Skip to contentSkip to conversation

2.4.10-pre5

DISCUSSION Garst R. Reese 10 participants 14 messages Sep 8, 2001 – Sep 10, 2001 Block layer

lore
Share & tools
lore
#0 lore raw

gcc-3.0
log should be self explanatory.
cc me if rqd.
Garst

make[2]: Entering directory `/usr/src/linux/drivers/block'
gcc -D__KERNEL__ -I/usr/src/linux/include -Wall -Wstrict-prototypes -Wno-trigraphs -O2 -fomit-frame-pointer -fno-strict-aliasing -fno-common -pipe -mpreferred-stack-boundary=2 -march=i586 -DMODULE -c -o rd.o rd.c
rd.c: In function `rd_ioctl':
rd.c:262: invalid type argument of `->'
rd.c: In function `rd_cleanup':
rd.c:375: too few arguments to function `blkdev_put'
make[2]: *** [rd.o] Error 1
make[2]: Leaving directory `/usr/src/linux/drivers/block' make[1]: *** [_modsubdir_block] Error 2
make[1]: Leaving directory `/usr/src/linux/drivers'
make: *** [_mod_drivers] Error 2
Command exited with non-zero status 2
4.13user 0.37system 0:04.50elapsed 99%CPU (0avgtext+0avgdata 0maxresident)k 0inputs+0outputs (3849major+4162minor)pagefaults 0swaps

#1 lore raw

Yes i got this too
anyone got a fix

----- Original Message -----
From: "Garst R. Reese" <[email protected]>
To: "linux-kernel" <[email protected]>
Cc: <[email protected]>
Sent: Saturday, September 08, 2001 2:42 PM
Subject: 2.4.10-pre5

Quoted 4 lines

gcc-3.0
log should be self explanatory.
cc me if rqd.
Garst

----------------------------------------------------------------------------
----

Quoted 2 lines

make[2]: Entering directory `/usr/src/linux/drivers/block'

gcc -D__KERNEL__ -I/usr/src/linux/include -Wall -Wstrict-prototypes -Wno-tri graphs -O2 -fomit-frame-pointer -fno-strict-aliasing -fno-common -pipe -mpre
ferred-stack-boundary=2 -march=i586 -DMODULE -c -o rd.o rd.c

Quoted 11 lines

rd.c: In function `rd_ioctl':
rd.c:262: invalid type argument of `->'
rd.c: In function `rd_cleanup':
rd.c:375: too few arguments to function `blkdev_put' make[2]: *** [rd.o] Error 1
make[2]: Leaving directory `/usr/src/linux/drivers/block' make[1]: *** [_modsubdir_block] Error 2
make[1]: Leaving directory `/usr/src/linux/drivers'
make: *** [_mod_drivers] Error 2
Command exited with non-zero status 2
4.13user 0.37system 0:04.50elapsed 99%CPU (0avgtext+0avgdata

0maxresident)k

Quoted 2 lines

0inputs+0outputs (3849major+4162minor)pagefaults 0swaps

#8 lore raw
In reply to Garst R. Reese
Quoted 2 lines · Kingsley Foreman

Yes i got this too
anyone got a fix

Don't use gcc-3.0

Quoted 45 lines

----- Original Message -----
From: "Garst R. Reese" <[email protected]>
To: "linux-kernel" <[email protected]>
Cc: <[email protected]>
Sent: Saturday, September 08, 2001 2:42 PM
Subject: 2.4.10-pre5

Quoted 4 lines

gcc-3.0
log should be self explanatory.
cc me if rqd.
Garst

----------------------------------------------------------------------------
----

Quoted 2 lines

make[2]: Entering directory `/usr/src/linux/drivers/block'

gcc -D__KERNEL__ -I/usr/src/linux/include -Wall -Wstrict-prototypes -Wno-tri graphs -O2 -fomit-frame-pointer -fno-strict-aliasing -fno-common -pipe -mpre
ferred-stack-boundary=2 -march=i586 -DMODULE -c -o rd.o rd.c

