how to set rss correctly, can let all queue receive packets ?
this is my rss configure as follows:
.rxmode.mq_mode = ETH_MQ_RX_RSS;
.rx_adv_conf.rss_conf.rss_key = NULL; .rx_adv_conf.rss_conf.rss_hf = ETH_RSS_IP | ETH_RSS_UDP | ETH_RSS_TCP;
.txmode.mq_mode = ETH_MQ_TX_NONE;
but only queue 0 receive packets, other queues display 0 packets;
command as follows:
/root/dpdk-18.11/app/test-pmd/testpmd -w 0000:3c:00.0 -l 0-9 -n 4 -- --rxq=10 --txq=10 --rxd=256 --txd=4096 -i --nb-cores=9
testpmd> show port xstats all
NIC extended statistics for port 0
rx_good_packets: 56
tx_good_packets: 18
rx_good_bytes: 4932
tx_good_bytes: 1767
rx_missed_errors: 0
rx_errors: 0
tx_errors: 0
rx_mbuf_allocation_errors: 0
rx_q0packets: 56
rx_q0bytes: 4932
rx_q0errors: 0
rx_q1packets: 0
rx_q1bytes: 0
rx_q1errors: 0
rx_q2packets: 0
rx_q2bytes: 0
rx_q2errors: 0
rx_q3packets: 0
rx_q3bytes: 0
rx_q3errors: 0
rx_q4packets: 0
rx_q4bytes: 0
rx_q4errors: 0
rx_q5packets: 0
rx_q5bytes: 0
rx_q5errors: 0
rx_q6packets: 0
rx_q6bytes: 0
rx_q6errors: 0
rx_q7packets: 0
rx_q7bytes: 0
rx_q7errors: 0
rx_q8packets: 0
rx_q8bytes: 0
rx_q8errors: 0
rx_q9packets: 0
rx_q9bytes: 0
rx_q9errors: 0
Related
Homework is to modify this script to take exec as an argument, but first I want to be able to run the script to try to figure out how to modify it
tcsh $ cat foreach_1
#!/bin/tcsh
# routine to zero-fill argv to 20 arguments
#
set buffer = (0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0)
set count = 1
#
if ($#argv > 20) goto toomany
#
foreach argument ($argv[*])
set buffer[$count] = $argument
# count++
end
# REPLACE command ON THE NEXT LINE WITH
# THE PROGRAM YOU WANT TO CALL.
exec command $buffer[*]
#
toomany:
echo "Too many arguments given."
echo "Usage: foreach_1 [up to 20 arguments]"
exit 1
But I get this error when trying to run it:
./foreach_1: line 5: syntax error near unexpected token `('
./foreach_1: line 5: `set buffer = (0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0)'
I don't have any extra quotes, so why is this happening?
In many shells (and I believe that tcsh is counted among the bourne compatibles), you must place the left-hand side of the expression, the =, and the right-hand side all directly adjacent to one another.
# shorten the ` = ` to `=` below:
set buffer=(0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0)
set count=1
if ($#argv > 20) goto toomany
#
foreach argument ($argv[*])
set buffer[$count] = $argument
# count++
end
# REPLACE command ON THE NEXT LINE WITH
# THE PROGRAM YOU WANT TO CALL.
exec command $buffer[*]
#
toomany:
echo "Too many arguments given."
echo "Usage: foreach_1 [up to 20 arguments]"
exit 1
I'm using Arch Linux. I have read about systemd, and as I understand it, systemd is the first process, and it starts the rest of the processes.
But when I use:
ps -aux
The result shows that /sbin/init has PID 1. And when I use:
pstree -Apn
The result shows that systemd has PID 1. Which is correct? Is /sbin/init starting systemd?
They're probably both right.
$ sudo ls -ltrh /proc/1/exe
[sudo] password for user:
lrwxrwxrwx 1 root root 0 May 30 21:22 /proc/1/exe -> /lib/systemd/systemd
$ echo $(tr '\0' ' ' < /proc/1/cmdline )
/sbin/init splash
$ stat /sbin/init
File: '/sbin/init' -> '/lib/systemd/systemd'
Size: 20 Blocks: 0 IO Block: 4096 symbolic link
Device: 801h/2049d Inode: 527481 Links: 1
Access: (0777/lrwxrwxrwx) Uid: ( 0/ root) Gid: ( 0/ root)
Access: 2017-05-30 21:27:12.058023583 -0500
Modify: 2016-10-26 08:04:58.000000000 -0500
Change: 2016-11-19 11:38:45.749226284 -0600
Birth: -
The commands above show us:
what is the file corresponding to pid 1's executable image?
what was invoked (passed to exec()) when pid 1 was started?
what are the characteristics of the path at /sbin/init?
