Note that one thing that we might do is to leave the remnants of /proc/ide
but _without_ the very verbose per-chipset reporting.
At least to me it looks like it's all the chipset reporting that causes
the huge kernel bloat, and it shouldn't be impossible to reinstate a
minimal /proc/ide without those parts - while still keeping most of the
backwards compatibility.
However, since I really don't much like the idea of having special
"ide-only" /proc files, I personally think any information people actually
used should be either in truly generic files (/proc/partitions as an
example), _or_ they should be in the generic device tree (talk to Pat
Mochel about that).
So my personal reaction to removal of /proc/ide is: "good riddance, but if
it turns out that we seriously need it for backwards compatibility, we can
add back a skeleton without the bloat".
(Side note: I'm afraid that don't think backwards compatibility weighs
very heavily on an embedded setup - I'm more thinking about things like "a
regular RedHat/SuSE/Debian/whatever install won't work any more".)
As to existing binaries (your list is interesting), I don't see what they
are doing about ide-specific stuff, since I sure hope those binaries are
happy with a SCSI-only system.
> For e.g. could the same arguments could be made for lspci only
> interface to pci info rather than /proc/bus/pci? The following
> references are made to /proc/bus/pci on my system:
I personally do like ASCII /proc files, as long as they don't add
maintainability problems etc.
Linus
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/