Trond Myklebust [Fri, 20 Oct 2006 06:28:40 +0000 (23:28 -0700)]
[PATCH] NFS: Deal with failure of invalidate_inode_pages2()
If invalidate_inode_pages2() fails, then it should in principle just be
because the current process was signalled. In that case, we just want to
ensure that the inode's page cache remains marked as invalid.
Also add a helper to allow the O_DIRECT code to simply mark the page cache as
invalid once it is finished writing, instead of calling
invalidate_inode_pages2() itself.
Trond Myklebust [Fri, 20 Oct 2006 06:28:40 +0000 (23:28 -0700)]
[PATCH] NFS: Fix NFSv4 callback regression
The change in semantics for nfs_find_client() introduced by David breaks the
NFSv4 callback channel.
Also, replace another completely broken BUG_ON() in nfs_find_client(). In
initialised clients, clp->cl_cons_state == 0, and callers of that function
should in any case never want to see clients that are uninitialised.
Trond Myklebust [Fri, 20 Oct 2006 06:28:38 +0000 (23:28 -0700)]
[PATCH] NFS: Fix oops in nfs_cancel_commit_list
Fix two bugs:
- nfs_inode_remove_request will call nfs_clear_request, so we cannot
reference req->wb_page after it. Move the call to dec_zone_page_state so
that it occurs while req->wb_page is still valid.
- Calling nfs_clear_page_writeback is unnecessary since the radix tree
tags will have been cleared by the call to nfs_inode_remove_request.
Replace with a simple call to nfs_unlock_request.
David Howells [Fri, 20 Oct 2006 06:28:36 +0000 (23:28 -0700)]
[PATCH] autofs3: Make sure all dentries refs are released before calling kill_anon_super()
Make sure all dentries refs are released before calling kill_anon_super()
so that the assumption that generic_shutdown_super() can completely destroy
the dentry tree for there will be no external references holds true.
What was being done in the put_super() superblock op, is now done in the
kill_sb() filesystem op instead, prior to calling kill_anon_super().
The call to shrink_dcache_sb() is removed as it is redundant since
shrink_dcache_for_umount() will now be called after the cleanup routine.
Signed-off-by: David Howells <dhowells@redhat.com> Acked-by: Ian Kent <raven@themaw.net> Cc: Andi Kleen <ak@suse.de> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
GENERIC_ACL shouldn't be under Network File Systems (which made it depend
on NET) as far as I can tell. Having it there and having many (FS) config
symbols disabled gives this (which the patch fixes):
mm/built-in.o: In function `shmem_check_acl':
shmem_acl.c:(.text.shmem_check_acl+0x33): undefined reference to `posix_acl_permission'
fs/built-in.o: In function `generic_acl_get':
(.text.generic_acl_get+0x30): undefined reference to `posix_acl_to_xattr'
fs/built-in.o: In function `generic_acl_set':
(.text.generic_acl_set+0x75): undefined reference to `posix_acl_from_xattr'
fs/built-in.o: In function `generic_acl_set':
(.text.generic_acl_set+0x94): undefined reference to `posix_acl_valid'
fs/built-in.o: In function `generic_acl_set':
(.text.generic_acl_set+0xc1): undefined reference to `posix_acl_equiv_mode'
fs/built-in.o: In function `generic_acl_init':
(.text.generic_acl_init+0x7a): undefined reference to `posix_acl_clone'
fs/built-in.o: In function `generic_acl_init':
(.text.generic_acl_init+0xb4): undefined reference to `posix_acl_clone'
fs/built-in.o: In function `generic_acl_init':
(.text.generic_acl_init+0xc8): undefined reference to `posix_acl_create_masq'
fs/built-in.o: In function `generic_acl_chmod':
(.text.generic_acl_chmod+0x49): undefined reference to `posix_acl_clone'
fs/built-in.o: In function `generic_acl_chmod':
(.text.generic_acl_chmod+0x76): undefined reference to `posix_acl_chmod_masq'
Signed-off-by: Randy Dunlap <randy.dunlap@oracle.com> Acked-by: Andreas Gruenbacher <agruen@suse.de> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Randy Dunlap [Fri, 20 Oct 2006 06:28:34 +0000 (23:28 -0700)]
[PATCH] cad_pid sysctl with PROC_FS=n
If CONFIG_PROC_FS=n:
