Well, klibc has -- apparently -- been rejected for 2.6.18. I guess that means now is the time to do additional work. I am hoping that this will be discussed as KS, and that there will be some additional direction from that. Right now I have to admit to being rather discouraged, not so much by the rejection itself as by the lack of rationale. -hpa
On Wed, Jul 05, 2006 at 10:38:49PM -0700, H. Peter Anvin wrote:> Well, klibc has -- apparently -- been rejected for 2.6.18. I guess that > means now is the time to do additional work.somehow expected due to the large number of discussions. cool concerning the additional work what are your plans?> I am hoping that this will be discussed as KS, and that there will be > some additional direction from that. Right now I have to admit to being > rather discouraged, not so much by the rejection itself as by the lack > of rationale.why couple the removal of in kernel code with klibc addition? i think the in kernel code should go through the feature-removal-schedule.txt procedure. such a patch should be the last of the klibc series. best regards -- maks
maximilian attems wrote:> > why couple the removal of in kernel code with klibc addition? > i think the in kernel code should go through the > feature-removal-schedule.txt procedure. > such a patch should be the last of the klibc series. >Well, first of all because maintaining multiple parallel pieces of code is a nightmare, and second because Linus stated that as a precondition for merging klibc several years ago. -hpa
maximilian attems wrote:> On Wed, Jul 05, 2006 at 10:38:49PM -0700, H. Peter Anvin wrote: >> Well, klibc has -- apparently -- been rejected for 2.6.18. I guess that >> means now is the time to do additional work. > > somehow expected due to the large number of discussions. > cool concerning the additional work what are your plans? >- Additional factoring of kinit; - Better integration with "make headers_install"; - Proper uClinux support. -hpa
On Wed, Jul 05, 2006 at 10:38:49PM -0700, H. Peter Anvin wrote:> Well, klibc has -- apparently -- been rejected for 2.6.18. I guess that > means now is the time to do additional work. > > I am hoping that this will be discussed as KS, and that there will be > some additional direction from that. Right now I have to admit to being > rather discouraged, not so much by the rejection itself as by the lack > of rationale.Keep up the good work! I cannot help when it comes to long term planning and the discusions about userspace versus kernel space. Expect that it is my impression that moving stuff out from the kernel to initramfs gives distributions much more flexibility. As for the userspace part the klcc support should maybe have been promoted a bit more? To show that it is relatively easy to build external programs for link with klibc. Sam