On my system, /sbin/init is a symlink to "/lib/systemd/systemd". This is likely similar to your system. We can see what information ps -aux is using by straceing it.
$ strace ps -aux
...
open("/proc/1/cmdline", O_RDONLY) = 6
read(6, "/sbin/init\0splash\0", 131072) = 18
read(6, "", 131054) = 0
close(6) = 0
...
and likewise for pstree:
$ strace pstree -Apn
...
getdents(3, /* 332 entries */, 32768) = 8464
open("/proc/1/stat", O_RDONLY) = 4
stat("/proc/1", {st_mode=S_IFDIR|0555, st_size=0, ...}) = 0
fstat(4, {st_mode=S_IFREG|0444, st_size=0, ...}) = 0
read(4, "1 (systemd) S 0 1 1 0 -1 4194560"..., 8192) = 192
read(4, "", 7168) = 0
open("/proc/1/task", O_RDONLY|O_NONBLOCK|O_DIRECTORY|O_CLOEXEC) = 5
...
So the difference in output is because they use different sources of information. /proc/1/cmdline tells us how the process was invoked. Whereas /proc/1/stat shows that the process' name is systemd.
$ cat /proc/1/stat
1 (systemd) S 0 1 1 0 -1 4194560 34371 596544 1358 3416 231 144 298 1758 20 0 1 0 4 190287872 772 18446744073709551615 1 1 0 0 0 0 671173123 4096 1260 0 0 0 17 2 0 0 12188 0 0 0 0 0 0 0 0 0 0
I use linux perf tool to collect branch info of programs, and the command and result is as follow:
$ sudo perf record -b /bin/ls
Error:
No hardware sampling interrupt available.
No APIC? If so then you can boot the kernel with the "lapic" boot parameter to force-enable it.
the content in /pro/cpuinfo is below:
processor : 0
vendor_id : GenuineIntel
cpu family : 6
model : 23
model name : Intel(R) Xeon(R) CPU E5405 # 2.00GHz
stepping : 10
microcode : 0xa07
cpu MHz : 1994.921
cache size : 6144 KB
physical id : 0
siblings : 4
core id : 0
cpu cores : 4
apicid : 0
initial apicid : 0
fpu : yes
fpu_exception : yes
cpuid level : 13
wp : yes
flags : fpu vme de pse tsc msr pae mce cx8 **apic** sep mtrr pge mca cmov pat pse36 clflush dts acpi strong text mmx fxsr sse sse2 ss ht tm
pbe syscall nx lm constant_tsc arch_perfmon pebs **bts** rep_good nopl aperfmperf pni dtes64 monitor ds_cpl vmx tm2 ssse3 cx16 xtpr pdcm
dca sse4_1 xsave lahf_lm dtherm tpr_shadow vnmi flexpriority
bugs :
bogomips : 3989.84
clflush size : 64
cache_alignment : 64
address sizes : 38 bits physical, 48 bits virtual
power management:
apic and bts in flags entry is strengthened(I want but just encapsuled by "**") and I don't know what else is import for this case. And the other 7 processors are same to processor 0.
The boot parameter "lapic" is added by modifying /boot/grub/grub.cfg:
menuentry 'Ubuntu' --class ubuntu --class gnu-linux --class gnu --class os $menuentry_id_option 'gnulinux-simple-0ed8a872-4eb7-4339-a0bb-6c0033da582e' {
recordfail
load_video
gfxmode $linux_gfx_mode
insmod gzio
insmod part_msdos
insmod ext2
set root='hd0,msdos1'
if [ x$feature_platform_search_hint = xy ]; then
search --no-floppy --fs-uuid --set=root --hint-bios=hd0,msdos1 --hint-efi=hd0,msdos1 --hint-baremetal=ahci0,msdos1 ced80bc6-08a9-4909-9717-97658cf0c4fd
else
search --no-floppy --fs-uuid --set=root ced80bc6-08a9-4909-9717-97658cf0c4fd
fi
linux /vmlinuz-4.2.0-42-generic root=/dev/mapper/fedora_hustyong-root ro **lapic** quiet splash $vt_handoff
initrd /initrd.img-4.2.0-42-generic
}
just add lapic in linux entry.
