9 Oct 2024 |
raitobezarius | or just use UFS | 14:52:20 |
emily | In reply to @raitobezarius:matrix.org or just use UFS do firmwares support that? | 14:54:12 |
raitobezarius | they could | 14:54:23 |
raitobezarius | just need to send a PR to EDK2 | 14:54:28 |
raitobezarius | actually, i think there's already support of it iirc | 14:54:36 |
raitobezarius | but it's not a "default" support | 14:54:42 |
raitobezarius | someone needs to deal with the UEFI forum | 14:54:46 |
Arian | this is above my pay grade | 14:54:57 |
ElvishJerricco | In reply to @emilazy:matrix.org someone should figure out a way to make FAT CoW zvols, my dude | 15:08:40 |
emily | In reply to @elvishjerricco:matrix.org zvols, my dude how does that fix UEFI? | 15:11:34 |
emily | I'm thinking of a way to reuse the on disk format but write to it in a way that is safe | 15:11:49 |
ElvishJerricco | we just need uefi to include zfs support /s | 15:12:05 |
emily | like maybe you can figure out how to atomically replace the pointer to the FAT table or something | 15:12:28 |
emily | you could implement it in FUSE | 15:12:46 |
ElvishJerricco | More seriously though, I wouldn't be surprised if FAT could be made quite a bit safer with the kenel's new atomic write feature. | 15:12:52 |
ElvishJerricco | requires hardware support, and I have no idea how common that is | 15:13:20 |
ElvishJerricco | but it lets you write one block atomically | 15:13:31 |
aloisw | In reply to @raitobezarius:matrix.org or just use UFS How much better is that compared to FAT? | 15:31:22 |
raitobezarius | much better; made for durability | 15:32:23 |
aloisw | Also, which UFS? | 15:58:51 |
aloisw | Or did you mean UDF? | 16:05:28 |
emily | oh, that makes more sense | 16:07:13 |
aloisw | I think to remember someone (maybe even you?) even talking in the past about how it's widely supported due to its usage on optical media anyway. | 16:10:42 |
raitobezarius | In reply to @aloisw:kde.org Or did you mean UDF? yes UDF | 16:11:24 |
raitobezarius | i should not write on matrix during meetings | 16:11:34 |
gdamjan | someone once said iso9660 should be supported as ESP, so perhaps .. :) | 17:56:13 |
raitobezarius | i said this | 18:14:24 |
raitobezarius | UDF is just the logical next step if you think ISO9660 should be supported | 18:14:33 |
raitobezarius | better skip legacy specs | 18:14:38 |
ElvishJerricco | In reply to @raitobezarius:matrix.org UDF is just the logical next step if you think ISO9660 should be supported why is that? | 19:07:50 |