kernel/sysctl.c:148: warning: 'proc_do_cad_pid' used but never defined
kernel/built-in.o:(.data+0x1228): undefined reference to `proc_do_cad_pid'
make: *** [.tmp_vmlinux1] Error 1
Signed-off-by: Randy Dunlap <randy.dunlap@oracle.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Ralf Baechle [Fri, 20 Oct 2006 06:28:34 +0000 (23:28 -0700)]
[PATCH] Fix warnings for WARN_ON if CONFIG_BUG is disabled
In most cases the return value of WARN_ON() is ignored. If the generic
definition for the !CONFIG_BUG case is used this will result in a warning:
CC kernel/sched.o
In file included from include/linux/bio.h:25,
from include/linux/blkdev.h:14,
from kernel/sched.c:39:
include/linux/ioprio.h: In function â\80\98task_ioprioâ\80\99:
include/linux/ioprio.h:50: warning: statement with no effect
kernel/sched.c: In function â\80\98context_switchâ\80\99:
kernel/sched.c:1834: warning: statement with no effect
Signed-off-by: Ralf Baechle <ralf@linux-mips.org> Cc: Jeremy Fitzhardinge <jeremy@goop.org> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Alexey Dobriyan [Fri, 20 Oct 2006 06:28:32 +0000 (23:28 -0700)]
[PATCH] OOM killer meets userspace headers
Despite mm.h is not being exported header, it does contain one thing
which is part of userspace ABI -- value disabling OOM killer for given
process. So,
a) create and export include/linux/oom.h
b) move OOM_DISABLE define there.
c) turn bounding values of /proc/$PID/oom_adj into defines and export
them too.
Note: mass __KERNEL__ removal will be done later.
Signed-off-by: Alexey Dobriyan <adobriyan@gmail.com> Cc: Nick Piggin <nickpiggin@yahoo.com.au> Cc: David Woodhouse <dwmw2@infradead.org> Cc: Christoph Hellwig <hch@infradead.org> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
I wrote a patch to avoid redundant memory hot-add call at boot time. This
was cause of strange fail message of memory hotplug like "ACPI: add_memory
failed". Memory is recognized by early boot code with EFI/E820.
But, if DSDT describes memory devices for them, then hot-add code is called
for already recognized memory, and it shows fail messages with -EEXIST.
So, sys admin will misunderstand this message as something wrong by it.
This patch avoids them by preventing redundant hot-add call until
completion of driver initialization.
Yasunori Goto [Fri, 20 Oct 2006 06:28:30 +0000 (23:28 -0700)]
[PATCH] Change log level of a message of acpi_memhotplug to KERN_DEBUG
I suppose this message seems quite useless except debugging. It just shows
"Hotplug Mem Device". System admin can't know anything by this message.
So, I would like to change it to KERN_DEBUG.
Darrick J. Wong [Fri, 20 Oct 2006 06:28:28 +0000 (23:28 -0700)]
[PATCH] fix "ACPI: Processor native C-states using MWAIT"
This patch breaks C-state discovery on my IBM IntelliStation Z30 because
the return value of acpi_processor_get_power_info_fadt is not assigned to
"result" in the case that acpi_processor_get_power_info_cst returns
-ENODEV. Thus, if ACPI provides C-state data via the FADT and not _CST (as
is the case on this machine), we incorrectly exit the function with -ENODEV
after reading the FADT. The attached patch sets the value of result so
that we don't exit early.
Signed-off-by: Darrick J. Wong <djwong@us.ibm.com> Acked-by: "Pallipadi, Venkatesh" <venkatesh.pallipadi@intel.com> Acked-by: "Brown, Len" <len.brown@intel.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Randy Dunlap [Fri, 20 Oct 2006 06:28:26 +0000 (23:28 -0700)]
[PATCH] Kconfig serial typos
Fix typo (repeated) in serial Kconfig.
Signed-off-by: Randy Dunlap <randy.dunlap@oracle.com> Cc: Russell King <rmk@arm.linux.org.uk> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
[PATCH] uml: mmapper - remove just added but wrong "const" attribute
When enabling the mmapper driver I got warnings because this "const"
miscdevice structure is passed to function as non-const pointer; unlike struct
tty_operations, however, I verified that misc_{de,}register _do_ modify their
parameter, so this const attribute must be removed.
