GitHub/LineageOS/android_kernel_samsung_universal7580.git
17 years agolguest.h declares a struct timespec, make it include linux/time.h
Jes Sorensen [Mon, 22 Oct 2007 01:03:33 +0000 (11:03 +1000)]
lguest.h declares a struct timespec, make it include linux/time.h

Signed-off-by: Jes Sorensen <jes@sgi.com>
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
17 years agoMove register setup into i386_core.c
Jes Sorensen [Mon, 22 Oct 2007 01:03:32 +0000 (11:03 +1000)]
Move register setup into i386_core.c

Move setup_regs() to lguest_arch_setup_regs() in i386_core.c given
that this is very architecture specific.

Signed-off-by: Jes Sorensen <jes@sgi.com>
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
17 years agoChange example launcher to use unsigned long not u32
Jes Sorensen [Mon, 22 Oct 2007 01:03:31 +0000 (11:03 +1000)]
Change example launcher to use unsigned long not u32

Apply Clue 2x4 to lguest userland<->kernel handling code and the
lguest launcher. Pointers are not to be passed in u32's!

Basic rule of thumb: Anything passing u32's back and forth should be
passing unsigned longs to be portable to 64 bit archs.

For those who forgotten already, I repeat: NO POINTERS IN u32!

Signed-off-by: Jes Sorensen <jes@sgi.com>
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
17 years agoMake hypercalls arch-independent.
Jes Sorensen [Mon, 22 Oct 2007 01:03:31 +0000 (11:03 +1000)]
Make hypercalls arch-independent.

Clean up the hypercall code to make the code in hypercalls.c
architecture independent. First process the common hypercalls and
then call lguest_arch_do_hcall() if the call hasn't been handled.
Rename struct hcall_ring to hcall_args.

This patch requires the previous patch which reorganize the layout of
struct lguest_regs on i386 so they match the layout of struct
hcall_args.

Signed-off-by: Jes Sorensen <jes@sgi.com>
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
17 years agoIntroduce "hcall" pointer to indicate pending hypercall.
Rusty Russell [Mon, 22 Oct 2007 01:03:30 +0000 (11:03 +1000)]
Introduce "hcall" pointer to indicate pending hypercall.

Currently we look at the "trapnum" to see if the Guest wants a
hypercall.  But once the hypercall is done we have to reset trapnum to
a bogus value, otherwise if we exit to userspace and return, we'd run
the same hypercall twice (that was a nasty bug to find!).

This has two main effects:

1) When Jes's patch changes the hypercall args to be a generic "struct
   hcall_args" we simply change the type of "lg->hcall".  It's set by
   arch code, so if it has to copy args or something it can do so, and
   point "hcall" into lg->arch somewhere.

2) Async hypercalls only get run when an actual hypercall is pending.
   This simplfies the code a little and is a more logical semantic.

Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
17 years agoReorder guest saved regs to match hyperall order
Jes Sorensen [Mon, 22 Oct 2007 01:03:29 +0000 (11:03 +1000)]
Reorder guest saved regs to match hyperall order

Move eax next to ebx/ecx/edx in struct lguest_regs on i386, so they
will be located together and allow it to map directly to a struct
hcall_ring entry (which will be renamed struct hcall_args as in a
subsequent patch).

This is in preparation for making the code hcall code architecture
independent.

Signed-off-by: Jes Sorensen <jes@sgi.com>
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
17 years agoMove i386 part of core.c to x86/core.c.
Jes Sorensen [Mon, 22 Oct 2007 01:03:28 +0000 (11:03 +1000)]
Move i386 part of core.c to x86/core.c.

Separate i386 architecture specific from core.c and move it to
x86/core.c and add x86/lguest.h header file to match.

Signed-off-by: Jes Sorensen <jes@sgi.com>
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
17 years agoMake shadow IDT a complete IDT with 256 entries.
Rusty Russell [Mon, 22 Oct 2007 01:03:28 +0000 (11:03 +1000)]
Make shadow IDT a complete IDT with 256 entries.

This simplifies the code a little, in preparation for allowing
alternate system call vectors in guests (Plan 9 uses 0x40).

Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
17 years agoRemove fixed limit on number of guests, and lguests array.
Rusty Russell [Mon, 22 Oct 2007 01:03:27 +0000 (11:03 +1000)]
Remove fixed limit on number of guests, and lguests array.

Back when we had all the Guest state in the switcher, we had a fixed
array of them.  This is no longer necessary.

If we switch the network code to using random_ether_addr (46 bits is
enough to avoid clashes), we can get rid of the concept of "guest id"
altogether.

Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
17 years agoIntroduce guest mem offset, static link example launcher
Rusty Russell [Mon, 22 Oct 2007 01:03:26 +0000 (11:03 +1000)]
Introduce guest mem offset, static link example launcher

In order to avoid problematic special linking of the Launcher, we give
the Host an offset: this means we can use any memory region in the
Launcher as Guest memory rather than insisting on mmap() at 0.

The result is quite pleasing: a number of casts are replaced with
simple additions.

Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
17 years agoAccept elf files that are valid but have sections that can not be mmap'ed for some...
Ronald G. Minnich [Tue, 28 Aug 2007 21:35:59 +0000 (14:35 -0700)]
Accept elf files that are valid but have sections that can not be mmap'ed for some reason.

Plan9 kernel binaries don't neatly align their ELF sections to our
page boundaries.

Signed-off-by: Ronald G. Minnich <rminnich@gmail.com>
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
17 years agoRename switcher.S to x86/switcher_32.S
Rusty Russell [Mon, 22 Oct 2007 01:03:25 +0000 (11:03 +1000)]
Rename switcher.S to x86/switcher_32.S

lguest uses a "switcher" shim mapped high to bounce between host and
guest.  As lguest becomes less i386-centric, we separate this code
into a subdir.

Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
17 years agoMove lguest guest support to arch/x86.
Rusty Russell [Mon, 22 Oct 2007 01:01:54 +0000 (11:01 +1000)]
Move lguest guest support to arch/x86.

Lguest has two sides: host support (to launch guests) and guest
support (replacement boot path and paravirt_ops).  This moves the
guest side to arch/x86/lguest where it's closer to related code.

Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Cc: Andi Kleen <ak@suse.de>
17 years agoMove lguest hcalls to arch-specific header
Jes Sorensen [Mon, 22 Oct 2007 00:56:26 +0000 (10:56 +1000)]
Move lguest hcalls to arch-specific header

Move architecture specific portion of lg_hcall code to asm-i386/lg_hcall.h
and have it included from linux/lguest.h.

[Changed to asm-i386/lguest_hcall.h so documentation finds it -RR]

Signed-off-by: Jes Sorensen <jes@sgi.com>
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Cc: Jes Sorensen <jes@sgi.com>
17 years agoClocksource is continuous regardless of the state of the host's TSC.
Tony Breeds [Mon, 22 Oct 2007 00:56:25 +0000 (10:56 +1000)]
Clocksource is continuous regardless of the state of the host's TSC.

Currently lguest will spend a lot of of time waking up the host, as it
cannot go tickless (if the [host] TSC has been marked unstable). On my
laptop I was getting ~40% of wakeups from lguest.

With this patch applied, my laptop is much happier!

Signed-off-by: Tony Breeds <tony@bakeyournoodle.com>
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
17 years agolguest_devices belongs in lguest_bus.c: it's not i386-specific.
Rusty Russell [Mon, 22 Oct 2007 00:56:24 +0000 (10:56 +1000)]
lguest_devices belongs in lguest_bus.c: it's not i386-specific.

Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
17 years agoMake lguest_launcher.h types userspace-friendly
Rusty Russell [Mon, 22 Oct 2007 00:56:24 +0000 (10:56 +1000)]
Make lguest_launcher.h types userspace-friendly

lguest_launcher.h uses "u32" not "__u32", which sets a bad example.  Fix that,
and include <linux/types.h>.

This means we need to use -I on the Launcher build line so types.h is found.

Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
17 years agolguest.txt update
Rusty Russell [Mon, 22 Oct 2007 00:56:23 +0000 (10:56 +1000)]
lguest.txt update

o Describe the new split configurations
o Highlight code documentation in drivers/lguest/README
o Point out necessity of having a getty on /dev/hvc0
o Remove gratuitous "m" in example
o Don't discuss I/O model here, stick to user documentation.

Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
17 years agoLguest currently depends on 32-bit x86, not just x86.
Rusty Russell [Mon, 22 Oct 2007 00:56:22 +0000 (10:56 +1000)]
Lguest currently depends on 32-bit x86, not just x86.

Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
17 years agoUse copy_to_user() not put_user for struct timespec
Jes Sorensen [Mon, 22 Oct 2007 00:56:22 +0000 (10:56 +1000)]
Use copy_to_user() not put_user for struct timespec

Use copy_to_user() when copying a struct timespec to the guest -
put_user() cannot handle two long's in one go on a 64bit arch.

Signed-off-by: Jes Sorensen <jes@sgi.com>
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Cc: Jes Sorensen <jes@sgi.com>
Cc: Al Viro <viro@ftp.linux.org.uk>
17 years agoturn err into errx in lguest call sites
Glauber de Oliveira Costa [Mon, 22 Oct 2007 00:56:21 +0000 (10:56 +1000)]
turn err into errx in lguest call sites

These two callsites should really be errx instead of err, since there is
no errno associated with them in the moment they are issued.

Signed-off-by: Glauber de Oliveira Costa <gcosta@redhat.com>
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Cc: Glauber de Oliveira Costa <gcosta@redhat.com>
17 years agoRemove binfmts.h include from lg.h
Rusty Russell [Tue, 23 Oct 2007 05:19:49 +0000 (15:19 +1000)]
Remove binfmts.h include from lg.h

It wasn't needed since a very early prototype of lguest.

Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
17 years agoMake asm-x86/bootparam.h includable from userspace.
Rusty Russell [Mon, 22 Oct 2007 00:56:19 +0000 (10:56 +1000)]
Make asm-x86/bootparam.h includable from userspace.

To actually write a bootloader (or, say, the lguest launcher)
currently requires duplication of these structures.  Making them
includable from userspace is much nicer.

We merge the common userspace-required definitions of e820_32/64.h
into e820.h for export.

Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
17 years agoConsolidate host virtualization support under Virtualization menu
Rusty Russell [Mon, 22 Oct 2007 00:55:43 +0000 (10:55 +1000)]
Consolidate host virtualization support under Virtualization menu

Move lguest under the virtualization menu.

Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Cc: Avi Kivity <avi@qumranet.com>
17 years agoNormalize config options for guest support
Rusty Russell [Mon, 22 Oct 2007 00:55:21 +0000 (10:55 +1000)]
Normalize config options for guest support

1) Group all the "guest OS" support options together, under a PARAVIRT_GUEST
   menu.
2) Make those options select CONFIG_PARAVIRT, as suggested by Andi.
3) Make kconfig help titles consistent.

Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Cc: Andi Kleen <ak@suse.de>
Cc: Zach Amsden <zach@vmware.com>
Cc: Jeremy Fitzhardinge <jeremy@goop.org>
Cc: Chris Wright <chrisw@sous-sol.org>
17 years agoprocfs: fix kernel-doc param warnings
Randy Dunlap [Mon, 22 Oct 2007 04:00:10 +0000 (21:00 -0700)]
procfs: fix kernel-doc param warnings

Fix mnt_flush_task() misplaced kernel-doc.
Fix typos in some of the doc text.

Warning(linux-2.6.23-git17//fs/proc/base.c:2280): No description found for parameter 'mnt'
Warning(linux-2.6.23-git17//fs/proc/base.c:2280): No description found for parameter 'pid'
Warning(linux-2.6.23-git17//fs/proc/base.c:2280): No description found for parameter 'tgid'

Signed-off-by: Randy Dunlap <randy.dunlap@oracle.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years agoauditsc: fix kernel-doc param warnings
Randy Dunlap [Mon, 22 Oct 2007 03:59:53 +0000 (20:59 -0700)]
auditsc: fix kernel-doc param warnings

Fix kernel-doc for auditsc parameter changes.

Warning(linux-2.6.23-git17//kernel/auditsc.c:1623): No description found for parameter 'dentry'
Warning(linux-2.6.23-git17//kernel/auditsc.c:1666): No description found for parameter 'dentry'

Signed-off-by: Randy Dunlap <randy.dunlap@oracle.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years agoExpand hwif->host_flags so that it fits new flags.
David Miller [Mon, 22 Oct 2007 09:09:00 +0000 (02:09 -0700)]
Expand hwif->host_flags so that it fits new flags.

Commit 238e4f142c33bb34440cc64029dde7b9fbc4e65f ("ide: add
IDE_HFLAG_NO_LBA48 and IDE_HFLAG_NO_LBA48_DMA host flags") caused a
regression because the host_flags in struct hwif_s wasn't expanded to
cope with the fact that the host flags no longer fit in 16 bits.

Signed-off-by: David S. Miller <davem@davemloft.net>
[ I hate having to add good commit descriptions.  - Linus ]
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years agoMerge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/dtor/input
Linus Torvalds [Tue, 23 Oct 2007 02:29:58 +0000 (19:29 -0700)]
Merge branch 'for-linus' of git://git./linux/kernel/git/dtor/input

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/dtor/input:
  Input: appletouch - apply idle reset logic to all touchpads
  Input: usbtouchscreen - add support for GoTop tablet devices
  Input: bf54x-keys - return real error when request_irq() fails
  Input: i8042 - export i8042_command()

17 years agoMerge branch 'for-linus' of master.kernel.org:/pub/scm/linux/kernel/git/cooloney...
Linus Torvalds [Tue, 23 Oct 2007 02:29:05 +0000 (19:29 -0700)]
Merge branch 'for-linus' of /linux/kernel/git/cooloney/blackfin-2.6

* 'for-linus' of master.kernel.org:/pub/scm/linux/kernel/git/cooloney/blackfin-2.6:
  Blackfin arch: use KBUILD_CFLAGS and KBUILD_AFLAGS in Makefile
  Blackfin arch: Javier Herrer writes: fix building when icache and dcache is disabled

17 years agoMerge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/avi/kvm
Linus Torvalds [Tue, 23 Oct 2007 02:24:17 +0000 (19:24 -0700)]
Merge branch 'for-linus' of git://git./linux/kernel/git/avi/kvm

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/avi/kvm:
  KVM: Use new smp_call_function_mask() in kvm_flush_remote_tlbs()
  sched: don't clear PF_VCPU in scheduler
  KVM: Improve local apic timer wraparound handling
  KVM: Fix local apic timer divide by zero
  KVM: Move kvm_guest_exit() after local_irq_enable()
  KVM: x86 emulator: fix access registers for instructions with ModR/M byte and Mod = 3
  KVM: VMX: Force vm86 mode if setting flags during real mode
  KVM: x86 emulator: implement 'movnti mem, reg'
  KVM: VMX: Reset mmu context when entering real mode
  KVM: VMX: Handle NMIs before enabling interrupts and preemption
  KVM: MMU: Set shadow pte atomically in mmu_pte_write_zap_pte()
  KVM: x86 emulator: fix repne/repnz decoding
  KVM: x86 emulator: fix merge screwup due to emulator split

17 years agoMerge branch 'for-linus' of git://git390.osdl.marist.edu/pub/scm/linux-2.6
Linus Torvalds [Tue, 23 Oct 2007 02:23:34 +0000 (19:23 -0700)]
Merge branch 'for-linus' of git://git390.osdl.marist.edu/linux-2.6

* 'for-linus' of git://git390.osdl.marist.edu/pub/scm/linux-2.6:
  [S390] 4level-fixup cleanup
  [S390] Cleanup page table definitions.
  [S390] Introduce follow_table in uaccess_pt.c
  [S390] Remove unused user_seg from thread structure.
  [S390] tlb flush fix.
  [S390] kernel: Fix dump on panic for DASDs under LPAR.
  [S390] struct class_device -> struct device conversion.
  [S390] cio: Fix incomplete commit for uevent suppression.
  [S390] cio: Use to_channelpath() for device to channel path conversion.
  [S390] Add per-cpu idle time / idle count sysfs attributes.
  [S390] Update default configuration.

17 years agoMerge branch 'master' of master.kernel.org:/pub/scm/linux/kernel/git/davem/sparc-2.6
Linus Torvalds [Tue, 23 Oct 2007 02:22:56 +0000 (19:22 -0700)]
Merge branch 'master' of /linux/kernel/git/davem/sparc-2.6

* 'master' of master.kernel.org:/pub/scm/linux/kernel/git/davem/sparc-2.6:
  [SPARC64]: ARRAY_SIZE() cleanup
  [SPARC64]: small Makefile cleanups
  [SPARC64]: Stop using __do_IRQ().
  [SPARC64]: Update defconfig.