But no sense after rebooting.
My questions:
1) What does the error info means?
2) Does the perf tool branch sampling use Intel Branch Trace Store(BTS)? Or Last Branch Record(LBR)?
3) How can I look up the LBR support?
4) what is different of the LBR and BTS support between x86 32bit and 64bit?
My OS is Ubuntu 14.04 64bit:
$ uname -a
Linux user-S5000VSA 4.2.0-42-generic #49~14.04.1-Ubuntu SMP Wed Jun 29 20:22:11 UTC 2016 x86_64 x86_64 x86_64 GNU/Linux
The perf install instructions:
$ sudo apt-get install linux-tools-common
$ sudo apt-get install linux-tools-4.2.0-27-generic linux-cloud-tools-4.2.0-27-generic
update:
the content of /proc/interrupts:
CPU0 CPU1 CPU2 CPU3 CPU4 CPU5 CPU6 CPU7
0: 127780 52084 127784 126729 127706 128431 127785 126822 IO-APIC 2-edge timer
1: 52 42 3 2 62 49 5 2 IO-APIC 1-edge i8042
8: 0 0 0 0 0 0 0 1 IO-APIC 8-edge rtc0
9: 0 0 0 0 0 0 0 0 IO-APIC 9-fasteoi acpi
12: 1428 1307 52 47 1424 1324 53 58 IO-APIC 12-edge i8042
14: 0 0 0 0 0 0 0 0 IO-APIC 14-edge ata_piix
15: 0 0 0 0 0 0 0 0 IO-APIC 15-edge ata_piix
17: 47 276 1004 49 52 295 993 50 IO-APIC 17-fasteoi radeon
20: 31062 4201 7533 29935 31080 4297 7540 29824 IO-APIC 20-fasteoi ata_piix
22: 0 0 0 0 0 0 0 0 IO-APIC 22-fasteoi uhci_hcd:usb3, uhci_hcd:usb5
23: 0 0 0 0 0 0 0 0 IO-APIC 23-fasteoi ehci_hcd:usb1, uhci_hcd:usb2, uhci_hcd:usb4
25: 2 755654 3 3 1 1 3 6 PCI-MSI 2621440-edge eth0
27: 0 0 0 0 0 0 0 1 PCI-MSI 131072-edge ioat-msi
NMI: 6756 678 6894 5867 861 2168 4994 3700 Non-maskable interrupts
LOC: 343554 578094 1736638 773135 219952 777567 1459249 689292 Local timer interrupts
SPU: 0 0 0 0 0 0 0 0 Spurious interrupts
PMI: 6756 678 6894 5867 861 2168 4994 3700 Performance monitoring interrupts
IWI: 6756 678 6894 5867 861 2168 4994 3700 IRQ work interrupts
RTR: 0 0 0 0 0 0 0 0 APIC ICR read retries
RES: 82594 294601 142535 259797 77845 316210 84927 261455 Rescheduling interrupts
CAL: 4749 9296 7358 31330 7560 8564 5751 20364 Function call interrupts
TLB: 5933 2044 12867 11215 6563 4682 8669 8272 TLB shootdowns
TRM: 0 0 0 0 0 0 0 0 Thermal event interrupts
THR: 0 0 0 0 0 0 0 0 Threshold APIC interrupts
DFR: 0 0 0 0 0 0 0 0 Deferred Error APIC interrupts
MCE: 0 0 0 0 0 0 0 0 Machine check exceptions
MCP: 292 292 292 292 292 292 292 292 Machine check polls
HYP: 0 0 0 0 0 0 0 0 Hypervisor callback interrupts
ERR: 0
MIS: 0
PIN: 0 0 0 0 0 0 0 0 Posted-interrupt notification event
PIW: 0 0 0 0 0 0 0 0 Posted-interrupt wakeup event
I install ubuntu 16.10 64bit in my PC and run perf record -b successfully. I think maybe it's wrong in kernel or linux-tools-4.2.0-27-generic or linux-cloud-tools-4.2.0-27-generic package.
I am using ReadFile to read a null character '0x00', but it is not able to see it.