Since the purpose of the change was to guarantee that no lock was needed, add
a comment to prove this differently.
Signed-off-by: Paolo 'Blaisorblade' Giarrusso <blaisorblade@yahoo.it> Cc: Jeff Dike <jdike@addtoit.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Silence useless warning about undefined symbol in Kconfig.
Signed-off-by: Paolo 'Blaisorblade' Giarrusso <blaisorblade@yahoo.it> Cc: Jeff Dike <jdike@addtoit.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
[PATCH] uml: cleanup run_helper() API to fix a leak
Freeing the stack is left uselessly to the caller of run_helper in some cases
- this is taken from run_helper_thread, but here it is useless, so no caller
needs it and the only place where this happens has a potential leak - in case
of error neither run_helper() nor xterm_open() call free_stack(). At this
point passing a pointer is not needed - the stack pointer should be passed
directly, but this change is not done here.
Signed-off-by: Paolo 'Blaisorblade' Giarrusso <blaisorblade@yahoo.it> Cc: Jeff Dike <jdike@addtoit.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
[PATCH] uml: use DEFCONFIG_LIST to avoid reading host's config
This should make sure that, for UML, host's configuration files are not
considered, which avoids various pains to the user. Our dependency are such
that the obtained Kconfig will be valid and will lead to successful
compilation - however they cannot prevent an user from disabling any boot
device, and if an option is not set in the read .config (say
/boot/config-XXX), with make menuconfig ARCH=um, it is not set. This always
disables UBD and all console I/O channels, which leads to non-working UML
kernels, so this bothers users - especially now, since it will happen on
almost every machine (/boot/config-`uname -r` exists almost on every machine).
It can be workarounded with make defconfig ARCH=um, but it is non-obvious and
can be avoided, so please _do_ merge this patch.
Given the existence of options, it could be interesting to implement
(additionally) "option required" - with it, Kconfig will refuse reading a
.config file (from wherever it comes) if the given option is not set. With
this, one could mark with it the option characteristic of the given
architecture (it was an old proposal of Roman Zippel, when I pointed out our
problem):
config UML
option required
default y
However this should be further discussed:
*) for x86, it must support constructs like:
==arch/i386/Kconfig==
config 64BIT
option required
default n
where Kconfig must require that CONFIG_64BIT is disabled or not present in the
read .config.
*) do we want to do such checks only for the starting defconfig or also for
.config? Which leads to:
*) I may want to port a x86_64 .config to x86 and viceversa, or even among more
different archs. Should that be allowed, and in which measure (the user may
force skipping the check for a .config or it is only given a warning by
default)?
Cc: Roman Zippel <zippel@linux-m68k.org> Cc: <kbuild-devel@lists.sourceforge.net> Signed-off-by: Paolo 'Blaisorblade' Giarrusso <blaisorblade@yahoo.it> Cc: Jeff Dike <jdike@addtoit.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
[PATCH] uml: reenable compilation of enable_timer, disabled by mistake
CONFIG_MODE_TT does not work there, the UML_ prefixed version must be used -
this causes a link-time failure when CONFIG_MODE_TT is enabled (i.e. always
here, never by Jeff).
Signed-off-by: Paolo 'Blaisorblade' Giarrusso <blaisorblade@yahoo.it> Cc: Jeff Dike <jdike@addtoit.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Fix coding conventions violations is arch/um/os-Linux/helper.c.
Signed-off-by: Paolo 'Blaisorblade' Giarrusso <blaisorblade@yahoo.it> Cc: Jeff Dike <jdike@addtoit.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
[PATCH] uml: split memory allocation prototypes out of user.h
user.h is too generic a header name. I've split out allocation routines from
it.
Signed-off-by: Paolo 'Blaisorblade' Giarrusso <blaisorblade@yahoo.it> Cc: Jeff Dike <jdike@addtoit.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
I happened to notice that this code is a leftover and it should be removed -
since there are sporadical efforts to revive the PPC port doing such cleanups
is not useless.
