Hi Peter,
Peter Turczak wrote:
first of all thanks for this nice release. Most of the annoying bugs of
the last official release are gone, like the problems of incremental
builds of busybox, and some kernel flaws.
Thanks for the feedback.
During experimentation with the truckload of new features i found the
following problems:
- When building busybox telnetd and integrating into inetd a connection
can be established, but will be immediately closed, i used the busybox
login and passwd system,too
This may be related to pty's setup.
If you enable the stand alone telnetd in the network apps
config (and enable the CONFIG_USER_TELNETD_DOES_NOT_USE_OPENPTY
switch along with it) it works on my 5208evb. I haven't tried the
busybox one yet.
- Selecting the system-shell it seems not to be possibe to choose the
busybox msh, either minix-shell and msh both yied a different
executable. This is more or less just cosmetic but it would make
configuring the system more consistent.
That would be nicer... But the problem is that the busybox shell
is configured within the busybox configs (and they are directly
from the busybox package). The other shells are separate packages,
and this switch is there enabler.
- The whole password/auth procedure is not explained very well
The dist essentially leaves that up to the platform build
(so the vendors/<VENDORS>/<BOARD>/Makefile), with some hooks
in amongst the packages involved.
- Maybe it would be smart to put /etc into some ramdisk/flash, too in
order to be able to do dns-lookups after an DHCP session or toy around
with auth
Again that is left upto the vendor/board specific Makefile.
There are various example targets in the tree that do it in
different ways (use a traditional ext2 ramdisk as the rootfs, etc).
- Something about the 520x SPI-Driver seems to be broken, its likely
that will try to fix it next week
Quite likely. I don't actively test the SPI interfaces on any
boards myself.
- The second serial port of my EVB does not work correctly when using
/dev/ttyS1, while /dev/ttyS0 works perfectly.
While porting it to my hardware which is similar to the MCF5208EVB by
Freescale i found out that some SST 39VF64xxxB NOR-Flashes seem to have
broken (or ,,improved'' in Marketing speech) tables. These show both
sector and blocksizes as distinct eraseblock regions. So these add up to
twice the device size. I did a fixup routine for this, which you will
find attached to this mail as a patch for linux-2.6.x/drivers/mtd. Hope
i got the XIP stuff right, it was a little confusing to be honest.
Looks quite involved. Let me have a closer look over it.
Regards
Greg
------------------------------------------------------------------------
Greg Ungerer -- Chief Software Dude EMAIL: [EMAIL PROTECTED]
Secure Computing Corporation PHONE: +61 7 3435 2888
825 Stanley St, FAX: +61 7 3891 3630
Woolloongabba, QLD, 4102, Australia WEB: http://www.SnapGear.com
_______________________________________________
uClinux-dev mailing list
[email protected]
http://mailman.uclinux.org/mailman/listinfo/uclinux-dev
This message was resent by [email protected]
To unsubscribe see:
http://mailman.uclinux.org/mailman/options/uclinux-dev
