Next attached sys-apps/hwdata-gentoo-0.4 pcitable "ide" to "ata" migration patch. There are 2 variants I have - with "experimental" drivers and no. I post only "experimental" (kernel CONFIG_EXPERIMENTAL). May change on request. Still not migrated only: By module name: delkin_cb, slc90e66, trm290. By CONFIG_: BLK_DEV_DELKIN, BLK_DEV_SLC90E66, BLK_DEV_TRM290. - there are no more reasons to compile device/ide/ (DEV_IDE) tree.
Created attachment 187482 [details, diff] sys-apps/hwdata-gentoo-0.4 pcitable ata
I have added this patch to the hwdata[1] git repository. I want to try to update some more of this before rolling another release, but this patch is sufficient itself to warrant a release if no other changes come down the pipe in a timely manner. Thanks a ton! 1. http://git.wolf31o2.org/gitweb/?p=projs/hwdata.git;a=commit;h=5b9557afbb1a3c0e2e487319f5d7f6563a369dd2
Should we wait for that release or is it useful to do a revision bump with that patch now?
;) Rumour form me is next: I don't try to use hwsetup (and hwdata) now. /lib/modules/.../modules.alias alredy have "pnp info" for many modules. For some of unlisted in aliases modules just must be fixed kernel sources to export modalias/pnp data (for example - ioatdma). To patch kernel you may use next script: http://raw.googlecode.com/svn/trunk/sys-kernel/genpnprd/files/modulesfix - "modulesfix /usr/src/linux" or "modulesfix /usr/src/linux f" - forced, without some of verification. This is bash-style wildcard info for "/sys/.../modalias" data. I use it in initrd bootloader (used on all of my machines, not "live media"), published in "raw" portage overlay, sys-kernel/genpnprd. This method is less "controllable" by third-party (like distribution team), but always up-to-date to kernel version. I remember about published here hwdata patch may be some unprecise for exotic devices, but same time I found it - I fully migrate to described tech and forget all. Just FYI. Sorry. PS I am ready to provide sys-kernel/genpnprd stuff in any other form (genkernel patch + kernel patch, for example) if there will be required. I think to talk about incomplete modaliases in kernel development list(s), but unsure in correct form of this. This is wide multi-target patch (bad for kernel team), etc...
Any news on this? Should hwdata-redhat be used instead?
