uml-patch-2.5.62-1

uml-patch-2.5.62-1

[Jeff Dike]

This patch updates UML to 2.5.62 and merges in the changes pending in the 2.4
UML pool.

Among the changes:
	More help entries in the config
	Build fixes, updates, and cleanups
	ubd driver cleanups and fixes
	uaccess fencepost fix

The 2.5.62-1 UML patch is available at
        http://uml-pub.ists.dartmouth.edu/uml/uml-patch-2.5.62-1.bz2
 
For the other UML mirrors and other downloads, see 
        http://user-mode-linux.sourceforge.net/dl-sf.html
 
Other links of interest:
 
        The UML project home page : http://user-mode-linux.sourceforge.net
        The UML Community site : http://usermodelinux.org

				Jeff

Re: uml-patch-2.5.62-1

[Oleg Drokin]

Hello!

On Wed, Feb 19, 2003 at 03:08:08PM -0500, Jeff Dike wrote:

> 	ubd driver cleanups and fixes

Ah, great. Except it introduced new breakage.
That hunk below from your diff adds add_disk() call.
Notice how a bit down we have another  call to add_disk(),
that is not removed. So we end up woth two add_disk() calls.
Of course sysfs gets upset immediately (probably not only it).

diff -Naur a/arch/um/drivers/ubd_kern.c b/arch/um/drivers/ubd_kern.c
--- a/arch/um/drivers/ubd_kern.c        Wed Feb 19 11:24:39 2003
+++ b/arch/um/drivers/ubd_kern.c        Wed Feb 19 11:29:49 2003
@@ -499,17 +516,22 @@
        disk->major = major;
        disk->first_minor = minor;
        disk->fops = &ubd_blops;
+       disk->private_data = dev;
+       disk->queue = &ubd_queue;
        set_capacity(disk, size / 512);
-       /* needs to be ubd -> /dev/ubd/discX/disc */
-       sprintf(disk->disk_name, "ubd");
+       sprintf(disk->disk_name, name);
        *disk_out = disk;
+       add_disk(disk);

-       /* /dev/ubd/N style names */
-       sprintf(devfs_name, "%d", unit);
-       *handle_out = devfs_register(dir_handle, devfs_name,
-                                    DEVFS_FL_REMOVABLE, major, minor,
-                                    S_IFBLK | S_IRUSR | S_IWUSR | S_IRGRP |
-                                    S_IWGRP, &ubd_blops, NULL);
+       if(handle_out != NULL){
+               /* /dev/ubd/N style names */
+               sprintf(devfs_name, "%d", unit);
+               *handle_out = devfs_register(dir_handle, devfs_name,
+                                            DEVFS_FL_DEFAULT, major, minor,
+                                            S_IFBLK | S_IRUSR | S_IWUSR |
+                                            S_IRGRP | S_IWGRP, &ubd_blops,
+                                            NULL);
+       }
        disk->private_data = &ubd_dev[unit];
        disk->queue = &ubd_queue;
        add_disk(disk);


Bye,
    Oleg

Re: [uml-devel] Re: uml-patch-2.5.62-1

[Jeff Dike]

green@namesys.com said:
> That hunk below from your diff adds add_disk() call. Notice how a bit
> down we have another  call to add_disk(), that is not removed. So we
> end up woth two add_disk() calls. Of course sysfs gets upset
> immediately (probably not only it). 

Oops, fixed.

		Jeff

uml-patch-2.5.62-1

[Jeff Dike]

This patch updates UML to 2.5.63.  The major change in this patch is the
inclusion of hostfs and hppfs, which were 2.4-only until someone figured
out the 2.5 vfs changes enough to forward port them.  This was done by Petr
Baudis, who ported hostfs.  I used those changes to bring hppfs forward.
hostfs seems to work OK, although there is a loose end or two that needs 
fixing.  hppfs is non-functional now, although it does mount and mirror
procfs.

There were also a bunch of bug fixes:
	some interrupt blocking bugs are fixed
	a ubd driver file locking bug is gone

The 2.5.63-1 UML patch is available at
        http://uml-pub.ists.dartmouth.edu/uml/uml-patch-2.5.63-1.bz2
 
For the other UML mirrors and other downloads, see 
        http://user-mode-linux.sourceforge.net/dl-sf.html
 
Other links of interest:
 
        The UML project home page : http://user-mode-linux.sourceforge.net
        The UML Community site : http://usermodelinux.org

    Jeff

Re: uml-patch-2.5.62-1

[Daniel Phillips]

On Wednesday 26 February 2003 20:05, Jeff Dike wrote:
> This patch updates UML to 2.5.63...

