[U-Boot] stuck after redirecting stdout to nc
Hello,
On uboot loading, I checking a specific environement (called consoleddev), to decide to rediect console to nc or staying in serial. I then see some strange behaviour: if the environment tell uboot to redirect to net console, everything works well, but only if the ethernet cable is connected. if it is not connected then I see that the kernel is not loading.
*relevant environemets: ncip 192.168.3.10 ipaddr 192.168.3.33
I would appreciate your advice,
Regards,
Ran
On Monday 02 January 2012 09:50:42 ran shalit wrote:
On uboot loading, I checking a specific environement (called consoleddev), to decide to rediect console to nc or staying in serial. I then see some strange behaviour: if the environment tell uboot to redirect to net console, everything works well, but only if the ethernet cable is connected. if it is not connected then I see that the kernel is not loading.
there could easily be a bug where netconsole does not work if the network is not available. patches welcome to add timeouts. -mike
On Fri, Jan 6, 2012 at 1:26 AM, Mike Frysinger vapier@gentoo.org wrote:
On Monday 02 January 2012 09:50:42 ran shalit wrote:
On uboot loading, I checking a specific environement (called
consoleddev),
to decide to rediect console to nc or staying in serial. I then see some strange behaviour: if the environment tell uboot to redirect to net console, everything
works
well, but only if the ethernet cable is connected. if it is not connected then I see that the kernel is not loading.
there could easily be a bug where netconsole does not work if the network is not available. patches welcome to add timeouts. -mike
Do you think such bug is independent of architecture type ?
Dear Ran Shalit,
In message CAJ2oMhKLcdT_NFjbfcJ4O7kGz+B-TK93MWynb4=F02iLnBB=sQ@mail.gmail.com you wrote:
Do you think such bug is independent of architecture type ?
This is impossible to say without knowing where and what the bug is.
Best regards,
Wolfgang Denk
On Fri, Jan 6, 2012 at 11:18 AM, Wolfgang Denk wd@denx.de wrote:
Dear Ran Shalit,
In message <CAJ2oMhKLcdT_NFjbfcJ4O7kGz+B-TK93MWynb4=F02iLnBB= sQ@mail.gmail.com> you wrote:
Do you think such bug is independent of architecture type ?
This is impossible to say without knowing where and what the bug is.
Best regards,
Wolfgang Denk
You're right.
What I did see in my tests is as follows: 1. only if there is connection to PC and the PC's interface ip is the same as the ncip environement, then and only then, I see that the u-boot finish its job and starts the bootcmd, and loading the kernel. I also see the u-boot prints in the ethernet if I start netconsole, but event without starting netconsole, the kernel starts (as expected) 2. Otherwise, if the network interface ip is not the same as ncip, or cable is disconnected the u-boot does not start bootcmd, and I actually do not know where it stoped and why.
Please tell me if you have any advice where to look, and what should be checked,
Best Regards, Ran
Dear Ran Shalit,
In message CAJ2oMh+D0MKGXYiAvjy-uGA3DCssJPCn=N0Fm5vC7j4h34XQ3g@mail.gmail.com you wrote:
- Otherwise, if the network interface ip is not the same as ncip, or cable
is disconnected the u-boot does not start bootcmd, and I actually do not know where it stoped and why.
Please tell me if you have any advice where to look, and what should be checked,
Attach a JTAG debugger and fire up GDB.
Content-Type: text/html; charset=ISO-8859-1 Content-Transfer-Encoding: quoted-printable
And please STOP posting HTML!!
Best regards,
Wolfgang Denk
participants (4)
-
Mike Frysinger -
Ran Shalit -
ran shalit
-
Wolfgang Denk