On Sun, 2019-06-02 at 11:46 -0400, Valdis Klētnieks wrote: > On Sun, 02 Jun 2019 08:48:15 +0100, S?rgio Basto said: > > > 'SUBDIRS' will be removed after Linux 5.3 > > Please use 'M=' or 'KBUILD_EXTMOD' instead > > - ifeq ($(shell if grep '"[432]\.' > > $(KERN_DIR)/include/linux/version.h > /dev/null 2>&1; then echo > > yes; fi),yes) > > + ifeq ($(shell if grep '"[98765432]\.' > > $(KERN_DIR)/include/linux/version.h /dev/null 2>&1; then echo yes; > > fi),yes) > > I submitted a more correct patch back in March, which did the right > thing > and changed all the SUBDIRS= to M=, though I admit I didn't go near > this > script (I only adressed the actual kernel module tree that ends up in > /usrshare/virtualbox/src/vboxhost) > > I admit not even being sure what this code used to be trying to do, > or why > this patch seems to work for you. Hint: It's looking for "any > number other than a 1, > followed by a period". At least in my kernel tree, what I have is: > > [/usr/src/linux-next] cat include/generated/uapi/linux/version.h > #define LINUX_VERSION_CODE 328192 > #define KERNEL_VERSION(a,b,c) (((a) << 16) + ((b) << 8) + (c)) > > Which means that neither [432]\. or [2-9]\. is ever going to match. > What's > in the version.h file your script is trying to work with? > > I'm not sure what version.h this script is looking at, but unless > there's a > '1.' and no other digits followed by a period anywhere in the file, > it's going > to find it. > > Hint: (a) your patch will mysteriously break if we ever get to an > 11.0 > kernel, and (b) It doesn't DTRT in distinguishing between a 2.4.X and > a 2.6.X > kernel, which is what actually matters for SUBDIRS/M. > > M= has been supported for forever. > > commit 0126be38d98815d25d9ec4573541ed4315bf6a88 > Author: Masahiro Yamada <[email protected]> > Date: Wed Nov 21 00:04:18 2018 +0900 > > kbuild: announce removal of SUBDIRS if used > > SUBDIRS has been kept as a backward compatibility since > commit ("[PATCH] kbuild: external module support") in 2002. > > In other words, since the 2.6.0 kernel. If somebody is trying to > build > VirtualBox on a 2.4 or earlier kernel, this is the *least* of their > problems. >
I wonder if Virtualbox 6.0.10 have any fix for this subject ? Since my patch still applies cleanly . Thanks, > -- Sérgio M. B. _______________________________________________ vbox-dev mailing list [email protected] https://www.virtualbox.org/mailman/listinfo/vbox-dev