17 years agoMerge branch 'master' of master.kernel.org:/pub/scm/linux/kernel/git/davem/net-2.6
Linus Torvalds [Tue, 23 Oct 2007 02:22:33 +0000 (19:22 -0700)]
Merge branch 'master' of /linux/kernel/git/davem/net-2.6

* 'master' of master.kernel.org:/pub/scm/linux/kernel/git/davem/net-2.6: (30 commits)
  [IPSEC] IPV6: Fix to add tunnel mode SA correctly.
  [NET]: Cut off the queue_mapping field from sk_buff
  [NET]: Hide the queue_mapping field inside netif_subqueue_stopped
  [NET]: Make and use skb_get_queue_mapping
  [NET]: Use the skb_set_queue_mapping where appropriate
  [INET]: Use MODULE_ALIAS_NET_PF_PROTO_TYPE where possible.
  [INET]: Let inet_diag and friends autoload
  [NIU]: Cleanup PAGE_SIZE checks a bit
  [NET]: Fix SKB_WITH_OVERHEAD calculation
  [ATM]: Fix clip module reload crash.
  [TG3]: Update version to 3.85
  [TG3]: PCI command adjustment
  [TG3]: Add management FW version to ethtool report
  [TG3]: Add 5723 support
  [Bluetooth] Convert RFCOMM to use kthread API
  [Bluetooth] Add constant for Bluetooth socket options level
  [Bluetooth] Add support for handling simple eSCO links
  [Bluetooth] Add address and channel attribute to RFCOMM TTY device
  [Bluetooth] Fix wrong argument in debug code of HIDP
  [Bluetooth] Add generic driver for Bluetooth USB devices
  ...

17 years agoMerge branch 'master' of git://git.kernel.org/pub/scm/linux/kernel/git/paulus/powerpc
Linus Torvalds [Tue, 23 Oct 2007 02:21:54 +0000 (19:21 -0700)]
Merge branch 'master' of git://git./linux/kernel/git/paulus/powerpc

* 'master' of git://git.kernel.org/pub/scm/linux/kernel/git/paulus/powerpc:
  [POWERPC] Enable restart support for lite5200 board
  [POWERPC] Add restart support for mpc52xx based platforms
  [POWERPC] Update device tree binding for mpc5200 gpt
  [POWERPC] Add mpc52xx_find_and_map_path(), refactor utility functions
  [POWERPC] bestcomm: Restrict bus prefetch bugfix to original mpc5200 silicon.

17 years agoMerge branch 'upstream' of git://ftp.linux-mips.org/pub/scm/upstream-linus
Linus Torvalds [Tue, 23 Oct 2007 02:21:08 +0000 (19:21 -0700)]
Merge branch 'upstream' of git://ftp.linux-mips.org/upstream-linus

* 'upstream' of git://ftp.linux-mips.org/pub/scm/upstream-linus:
  [MIPS] time: Make c0_compare_int_usable more bullet proof
  [MIPS] Kbuild: Use the new cc-cross-prefix feature.
  [MIPS] Fix include wrapper symbol to something sane.
  [MIPS] Malta: Delete dead code.
  [MIPS] time: Add GT641xx timer0 clockevent driver
  [MIPS] time: SMP-proofing of Sibyte clockevent/clocksource code.
  [MIPS] time: SMP/NUMA-proofing of IP27 HUB RT timer code.
  [MIPS] time: Fix calculation in clockevent_set_clock()

17 years agoMerge git://git.infradead.org/battery-2.6
Linus Torvalds [Tue, 23 Oct 2007 02:20:52 +0000 (19:20 -0700)]
Merge git://git.infradead.org/battery-2.6

* git://git.infradead.org/battery-2.6:
  apm_power: calculate to_full/to_empty time using energy
  apm_power: improve battery finding algorithm
  apm_power: fix obviously wrong logic for time reporting

17 years agoMerge branch 'master' of ssh://master.kernel.org/pub/scm/linux/kernel/git/mchehab...
Linus Torvalds [Tue, 23 Oct 2007 02:20:22 +0000 (19:20 -0700)]
Merge branch 'master' of ssh:///linux/kernel/git/mchehab/v4l-dvb

* 'master' of ssh://master.kernel.org/pub/scm/linux/kernel/git/mchehab/v4l-dvb: (37 commits)
  V4L/DVB (6382): saa7134: fix NULL dereference at suspend time for cards without IR receiver
  V4L/DVB (6380): ivtvfb: Removal of the 'osd_compat' module option
  V4L/DVB (6379): patch which improves GotView Saa7135 remote control
  V4L/DVB (6378b): Updates info about the removal of V4L1 at feature-removal-schedule.txt
  V4L/DVB (6378a): Removal of VIDIOC_[G|S]_MPEGCOMP from feature-removal-schedule.txt
  V4L/DVB (6378): DiB0700-device: Using 1.10 firmware
  V4L/DVB (6357): pvrusb2: Improve encoder chip health tracking
  V4L/DVB (6356): "while (!ca->wakeup)" breaks the CAM initialisation
  V4L/DVB (6352): ir-kbd-i2c: Missing break statement
  V4L/DVB (6350): V4L: possible leak in em28xx_init_isoc
  V4L/DVB (6348): ivtv: undo video mute when closing the radio
  V4L/DVB (6347): ivtv: fix video mute when radio is used
  V4L/DVB (6346): ivtvfb: YUV output size fix when ivtvfb is not loaded
  V4L/DVB (6345): ivtvfb: YUV handling of an image which is not visible in the display area
  V4L/DVB (6343): ivtvfb: check return value of unregister_framebuffer
  V4L/DVB (6342): ivtv: fix circular locking (bug 9037)
  V4L/DVB (6341): ivtv: fix resizing MPEG1 streams
  V4L/DVB (6340): ivtvfb: screen mode change sometimes goes wrong
  V4L/DVB (6339): ivtv: set the video color to black instead of green when capturing from the radio
  V4L/DVB (6338): ivtv: fix incorrect EBUSY return
  ...

17 years agoppc: fix AT_VECTOR_SIZE on arch/ppc
Grant Likely [Mon, 22 Oct 2007 22:38:32 +0000 (16:38 -0600)]
ppc: fix AT_VECTOR_SIZE on arch/ppc

Commit 4f9a58d75bfe82ab2b8ba5b8506dfb190a267834 ("increase
AT_VECTOR_SIZE to terminate saved_auxv properly") changes the size of
AT_VECTOR_SIZE from hard coded '44' to a calculation based on the value
of AT_VECTOR_SIZE_ARCH and AT_VECTOR_SIZE_BASE.

The change works for arch/powerpc, but it breaks arch/ppc because the
needed AT_VECTOR_SIZE_ARCH is not present in include/asm-ppc/system.h
and a default value of 0 is used instead.  This results in
AT_VECTOR_SIZE being too small and it causes a kernel crash on loading
init.

Signed-off-by: Grant Likely <grant.likely@secretlab.ca>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years agoMerge ssh://master.kernel.org/pub/scm/linux/kernel/git/sam/kbuild
Linus Torvalds [Tue, 23 Oct 2007 02:16:17 +0000 (19:16 -0700)]
Merge ssh:///pub/scm/linux/kernel/git/sam/kbuild

* ssh://master.kernel.org/pub/scm/linux/kernel/git/sam/kbuild:
  kbuild: cc-cross-prefix spelling
  kbuild: allow depmod in cross builds again
  kbuild: fix modules_install after a 'make vmlinux'

17 years agoFix appletalk sysctl entry name
Eric W. Biederman [Mon, 22 Oct 2007 18:55:36 +0000 (12:55 -0600)]
Fix appletalk sysctl entry name

Gabriel C reported that modprobing appletalk on current git gives a
warning in dmesg :

   "sysctl table check failed: /net/appletalk .3.7 procname does not match binary path procname"

Oops.  My apologies it appears I made a mistake when creating my table
to check up on sysctl values.

