Re: addendum - several modules failing to be built - all relatedto pccarddevs.h

From: P.D. Seniura <pdseniura_at_techie.com>
Date: Fri, 04 Jun 2004 17:56:46 -0600
----- Original Message -----
From: Bill Fumerola <billf_at_FreeBSD.org>
Date: Fri, 4 Jun 2004 13:58:38 -0700
To: "P.D. Seniura" <pdseniura_at_techie.com>
Subject: Re: addendum - several modules failing to be built - all related to pccarddevs.h

> On Fri, Jun 04, 2004 at 09:42:25PM +0300, Ruslan Ermilov wrote:
> > FWIW, running ``make depend'' is almost always optional.
> 
> except when it's not, apparently.
> 
> On Fri, Jun 04, 2004 at 12:38:01PM -0600, P.D. Seniura wrote:
> > The 'make buildworld' did that.
> > The log shown was a "re-run".
> > No other patches/updates were applied in-between.
> 
> make buildworld builds in the objdir (/usr/obj/blah). when you build
> within /sys/modules/ you're no longer building in the objdir.
> 
> so here's a thought: try it before telling me i'm wrong. i did.

I didn't get a spot in to reply to you 'til now.

I did try 'make depend' before another 'make all'
while under /sys/modules -- your msg showed up
before John's; I'd try anything at that point.

Simply doing the 'make depend' didn't fix the
'make all', it still didn't know how to make the
pccard stuff.

Before all this, tho, to discuss another point you
just brought up.  My 'make buildworld' under /src
left _something_ in the /sys/modules such that a 'make'
under /sys/modules did not need to remake everything --
it just remade the modules that failed w/r/t pccard. 
I didn't show that log for brevity.  But it acted as
if building modules either way were in sync with
one-another.  What I'm trying to say is that the
modules that were okay from the buildworld (no pccard
error) did not (need to be) remade when I did the make
under /sys/modules.  It all acted as I'd expect,
either way.  I _think_ this is because of the NO_CLEAN*
settings I've been using for months to save time.  Even
with NO_CLEAN*, I _always_ blow away /usr/obj/* before
building world _or_ kernel -- except when tracking
down these kinds of problems, when you'd want to see
what's happening where/why/etc.

After that, John's 2nd msg came thru. 
Commenting-out the NO_CLEAN* settings fixed it. 
Did 'make cleandir' then your 'make depend' then
'make all', all under /sys/modules.  Got fresh updated
modules installed now.  Of course the cleandir forced
even the previously-good modules to get rebuilt this
time around.  'Cleaning' might be the focus as to why
this particular problem cropped up here, and worth
investigating further.

Not sure what follows this on a buildworld, tho (where
pccard originally failed it), so I'm going to blow it
all away again and let it run over the weekend & let
it clean house all it wants to.  ;)  I'll re-enable
the NO_CLEAN* options next week and give it another go.

I'm still curious if 'make cleandir' is suppose to
override those make.conf settings, tho.

I _do_ appreciate your help & everyone else's. 

I'm wondering if your system has the NO_CLEAN*
settings enabled?  I think trying them could help
narrow down further why this problem might happen.

I'm going to let it do a full house cleaning while
it builds over the weekend.  The CTM src-cur bucket
just came thru, still waiting on the ports-cur bucket.

I never know when the build shortcuts will cause
trouble like this.  I've been trying to find ways to
save time, on a slow puny pentium2, as Da Bosses are
wondering why it takes so long.  Well, when they give
me a snail, it *will* be slow.
So there ya go.
Your taxes hard at work.  ;)