Built and booted.  However, without CONFIG_MODULES=y it doesn't build:

gcc -Wp,-MD,arch/um/sys-i386/.module.o.d -D__KERNEL__ -Iinclude -Wall 
-Wstrict-prototypes -Wno-trigraphs -O2 -fno-strict-aliasing -fno-common 
-U__i386__ -Ui386 -g -D__arch_um__ -DSUBARCH=\"i386\" -D_LARGEFILE64_SOURCE 
-Iarch/um/include -Derrno=kernel_errno -Dsigprocmask=kernel_sigprocmask 
-I/m/src/uml.2.5.63/arch/um/kernel/tt/include 
-I/m/src/uml.2.5.63/arch/um/kernel/skas/include -nostdinc -iwithprefix 
include    -DKBUILD_BASENAME=module -DKBUILD_MODNAME=module -c -o 
arch/um/sys-i386/module.o arch/um/sys-i386/module.c
arch/um/sys-i386/module.c: In function `apply_relocate':
arch/um/sys-i386/module.c:89: dereferencing pointer to incomplete type
arch/um/sys-i386/module.c: In function `apply_relocate_add':
arch/um/sys-i386/module.c:103: dereferencing pointer to incomplete type
make[1]: *** [arch/um/sys-i386/module.o] Error 1
make: *** [arch/um/sys-i386] Error 2

Native 2.5.63 (i386) is ok with or without CONFIG_MODULES=y.

Regards,

Daniel

Re: uml-patch-2.5.62-1

[Oleg Drokin]

Hello!

On Thu, Feb 27, 2003 at 06:45:27AM +0100, Daniel Phillips wrote:
> > This patch updates UML to 2.5.63...
> Built and booted.  However, without CONFIG_MODULES=y it doesn't build:
> gcc -Wp,-MD,arch/um/sys-i386/.module.o.d -D__KERNEL__ -Iinclude -Wall 
> -Wstrict-prototypes -Wno-trigraphs -O2 -fno-strict-aliasing -fno-common 
> -U__i386__ -Ui386 -g -D__arch_um__ -DSUBARCH=\"i386\" -D_LARGEFILE64_SOURCE 
> -Iarch/um/include -Derrno=kernel_errno -Dsigprocmask=kernel_sigprocmask 
> -I/m/src/uml.2.5.63/arch/um/kernel/tt/include 
> -I/m/src/uml.2.5.63/arch/um/kernel/skas/include -nostdinc -iwithprefix 
> include    -DKBUILD_BASENAME=module -DKBUILD_MODNAME=module -c -o 
> arch/um/sys-i386/module.o arch/um/sys-i386/module.c
> arch/um/sys-i386/module.c: In function `apply_relocate':
> arch/um/sys-i386/module.c:89: dereferencing pointer to incomplete type
> arch/um/sys-i386/module.c: In function `apply_relocate_add':
> arch/um/sys-i386/module.c:103: dereferencing pointer to incomplete type
> make[1]: *** [arch/um/sys-i386/module.o] Error 1
> make: *** [arch/um/sys-i386] Error 2
> Native 2.5.63 (i386) is ok with or without CONFIG_MODULES=y.

Patch below (that Jeff have not picked up yet) fixes that.

Bye,
    Oleg

# This is a BitKeeper generated patch for the following project:
# Project Name: Linux kernel tree
# This patch format is intended for GNU patch command version 2.5 or higher.
# This patch includes the following deltas:
#	           ChangeSet	1.949   -> 1.950  
#	arch/um/sys-i386/Makefile	1.15    -> 1.16   
#
# The following is the BitKeeper ChangeSet Log
# --------------------------------------------
# 03/02/13	green@angband.namesys.com	1.950
# Only build module.c if we have modules support selected
# --------------------------------------------
#
diff -Nru a/arch/um/sys-i386/Makefile b/arch/um/sys-i386/Makefile
--- a/arch/um/sys-i386/Makefile	Thu Feb 27 10:38:20 2003
+++ b/arch/um/sys-i386/Makefile	Thu Feb 27 10:38:20 2003
@@ -1,7 +1,8 @@
-obj-y = bugs.o checksum.o extable.o fault.o ksyms.o ldt.o module.o \
+obj-y = bugs.o checksum.o extable.o fault.o ksyms.o ldt.o \
 	ptrace.o ptrace_user.o semaphore.o sigcontext.o syscalls.o sysrq.o
 
 obj-$(CONFIG_HIGHMEM) += highmem.o
+obj-$(CONFIG_MODULES) += module.o
 
 USER_OBJS := bugs.o ptrace_user.o sigcontext.o fault.o
 USER_OBJS := $(foreach file,$(USER_OBJS),$(obj)/$(file))