✎ soc:2009:asdlkf:project_plan:start

This page is read only. You can view the source, but not change it. Ask your administrator if you think this is wrong.

====== Chris Kluka: Network driver development ======

===== Project Plan =====

==== Summary ====

==== Outline ====

==== Milestones and Timeline ====
Given that drivers for this card exist for BSD and Linux, implementing a
driver for gPXE should not be too difficult. After looking through the gPXE API
and some of the other drivers in gpxe/src/drivers/net/, I feel confident that I
would be able to implement the 5 core functions of the gPXE API. 

I speculate that almost (if not) all code will be written in C. The
possibility exists to compose a few snippets in ASM to support one of last
year's SOC project which was based around code optimization. However, at this
point, I have not been able to identify any specific areas of the driver which
would benefit significantly from assembler optimization. 

Then, once the driver is implemented, (time pending) I would write a basic
testing class and polish some documentation around the driver. As driver
implementation is relatively straight forward, I do not expect more than 3-4
pages of documentation for this task. 

Timeline:

May 4-20: Read through the gPXE code base, familiarize self with project and
project's expectations

May 23: begin creating class diagrams of existing drivers. This is to
include the Linux, the BSD, and (if possible), some other OS's drivers Begin
designing driver outline.

May 23: Compose emails to the authors of the previous drivers requesting any
documentation on hardware specifics. I was unable to find a datasheet on this
device, but was able to find a datasheet on what I believe is its IC
(http://people.freebsd.org/~wpaul/SysKonnect/xmacii_datasheet_rev_c_9-29.pdf).
I will have to verify that this is the correct datasheet.

May 25: Completed class diagram of at least one of the above OS's.

May 30: Completed class diagram of at least two of the above. Specific
attention to hardware API is the most important part. 

June 4: Hopefully by this point I have collected sufficient information to
compose a definitive hardware API for this device. 

June 4: Compose a list of functions available on the card as well as DMA
call documentation. Basically, have a completed summary of how the card wants
to be 'talked to'. 

June 5-9: Compose a class diagram of the whole driver. This will largely be tinplating
off of the existing gPXE drivers for other NIC's and the other OS's drivers for
this NIC. 

June 10-20: Write the driver.

June 21: Cross fingers and hopefully it's pretty much done.

June 22-30: TBA; hopefully moving on to documentation/testing, but available
for rewriting the driver/bug stomping/etc...

July 1-5: Collaboration with mentor as to progress up until this point,
speculation on the other half of the time. Basically, reassess my
position/timeline.

July 6: midterm review

July 7-15: If not done, documentation/testing.

July 15-august 10: At this point, if I'm done, I'd like to start work on a
2nd driver which hopefully can be completed in 1/2 to 1/3 of the time of the
first driver. If I'm not done, reassess my position and timeline and schedule
accordingly. 

August 10-17, more testing, more documentation. Hopefully end up with (a)
completed driver(s) with testing units and complete documentation including a
hardware API for this device.

Trace: • dverkamp • howtos • 2006 • mdc_sandbox • audit-the-shifts • qemupxe • 2008 • iscsibootprepw2k3 • qemu • win2k8_physical • memsearch • fedora11bko • ubuntu_aoe • debian_and_ubuntu • week1 • week12 • qemupxeknifehttp • week10 • week8 • 2009