TODO

[namjaejeon]

1. FITRIM ioctl
2. NFS support
3. fallocate
4. symlink ?

[inste]

To have 2 and 3 would be nice :)
Ready to test patches.

[namjaejeon]

Thanks for your opinion, I will try to work 2, 3 first if I make the time for these.
And I can't start working these jobs right now. I would be happy if someone try to implement it.

[piranna]

Driver provided by Samsung did included symlink support by using reserved (unused) bits, although it's not standard and it got removed when merged in Linux kernel.

I would also be interested on having executable bit support using that reserved bits too.

[Erichorng]

Hi! I'd like to ask a question.
I've recently encountered some issues when transferring large files using Samba, and it seems to be related to the lack of support for fallocate in exFAT.
According to the bellow discussion, I'm not sure if your plan refers to implementing the '--keep-size' functionality or the entire set of features.
From my understanding, to address the problem of transferring large files with Samba, it would be necessary to support at least the keep size flag and the default operation(mode=0).

https://lists.samba.org/archive/samba-technical/2020-December/136084.html

[namjaejeon]

@Erichorng Currently @YuezhangMo have been working valid-size support for exfat. I think that this patch fix your issue by speeding up ftruncate() of exfat. I will inform you if it is merged into exfat mainline.

[Erichorng]

Thank you for your response. Looking forward to it!

[namjaejeon]

@Erichorng Can you check if your problem is fixed with the following exfat branch ?

git clone https://github.com/namjaejeon/linux-exfat-oot/

[Erichorng]

@namjaejeon Thank you for the update.
I just tried it, and ftruncate on the new version of exFAT seems much faster! Also I found that it actually allocates disk space; it feels similar to the behavior of fallocate.