Re: Linux 2.5.12-dj1

Re: Linux 2.5.12-dj1

[rwhron]

This is the output while compiling aic7xxx_old.c:

aic7xxx_old.c:11950: unknown field `abort' specified in initializer
aic7xxx_old.c:11950: warning: initialization from incompatible pointer type
aic7xxx_old.c:11950: unknown field `reset' specified in initializer
aic7xxx_old.c:11950: warning: initialization from incompatible pointer type
aic7xxx_old.c:11950: duplicate initializer
aic7xxx_old.c:11950: (near initialization for `driver_template.slave_attach')
aic7xxx_old.c:11950: duplicate initializer
aic7xxx_old.c:11950: (near initialization for `driver_template.bios_param')
make[3]: *** [aic7xxx_old.o] Error 1

aic7xxxx_old.c compiles in 2.5.12 and 2.5.7-dj3, although it may
be broken in some other way.

Is it appropriate to edit .config to use the new driver by setting:
CONFIG_SCSI_AIC7XXX=y
CONFIG_AIC7XXX_CMDS_PER_DEVICE=253
CONFIG_AIC7XXX_RESET_DELAY_MS=15000

Is the new driver experimental, or ?

-- 
Randy Hron

Re: Linux 2.5.12-dj1

[Dave Jones]

On Thu, May 02, 2002 at 07:20:10AM -0400, rwhron@earthlink.net wrote:
 > This is the output while compiling aic7xxx_old.c:
 > 
 > aic7xxx_old.c:11950: unknown field `abort' specified in initializer
 > aic7xxx_old.c:11950: unknown field `reset' specified in initializer

Stereotypical "needs error handling code" errors.

 > aic7xxxx_old.c compiles in 2.5.12 and 2.5.7-dj3, although it may
 > be broken in some other way.

Yep. Removal of the abort/reset methods shows us which SCSI drivers
don't have any error handling. So they 'worked' up until you got
an error, and then...

 > Is it appropriate to edit .config to use the new driver by setting:
 > CONFIG_SCSI_AIC7XXX=y
 > CONFIG_AIC7XXX_CMDS_PER_DEVICE=253
 > CONFIG_AIC7XXX_RESET_DELAY_MS=15000
 > Is the new driver experimental, or ?

I've no experience of either driver in practice, so I'm not actually
sure what the game plan is here. Someone want to fill us in? Justin?

-- 
| Dave Jones.        http://www.codemonkey.org.uk
| SuSE Labs

Re: Linux 2.5.12-dj1

[Justin T. Gibbs]

> > Is it appropriate to edit .config to use the new driver by setting:
> > CONFIG_SCSI_AIC7XXX=y
> > CONFIG_AIC7XXX_CMDS_PER_DEVICE=253
> > CONFIG_AIC7XXX_RESET_DELAY_MS=15000
> > Is the new driver experimental, or ?
>
>I've no experience of either driver in practice, so I'm not actually
>sure what the game plan is here. Someone want to fill us in? Justin?

Since I have not reviewed how the driver has been modified for 2.5,
I can't really say much about its status in that branch.  The driver
is not "experimental" in the 2.4 branch.

--
Justin

Re: Linux 2.5.12-dj1

[Maksim (Max) Krasnyanskiy]

> >...
> > o   Bluetooth update.                         (Maksim (Max) Krasnyanskiy)
> >...
>
>"dev_list" is now defined in two files and I got the following compile
>error when building a kernel with all Bluetooth drivers compiled
>statically into the kernel:
You probably don't want to compile all _cs.o modules into the kernel.
But it's a bug in any case. dev_list should be static.  I'll fix that.

Thanks
Max

Linux 2.5.12-dj1

[Dave Jones]

Back up to date against Linus. Lots of resyncing going on, still
quite a lot left to go too.

It's lost ~2MB in overall size over the last month, which isn't
bad considering quite a bit has also been going in at the same time.
The framebuffer work I'm leaving to James. Eventually it will all
disappear from my tree magically. Hopefully.

As usual,..
Patch against 2.5.12 vanilla is available from:
ftp://ftp.kernel.org/pub/linux/kernel/people/davej/patches/2.5/

Merged patch archive: http://www.codemonkey.org.uk/patches/merged/

Check http://www.codemonkey.org.uk/Linux-2.5.html before reporting
known bugs that are also in mainline.