Signed-off-by: Paolo 'Blaisorblade' Giarrusso <blaisorblade@yahoo.it> Cc: Jeff Dike <jdike@addtoit.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Jeff Dike [Fri, 20 Oct 2006 06:28:18 +0000 (23:28 -0700)]
[PATCH] uml: MODE_TT is bust
arch/um/sys-x86_64/ptrace.c:20:1: warning: "SC_SS" redefined
In file included from arch/um/include/sysdep/ptrace.h:18,
from include/asm/ptrace-generic.h:12,
from include/asm/ptrace.h:15,
from arch/um/sys-x86_64/ptrace.c:8:
arch/um/include/sysdep/sc.h:38:1: warning: this is the location of the previous definition
arch/um/sys-x86_64/ptrace.c: In function 'putreg':
arch/um/sys-x86_64/ptrace.c:63: warning: implicit declaration of function 'SC_FS_BASE'
arch/um/sys-x86_64/ptrace.c:63: error: invalid lvalue in unary '&'
arch/um/sys-x86_64/ptrace.c:63: warning: implicit declaration of function 'SC_GS_BASE'
arch/um/sys-x86_64/ptrace.c:63: error: invalid lvalue in unary '&'
arch/um/sys-x86_64/ptrace.c: In function 'getreg':
arch/um/sys-x86_64/ptrace.c:101: error: invalid lvalue in unary '&'
arch/um/sys-x86_64/ptrace.c:101: error: invalid lvalue in unary '&'
I'd have to say that the fix for this, for now, is this:
Cc: Paolo 'Blaisorblade' Giarrusso <blaisorblade@yahoo.it> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Ralf Baechle [Fri, 20 Oct 2006 06:28:17 +0000 (23:28 -0700)]
[PATCH] Make <linux/personality.h> userspace proof
<linux/personality.h> contains the constants for personality(2) but also
some defintions that are useless or even harmful in userspace such as the
personality() macro.
Jeff Moyer [Fri, 20 Oct 2006 06:28:13 +0000 (23:28 -0700)]
[PATCH] direct-io: sync and invalidate file region when falling back to buffered write
When direct-io falls back to buffered write, it will just leave the dirty data
floating about in pagecache, pending regular writeback.
But normal direct-io semantics are that IO is synchronous, and that it leaves
no pagecache behind.
So change the fallback-to-buffered-write code to sync the file region and to
then strip away the pagecache, just as a regular direct-io write would do.
Acked-by: Jeff Moyer <jmoyer@redhat.com> Cc: Zach Brown <zach.brown@oracle.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Auke Kok [Fri, 20 Oct 2006 06:28:12 +0000 (23:28 -0700)]
[PATCH] e100: fix reboot -f with netconsole enabled
When rebooting with netconsole over e100, the driver shutdown code would
deadlock with netpoll. Reduce shutdown code to a bare minimum while retaining
WoL and suspend functionality.
Signed-off-by: Auke Kok <auke-jan.h.kok@intel.com> Cc: Jeff Garzik <jeff@garzik.org> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Linus Torvalds [Thu, 19 Oct 2006 01:30:00 +0000 (18:30 -0700)]
Merge branch 'ubuntu-updates' of master.kernel.org:/pub/scm/linux/kernel/git/bcollins/ubuntu-2.6
* 'ubuntu-updates' of master.kernel.org:/pub/scm/linux/kernel/git/bcollins/ubuntu-2.6:
[pci_ids] Add Quicknet XJ vendor/device ID's.
[valkyriefb] Ifdef for when CONFIG_NVRAM isn't enabled.
[platinumfb] Ifdef for when CONFIG_NVRAM isn't enabled.
[igafb] Add pci dev table for module auto loading.
[controlfb] Ifdef for when CONFIG_NVRAM isn't enabled.
[hid-core] TurboX Keyboard needs NOGET quirk.
[ixj] Add pci dev table for module auto loading.
[initio] Add pci dev table for module auto loading.
[fdomain] Add pci dev table for module auto loading.
[BusLogic] Add pci dev table for auto module loading.
[mv643xx] Add pci device table for auto module loading.
[alim7101] Add pci dev table for auto module loading.
* git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/hwmon-2.6:
hwmon: Fix debug messages in w83781d
hwmon: Let w83781d and lm78 load again
w83627ehf: Fix the detection of fan5
k8temp: Documentation update
smsc47m1: List the SMSC LPC47M112 as supported
hwmon: Fix documentation typos
adm9240: Update Grant Coady's email address
w83791d: Fix unchecked return status
Jean Delvare [Fri, 13 Oct 2006 14:56:28 +0000 (16:56 +0200)]
hwmon: Let w83781d and lm78 load again
Let the w83781d and lm78 hardware monitoring drivers load even when
no chip was detected at the ISA address. There can still be supported
chips connected to an I2C bus or SMBus.