Signed-off-by: "Eric W. Biederman" <ebiederm@xmission.com>
Tested-by: Gabriel C <nix.or.die@googlemail.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years agoMerge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/ieee1394...
Linus Torvalds [Tue, 23 Oct 2007 02:14:05 +0000 (19:14 -0700)]
Merge branch 'for-linus' of git://git./linux/kernel/git/ieee1394/linux1394-2.6

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/ieee1394/linux1394-2.6:
  firewire: fw-ohci: shut up a superfluous compiler warning
  firewire: fw-ohci: log a note about unsupported features

17 years agoMerge branch 'sg' of git://git.kernel.dk/linux-2.6-block
Linus Torvalds [Tue, 23 Oct 2007 02:11:06 +0000 (19:11 -0700)]
Merge branch 'sg' of git://git.kernel.dk/linux-2.6-block

* 'sg' of git://git.kernel.dk/linux-2.6-block:
  Add CONFIG_DEBUG_SG sg validation
  Change table chaining layout
  Update arch/ to use sg helpers
  Update swiotlb to use sg helpers
  Update net/ to use sg helpers
  Update fs/ to use sg helpers
  [SG] Update drivers to use sg helpers
  [SG] Update crypto/ to sg helpers
  [SG] Update block layer to use sg helpers
  [SG] Add helpers for manipulating SG entries

17 years ago[SPARC64]: ARRAY_SIZE() cleanup
Alejandro Martinez Ruiz [Tue, 23 Oct 2007 00:24:19 +0000 (17:24 -0700)]
[SPARC64]: ARRAY_SIZE() cleanup

Signed-off-by: Alejandro Martinez Ruiz <alex@flawedcode.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
17 years agoMerge branch 'for-2.6.24' of git://git.secretlab.ca/git/linux-2.6-mpc52xx
Paul Mackerras [Mon, 22 Oct 2007 22:45:23 +0000 (08:45 +1000)]
Merge branch 'for-2.6.24' of git://git.secretlab.ca/git/linux-2.6-mpc52xx

17 years ago[MIPS] time: Make c0_compare_int_usable more bullet proof
Atsushi Nemoto [Mon, 22 Oct 2007 16:14:06 +0000 (01:14 +0900)]
[MIPS] time: Make c0_compare_int_usable more bullet proof

Use write_c0_compare(read_c0_count()) to clear interrupt.

Signed-off-by: Atsushi Nemoto <anemo@mba.ocn.ne.jp>
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
17 years ago[MIPS] Kbuild: Use the new cc-cross-prefix feature.
Ralf Baechle [Mon, 22 Oct 2007 18:10:57 +0000 (19:10 +0100)]
[MIPS] Kbuild: Use the new cc-cross-prefix feature.

Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
17 years ago[MIPS] Fix include wrapper symbol to something sane.
Ralf Baechle [Mon, 22 Oct 2007 14:17:39 +0000 (15:17 +0100)]
[MIPS] Fix include wrapper symbol to something sane.

And why are there i8253.h and 8253pit.h ...

Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
17 years ago[MIPS] Malta: Delete dead code.
Ralf Baechle [Mon, 22 Oct 2007 14:13:22 +0000 (15:13 +0100)]
[MIPS] Malta: Delete dead code.

Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
17 years ago[MIPS] time: Add GT641xx timer0 clockevent driver
Yoichi Yuasa [Mon, 22 Oct 2007 10:43:15 +0000 (19:43 +0900)]
[MIPS] time: Add GT641xx timer0 clockevent driver

And make use of it for Cobalt.  A few others such as the Malta could make
use of it as well.

Signed-off-by: Yoichi Yuasa <yoichi_yuasa@tripeaks.co.jp>
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
17 years ago[MIPS] time: SMP-proofing of Sibyte clockevent/clocksource code.
Ralf Baechle [Mon, 22 Oct 2007 09:38:44 +0000 (10:38 +0100)]
[MIPS] time: SMP-proofing of Sibyte clockevent/clocksource code.

The BCM148 has 4 cores but there are also just 4 generic timers available
so use the ZBbus cycle counter instead of it.  In addition the ZBbus
counter also offers a much higher resolution and 64-bit counting so I'm
considering a later complete conversion to it once I figure out if all
members of the Sibyte SOC family support it - the docs seem to agree but
the headers files seem to disagree ...

Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
17 years ago[MIPS] time: SMP/NUMA-proofing of IP27 HUB RT timer code.
Ralf Baechle [Mon, 22 Oct 2007 09:34:13 +0000 (10:34 +0100)]
[MIPS] time: SMP/NUMA-proofing of IP27 HUB RT timer code.

Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
17 years ago[MIPS] time: Fix calculation in clockevent_set_clock()
Atsushi Nemoto [Fri, 19 Oct 2007 15:28:33 +0000 (00:28 +0900)]
[MIPS] time: Fix calculation in clockevent_set_clock()

Signed-off-by: Atsushi Nemoto <anemo@mba.ocn.ne.jp>
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
17 years agoAdd CONFIG_DEBUG_SG sg validation
Jens Axboe [Mon, 22 Oct 2007 18:01:06 +0000 (20:01 +0200)]
Add CONFIG_DEBUG_SG sg validation

Add a Kconfig entry which will toggle some sanity checks on the sg
entry and tables.

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
17 years agoChange table chaining layout
Jens Axboe [Mon, 22 Oct 2007 17:57:20 +0000 (19:57 +0200)]
Change table chaining layout

Change the page member of the scatterlist structure to be an unsigned
long, and encode more stuff in the lower bits:

- Bits 0 and 1 zero: this is a normal sg entry. Next sg entry is located
  at sg + 1.
- Bit 0 set: this is a chain entry, the next real entry is at ->page_link
  with the two low bits masked off.
- Bit 1 set: this is the final entry in the sg entry. sg_next() will return
  NULL when passed such an entry.

It's thus important that sg table users use the proper accessors to get
and set the page member.

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
17 years agoUpdate arch/ to use sg helpers
Jens Axboe [Mon, 22 Oct 2007 18:02:46 +0000 (20:02 +0200)]
Update arch/ to use sg helpers

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
17 years agoUpdate swiotlb to use sg helpers
Jens Axboe [Mon, 22 Oct 2007 17:44:53 +0000 (19:44 +0200)]
Update swiotlb to use sg helpers

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
17 years agoUpdate net/ to use sg helpers
Jens Axboe [Mon, 22 Oct 2007 17:44:26 +0000 (19:44 +0200)]
Update net/ to use sg helpers

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
17 years agoUpdate fs/ to use sg helpers
Jens Axboe [Mon, 22 Oct 2007 17:43:30 +0000 (19:43 +0200)]
Update fs/ to use sg helpers

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
17 years ago[SG] Update drivers to use sg helpers
Jens Axboe [Mon, 22 Oct 2007 19:19:53 +0000 (21:19 +0200)]
[SG] Update drivers to use sg helpers

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
17 years agokbuild: cc-cross-prefix spelling
Geert Uytterhoeven [Sun, 21 Oct 2007 17:55:00 +0000 (19:55 +0200)]
kbuild: cc-cross-prefix spelling

Spelling fixes for cc-cross-prefix documentation

Signed-off-by: Geert Uytterhoeven <geert@linux-m68k.org>
Signed-off-by: Sam Ravnborg <sam@ravnborg.org>
17 years agokbuild: allow depmod in cross builds again
Sam Ravnborg [Mon, 22 Oct 2007 09:30:45 +0000 (11:30 +0200)]
kbuild: allow depmod in cross builds again

depmod from module-init-tools 3.3-pre2 are reported
to work fine in cross build.
depmod from module-init-tools 3.1-pre5 are known to SEGV

Do not workaround older module-init-tools bugs here.
The right fix is for users to upgrade module-init-tools.

Signed-off-by: Sam Ravnborg <sam@ravnborg.org>
Cc: Geert Uytterhoeven <geert@linux-m68k.org>
17 years agokbuild: fix modules_install after a 'make vmlinux'
Sam Ravnborg [Mon, 22 Oct 2007 09:23:07 +0000 (11:23 +0200)]
kbuild: fix modules_install after a 'make vmlinux'

make vmlinux would delete the content of $(MODVERDIR)
equals .tmp_versions. This caused a subsequent
make modules_install to fail.

Fix it so we clean the directory only for the
modules build - but we still unconditionally create it so
we can do:
make dir/file.ko
without a preceeding make modules.

