On Tue, 2011-08-23 at 11:51 -0300, Christian Robottom Reis wrote:
On Mon, Aug 22, 2011 at 02:45:21PM +0100, Tixy wrote:
=== Highlights ===
Trying to decipher algorithms used by SD cards. (Turns out Arnd had already documented some SanDisk peculiarities which probably explains a lot of what I was seeing [1])
Re-running SD media usecases and benchmarks after is became apparent that the initial btrfs results were bogus due to environment and scripting bugs.
I am not sure how applicable this is, but one possibly interesting distraction is looking at how David W. benchmarked a couple of flash filesystems here:
http://article.gmane.org/gmane.linux.kernel.embedded/3650
Thanks, that's similar in some ways and has some useful ideas. My work was mostly aimed at getting traces of filesystem activity to feed into a simulation tool. So we can compare how different filesystems behave on different real and hypothetical cards without actually having to run all combinations of these.
- Discovered that jprobes no longer work on Thumb kernels. This seems to be a new or previously overlooked bug with symbol lookup. Needs investigating...
Once we know how to test for this, we could have a test run as part of the CI loop (which is beginning to take shape from what I can tell on ci.linaro.org) to ensure it stays running for both Thumb and non-Thumb builds.
I found this bug because my kprobes test code failed. I'm going to try and get this code accepted in the mainline kernel tree, so it can be used by anyone working on kprobes in the future. Either way, it will be useful if it runs as part of any automated testing.