This fixes bug #7293.
Signed-off-by: Jean Delvare <khali@linux-fr.org> Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
Jean Delvare [Sun, 8 Oct 2006 20:00:44 +0000 (22:00 +0200)]
smsc47m1: List the SMSC LPC47M112 as supported
The SMSC LPC47M112 Super-I/O chip appears to be compatible with the
LPC47M10x and LPC47M13x as far as hardware monitoring is concerned.
The device ID is even the same, so it's really only a documentation
update.
Signed-off-by: Jean Delvare <khali@linux-fr.org> Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
Jim Cromie [Sun, 8 Oct 2006 19:56:29 +0000 (21:56 +0200)]
w83791d: Fix unchecked return status
Replace all unchecked calls to device_create_file with a single group
declaration, and one call to sysfs_create_group, and check that one
return status. Also remove the files on device detach.
Signed-off-by: Jim Cromie <jim.cromie@gmail.com>
Signed-off by: Charles Spirakis <bezaur@gmail.com> Signed-off-by: Jean Delvare <khali@linux-fr.org> Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
Moved the attributes into a group, making the compiler be quiet about
ignoring the return value of the file create calls. This also also
fixed a bug when removing the files, which were not symlinks.
Cc: "Ed L. Cashin" <ecashin@coraid.com> Cc: Alan Cox <alan@redhat.com> Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
Ed L. Cashin [Wed, 20 Sep 2006 18:36:51 +0000 (14:36 -0400)]
aoe: revert printk macros
This patch addresses the concern that the aoe driver should
not introduce unecessary conventions that must be learned by
the reader. It reverts patch 6.
Signed-off-by: "Ed L. Cashin" <ecashin@coraid.com> Acked-by: Alan Cox <alan@redhat.com> Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
Ed L. Cashin [Wed, 20 Sep 2006 18:36:49 +0000 (14:36 -0400)]
aoe: zero copy write 1 of 2
Avoid memory copy on writes.
(This patch depends on fixes in patch 9 to follow.)
Although skb->len should not be set when working with linear skbuffs,
the skb->tail pointer maintained by skb_put/skb_trim is not relevant
to what happens when the skb_fill_page_desc function is called. This
issue was raised without comment in linux-kernel and netdev earlier
this month:
So until there is something analogous to skb_put that works for
zero-copy write skbuffs, we will do what the other callers of
skb_fill_page_desc are doing.
Signed-off-by: "Ed L. Cashin" <ecashin@coraid.com> Acked-by: Alan Cox <alan@redhat.com> Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
Alan Stern [Thu, 5 Oct 2006 21:03:24 +0000 (17:03 -0400)]
Driver core: Don't ignore error returns from probing
This patch (as797) fixes device_add() in the driver core. It needs to
pay attention when the driver for a new device reports an error.
At the same time, since bus_remove_device() undoes the effects of both
bus_add_device() and bus_attach_device(), it needs to check whether
the bus_attach_device step failed.
Signed-off-by: Alan Stern <stern@rowland.harvard.edu> Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
sysfs: remove duplicated dput in sysfs_update_file
Following function can drops d_count twice against one reference
by lookup_one_len.
<SOURCE>
/**
* sysfs_update_file - update the modified timestamp on an object attribute.
* @kobj: object we're acting for.
* @attr: attribute descriptor.
*/
int sysfs_update_file(struct kobject * kobj, const struct attribute * attr)
{
struct dentry * dir = kobj->dentry;
struct dentry * victim;
int res = -ENOENT;
mutex_lock(&dir->d_inode->i_mutex);
victim = lookup_one_len(attr->name, dir, strlen(attr->name));
if (!IS_ERR(victim)) {
/* make sure dentry is really there */
if (victim->d_inode &&
(victim->d_parent->d_inode == dir->d_inode)) {
victim->d_inode->i_mtime = CURRENT_TIME;
fsnotify_modify(victim);
/**
* Drop reference from initial sysfs_get_dentry().
*/
dput(victim);
res = 0;
} else
d_drop(victim);
/**
* Drop the reference acquired from sysfs_get_dentry() above.