 -- Davej.
 

2.5.12-dj1
o   Re-add dropped patch for hd.c to fix compile.
o   Give ide-scsi new style error handling.		(Doug Gilbert)
o   Fix up some broken comment delimiters.		(John Kim)
o   Add missing checks to exec_permission_light		(Alexander Viro)
o   Futex update.					(Rusty Russell)
o   Make tasklets use per-cpu.				(Rusty Russell)
o   Orinoco update.					(David Gibson)
o   DEC Lance driver endianism fix.			(Maciej W. Rozycki)
o   Export block_flushpage()				(Andrew Morton)
o   Export path_lookup()				(Paul Menage)
o   Fix invalid rdev oops in md code.			(Roman Zippel)
o   Bluetooth update.				(Maksim (Max) Krasnyanskiy)
o   opl3sa2 resource free on probe failure.		(Zwane Mwaikambo)
o   Further sonypi updates.				(Stelian Pop)
o   Make IKConfig work when CONFIG_MODULES=n		(Adrian Bunk)
o   set_bit fixups in irda code.			(Paul Mackerras)


-- 
| Dave Jones.        http://www.codemonkey.org.uk
| SuSE Labs

Re: Linux 2.5.12-dj1

[Adrian Bunk]

On Wed, 1 May 2002, Dave Jones wrote:

>...
> o   Bluetooth update.				(Maksim (Max) Krasnyanskiy)
>...

"dev_list" is now defined in two files and I got the following compile
error when building a kernel with all Bluetooth drivers compiled
statically into the kernel:

<--  snip  -->

...
make[3]: Entering directory
`/home/bunk/linux/kernel-2.5/linux-2.5.12/drivers/bluetooth'
gcc -D__KERNEL__ -I/home/bunk/linux/kernel-2.5/linux-2.5.12/include -Wall
-Wstrict-prototypes -Wno-trigraphs -O2 -fno-strict-aliasing -fno-common
-pipe -mpreferred-stack-boundary=2 -march=k6   -nostdinc -I
/usr/lib/gcc-lib/i386-linux/2.95.4/include -DKBUILD_BASENAME=hci_usb  -c
-o hci_usb.o hci_usb.c
gcc -D__KERNEL__ -I/home/bunk/linux/kernel-2.5/linux-2.5.12/include -Wall
-Wstrict-prototypes -Wno-trigraphs -O2 -fno-strict-aliasing -fno-common
-pipe -mpreferred-stack-boundary=2 -march=k6   -nostdinc -I
/usr/lib/gcc-lib/i386-linux/2.95.4/include -DKBUILD_BASENAME=hci_vhci  -c
-o hci_vhci.o hci_vhci.c
gcc -D__KERNEL__ -I/home/bunk/linux/kernel-2.5/linux-2.5.12/include -Wall
-Wstrict-prototypes -Wno-trigraphs -O2 -fno-strict-aliasing -fno-common
-pipe -mpreferred-stack-boundary=2 -march=k6   -nostdinc -I
/usr/lib/gcc-lib/i386-linux/2.95.4/include -DKBUILD_BASENAME=hci_ldisc  -c
-o hci_ldisc.o hci_ldisc.c
gcc -D__KERNEL__ -I/home/bunk/linux/kernel-2.5/linux-2.5.12/include -Wall
-Wstrict-prototypes -Wno-trigraphs -O2 -fno-strict-aliasing -fno-common
-pipe -mpreferred-stack-boundary=2 -march=k6   -nostdinc -I
/usr/lib/gcc-lib/i386-linux/2.95.4/include -DKBUILD_BASENAME=hci_h4  -c -o
hci_h4.o hci_h4.c
ld -m elf_i386 -r -o hci_uart.o hci_ldisc.o hci_h4.o
gcc -D__KERNEL__ -I/home/bunk/linux/kernel-2.5/linux-2.5.12/include -Wall
-Wstrict-prototypes -Wno-trigraphs -O2 -fno-strict-aliasing -fno-common
-pipe -mpreferred-stack-boundary=2 -march=k6   -nostdinc -I
/usr/lib/gcc-lib/i386-linux/2.95.4/include -DKBUILD_BASENAME=dtl1_cs  -c
-o dtl1_cs.o dtl1_cs.c
gcc -D__KERNEL__ -I/home/bunk/linux/kernel-2.5/linux-2.5.12/include -Wall
-Wstrict-prototypes -Wno-trigraphs -O2 -fno-strict-aliasing -fno-common
-pipe -mpreferred-stack-boundary=2 -march=k6   -nostdinc -I
/usr/lib/gcc-lib/i386-linux/2.95.4/include -DKBUILD_BASENAME=bluecard_cs
-c -o bluecard_cs.o bluecard_cs.c
rm -f bluetooth.o
ld -m elf_i386  -r -o bluetooth.o hci_usb.o hci_vhci.o hci_uart.o
dtl1_cs.o bluecard_cs.o
bluecard_cs.o(.data+0x40): multiple definition of `dev_list'
dtl1_cs.o(.data+0x40): first defined here
make[3]: *** [bluetooth.o] Error 1
make[3]: Leaving directory `/home/bunk/linu