> -- 
> - bill fumerola / fumerola_at_yahoo-inc.com / billf_at_FreeBSD.org
> 
> 
> [ SECURE !6 ]% make clean                                    (/sys/modules/aic)
> rm -f /usr/home/fumerola/malaise/vendor/freebsd_5/src/sys/modules/aic/export_syms aic.ko aic.kld aic.o aic_pccard.o aic_isa.o _at_ machine symb.tmp tmp.o opt_cam.h opt_scsi.h bus_if.h device_if.h isa_if.h card_if.h pci_if.h pccarddevs.h
> [choker.corp.yahoo(p4)-fumerola Fri04/13:33:23]
> [ SECURE !7 ]% make                                          (/sys/modules/aic)
> Warning: Object directory not changed from original /usr/home/fumerola/malaise/vendor/freebsd_5/src/sys/modules/aic
> _at_ -> /usr/home/fumerola/malaise/vendor/freebsd_5/src/sys
> machine -> /usr/home/fumerola/malaise/vendor/freebsd_5/src/sys/i386/include
> awk -f _at_/tools/makeobjops.awk _at_/kern/device_if.m -h
> awk -f _at_/tools/makeobjops.awk _at_/kern/bus_if.m -h
> touch opt_cam.h
> cc -O -pipe -march=pentium4  -D_KERNEL -DKLD_MODULE -nostdinc -I-   -I. -I_at_ -I_at_/../include -I/usr/include -finline-limit=8000 -fno-common  -mno-align-long-strings -mpreferred-stack-boundary=2 -ffreestanding -Wall -Wredundant-decls -Wnested-externs -Wstrict-prototypes  -Wmissing-prototypes -Wpointer-arith -Winline -Wcast-qual  -fformat-extensions -std=c99 -c /usr/home/fumerola/malaise/vendor/freebsd_5/src/sys/dev/aic/aic.c
> awk -f _at_/tools/makeobjops.awk _at_/dev/pccard/card_if.m -h
> make: don't know how to make _at_/dev/pccard/pccarddevs.h. Stop
> [choker.corp.yahoo(p4)-fumerola Fri04/13:33:28]
> [ SECURE !8 ]% make depend                                   (/sys/modules/aic)
> awk -f _at_/tools/makeobjops.awk _at_/dev/pci/pci_if.m -h
> awk -f _at_/tools/makeobjops.awk _at_/isa/isa_if.m -h
> awk -f _at_/tools/pccarddevs2h.awk _at_/dev/pccard/pccarddevs
> touch opt_scsi.h
> rm -f .depend
> mkdep -f .depend -a   -nostdinc -D_KERNEL -DKLD_MODULE -I- -I. -I_at_ -I_at_/../include /usr/home/fumerola/malaise/vendor/freebsd_5/src/sys/modules/aic/../../dev/aic/aic.c /usr/home/fumerola/malaise/vendor/freebsd_5/src/sys/modules/aic/../../dev/aic/aic_pccard.c /usr/home/fumerola/malaise/vendor/freebsd_5/src/sys/modules/aic/../../dev/aic/aic_isa.c
> [choker.corp.yahoo(p4)-fumerola Fri04/13:33:32]
> [ SECURE !9 ]% make                                          (/sys/modules/aic)
> Warning: Object directory not changed from original /usr/home/fumerola/malaise/vendor/freebsd_5/src/sys/modules/aic
> cc -O -pipe -march=pentium4  -D_KERNEL -DKLD_MODULE -nostdinc -I-   -I. -I_at_ -I_at_/../include -finline-limit=8000 -fno-common  -mno-align-long-strings -mpreferred-stack-boundary=2 -ffreestanding -Wall -Wredundant-decls -Wnested-externs -Wstrict-prototypes  -Wmissing-prototypes -Wpointer-arith -Winline -Wcast-qual  -fformat-extensions -std=c99 -c /usr/home/fumerola/malaise/vendor/freebsd_5/src/sys/dev/aic/aic_pccard.c
> cc -O -pipe -march=pentium4  -D_KERNEL -DKLD_MODULE -nostdinc -I-   -I. -I_at_ -I_at_/../include -finline-limit=8000 -fno-common  -mno-align-long-strings -mpreferred-stack-boundary=2 -ffreestanding -Wall -Wredundant-decls -Wnested-externs -Wstrict-prototypes  -Wmissing-prototypes -Wpointer-arith -Winline -Wcast-qual  -fformat-extensions -std=c99 -c /usr/home/fumerola/malaise/vendor/freebsd_5/src/sys/dev/aic/aic_isa.c
> ld  -d -warn-common -r -d -o aic.kld aic.o aic_pccard.o aic_isa.o
> touch /usr/home/fumerola/malaise/vendor/freebsd_5/src/sys/modules/aic/export_syms
> awk -f /usr/home/fumerola/malaise/vendor/freebsd_5/src/sys/modules/aic/../../conf/kmod_syms.awk aic.kld  /usr/home/fumerola/malaise/vendor/freebsd_5/src/sys/modules/aic/export_syms |  xargs -J% objcopy % aic.kld
> ld -Bshareable  -d -warn-common -o aic.ko aic.kld
> objcopy --strip-debug aic.ko
> [choker.corp.yahoo(p4)-fumerola Fri04/13:34:56]
> [ SECURE !10 ]%    
> 

-- 
___________________________________________________________
Sign-up for Ads Free at Mail.com
http://promo.mail.com/adsfreejump.htm
Received on Fri Jun 04 2004 - 14:57:51 UTC

This archive was generated by hypermail 2.4.0 : Wed May 19 2021 - 11:37:56 UTC