If I send "0x01 0x02 0x00 0x03" to my program. It will receive "0x01 0x02 0x03". I know for a fact that my sending side is sending the null character (I used a scope to see the data).
I have checked the DCB structure and have fNull set to false. (which is supposed to allow for null bytes to be received and not discarded)
This is the configuration of the DCB shown from GetCommState:
DCBlength 28
BaudRate 9600
fBinary 1
fParity 0
fOutxCtsFlow 0
fOutxDsrFlow 0
fDtrControl 0
fDsrSensitivity 0
fTXContinueOnXoff 0
fOutX 1
fInX 1
fErrorChar 0
fNull 0
fRtsControl 0
fAbortOnError 0
fDummy2 0
wReserved 0
XonLim 2048
XoffLim 512
ByteSize 7
Parity 0
StopBits 2
XonChar 0 '\0'
XoffChar 0 '\0'
ErrorChar 0 '\0'
EofChar 0 '\0'
EvtChar 0 '\0'
wReserved1 0
Am I missing something in the structure that might be blocking the null character as well? or is it possible it's a driver configuration or registry configuration that I need to change?
You have XON/XOFF flow control enabled and have specified null as the XON/XOFF characters. That is why you are not seeing null data bytes. You need to set the fOutX and fInX fields to 0 instead of 1.
In Solaris, the output of 'netstat -i' gives something like the following:
root# netstat -i
Name Mtu Net/Dest Address Ipkts Ierrs Opkts Oerrs Collis Queue
lo0 8232 loopback localhost 136799 0 136799 0 0 0
igb0 1500 vulture vulture 1272272 0 347277 0 0 0
Note that there is a Queue field on the end.
In Linux, 'netstat -i' gives output with no Queue field:
[root#roseate ~]# netstat -i
Kernel Interface table
Iface MTU Met RX-OK RX-ERR RX-DRP RX-OVR TX-OK TX-ERR TX-DRP TX-OVR Flg
eth0 1500 0 2806170 0 0 0 791768 0 0 0 BMRU
eth1 1500 0 0 0 0 0 0 0 0 0 BMU
eth2 1500 0 0 0 0 0 0 0 0 0 BMU
eth3 1500 0 0 0 0 0 0 0 0 0 BMU
lo 16436 0 1405318 0 0 0 1405318 0 0 0 LRU
I've figured out how to get collisions in Linux by adding the -e option, but is there a way to get the Queue in Linux?
The only reference to queue I ever saw in netstat on Linux was when using -s, but that's probably too garrulous for your use-case?
$ netstat -na | awk 'BEGIN { RecvQ=0; SendQ=0; } { RecvQ+=$2; SendQ+=$3; } END { print "RecvQ " RecvQ/1024; print "SendQ " SendQ/1024; }'
RecvQ 255.882
SendQ 0.0507812
For per interface, I have dirty way
[spatel#us04 ~]$ for qw in `/sbin/ifconfig | grep 'inet addr:' | cut -d: -f2 | awk '{print $1}'`; do echo `/sbin/ip addr | grep $qw | awk '{print $7}'` : ; echo `netstat -na | grep $qw | awk 'BEGIN { RecvQ=0; SendQ=0; } { RecvQ+=$2; SendQ+=$3; } END { print "RecvQ " RecvQ/1024; print "SendQ " SendQ/1024; }'`; done
eth0 :
RecvQ 0 SendQ 0
eth2 :
RecvQ 0.0703125 SendQ 1.56738
:
RecvQ 0 SendQ 0
I ended up using
tc -s -d qdisc
[root#roseate ~]# tc -s -d qdisc
qdisc mq 0: dev eth2 root
Sent 0 bytes 0 pkt (dropped 0, overlimits 0 requeues 0)
rate 0bit 0pps backlog 0b 0p requeues 0
qdisc mq 0: dev eth3 root
Sent 0 bytes 0 pkt (dropped 0, overlimits 0 requeues 0)
rate 0bit 0pps backlog 0b 0p requeues 0
qdisc mq 0: dev eth0 root
Sent 218041403 bytes 1358829 pkt (dropped 0, overlimits 0 requeues 1)
rate 0bit 0pps backlog 0b 0p requeues 1
qdisc mq 0: dev eth1 root
Sent 0 bytes 0 pkt (dropped 0, overlimits 0 requeues 0)
rate 0bit 0pps backlog 0b 0p requeues 0
which gives backlog bytes and packets.
Source