<-- snip  -->

cu
Adrian

-- 

You only think this is a free country. Like the US the UK spends a lot of
time explaining its a free country because its a police state.
								Alan Cox

Re: Linux 2.5.12-dj1

[Rudmer van Dijk]

On Wednesday 01 May 2002 22:36, Dave Jones wrote:
> Back up to date against Linus. Lots of resyncing going on, still
> quite a lot left to go too.
>
Got this link error:

ld -m elf_i386 -T /usr/src/kernel/linux-2.5.12-dj1/arch/i386/vmlinux.lds -e 
stext arch/i386/kernel/head.o arch/i386/kernel/init_task.o init/main.o 
init/version.o init/do_mounts.o \
        --start-group \
        arch/i386/kernel/kernel.o arch/i386/mm/mm.o kernel/kernel.o mm/mm.o 
fs/fs.o ipc/ipc.o \
        /usr/src/kernel/linux-2.5.12-dj1/arch/i386/lib/lib.a 
/usr/src/kernel/linux-2.5.12-dj1/lib/lib.o 
/usr/src/kernel/linux-2.5.12-dj1/arch/i386/lib/lib.a \
         drivers/base/base.o drivers/char/char.o drivers/block/block.o 
drivers/misc/misc.o drivers/net/net.o drivers/media/media.o 
drivers/char/agp/agp.o drivers/char/drm/drm.o drivers/ide/idedriver.o 
drivers/cdrom/driver.o sound/sound.o drivers/pci/driver.o 
drivers/video/video.o drivers/input/inputdrv.o drivers/input/serio/seriodrv.o 
\
        net/network.o \
        --end-group \
        -o vmlinux
drivers/block/block.o: In function `ramdisk_updatepage':
drivers/block/block.o(.text+0x3740): undefined reference to 
`mark_buffer_uptodate'
make: *** [vmlinux] Error 1

Grepping through the sources listed only one occurance of 
mark_buffer_uptodate:
# find . -type f | xargs grep -w mark_buffer_uptodate
./drivers/block/rd.c:                           mark_buffer_uptodate(tmp, 1);
Binary file ./drivers/block/rd.o matches
Binary file ./drivers/block/block.o matches

mark_buffer_uptodate was introduced in -2.5.12-dj1

	Rudmer

Re: Linux 2.5.12-dj1

[Dave Jones]

On Thu, May 02, 2002 at 01:06:07PM +0200, Rudmer van Dijk wrote:
 > drivers/block/block.o: In function `ramdisk_updatepage':
 > drivers/block/block.o(.text+0x3740): undefined reference to 
 > `mark_buffer_uptodate'

The chunk of code introduced to the ramdisk code in 2.5.12-dj1 is
a forward port of what happened in 2.4.  It probably needs fixing up
again (I've rewritten and jiggled it a little since the initial
forward port).  I'm not entirely sure it works as advertised anymore.

It's on my radar, I'll take a look for the next patch I put out.

    Dave.

-- 
| Dave Jones.        http://www.codemonkey.org.uk
| SuSE Labs