On Sun, Jul 14, 2013 at 02:17:02PM -0500, Dave Wagler wrote: > I have just completed my build of LFS 7.3, that is, just fixed my last > known problem, and have a working system that talks to the internet. I ran > into two problems that I believe were not of my own making. > > 1. Missing letters on the keyboard: > > After I did the first chroot type login during the build process, the > lower-case 'c' and the upper-case 'E' were not being recognized. Neither > typing nor copy/paste would get either letter into a terminal command. Some > searching found a suggestion that the inputrc file could be bad. After > locating the file and seeing what it contained, replacing it seemed low > risk. So I copied in the file from the Linux Mint 14 host system, and the > problem disappeared. There are a number of differences between the two > files, and I have no idea which fixed my problem.
No idea on this - I load my own keymap (uk with somne UTF-8 extensions and completely different compose table), but I don't recall anyone mentioning this sort of problem in recent times. > > 2. Missing Realtek ethernet controller firmware patches: > > My system has a Realtek Semiconductor Co., Ltd. RTL8111/8168B PCI Express > Gigabit Ethernet controller (rev 09). It requires the r8169 kernel driver. > When I booted the LFS system, this error occurred: > > r8169 0000:03:00.0 eth0: unable to load firmware patch > rtl_nic/rtl8168f-1.fw (-2) > > I found several patch files in the Linux Mint 14 host system, copied them > into LFS, and the error disappeared. > Yeah. If it's like my two machines using that, it does seem to work without the firmware! See the example under the xorg ati video driver in the BLFS book. If you have an excessive amount of memory, building in the firmware does no harm. Otherwise, put it in /lib/firmware, point the kernel to which firmware(s) to use, and point the kernel to /lib/firmware. The downside to building it in is that it takes up kernel memory, which cannot be freed, but is only used when the device is initialized. So, more of an issue on 32-bit (on my latest machine, with 4GB, I get about 3.5GB of memory in x86_64 after deducting video and whatever, but only 2.5GB in 32-bit). I did think about suggesting that we add a general firmware page in BLFS, but I'm not convinced that I know enough about it and Armin made major improvements to the ati firmware notes while I was thinking about this. Might change my mind about a general page if the amd firmware update last week applies to my machines (previous version was older than my machines), haven't looked at it so far. I've noticed that messages about missingi nic firmware seem to come and go with different kernel versions - or perhaps I just don't notice because they only appear on the screen during booting and don't get logged (video firmware messages, for kms, might make it to the logs - if you are lucky). For the ATI video (R600+) it *does* change with newer kernel versions - I had to add something to my newest machine for 3.10.0. > Just thought I'd let you know. I don't know that they can actually be > considered bugs. May just be my unique system. > > Dave No doubt others will have different views from me - firmware for intel CPUs has apparently been available for some years (I've mostly been AMD), but I first came across missing firmware last year - and I've been using wholly/mostly LFS for about 10 years. I've seen comments on lkml, of course, but it didn't affect me. Of course, CPU firmware mostly seems to fix early revisions of a particular processor model. ĸen -- das eine Mal als Tragödie, das andere Mal als Farce -- http://linuxfromscratch.org/mailman/listinfo/lfs-support FAQ: http://www.linuxfromscratch.org/lfs/faq.html Unsubscribe: See the above information page