Reported by David Miller <davem@davemloft.net>

Signed-off-by: Sam Ravnborg <sam@ravnborg.org>
Cc: David Miller <davem@davemloft.net>
17 years agofirewire: fw-ohci: shut up a superfluous compiler warning
Stefan Richter [Sun, 21 Oct 2007 09:20:07 +0000 (11:20 +0200)]
firewire: fw-ohci: shut up a superfluous compiler warning

New warning since commit ab88ca488b8af66c3defa165874e81e695319a19,
"firewire: fw-ohci: missing dma_unmap_single":
drivers/firewire/fw-ohci.c: In function 'at_context_transmit':
drivers/firewire/fw-ohci.c:609: warning: 'payload_bus' may be used
 uninitialized in this function

Access to payload_bus is conditional on packet->payload_length > 0,
and that won't change while in at_context_queue_packet.

Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>
17 years agofirewire: fw-ohci: log a note about unsupported features
Stefan Richter [Sun, 21 Oct 2007 08:43:11 +0000 (10:43 +0200)]
firewire: fw-ohci: log a note about unsupported features

because there seems to be more time needed to implement this.
Also, change related error return values to more appropriate ones.

Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>
17 years ago[SG] Update crypto/ to sg helpers
Jens Axboe [Mon, 22 Oct 2007 17:40:16 +0000 (19:40 +0200)]
[SG] Update crypto/ to sg helpers

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
17 years ago[SG] Update block layer to use sg helpers
Jens Axboe [Mon, 22 Oct 2007 17:39:33 +0000 (19:39 +0200)]
[SG] Update block layer to use sg helpers

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
17 years agoKVM: Use new smp_call_function_mask() in kvm_flush_remote_tlbs()
Laurent Vivier [Mon, 22 Oct 2007 14:33:07 +0000 (16:33 +0200)]
KVM: Use new smp_call_function_mask() in kvm_flush_remote_tlbs()

In kvm_flush_remote_tlbs(), replace a loop using smp_call_function_single()
by a single call to smp_call_function_mask() (which is new for x86_64).

Signed-off-by: Laurent Vivier <Laurent.Vivier@bull.net>
Signed-off-by: Avi Kivity <avi@qumranet.com>
17 years agoexportfs: update documentation
Christoph Hellwig [Sun, 21 Oct 2007 23:42:19 +0000 (16:42 -0700)]
exportfs: update documentation

Update documentation to the current state of affairs.  Remove duplicated
method descruptions in exportfs.h and point to Documentation/filesystems/
Exporting instead.  Add a little file header comment in expfs.c describing
what's going on and mentioning Neils and my copyright [1].

Signed-off-by: Christoph Hellwig <hch@lst.de>
Cc: Neil Brown <neilb@suse.de>
Cc: "J. Bruce Fields" <bfields@fieldses.org>
Cc: <linux-ext4@vger.kernel.org>
Cc: Dave Kleikamp <shaggy@austin.ibm.com>
Cc: Anton Altaparmakov <aia21@cantab.net>
Cc: David Chinner <dgc@sgi.com>
Cc: Timothy Shimmin <tes@sgi.com>
Cc: OGAWA Hirofumi <hirofumi@mail.parknet.co.jp>
Cc: Hugh Dickins <hugh@veritas.com>
Cc: Chris Mason <mason@suse.com>
Cc: Jeff Mahoney <jeffm@suse.com>
Cc: "Vladimir V. Saveliev" <vs@namesys.com>
Cc: Steven Whitehouse <swhiteho@redhat.com>
Cc: Mark Fasheh <mark.fasheh@oracle.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years agoexportfs: make struct export_operations const
Christoph Hellwig [Sun, 21 Oct 2007 23:42:17 +0000 (16:42 -0700)]
exportfs: make struct export_operations const

Now that nfsd has stopped writing to the find_exported_dentry member we an
mark the export_operations const

Signed-off-by: Christoph Hellwig <hch@lst.de>
Cc: Neil Brown <neilb@suse.de>
Cc: "J. Bruce Fields" <bfields@fieldses.org>
Cc: <linux-ext4@vger.kernel.org>
Cc: Dave Kleikamp <shaggy@austin.ibm.com>
Cc: Anton Altaparmakov <aia21@cantab.net>
Cc: David Chinner <dgc@sgi.com>
Cc: Timothy Shimmin <tes@sgi.com>
Cc: OGAWA Hirofumi <hirofumi@mail.parknet.co.jp>
Cc: Hugh Dickins <hugh@veritas.com>
Cc: Chris Mason <mason@suse.com>
Cc: Jeff Mahoney <jeffm@suse.com>
Cc: "Vladimir V. Saveliev" <vs@namesys.com>
Cc: Steven Whitehouse <swhiteho@redhat.com>
Cc: Mark Fasheh <mark.fasheh@oracle.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years agoexportfs: remove old methods
Christoph Hellwig [Sun, 21 Oct 2007 23:42:16 +0000 (16:42 -0700)]
exportfs: remove old methods

Now that all filesystems are converted remove support for the old methods.

Signed-off-by: Christoph Hellwig <hch@lst.de>
Cc: Neil Brown <neilb@suse.de>
Cc: "J. Bruce Fields" <bfields@fieldses.org>
Cc: <linux-ext4@vger.kernel.org>
Cc: Dave Kleikamp <shaggy@austin.ibm.com>
Cc: Anton Altaparmakov <aia21@cantab.net>
Cc: David Chinner <dgc@sgi.com>
Cc: Timothy Shimmin <tes@sgi.com>
Cc: OGAWA Hirofumi <hirofumi@mail.parknet.co.jp>
Cc: Hugh Dickins <hugh@veritas.com>
Cc: Chris Mason <mason@suse.com>
Cc: Jeff Mahoney <jeffm@suse.com>
Cc: "Vladimir V. Saveliev" <vs@namesys.com>
Cc: Steven Whitehouse <swhiteho@redhat.com>
Cc: Mark Fasheh <mark.fasheh@oracle.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years agoocfs2: new export ops
Christoph Hellwig [Sun, 21 Oct 2007 23:42:15 +0000 (16:42 -0700)]
ocfs2: new export ops

OCFS2 has it's own 64bit-firendly filehandle format so we can't use the
generic helpers here.  I'll add a struct for the types later.

Signed-off-by: Christoph Hellwig <hch@lst.de>
Cc: Neil Brown <neilb@suse.de>
Cc: "J. Bruce Fields" <bfields@fieldses.org>
Cc: Mark Fasheh <mark.fasheh@oracle.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years agogfs2: new export ops
Christoph Hellwig [Sun, 21 Oct 2007 23:42:14 +0000 (16:42 -0700)]
gfs2: new export ops

Convert gfs2 to the new ops.  Uses a similar structure to the generic helpers,
but gfs2 has it's own file handle formats.

Signed-off-by: Christoph Hellwig <hch@lst.de>
Cc: Neil Brown <neilb@suse.de>
Cc: "J. Bruce Fields" <bfields@fieldses.org>
Cc: Steven Whitehouse <swhiteho@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years agoreiserfs: new export ops
Christoph Hellwig [Sun, 21 Oct 2007 23:42:13 +0000 (16:42 -0700)]
reiserfs: new export ops

Another nice little cleanup by using the new methods.

Signed-off-by: Christoph Hellwig <hch@lst.de>
Cc: Neil Brown <neilb@suse.de>
Cc: "J. Bruce Fields" <bfields@fieldses.org>
Cc: Chris Mason <mason@suse.com>
Cc: Jeff Mahoney <jeffm@suse.com>
Cc: "Vladimir V. Saveliev" <vs@namesys.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years agoshmem: new export ops
Christoph Hellwig [Sun, 21 Oct 2007 23:42:13 +0000 (16:42 -0700)]
shmem: new export ops

I'm not sure what people were thinking when adding support to export tmpfs,
but here's the conversion anyway:

Signed-off-by: Christoph Hellwig <hch@lst.de>
Cc: Neil Brown <neilb@suse.de>
Cc: "J. Bruce Fields" <bfields@fieldses.org>
Cc: Hugh Dickins <hugh@veritas.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years agoisofs: new export ops
Christoph Hellwig [Sun, 21 Oct 2007 23:42:12 +0000 (16:42 -0700)]
isofs: new export ops

Nice little cleanup by consolidating things a little and using a structure for
the special file handle format.

Signed-off-by: Christoph Hellwig <hch@lst.de>
Cc: Neil Brown <neilb@suse.de>
Cc: "J. Bruce Fields" <bfields@fieldses.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years agofat: new export ops
Christoph Hellwig [Sun, 21 Oct 2007 23:42:11 +0000 (16:42 -0700)]
fat: new export ops

Very little changes here, fat had a mostly no op decode_fh before and does not
store any parent information.

