Since my quarter term report I have made a great deal of progress. The boot loader x86_64 support is finished, and the kernel can now be booted to the point of searching for the boot volume. A screenshot of this:
Usually when you arrive in KDL (Kernel Debugging Land) it means that something bad happened. KDL provides a lot of tools to investigate what might have taken place. Still, it is quite possible that, even though you have that arsenal of tools available, you may not be able to immediately make sense of what is going on. In that case you ideally either continue debugging yourself or you write up a nice bug report over at the Haiku bug tracker so that someone else can take a closer look.
For the 1/4 term milestone, my goal was to have inode-moving working. This is mostly completed, you can view the code at http://web.student.chalmers.se/~andrhen/move_inode_v2.patch
For this period, I have the following things planned:
Allocation of new block positions: I have a good grasp of what needs to be done for this, and it’s not a lot of work.
Moving file data: Last week I thought I had this nailed down, but it turned out to be a little more involved than that. Still, there has been some good progress made, and I’m sure it’ll be completed within the period.
I have already implemented all mandatory hooks (and several others), what means that NFSv4 client now allows to browse directories and read files on remote filesystems. Last several days I spent on improving the existing code and supporting some less usual NFSv4 describes, that includes reclaiming share reservations, support for server migration and volatile filehandles. I also needed to deal with NFSv4 that do not provide file's inode number, that was solved only partially since proper workaround will be much easier to implement when file metadata and directory contents are cached.
I completed my 1/4 goal ahead of proposal schedule. By the original plan, I should investigate whether we need necessary changes to x86 idle routine and x86 architecture specific instruction usage. The results were reported to gsoc maillist on 3rd Jun. Here are the copied results:
- no need to change x86 idle routine
- monitor/mwait works perfectly. I have measured the power consumption when using idle implemented with “monitor/mwait”, it’s the same as the version implemented with “hlt”.
As I mentioned in my first blog post, I had exams until a couple of weeks ago, so I didn’t start working on my project properly until then. However, I am now working full time so I expect to make a lot more progress in the coming weeks.
So far I have made it possible to compile the kernel for x86_64 by adding stub implementations of all the architecture functions and fixing compilation errors/warnings (all architectures that Haiku supports at the moment are 32-bit, so there were various problems in the code when compiling for 64-bit). I have also made changes to the kernel_args structure (the structure which the bootloader passes to the kernel containing information such as the boot volume, loaded modules and the memory map). As I am making the x86 bootloader able to load both a 32-bit and 64-bit kernel, this structure must be compatible between both. The changes I have made make all members the same size regardless of whether compiling for 32-bit or 64-bit.
Since my last blog entry I have mostly completed the implementation of the AWT/Java2D port. It is still in need of a lot of testing, but it is stable enough to run a lot of Swing apps out of the box. For example, here’s jEdit and SwingSet:
After a week of struggling with my friend the compiler, I finally managed to bring the port into a buildable state again. Porting the latest changes to libwebkit, the last of the four WebKit subcomponents (libwtf, libjavascriptcore, libwebcore, libwebkit), took me more time than I anticipated but to my great surprise, after I fixed the first runtime crash, HaikuLauncher and WebPositive would run and be almost usable. Yes, there are visual glitches, and it is quite easy to crash, but you can search google, use maps, facebook, haiku-os.org and post this blog post.
Hello fellow Haiku’ers, as promised I’m posting a quick update on my WebKit / WebPositive contract work. It’s been a little more than a week already, and a small report is due!
Welcome WebKit r115944 ! As you may know, WebKit is a really big project, in the last two years, 70000 revisions have passed and the file count has almost doubled. The approach I took was to start by checking out a recent WebKit revision and try building the components one by one, re-applying our changes. The idea was to add only the strict necessary for Haiku and at the same time try to include as many features as possible, ignoring assumptions and workarounds that aren’t needed anymore. As many things have changed in WebKit and as I needed to get familiar with this huge codebase anyway, I decided to dismantle our port and put it back together again, like one would have done with a complex piece of mechanics. Thus I did a Jamfile from scratch, based on other platforms buildsystems, and replayed our changes one by one, as the compiler asked. Each time trying to document my changes and research the reasons and implications of the changes.
During community bonding period, apart from reading specifications, I also analyzed other NFSv4 client implementations, looking for interesting ideas and solutions. One thing I noticed which may be worth using in my NFSv4 client implementation is caching of some parts of generated RPC request.
In addition to that I also got more familiar with what Haiku expects from file system modules. I have written a very simple dummy file system and check how things work. The problem I encountered here is that Haiku when creating a vnode identifies the file using its inode number. NFSv4 server provides its clients inode numbers but does not provide any way to get file using its inode number. That is why the client will have to keep some cache of inode numbers it is aware of in order to map them to NFSv4 file handles. Self-balancing BST seems to be a good idea.