*/
dput(victim);
}
mutex_unlock(&dir->d_inode->i_mutex);
return res;
}
</SOURCE>
PCI-hotplug (drivers/pci/hotplug/pci_hotplug_core.c) is only user of
this function. I confirmed that dentry of /sys/bus/pci/slots/XXX/*
have negative d_count value.
Matthew Wilcox [Sun, 24 Sep 2006 05:35:04 +0000 (23:35 -0600)]
Fix dev_printk() is now GPL-only
Make dev_printk usable from non-GPL modules again
dev_printk now calls dev_driver_string. We want even proprietary modules
to be calling dev_printk, so the export of dev_driver_string needs to be
non-GPL-only.
Signed-off-by: Matthew Wilcox <matthew@wil.cx> Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
This makes it possible to build pci hotplug drivers outside of the main
kernel tree, and Sam keeps telling me to move local header files to
their proper places...
Matt Domsch [Fri, 29 Sep 2006 20:23:23 +0000 (15:23 -0500)]
PCI: optionally sort device lists breadth-first
Problem:
New Dell PowerEdge servers have 2 embedded ethernet ports, which are
labeled NIC1 and NIC2 on the chassis, in the BIOS setup screens, and
in the printed documentation. Assuming no other add-in ethernet ports
in the system, Linux 2.4 kernels name these eth0 and eth1
respectively. Many people have come to expect this naming. Linux 2.6
kernels name these eth1 and eth0 respectively (backwards from
expectations). I also have reports that various Sun and HP servers
have similar behavior.
Root cause:
Linux 2.4 kernels walk the pci_devices list, which happens to be
sorted in breadth-first order (or pcbios_find_device order on i386,
which most often is breadth-first also). 2.6 kernels have both the
pci_devices list and the pci_bus_type.klist_devices list, the latter
is what is walked at driver load time to match the pci_id tables; this
klist happens to be in depth-first order.
On systems where, for physical routing reasons, NIC1 appears on a
lower bus number than NIC2, but NIC2's bridge is discovered first in
the depth-first ordering, NIC2 will be discovered before NIC1. If the
list were sorted breadth-first, NIC1 would be discovered before NIC2.
A PowerEdge 1955 system has the following topology which easily
exhibits the difference between depth-first and breadth-first device
lists.
-[0000:00]-+-00.0 Intel Corporation 5000P Chipset Memory Controller Hub
+-02.0-[0000:03-08]--+-00.0-[0000:04-07]--+-00.0-[0000:05-06]----00.0-[0000:06]----00.0 Broadcom Corporation NetXtreme II BCM5708S Gigabit Ethernet (labeled NIC2, 2.4 kernel name eth1, 2.6 kernel name eth0)
+-1c.0-[0000:01-02]----00.0-[0000:02]----00.0 Broadcom Corporation NetXtreme II BCM5708S Gigabit Ethernet (labeled NIC1, 2.4 kernel name eth0, 2.6 kernel name eth1)
Other factors, such as device driver load order and the presence of
PCI slots at various points in the bus hierarchy further complicate
this problem; I'm not trying to solve those here, just restore the
device order, and thus basic behavior, that 2.4 kernels had.
Solution:
The solution can come in multiple steps.
Suggested fix #1: kernel
Patch below optionally sorts the two device lists into breadth-first
ordering to maintain compatibility with 2.4 kernels. It adds two new
command line options:
pci=bfsort
pci=nobfsort
to force the sort order, or not, as you wish. It also adds DMI checks
for the specific Dell systems which exhibit "backwards" ordering, to
make them "right".
Suggested fix #2: udev rules from userland
Many people also have the expectation that embedded NICs are always
discovered before add-in NICs (which this patch does not try to do).
Using the PCI IRQ Routing Table provided by system BIOS, it's easy to
determine which PCI devices are embedded, or if add-in, which PCI slot
they're in. I'm working on a tool that would allow udev to name
ethernet devices in ascending embedded, slot 1 .. slot N order,
subsort by PCI bus/dev/fn breadth-first. It'll be possible to use it
independent of udev as well for those distributions that don't use
udev in their installers.
