I'm Andrew Bradford. These are some things that I've accomplished:
- Bought my first Android phone, an LG G5 model RS988, and successfully was able to request my bootloader unlock key from LG.
- Got a whole bunch of updates into the CLFS embedded book that I'd accumulated over the past year for the 20171012 edition of the book.
- Switched DNS providers for bradfordembedded.com from Fastmail to Hurricane Electric's Free DNS service.
- Provided feedback to NXP via their community about confusing statements regarding the main system bus connections for the various Cortex-M cores within the datasheets for the LPC43xx line of microcontrollers and received conflicting feedback from NXP representatives: https://community.nxp.com/thread/462040
- Wrote up rework instructions to have a sane FTDI UART interface on an ArchiTech Hachiko development board: http://www.bradfordembedded.com/2017/07/hachiko-uart-sanity-rework
- Shutdown my NTP servers in India. Some network routing changed a month or two back providing very poor connectivity between the NTP Pool monitoring station and my servers such that they were rarely ever in the pool.
- Installed Fedora 25 on my laptop to try it out.
- Updated the NEC uPD720200a USB 3.0 host controller firmware in my Lenovo T420s laptop so that it now correctly detects USB devices which are hot-plugged into it during run-time.
- Setup a partial Debian mirror at home that syncs every night as my home Internet connection is rather slow.
- Setup an NTP pool vendor zone for work.
- Sent feedback to TI regarding discrepancies between the TPS65916 PMIC documentation regarding PRIMARY_SECONDARY_PAD1 register bitfields. The TRM does not match the user's guide for how the bitfields are split with PRIMARY_SECONDARY_PAD2.
- Sent feedback to TI regarding discrepancies in the AM570x/AM571x TRM about the size if the McSPI FIFOs, it is stated that they are both 32 bytes and 64 bytes for each of Tx and Rx when split but being 128 bytes if dedicated to either Tx xor Rx.
- Learned that fetching a github pull request's changes is much simpler than I had realized, you can just do: git fetch $REMOTE +refs/pull/$PRNUM/merge
- Applied some simple CSS to this site based on the concepts from http://bettermotherfuckingwebsite.com/ which should make rendering on mobile better and improve readability for all clients.
- Added a html charset to this page which fixed a warning from Firefox dev tools.
- Removed concept of "private" accomplishments from this site generator.
- Made the gitlab git repo hosting this website generator and files "public" so others can view it: https://gitlab.com/bradfa/bradfa.gitlab.io
- Found, after much banging of my head on the desk, that a systemd change caused gnupg-agent to be unable to properly spawn pinentry because I didn't have the dbus-user-session package installed as per https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=851440
- Created a KiCad schematic symbol for the SiLabs CP2102N 24 pin variant.
- Documented setting up a Jenkins slave using the SSH method. Learned a bit about Jenkins in the process.
- Got USB DFU working within u-boot 2017.01 for a custom board. Learned a bit about how DFU works.
- Setup syncthing https://syncthing.net/ on some of my computers to replace my uses of dropbox. It seems to do everything that the bittorrent sync did but also fill the gaps that bittorrent sync had which made me shy away from it. I'm liking it!
- Submitted a correction to NXP for the SJA1105 Ethernet switch datasheet.
- Found that round trip time and dns lookup time were about equivalent using Google's public DNS or OpenDNS on my home Internet connection.
- Signed up for and switched my home Internet connection over to using OpenDNS. Currently using their OpenDNS Home level service.
- Installed the OpenDNS dynamic IP updater tool on the family computer.
- Using the statistics generated from just last night as shown on the OpenDNS control panel, found that one of the top 5 most looked up domains was *.doubleclick.net, so updated my router's DNS resolver to return localhost for any lookup matching that. Some websites I visit load substantially faster, now! :)
- Unsubscribed from the debian-mentors, debian-devel, and lede-dev mailing lists. I currently do not have time to wade through the traffic on these and it drowns out the mail I do get related to these projects which I do care about. This may be an indication that my desire to move towards being a Debian Developer might be a stretch goal for 2017.
- Wrote up a README and provided license text for my USB Cereal hardware design: https://github.com/bradfordembedded/usb-cereal
- Read chapter 12 (PCI) of Linux Device Drivers v3, again.
- Updated the build script for this website to automatically turn hyperlinks into html hyperlinks like the next entry:
- This is just an example line showing more than https://www.google.com one hyperlink http://www.github.com on a line
- Got a basic Linux driver written for the TI TSC2013 based on the TSC2004/TSC200x driver code. The current state of it isn't likely to be up to the standards for upstream acceptance but it works.
- Learned a little about tslib https://github.com/kergoth/tslib
- Spent the morning learning about how Linux deals with uncached memory areas and memory mapping.
- Continued to work on writing a Linux driver for the TI TSC2013 touch screen controller.
- Tried to update our Ubuntu 12.04 based mythtv box to a more up-to-date hardware support release. That ended up bringing in a newer kernel, lots of new xorg packages, and basically broke everything related to video playback in mythtv. Rolling back to the quantal hardware support release and nvidia-304 proprietary drivers fixed the problems but was frustrating and doesn't give me hope at being able to easily upgrade from 12.04 to 14.04 or 16.04 directly.
- Learned how to use the `go get` command for golang so I can now fetch and build the do-agent code. Go uses an interesting way of forcing you to organize your code.
- Learned more about writing systemd service and socket units, it seems like quite a neat trick and very useful.
- Reported a potential bug with Digital Ocean's "new graphs" which use do-agent as the network throughput graphs seem to be off by a factor of 8-10, likely they are showing a units of Mbps but are collecting data as MBps. Opened a support ticket for this.
- Spun up another NTP server in India by cloning my existing server.
- Installed the Digital Ocean do-agent into both servers to get the neat new graphs.
- Found that some of the do-agent commands don't work how they should, tried to get and build the do-agent source code but had no luck. It's in Go but I don't understand how the build system works.
- Fixed up some firmware update scripting.
- Captured 1023 seconds of NTP traffic from my server when there were 3 IPv6 servers in the India pool and when I was the only IPv6 server in the pool. This generated over 6 GB of tcpdump output just for port 123.
- Started investigating the difference between the TI TSC2004 and TSC2013 touch screen controller chips, hopefully allowing reuse of the existing tsc200x driver in Linux for TSC2013.
- Created this accomplishments website and generator scripting.
- Subscribed to debian-mentors and debian-devel mailing lists.
For more info on me, see my blog.
Copyright (C) Andrew Bradford