Signed-off-by: Christoph Hellwig <hch@lst.de>
Cc: Neil Brown <neilb@suse.de>
Cc: "J. Bruce Fields" <bfields@fieldses.org>
Cc: OGAWA Hirofumi <hirofumi@mail.parknet.co.jp>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years agoxfs: new export ops
Christoph Hellwig [Sun, 21 Oct 2007 23:42:11 +0000 (16:42 -0700)]
xfs: new export ops

This one is a lot more complicated than the previous ones.  XFS already had a
very clever scheme for supporting 64bit inode numbers in filehandles, and I've
reworked this to be some kind of a prototype for the generic 64bit inode
filehandle support.

Signed-off-by: Christoph Hellwig <hch@lst.de>
Cc: Neil Brown <neilb@suse.de>
Cc: "J. Bruce Fields" <bfields@fieldses.org>
Cc: David Chinner <dgc@sgi.com>
Cc: Timothy Shimmin <tes@sgi.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years agontfs: new export ops
Christoph Hellwig [Sun, 21 Oct 2007 23:42:10 +0000 (16:42 -0700)]
ntfs: new export ops

Trivial switch over to the new generic helpers.

Signed-off-by: Christoph Hellwig <hch@lst.de>
Cc: Neil Brown <neilb@suse.de>
Cc: "J. Bruce Fields" <bfields@fieldses.org>
Cc: Anton Altaparmakov <aia21@cantab.net>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years agojfs: new export ops
Christoph Hellwig [Sun, 21 Oct 2007 23:42:09 +0000 (16:42 -0700)]
jfs: new export ops

Trivial switch over to the new generic helpers.

Signed-off-by: Christoph Hellwig <hch@lst.de>
Cc: Neil Brown <neilb@suse.de>
Cc: "J. Bruce Fields" <bfields@fieldses.org>
Cc: Dave Kleikamp <shaggy@austin.ibm.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years agoefs: new export ops
Christoph Hellwig [Sun, 21 Oct 2007 23:42:09 +0000 (16:42 -0700)]
efs: new export ops

Trivial switch over to the new generic helpers.

Signed-off-by: Christoph Hellwig <hch@lst.de>
Cc: Neil Brown <neilb@suse.de>
Cc: "J. Bruce Fields" <bfields@fieldses.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years agoext4: new export ops
Christoph Hellwig [Sun, 21 Oct 2007 23:42:08 +0000 (16:42 -0700)]
ext4: new export ops

Trivial switch over to the new generic helpers.

Signed-off-by: Christoph Hellwig <hch@lst.de>
Cc: Neil Brown <neilb@suse.de>
Cc: "J. Bruce Fields" <bfields@fieldses.org>
Cc: <linux-ext4@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years agoext3: new export ops
Christoph Hellwig [Sun, 21 Oct 2007 23:42:07 +0000 (16:42 -0700)]
ext3: new export ops

Trivial switch over to the new generic helpers.

Signed-off-by: Christoph Hellwig <hch@lst.de>
Cc: Neil Brown <neilb@suse.de>
Cc: "J. Bruce Fields" <bfields@fieldses.org>
Cc: <linux-ext4@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years agoext2: new export ops
Christoph Hellwig [Sun, 21 Oct 2007 23:42:07 +0000 (16:42 -0700)]
ext2: new export ops

Trivial switch over to the new generic helpers.

Signed-off-by: Christoph Hellwig <hch@lst.de>
Cc: Neil Brown <neilb@suse.de>
Cc: "J. Bruce Fields" <bfields@fieldses.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years agoexportfs: add new methods
Christoph Hellwig [Sun, 21 Oct 2007 23:42:05 +0000 (16:42 -0700)]
exportfs: add new methods

Add the guts for the new filesystem API to exportfs.

There's now a fh_to_dentry method that returns a dentry for the object looked
for given a filehandle fragment, and a fh_to_parent operation that returns the
dentry for the encoded parent directory in case the file handle contains it.

There are default implementations for these methods that only take a callback
for an nfs-enhanced iget variant and implement the rest of the semantics.

Signed-off-by: Christoph Hellwig <hch@lst.de>
Cc: Neil Brown <neilb@suse.de>
Cc: "J. Bruce Fields" <bfields@fieldses.org>
Cc: <linux-ext4@vger.kernel.org>
Cc: Dave Kleikamp <shaggy@austin.ibm.com>
Cc: Anton Altaparmakov <aia21@cantab.net>
Cc: David Chinner <dgc@sgi.com>
Cc: Timothy Shimmin <tes@sgi.com>
Cc: OGAWA Hirofumi <hirofumi@mail.parknet.co.jp>
Cc: Hugh Dickins <hugh@veritas.com>
Cc: Chris Mason <mason@suse.com>
Cc: Jeff Mahoney <jeffm@suse.com>
Cc: "Vladimir V. Saveliev" <vs@namesys.com>
Cc: Steven Whitehouse <swhiteho@redhat.com>
Cc: Mark Fasheh <mark.fasheh@oracle.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years agoexportfs: add fid type
Christoph Hellwig [Sun, 21 Oct 2007 23:42:03 +0000 (16:42 -0700)]
exportfs: add fid type

This patchset is a medium scale rewrite of the export operations interface.
The goal is to make the interface less complex, and easier to understand from
the filesystem side, aswell as preparing generic support for exporting of
64bit inode numbers.

This touches all nfs exporting filesystems, and I've done testing on all of
the filesystems I have here locally (xfs, ext2, ext3, reiserfs, jfs)

This patch:

Add a structured fid type so that we don't have to pass an array of u32 values
around everywhere.  It's a union of possible layouts.

As a start there's only the u32 array and the traditional 32bit inode format,
but there will be more in one of my next patchset when I start to document the
various filehandle formats we have in lowlevel filesystems better.

Also add an enum that gives the various filehandle types human- readable
names.

Note: Some people might think the struct containing an anonymous union is
ugly, but I didn't want to pass around a raw union type.

Signed-off-by: Christoph Hellwig <hch@lst.de>
Cc: Neil Brown <neilb@suse.de>
Cc: "J. Bruce Fields" <bfields@fieldses.org>
Cc: <linux-ext4@vger.kernel.org>
Cc: Dave Kleikamp <shaggy@austin.ibm.com>
Cc: Anton Altaparmakov <aia21@cantab.net>
Cc: David Chinner <dgc@sgi.com>
Cc: Timothy Shimmin <tes@sgi.com>
Cc: OGAWA Hirofumi <hirofumi@mail.parknet.co.jp>
Cc: Hugh Dickins <hugh@veritas.com>
Cc: Chris Mason <mason@suse.com>
Cc: Jeff Mahoney <jeffm@suse.com>
Cc: "Vladimir V. Saveliev" <vs@namesys.com>
Cc: Steven Whitehouse <swhiteho@redhat.com>
Cc: Mark Fasheh <mark.fasheh@oracle.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years agokexec: add BSS to resource tree
Bernhard Walle [Sun, 21 Oct 2007 23:42:01 +0000 (16:42 -0700)]
kexec: add BSS to resource tree

Add the BSS to the resource tree just as kernel text and kernel data are in
the resource tree.  The main reason behind this is to avoid crashkernel
reservation in that area.

While it's not strictly necessary to have the BSS in the resource tree (the
actual collision detection is done in the reserve_bootmem() function before),
the usage of the BSS resource should be presented to the user in /proc/iomem
just as Kernel data and Kernel code.

Note: The patch currently is only implemented for x86 and ia64 (because
efi_initialize_iomem_resources() has the same signature on i386 and ia64).

[akpm@linux-foundation.org: coding-style fixes]
Signed-off-by: Bernhard Walle <bwalle@suse.de>
Cc: "Eric W. Biederman" <ebiederm@xmission.com>
Cc: Vivek Goyal <vgoyal@in.ibm.com>
Cc: <linux-arch@vger.kernel.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years agointel-iommu sg chaining support
FUJITA Tomonori [Sun, 21 Oct 2007 23:42:00 +0000 (16:42 -0700)]
intel-iommu sg chaining support

x86_64 defines ARCH_HAS_SG_CHAIN. So if IOMMU implementations don't
support sg chaining, we will get data corruption.