Suggested fix #3: system board routing rules
One can constrain the system board layout to put NIC1 ahead of NIC2
regardless of breadth-first or depth-first discovery order. This adds
a significant level of complexity to board routing, and may not be
possible in all instances (witness the above systems from several
major manufacturers). I don't want to encourage this particular train
of thought too far, at the expense of not doing #1 or #2 above.
Feedback appreciated. Patch tested on a Dell PowerEdge 1955 blade
with 2.6.18.
You'll also note I took some liberty and temporarily break the klist
abstraction to simplify and speed up the sort algorithm. I think
that's both safe and appropriate in this instance.
Signed-off-by: Matt Domsch <Matt_Domsch@dell.com> Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
Akinobu Mita [Fri, 13 Oct 2006 18:07:30 +0000 (03:07 +0900)]
cpcihp_generic: prevent loading without "bridge" parameter
cpcihp_generic module requires configured "bridge" module parameter.
But it can be loaded successfully without that parameter.
Because module init call ends up returning positive value.
This patch prevents from loading without setting "bridge" module parameter.
Signed-off-by: Akinbou Mita <akinobu.mita@gmail.com> Signed-off-by: Scott Murray <scottm@somanetworks.com> Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
PCI: fix pcie_portdrv_restore_config undefined without CONFIG_PM error
On Thu, 2006-09-28 at 03:42, Olaf Hering wrote:
> PCI-Express AER implemetation: pcie_portdrv error handler
>
> This patch breaks if CONFIG_PM is not enabled,
> pcie_portdrv_restore_config() will be undefined.
I move the definition of pcie_portdrv_restore_config
out of CONFIG_PM.
Below patch is against 2.6.18-mm1. Could you try it?
The ACPIPnP implementation had the understanding of Linux resource flags very
wrong, resulting in a nonfunctional implementation of DMA resource
allocation.
This was usually not a problem, since almost no on-board PnP devices use ISA
DMA, with the exception of ECP parallel ports. Even with that, parallel port
DMA is preconfigured by the BIOS, so this routine isn't normally called.
Except in the case where somebody does 'rmmod parport_pc; modprobe
parport_pc', where the rmmod case disables the ECP parallel port resources,
and they need to be enabled again to initialize the module. This didn't
work, resulting in a non-printing printer.
The application doing exactly the above to force reprobing of printers is
the YaST printer module. Thus without this fix YaST wedged the printer when
configuring it, and was not able to print a test page.
PCI: Turn pci_fixup_video into generic for embedded VGA
pci_fixup_video turns into generic code because there are many platforms need this fixup
for embedded VGA as well as x86. The Video BIOS integrates into System BIOS on a machine
has embedded VGA although embedded VGA generally don't have PCI ROM. As a result,
embedded VGA need the way that the sysfs rom points to the Video BIOS of System
RAM (0xC0000). PCI-to-PCI Bridge Architecture specification describes the condition whether
or not PCI ROM forwards VGA compatible memory address. fixup_video suits this specification.
Although the Video ROM generally implements in x86 code regardless of platform, some
application such as X Window System can run this code by dosemu86. Therefore,
pci_fixup_video should turn into generic code.
Daniel Drake [Mon, 25 Sep 2006 23:52:19 +0000 (16:52 -0700)]
PCI: VIA IRQ quirk behaviour change
The most recent VIA IRQ quirk changes have broken various VIA devices for
some users. We are not able to add these devices to the blacklist as they
are also available in PCI-card form, and running the quirk on these devices
brings us back to square one (running the VIA quirk on non-VIA boards where
the quirk is not needed).
This patch, based on suggestions from Sergey Vlasov, implements a scheme
similar to but more restrictive than the scheme we had in 2.6.16 and
earlier. It runs the quirk on all VIA hardware, but *only* if a VIA
southbridge was detected on the system.
To further reduce the amount of quirked devices, this patch includes a
change suggested by Linus at http://lkml.org/lkml/2005/9/27/113 This
ensures that devices bound to non-legacy IO-APIC interrupt lines are not
quirked. We have made one change to Linus' suggestion: we do a comparison
of ">15" rather than ">=15", as 15 is still in the legacy interrupt range.
There is still a downside to this patch: if the user inserts a VIA PCI card
into a VIA-based motherboard, in some circumstances the quirk will also run
on the VIA PCI card. This corner case is hard to avoid.
Signed-off-by: Daniel Drake <dsd@gentoo.org> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
projects / linux-2.6 / log