[haiku-development] Re: XFS support for Haiku - updates
- From: Abhinand N <01abhinandn@xxxxxxxxx>
- To: haiku-development@xxxxxxxxxxxxx
- Date: Wed, 13 Jun 2018 01:00:08 +0530
Hi,
On Tue, Jun 12, 2018, 7:35 PM Axel Dörfler <axeld@xxxxxxxxxxxxxxxx> wrote:
First: the minimum thing you should do is to check the magic number of
your super block after you read it; it really doesn't take long to
implement it.
I just figured that out while writing the blog post, that I was not
checking that with the default magic number.
Then you should dump the contents, so that you know what
the file system actually sees.
CheckSuperBlock() seems to be a direct copy from BFS, and is completely
wrong for your needs.
I tried adding a couple of more loggers and tried printing the magic
number, which is not that as XFS. As you said I am wrong at the
checkSuperBlock(), where it is not able to initialise offset.
As I read through XFS doc, the superblock is a sector on length which is
512 in most cases. So I modified Identify() and used read_pos(fd, 512,
superBlock, sizeof(superBlock)) to read the superblock. I am not sure if I
made a mistake in case of offset.
But to your actual problem: Please look at the code of other file
systems, and just read the documentation [1] there is, it's actually
quite helpful in this case!
The mount-hook is supposed to publish the vnode. Only then it can later
access it.
I'll look into this.
You have to put a lot more time and effort into your project if you want
to let it go somewhere. I cannot be more direct than that: you *will*
fail if you don't spend more time on GSoC.
Yes, I understand. I had a tight schedule last week and was not able to
spend more time, where I could focus and work. I am not saying this as an
excuse. I'll try my best to cover up the work by putting in extra time.
Thanks
Other related posts:
