This is the mail archive of the
ecos-discuss@sourceware.cygnus.com
mailing list for the eCos project.
Release testing.
- To: ecos-discuss at sourceware dot cygnus dot com
- Subject: [ECOS] Release testing.
- From: "Patrick O'Grady" <patrick at softprocess dot com>
- Date: Mon, 7 Feb 2000 08:53:47 -0800 (PST)
Hi, All--
I've got a new version of the i386/PC HAL which I'd like to post... it
has drivers for the serial ports (special thanks to the ARM AEB which also
uses 16550s), has a bug in calling DSRs fixed, includes break character
support for GDB so you can stop your program by pressing ^C, configures
the A20 line properly, supports software reset, and has a much cleaner
installation. But before posting, I've got two questions:
- The 16550As have 16-character FIFOs, but the serial_DSR function I
borrowed from the AEB HAL only supports transmitting and receiving one
character at a time. Is there a more efficient way of sending more than
one character than putting a loop around the (chan->callbacks->xmt_char)
and (chan->callbacks->rcv_char) in the serial port DSR function? Fewer
function calls==better performance, and probably isn't that hard to do.
- Release testing--there are a bunch of test programs which come with the
OS, each of which individually tests a particular facility. I can run
these individually and show that the parts of the system are working...
which is fine the first time, but becomes really cumbersome if I want to
post an update. Is there a wide-coverage soak test which I can run to
verify that my update hasn't really broken anything? The tm_basic test
code seems to be the most thorough so far... is there anything better?
It's always a drag posting something that you *know* needs more testing...
Let me know what you think. Cheers...
-patrick