Signed-off-by: FUJITA Tomonori <fujita.tomonori@lab.ntt.co.jp>
Acked-by: Anil S Keshavamurthy <anil.s.keshavamurthy@intel.com>
Cc: Jens Axboe <jens.axboe@oracle.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years agointel-iommu: fix for IOMMU early crash
Keshavamurthy, Anil S [Sun, 21 Oct 2007 23:41:59 +0000 (16:41 -0700)]
intel-iommu: fix for IOMMU early crash

pci_dev's->sysdata is highly overloaded and currently IOMMU is broken due
to IOMMU code depending on this field.

This patch introduces new field in pci_dev's dev.archdata struct to hold
IOMMU specific per device IOMMU private data.

Signed-off-by: Anil S Keshavamurthy <anil.s.keshavamurthy@intel.com>
Acked-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: Greg KH <greg@kroah.com>
Cc: Jeff Garzik <jeff@garzik.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years agointel-iommu: optimize sg map/unmap calls
Keshavamurthy, Anil S [Sun, 21 Oct 2007 23:41:58 +0000 (16:41 -0700)]
intel-iommu: optimize sg map/unmap calls

This patch adds PageSelectiveInvalidation support replacing existing
DomainSelectiveInvalidation for intel_{map/unmap}_sg() calls and also
enables to mapping one big contiguous DMA virtual address which is mapped
to discontiguous physical address for SG map/unmap calls.

"Doamin selective invalidations" wipes out the IOMMU address translation
cache based on domain ID where as "Page selective invalidations" wipes out
the IOMMU address translation cache for that address mask range which is
more cache friendly when compared to Domain selective invalidations.

Here is how it is done.
1) changes to iova.c
alloc_iova() now takes a bool size_aligned argument, which
when when set, returns the io virtual address that is
naturally aligned to 2 ^ x, where x is the order
of the size requested.

Returning this io vitual address which is naturally
aligned helps iommu to do the "page selective
invalidations" which is IOMMU cache friendly
over "domain selective invalidations".

2) Changes to driver/pci/intel-iommu.c
Clean up intel_{map/unmap}_{single/sg} () calls so that
s/g map/unamp calls is no more dependent on
intel_{map/unmap}_single()

intel_map_sg() now computes the total DMA virtual address
required and allocates the size aligned total DMA virtual address
and maps the discontiguous physical address to the allocated
contiguous DMA virtual address.

In the intel_unmap_sg() case since the DMA virtual address
is contiguous and size_aligned, PageSelectiveInvalidation
is used replacing earlier DomainSelectiveInvalidations.

Signed-off-by: Anil S Keshavamurthy <anil.s.keshavamurthy@intel.com>
Cc: Greg KH <greg@kroah.com>
Cc: Ashok Raj <ashok.raj@intel.com>
Cc: Suresh B <suresh.b.siddha@intel.com>
Cc: Andi Kleen <ak@suse.de>
Cc: Arjan van de Ven <arjan@infradead.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years agoIntel IOMMU: Iommu floppy workaround
Keshavamurthy, Anil S [Sun, 21 Oct 2007 23:41:57 +0000 (16:41 -0700)]
Intel IOMMU: Iommu floppy workaround

This config option (DMAR_FLPY_WA) sets up 1:1 mapping for the floppy device so
that the floppy device which does not use DMA api's will continue to work.

Once the floppy driver starts using DMA api's this config option can be turn
off or this patch can be yanked out of kernel at that time.

[akpm@linux-foundation.org: cleanups, rename things, build fix]
[jengelh@computergmbh.de: Kconfig fixes]
Signed-off-by: Anil S Keshavamurthy <anil.s.keshavamurthy@intel.com>
Cc: Andi Kleen <ak@suse.de>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Muli Ben-Yehuda <muli@il.ibm.com>
Cc: "Siddha, Suresh B" <suresh.b.siddha@intel.com>
Cc: Arjan van de Ven <arjan@infradead.org>
Cc: Ashok Raj <ashok.raj@intel.com>
Cc: "David S. Miller" <davem@davemloft.net>
Cc: Christoph Lameter <clameter@sgi.com>
Cc: Greg KH <greg@kroah.com>
Signed-off-by: Jan Engelhardt <jengelh@gmx.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years agoIntel IOMMU: Iommu Gfx workaround
Keshavamurthy, Anil S [Sun, 21 Oct 2007 23:41:55 +0000 (16:41 -0700)]
Intel IOMMU: Iommu Gfx workaround

When we fix all the opensource gfx drivers to use the DMA api's, at that time
we can yank this config options out.

[jengelh@computergmbh.de: Kconfig fixes]
Signed-off-by: Anil S Keshavamurthy <anil.s.keshavamurthy@intel.com>
Cc: Andi Kleen <ak@suse.de>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Muli Ben-Yehuda <muli@il.ibm.com>
Cc: "Siddha, Suresh B" <suresh.b.siddha@intel.com>
Cc: Arjan van de Ven <arjan@infradead.org>
Cc: Ashok Raj <ashok.raj@intel.com>
Cc: "David S. Miller" <davem@davemloft.net>
Cc: Christoph Lameter <clameter@sgi.com>
Cc: Greg KH <greg@kroah.com>
Signed-off-by: Jan Engelhardt <jengelh@gmx.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years agoIntel IOMMU: DMAR fault handling support
Keshavamurthy, Anil S [Sun, 21 Oct 2007 23:41:54 +0000 (16:41 -0700)]
Intel IOMMU: DMAR fault handling support

MSI interrupt handler registrations and fault handling support for Intel-IOMMU
hadrware.

This patch enables the MSI interrupts for the DMA remapping units and in the
interrupt handler read the fault cause and outputs the same on to the console.

Signed-off-by: Anil S Keshavamurthy <anil.s.keshavamurthy@intel.com>
Cc: Andi Kleen <ak@suse.de>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Muli Ben-Yehuda <muli@il.ibm.com>
Cc: "Siddha, Suresh B" <suresh.b.siddha@intel.com>
Cc: Arjan van de Ven <arjan@infradead.org>
Cc: Ashok Raj <ashok.raj@intel.com>
Cc: "David S. Miller" <davem@davemloft.net>
Cc: Christoph Lameter <clameter@sgi.com>
Cc: Greg KH <greg@kroah.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years agoIntel IOMMU: Intel iommu cmdline option - forcedac
Keshavamurthy, Anil S [Sun, 21 Oct 2007 23:41:53 +0000 (16:41 -0700)]
Intel IOMMU: Intel iommu cmdline option - forcedac

Introduce intel_iommu=forcedac commandline option.  This option is helpful to
verify the pci device capability of handling physical dma'able address greater
than 4G.

Signed-off-by: Anil S Keshavamurthy <anil.s.keshavamurthy@intel.com>
Cc: Andi Kleen <ak@suse.de>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Muli Ben-Yehuda <muli@il.ibm.com>
Cc: "Siddha, Suresh B" <suresh.b.siddha@intel.com>
Cc: Arjan van de Ven <arjan@infradead.org>
Cc: Ashok Raj <ashok.raj@intel.com>
Cc: "David S. Miller" <davem@davemloft.net>
Cc: Christoph Lameter <clameter@sgi.com>
Cc: Greg KH <greg@kroah.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years agoIntel IOMMU: Avoid memory allocation failures in dma map api calls
Keshavamurthy, Anil S [Sun, 21 Oct 2007 23:41:52 +0000 (16:41 -0700)]
Intel IOMMU: Avoid memory allocation failures in dma map api calls

Intel IOMMU driver needs memory during DMA map calls to setup its internal
page tables and for other data structures.  As we all know that these DMA map
calls are mostly called in the interrupt context or with the spinlock held by
the upper level drivers(network/storage drivers), so in order to avoid any
memory allocation failure due to low memory issues, this patch makes memory
allocation by temporarily setting PF_MEMALLOC flags for the current task
before making memory allocation calls.

We evaluated mempools as a backup when kmem_cache_alloc() fails
and found that mempools are really not useful here because
 1) We don't know for sure how much to reserve in advance
 2) And mempools are not useful for GFP_ATOMIC case (as we call
    memory alloc functions with GFP_ATOMIC)

(akpm: point 2 is wrong...)

With PF_MEMALLOC flag set in the current->flags, the VM subsystem avoids any
watermark checks before allocating memory thus guarantee'ing the memory till
the last free page.  Further, looking at the code in mm/page_alloc.c in
__alloc_pages() function, looks like this flag is useful only in the
non-interrupt context.

If we are in the interrupt context and memory allocation in IOMMU driver fails
for some reason, then the DMA map api's will return failure and it is up to
the higher level drivers to retry.  Suppose, if upper level driver programs
the controller with the buggy DMA virtual address, the IOMMU will block that
DMA transaction when that happens thus preventing any corruption to main
memory.