Quoted 11 lines

rd.c: In function `rd_ioctl':
rd.c:262: invalid type argument of `->'
rd.c: In function `rd_cleanup':
rd.c:375: too few arguments to function `blkdev_put' make[2]: *** [rd.o] Error 1
make[2]: Leaving directory `/usr/src/linux/drivers/block' make[1]: *** [_modsubdir_block] Error 2
make[1]: Leaving directory `/usr/src/linux/drivers'
make: *** [_mod_drivers] Error 2
Command exited with non-zero status 2
4.13user 0.37system 0:04.50elapsed 99%CPU (0avgtext+0avgdata

0maxresident)k

Quoted 2 lines

0inputs+0outputs (3849major+4162minor)pagefaults 0swaps

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/

Signature · 5 lines
Linux, the choice          | Stupidity is its own reward. 
of a GNU generation   -o)  | 
Kernel 2.4.9-ac1       /\  | 
on a i586             _\_v | 
                           | 
#9 lore raw
In reply to Garst R. Reese
Quoted 1 line · Josh McKinney

On approximately Sat, Sep 08, 2001 at 03:44:31PM +0930, Kingsley

Quoted 4 lines · Foreman
Quoted 2 lines

Yes i got this too
anyone got a fix

Don't use gcc-3.0

Excuse me, but shouldn't it be a goal to have the kernel compiling cleanly on gcc-3.0 as soon as is reasonable?
Telling people to use older versions of GCC for the kernel is not really a "good" thing if it can be avoided.
Personaly I don't use gcc-3.0 yet, but some do, and many will follow, and when they do, it'd be nice to have the kernel compiling cleanly, would it not?
They want a fix, not a one-line editorial. They're reporting a "bug", so that it may be fixed.

#11 lore raw
In reply to Garst R. Reese
Quoted 4 lines · Garst R. Reese

Garstmake[2]: Entering directory `/usr/src/linux/drivers/block'
gcc -D__KERNEL__ -I/usr/src/linux/include -Wall -Wstrict-prototypes -Wno-trigraphs -O2 -fomit-frame-pointer -fno-strict-aliasing -fno-common -pipe -mpreferred-stack-boundary=2 -march=i586 -DMODULE -c -o rd.o rd.c
rd.c: In function `rd_ioctl':
rd.c:262: invalid type argument of `->'

Quoted 8 lines

rd.c: In function `rd_cleanup':
rd.c:375: too few arguments to function `blkdev_put' make[2]: *** [rd.o] Error 1
make[2]: Leaving directory `/usr/src/linux/drivers/block' make[1]: *** [_modsubdir_block] Error 2
make[1]: Leaving directory `/usr/src/linux/drivers'
make: *** [_mod_drivers] Error 2
Command exited with non-zero status 2

This patch seems to work:

drivers/block/rd.c renamed from drivers/block/rd.c.orig +2 -2
@@ -259,7 +259,7 @@
 			/* special: we want to release the ramdisk memory,
 			   it's not like with the other blockdevices where
 			   this ioctl only flushes away the buffer cache. */
-			if ((atomic_read(rd_bdev[minor]->bd_openers) > 2))
+			if ((atomic_read(&rd_bdev[minor]->bd_openers) > 2))
 				return -EBUSY;
 			destroy_buffers(inode->i_rdev);
 			rd_blocksizes[minor] = 0;
@@ -372,7 +372,7 @@
 		struct block_device *bdev = rd_bdev[i];
 		rd_bdev[i] = NULL;
 		if (bdev) {
-			blkdev_put(bdev);
+			blkdev_put(bdev, BDEV_FILE);
 			bdput(bdev);
 		}
 		destroy_buffers(MKDEV(MAJOR_NR, i));
Signature · 4 lines
Peter Österlund             [email protected]
Sköndalsvägen 35            http://w1.894.telia.com/~u89404340
S-128 66 Sköndal            +46 8 942647
Sweden
#12 lore raw
In reply to Garst R. Reese
Quoted 16 lines · Peter Osterlund
Quoted 4 lines · Garst R. Reese

