Denys Vlasenko wrote:
On Monday 24 January 2011 21:03, Peter Korsgaard wrote:
"Jim" == Cathey, Jim <[email protected]> writes:
This thread started by Michal, who's trying to use a very slow
serial-over-jtag setup on microblaze, and had issues with characters
getting lost when getty started up (because of the tcflush), so we
wanted to add a tcdrain() instead.
Unfortunately, Michal did not provide much of information:
what is the speed of his serial link?
There is no HW setting about baudrate in HW. It is Jtag hw console -
xilinx mdm IP. It is compatible with uartlite Kernel driver.
http://www.xilinx.com/support/documentation/ip_documentation/mdm.pdf
how many characters
are lost on average?
Every boot is different but approximately 200 chars are lost.
How he starts getty?
getty: ttyUL0::respawn:/bin/getty -L 50 ttyUL0 vt100 (of course I
changed it to 0, 50 and 300 according to libbb/speed_table.c)
What architecture and CPU?
architecture: Microblaze (I think you don't need to know which exactly
version/configuration I used) 90MHz (Some people use 25MHz too but without
console on mdm)
Anything?
More info in https://lkml.org/lkml/2011/1/16/13
or the first on linux serial because I thought that the problem is in
Peter's uarlite Linux kernel driver
http://www.spinics.net/lists/linux-serial/msg03173.html
His last mail says that sleep(1) doesn't help, which I find
hard to believe.
sleep just postpone flush. As you can see in that lkml thread. Peter
suggested to try tcdrain instead of tcflush based on my debugging/output
messages.
CPIO is available at
http://www.monstr.eu/20110125-busybox-rootfs.cpio
Some logs below.
Thanks,
Michal
1. Current behavior:
[ 42.297299] < 0.417384> Running local start scripts.
[ 42.499939] < 0.202640> Mounting sysfs:
[ 42.912690] < 0.412751> mdev: initialising /dev
[ 43.120934] < 0.208244> Mounting /etc/config:
[ 43.523944] < 0.403010> Populating /etc/config:
[ 44.528967] < 1.005023> flatfsd: Created 8 configuration files (168 net eth0: Promiscuous mode
disabled.
[ 44.529072] < 0.000105> bytes)
[ 44.736594] < 0.207522> Mounting devpts:
[ 45.137225] < 0.400631> Setting hostname:
[ 45.536748] < 0.399523> Bringing up network interfaces:
[ 45.739614] < 0.202866> udhcpc (v1.18.2) started
[ 46.137376] < 0.397762> Sending discover.
[ 47.229770] < 1.092394> Xilinx-SP605-LTP-full-WB-pc-msr-next-22-watch-wt-mdm
login:
2. sleep(3):
[ 45.882825] < 0.405516> Bringing up network interfaces:
[ 46.082834] < 0.200009> udhcpc (v1.18.2) started
[ 46.483630] < 0.400796> Sending discover...
[ 46.884336] < 0.400643> Sending select for 192.168.0.100...
[ 47.482714] < 0.598378> Lease of 192.168.0.100 obtained, lease time 7200
[ 47.878229] < 0.395515> Starting portmap:
[ 47.878281] < 0.000052>
[ 47.878298] < 0.000017> Welcome to
[ 48.474868] < 0.596570> _____ _ _ _
[ 48.877371] < 0.402503> | ___ \ | | | | (_)
[ 49.264710] < 0.387339> | |_/ / ___ | |_ __ _ |
[ 50.058137] < 0.793427> Xilinx-SP605-LTP-full-WB-pc-msr-next-22-watch-wt-mdm
login:
3. with tcdrain(from origin our patch)
[ 47.092812] < 0.591759> Lease of 192.168.0.100 obtained, lease time 7200
[ 47.332540] < 0.239728> Starting portmap:
[ 47.332649] < 0.000109>
[ 47.567536] < 0.234887> Welcome to
[ 48.045421] < 0.477885> _____ _ _ _
[ 48.670993] < 0.625572> | ___ \ | | | | (_)
[ 49.266218] < 0.595225> | |_/ / ___ | |_ __ _ | | _ _ __ _ _ __ __
[ 49.970985] < 0.704767> | __/ / _ \| __| / _` || | | || '_ \ | | | |\ \/ /
[ 50.494000] < 0.523015> | | | __/| |_ | (_| || |____| || | | || |_| | > <
[ 51.198863] < 0.704863> \_| \___| \__| \__,_|\_____/|_||_| |_| \__,_|/_/\_\
[ 51.198969] < 0.000106>
[ 51.714875] < 0.515906> on
Xilinx-SP605-LTP-full-WB-pc-msr-next-22-watch-wt-mdm
[ 51.714921] < 0.000046>
[ 52.621779] < 0.702838> Xilinx-SP605-LTP-full-WB-pc-msr-next-22-watch-wt-mdm
login:
4. with tcdrain with alarms has the same behavior as above.
--
Michal Simek, Ing. (M.Eng)
w: www.monstr.eu p: +42-0-721842854
Maintainer of Linux kernel 2.6 Microblaze Linux - http://www.monstr.eu/fdt/
Microblaze U-BOOT custodian
_______________________________________________
busybox mailing list
[email protected]
http://lists.busybox.net/mailman/listinfo/busybox
