
0000000000000011
[  105.392667] R13: 000000000000b826 R14: ffff88007b403dd8 R15: 
ffff880078428800
[  105.393781] FS:  00007f11b3900700(0000) GS:ffff88007b400000(0000) 
knlGS:0000000000000000
[  105.394907] CS:  e033 DS: 0000 ES: 0000 CR0: 000000008005003b
[  105.395840] CR2: ffff880078c6f000 CR3: 00000000797f5000 CR4: 
0000000000002660
[  105.395840] DR0: 0000000000000000 DR1: 0000000000000000 DR2: 
0000000000000000
[  105.395840] DR3: 0000000000000000 DR6: 00000000ffff0ff0 DR7: 
0000000000000400
[  105.395840] Process swapper/0 (pid: 0, threadinfo ffffffff808b8000, task 
ffffffff808c3960)
[  105.395840] Stack:
[  105.395840]  ffffffff00000269 0000002200000022 ffff8800784288d0 
ffff88007b403db8
[  105.395840]  ffff88007b403d98 000000150000b815 ffff880078429d50 
0000000000000000
[  105.395840]  ffff880078428858 ffff88007b40e878 0000000000000012 
000000400000b82a
[  105.395840] Call Trace:
[  105.395840]  [<ffffffff804419c0>] net_rx_action+0x170/0x2c0
[  105.395840]  [<ffffffff80036dae>] __do_softirq+0xee/0x230
[  105.395840]  [<ffffffff80037075>] irq_exit+0x95/0xa0
[  105.395840]  [<ffffffff803937cd>] evtchn_do_upcall+0x2ad/0x2f0
[  105.395840]  [<ffffffff80535a0e>] do_hypervisor_callback+0x1e/0x30
[  105.395840]  [<ffffffff800033aa>] HYPERVISOR_sched_op_new+0xa/0x20
[  105.395840]  [<ffffffff8000e671>] xen_idle+0x41/0x110
[  105.395840]  [<ffffffff8000e7ef>] cpu_idle+0xaf/0x110
[  105.395840]  [<ffffffff80946b1f>] start_kernel+0x424/0x42f
[  105.395840] Code: 44 21 ea 48 8d 54 d0 40 8b 87 d8 00 00 00 44 0f bf 
42 06 44 0f b7 4a 02 48 8b 44 01 30 49 63 cc 48 83 c1 03 48 c1 e1 04 48 
01 f1 <48> 89 01 44 89 49 08 44 89 41 0c 48 8b 08 80 e5 80 0f 85 54 09 
[  105.395840] RIP  [<ffffffffa001a75c>] netif_poll+0x49c/0xe80 [xennet]
[  105.395840]  RSP <ffff88007b403d18>
[  105.395840] CR2: ffff880078c6f000
[  105.395840] ---[ end trace 67dcc8bdc485cab5 ]---
[  105.395840] Kernel panic - not syncing: Fatal exception in interrupt


-- 
*/Best regards,/*
/Eugene Istomin/



> >>> On 17.05.13 at 15:00, Eugene Istomin <e.istomin@edss.ee> 
wrote:
> > Bump, here it is:
> Okay, but I think we're still lacking information on what your
> Dom0 kernel is.
> 
> Ian, Wei - looking at the forward ported as well as the upstream
> frontends, I'm wondering if there isn't a fundamental flaw in
> ..._get_responses(): It allows up to MAX_SKB_FRAGS + 1 slots/
> frags, and ..._fill_frags() then fills as many fragments as got
> queued onto the respective list. Only after both of them are done,
> __pskb_pull_tail() gets invoked reducing the fragment count by
> one if the condition is met that made ..._get_responses() bump