Garstmake[2]: Entering directory `/usr/src/linux/drivers/block'
gcc -D__KERNEL__ -I/usr/src/linux/include -Wall -Wstrict-prototypes -Wno-trigraphs -O2 -fomit-frame-pointer -fno-strict-aliasing -fno-common -pipe -mpreferred-stack-boundary=2 -march=i586 -DMODULE -c -o rd.o rd.c
rd.c: In function `rd_ioctl':
rd.c:262: invalid type argument of `->'

Quoted 8 lines

rd.c: In function `rd_cleanup':
rd.c:375: too few arguments to function `blkdev_put' make[2]: *** [rd.o] Error 1
make[2]: Leaving directory `/usr/src/linux/drivers/block' make[1]: *** [_modsubdir_block] Error 2
make[1]: Leaving directory `/usr/src/linux/drivers'
make: *** [_mod_drivers] Error 2
Command exited with non-zero status 2

Quoted 1 line

This patch seems to work:

Yes, it compiled, but Netscape-4.75 either kills it, or locks up. So there are other problems.
Nothing in logs. booting back to pre2 is ok
Garst

#2 lore raw
In reply to Kingsley Foreman
Quoted 5 lines

ferred-stack-boundary=2 -march=i586 -DMODULE -c -o rd.o rd.c

Quoted 4 lines

rd.c: In function `rd_ioctl':
rd.c:262: invalid type argument of ->' rd.c: In function rd_cleanup':
rd.c:375: too few arguments to function `blkdev_put'

2.4.10pre5 doesnt compile for rd. It looks like the same error I got when I applied Al's patch to -ac (and thus took it back out)

#3 lore raw
Quoted 9 lines · Alan Cox
Quoted 6 lines

ferred-stack-boundary=2 -march=i586 -DMODULE -c -o rd.o rd.c

Quoted 4 lines

rd.c: In function `rd_ioctl':
rd.c:262: invalid type argument of ->' rd.c: In function rd_cleanup':
rd.c:375: too few arguments to function `blkdev_put'

2.4.10pre5 doesnt compile for rd. It looks like the same error I got when I applied Al's patch to -ac (and thus took it back out)

Just in case anybody wondered it doesn't compile with gcc-3.0.1 either.

--alessandro

"this is no time to get cute, it's a mad dog's promenade
so walk tall, or baby don't walk at all"
(Bruce Springsteen, 'New York City Serenade')

#4 lore raw
In reply to Alan Cox

Since I have gotten a decent amount of flame mail from my one-liner posted yesterday, I am just curious. Has anyone really been able to successfully compile their kernel with gcc-3*. I did once or twice long before it was released, but it dies on the same error everytime. I posted it to the GCC mailing list because it was an internal compiler error, but it went unanswered, along with other reports of the same bug. Anyway, I just want to see if someone really is able to use it as a reliable compiler.

Josh

Quoted 23 lines · Alessandro Suardi
Quoted 9 lines · Alan Cox
Quoted 6 lines

ferred-stack-boundary=2 -march=i586 -DMODULE -c -o rd.o rd.c

Quoted 4 lines

rd.c: In function `rd_ioctl':
rd.c:262: invalid type argument of ->' rd.c: In function rd_cleanup':
rd.c:375: too few arguments to function `blkdev_put'

2.4.10pre5 doesnt compile for rd. It looks like the same error I got when I applied Al's patch to -ac (and thus took it back out)

Just in case anybody wondered it doesn't compile with gcc-3.0.1 either.

--alessandro

"this is no time to get cute, it's a mad dog's promenade
so walk tall, or baby don't walk at all"
(Bruce Springsteen, 'New York City Serenade')
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/

Signature · 5 lines
Linux, the choice          | Magic is always the best solution --
of a GNU generation   -o)  | especially reliable magic. 
Kernel 2.4.9-ac1       /\  | 
on a i586             _\_v | 
                           | 
#6 lore raw
In reply to Alan Cox

I've used it several times. I thought that I experienced a compiler bug at one point, but then I found that the same behavior occurs with 2.95.3, so...

Quoted 9 lines · Josh McKinney

Since I have gotten a decent amount of flame mail from my one-liner posted yesterday, I am just curious. Has anyone really been able to successfully compile their kernel with gcc-3*. I did once or twice long before it was released, but it dies on the same error everytime. I posted it to the GCC mailing list because it was an internal compiler error, but it went unanswered, along with other reports of the same bug. Anyway, I just want to see if someone really is able to use it as a reliable compiler.

