Is anybody actually advocating removing _user_visible_ devfs functionality? my understanding of the argument over on lkml was that people were advocating (1) leaving it in for 2.6 / including a smaller devfs implementation (2) Developing a user-space solution for 2.7 in which case, we would still use devfs names and autodetection, which are what packages/apps care about, and whatever the userspace tool to set up names would be included in the d-i. My experience in adding devfs to to apps is that, if we could rely on devfs always present, it drops the size of apps (as it simplifies hardware detection). Unfortunately, we can't depend on devfs, and so we have to include code to handle _both_ namespaces, which bloats rather than shrinks code .. In the limited case of d-i, we probably _could_ rely on devfs-only being present: (or "pseudo-devfs": something that sets up devices with the same names, only including /dev entries for hardware present); I'm not sure how this affects code size, as devfs is not small, (but a replacement talked on lkml is smaller, but 2.6 only if it happens) Regards, Alastair On Sun, 2003-02-09 at 09:28, Bastian Blank wrote: > On Sun, Feb 09, 2003 at 10:07:39AM +0100, Petter Reinholdtsen wrote: > > It might give us a few new problems: > > > > - Some platforms do not have working devfs support. drop or add > > workaround? > > this archs uses 2.2 kernels, so we can't currently use d-i on it. there > are devfs patches for 2.2 but i don't know there state. > > > - Will all kernel devices handle devfs properly? > > i know one module with problems, ethertap. but this driver is marked as > depreaced. > > > In skolelinux, we have enabled devfs by default (by using devfsd) to > > get LVM working. We haven't experienced any problems yet. I believe > > enabling devfs by default in Debian would be a good thing to do. > > i use devfs on any machines and i won't use a regular dev on my settop > boxes with the ro root. > > bastian > > -- > No one wants war. > -- Kirk, "Errand of Mercy", stardate 3201.7 -- Alastair McKinstry <mckinstry@computer.org> GPG Key fingerprint = 9E64 E714 8E08 81F9 F3DC 1020 FA8E 3790 9051 38F4 He that would make his own liberty secure must guard even his enemy from oppression; for if he violates this duty he establishes a precedent that will reach to himself. - --Thomas Paine
Attachment:
signature.asc
Description: This is a digitally signed message part