So far in our test scenario, we were unable to create any memory allocation
failure inside dma map api calls.

Signed-off-by: Anil S Keshavamurthy <anil.s.keshavamurthy@intel.com>
Cc: Andi Kleen <ak@suse.de>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Muli Ben-Yehuda <muli@il.ibm.com>
Cc: "Siddha, Suresh B" <suresh.b.siddha@intel.com>
Cc: Arjan van de Ven <arjan@infradead.org>
Cc: Ashok Raj <ashok.raj@intel.com>
Cc: "David S. Miller" <davem@davemloft.net>
Cc: Christoph Lameter <clameter@sgi.com>
Cc: Greg KH <greg@kroah.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years agoIntel IOMMU: Intel IOMMU driver
Keshavamurthy, Anil S [Sun, 21 Oct 2007 23:41:49 +0000 (16:41 -0700)]
Intel IOMMU: Intel IOMMU driver

Actual intel IOMMU driver.  Hardware spec can be found at:
http://www.intel.com/technology/virtualization

This driver sets X86_64 'dma_ops', so hook into standard DMA APIs.  In this
way, PCI driver will get virtual DMA address.  This change is transparent to
PCI drivers.

[akpm@linux-foundation.org: remove unneeded cast]
[akpm@linux-foundation.org: build fix]
[bunk@stusta.de: fix duplicate CONFIG_DMAR Makefile line]
Signed-off-by: Anil S Keshavamurthy <anil.s.keshavamurthy@intel.com>
Cc: Andi Kleen <ak@suse.de>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Muli Ben-Yehuda <muli@il.ibm.com>
Cc: "Siddha, Suresh B" <suresh.b.siddha@intel.com>
Cc: Arjan van de Ven <arjan@infradead.org>
Cc: Ashok Raj <ashok.raj@intel.com>
Cc: "David S. Miller" <davem@davemloft.net>
Cc: Christoph Lameter <clameter@sgi.com>
Cc: Greg KH <greg@kroah.com>
Signed-off-by: Adrian Bunk <bunk@stusta.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years agoIntel IOMMU: IOVA allocation and management routines
Keshavamurthy, Anil S [Sun, 21 Oct 2007 23:41:48 +0000 (16:41 -0700)]
Intel IOMMU: IOVA allocation and management routines

This code implements a generic IOVA allocation and management.  As per Dave's
suggestion we are now allocating IO virtual address from Higher DMA limit
address rather than lower end address and this eliminated the need to preserve
the IO virtual address for multiple devices sharing the same domain virtual
address.

Also this code uses red black trees to store the allocated and reserved iova
nodes.  This showed a good performance improvements over previous linear
linked list.

[akpm@linux-foundation.org: remove inlines]
[akpm@linux-foundation.org: coding style fixes]
Signed-off-by: Anil S Keshavamurthy <anil.s.keshavamurthy@intel.com>
Cc: Andi Kleen <ak@suse.de>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Muli Ben-Yehuda <muli@il.ibm.com>
Cc: "Siddha, Suresh B" <suresh.b.siddha@intel.com>
Cc: Arjan van de Ven <arjan@infradead.org>
Cc: Ashok Raj <ashok.raj@intel.com>
Cc: "David S. Miller" <davem@davemloft.net>
Cc: Christoph Lameter <clameter@sgi.com>
Cc: Greg KH <greg@kroah.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years agoIntel IOMMU: clflush_cache_range now takes size param
Keshavamurthy, Anil S [Sun, 21 Oct 2007 23:41:47 +0000 (16:41 -0700)]
Intel IOMMU: clflush_cache_range now takes size param

Introduce the size param for clflush_cache_range().

Signed-off-by: Anil S Keshavamurthy <anil.s.keshavamurthy@intel.com>
Cc: Andi Kleen <ak@suse.de>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Muli Ben-Yehuda <muli@il.ibm.com>
Cc: "Siddha, Suresh B" <suresh.b.siddha@intel.com>
Cc: Arjan van de Ven <arjan@infradead.org>
Cc: Ashok Raj <ashok.raj@intel.com>
Cc: "David S. Miller" <davem@davemloft.net>
Cc: Christoph Lameter <clameter@sgi.com>
Cc: Greg KH <greg@kroah.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years agoIntel IOMMU: PCI generic helper function
Keshavamurthy, Anil S [Sun, 21 Oct 2007 23:41:46 +0000 (16:41 -0700)]
Intel IOMMU: PCI generic helper function

When devices are under a p2p bridge, upstream transactions get replaced by the
device id of the bridge as it owns the PCIE transaction.  Hence its necessary
to setup translations on behalf of the bridge as well.  Due to this limitation
all devices under a p2p share the same domain in a DMAR.

We just cache the type of device, if its a native PCIe device
or not for later use.

[akpm@linux-foundation.org: BUG_ON -> WARN_ON+recover]
Signed-off-by: Anil S Keshavamurthy <anil.s.keshavamurthy@intel.com>
Cc: Andi Kleen <ak@suse.de>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Muli Ben-Yehuda <muli@il.ibm.com>
Cc: "Siddha, Suresh B" <suresh.b.siddha@intel.com>
Cc: Arjan van de Ven <arjan@infradead.org>
Cc: Ashok Raj <ashok.raj@intel.com>
Cc: "David S. Miller" <davem@davemloft.net>
Cc: Christoph Lameter <clameter@sgi.com>
Cc: Greg KH <greg@kroah.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years agoIntel IOMMU: DMAR detection and parsing logic
Keshavamurthy, Anil S [Sun, 21 Oct 2007 23:41:41 +0000 (16:41 -0700)]
Intel IOMMU: DMAR detection and parsing logic

This patch supports the upcomming Intel IOMMU hardware a.k.a.  Intel(R)
Virtualization Technology for Directed I/O Architecture and the hardware spec
for the same can be found here
http://www.intel.com/technology/virtualization/index.htm

FAQ! (questions from akpm, answers from ak)

> So...  what's all this code for?
>
> I assume that the intent here is to speed things up under Xen, etc?

Yes in some cases, but not this code.  That would be the Xen version of this
code that could potentially assign whole devices to guests.  I expect this to
be only useful in some special cases though because most hardware is not
virtualizable and you typically want an own instance for each guest.

Ok at some point KVM might implement this too; i likely would use this code
for this.

> Do we
> have any benchmark results to help us to decide whether a merge would be
> justified?

The main advantage for doing it in the normal kernel is not performance, but
more safety.  Broken devices won't be able to corrupt memory by doing random
DMA.

Unfortunately that doesn't work for graphics yet, for that need user space
interfaces for the X server are needed.

There are some potential performance benefits too:

- When you have a device that cannot address the complete address range an
  IOMMU can remap its memory instead of bounce buffering.  Remapping is likely
  cheaper than copying.

- The IOMMU can merge sg lists into a single virtual block.  This could
  potentially speed up SG IO when the device is slow walking SG lists.  [I
  long ago benchmarked 5% on some block benchmark with an old MPT Fusion; but
  it probably depends a lot on the HBA]

And you get better driver debugging because unexpected memory accesses from
the devices will cause a trappable event.

>
> Does it slow anything down?

It adds more overhead to each IO so yes.

This patch:

Add support for early detection and parsing of DMAR's (DMA Remapping) reported
to OS via ACPI tables.

DMA remapping(DMAR) devices support enables independent address translations
for Direct Memory Access(DMA) from Devices.  These DMA remapping devices are
reported via ACPI tables and includes pci device scope covered by these DMA
remapping device.

For detailed info on the specification of "Intel(R) Virtualization Technology
for Directed I/O Architecture" please see
http://www.intel.com/technology/virtualization/index.htm

Signed-off-by: Anil S Keshavamurthy <anil.s.keshavamurthy@intel.com>
Cc: Andi Kleen <ak@suse.de>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Muli Ben-Yehuda <muli@il.ibm.com>
Cc: "Siddha, Suresh B" <suresh.b.siddha@intel.com>
Cc: Arjan van de Ven <arjan@infradead.org>
Cc: Ashok Raj <ashok.raj@intel.com>
Cc: "David S. Miller" <davem@davemloft.net>
Cc: Christoph Lameter <clameter@sgi.com>
Cc: Greg KH <greg@kroah.com>
Cc: Len Brown <lenb@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>