Josh

Signature · 5 lines
-Steven
In a time of universal deceit, telling the truth is a revolutionary act.
			-- George Orwell
Freedom is slavery. Ignorance is strength. War is peace.
			-- George Orwell
#13 lore raw

Re: 2.4.10-pre5/pre6

In reply to Garst R. Reese

aar:/lib/modules/2.4.10-pre6#
Pre5 problems cured but,
dhinds pcmcia-cs-3.1.29 gives:
depmod -ae
depmod: *** Unresolved symbols in /lib/modules/2.4.10-pre6/kernel/drivers/scsi/sd_mod.o

depmod:         del_gendisk
depmod:         add_gendisk

This patch is still needed for pre6

Quoted 28 lines · Peter Osterlund

This patch seems to work:

drivers/block/rd.c Sat Sep 8 12:21:50 2001 renamed from drivers/block/rd.c.orig Sat Sep 8 11:58:19 2001 +2 -3
@@ -259,7 +259,7 @@
                        /* special: we want to release the ramdisk memory,
                           it's not like with the other blockdevices where
                           this ioctl only flushes away the buffer cache. */
-                       if ((atomic_read(rd_bdev[minor]->bd_openers) > 2))
+                       if ((atomic_read(&rd_bdev[minor]->bd_openers) > 2))
                                return -EBUSY;
                        destroy_buffers(inode->i_rdev);
                        rd_blocksizes[minor] = 0;
@@ -372,7 +372,7 @@
                struct block_device *bdev = rd_bdev[i];
                rd_bdev[i] = NULL;
                if (bdev) {
-                       blkdev_put(bdev);
+                       blkdev_put(bdev, BDEV_FILE);
                        bdput(bdev);
                }
                destroy_buffers(MKDEV(MAJOR_NR, i));
 
--

Peter Österlund [email protected]
Sköndalsvägen 35 http://w1.894.telia.com/~u89404340
S-128 66 Sköndal +46 8 942647
Sweden

#7 lore raw

Re: compiling kernel with gcc-3 (was: 2.4.10-pre5)

In reply to Alan Cox
Quoted 9 lines · Josh McKinney

Since I have gotten a decent amount of flame mail from my one-liner posted yesterday, I am just curious. Has anyone really been able to successfully compile their kernel with gcc-3*. I did once or twice long before it was released, but it dies on the same error everytime. I posted it to the GCC mailing list because it was an internal compiler error, but it went unanswered, along with other reports of the same bug. Anyway, I just want to see if someone really is able to use it as a reliable compiler.

I've been compiling kernels since June 22 with 3.0, never a
problem. -pre5 and -pre6 compiled with 3.0.1 backing out the rd.c changes.

--alessandro

"this is no time to get cute, it's a mad dog's promenade
so walk tall, or baby don't walk at all"
(Bruce Springsteen, 'New York City Serenade')

#5 lore raw

Re: 2.4.10-pre5 (and gcc-3)

In reply to Josh McKinney
Quoted 4 lines · Josh McKinney

Since I have gotten a decent amount of flame mail from my one-liner posted yesterday, I am just curious. Has anyone really been able to successfully compile their kernel with gcc-3*. I did once or twice long before it was released, but it dies on the same error everytime.

I had problems with pre-release versions, but gcc-3.0 works fine for me. I've got two boxes with K6s, this one running 2.4.6 (gcc-3) and the other was running various fairly recent kernels through to 2.4.8-pre3 compiled with gcc-3. I've seen strange errors with 2.4.7-ac5 (messages about loss of clock data, and trying to compile gcc-2.95.3 gave sig 11), but I'm not ready to lay those at the compiler's door. (The test box has turned into a linux from scratch system using gcc-2.95.3, it will be a little while before I get gcc-3 back up on it).

Ken

Signature · 3 lines
         If a six turned out to be nine, I don't mind.

         Home page : http://www.kenmoffat.uklinux.net