On machines supporting the OPAL firmware version 1, the system
is initially booted under pHyp. We then use a special hypercall
to verify if OPAL is available and if it is, we then trigger
a "takeover" which disables pHyp and loads the OPAL runtime
firmware, giving control to the kernel in hypervisor mode.
This patch add the necessary code to detect that the OPAL takeover
capability is present when running under PowerVM (aka pHyp) and
perform said takeover to get hypervisor control of the processor.
To perform the takeover, we must first use RTAS (within Open
Firmware runtime environment) to start all processors & threads,
in order to give control to OPAL on all of them. We then call
the takeover hypercall on everybody, OPAL will re-enter the kernel
main entry point passing it a flat device-tree.
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
This adds a skeletton for the new Power "Non Virtualized"
platform which will be used by machines supporting running
without an hypervisor, for example in order to run KVM.
These machines will be using a new firmware called OPAL
for which the support will be provided by later patches.
The PowerNV platform is intended to be also usable under
the BML environment used internally for early CPU bringup
which is why the code also supports using RTAS instead of
OPAL in various places.
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
With OPAL, r8 and r9 will be used to pass the OPAL base and entry
for debugging purposes (those informations are also in the
device-tree). We don't want to clobber those registers that
early.
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
This new function is used to properly setup the PCI Express Max Payload Size
(and in some circumstances Max Read Request Size).
Some systems will not operate properly if these aren't set correctly and
the firmware doesn't always do it.
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
This adds more generic support for doing CPU hotplug with a simple
idle loop and no actual reset of the processors. The generic
smp_generic_kick_cpu() does the hotplug bringup trick if the PACA
shows that the CPU has already been started at boot and we provide
an accessor for the CPU state.
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
It was preventing the global early debug selection whenever KVM was enabled
instead of only preventing the 440 specific one.
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
The icswx code introduced an A-B B-A deadlock:
CPU0 CPU1
---- ----
lock(&anon_vma->mutex);
lock(&mm->mmap_sem);
lock(&anon_vma->mutex);
lock(&mm->mmap_sem);
Instead of using the mmap_sem to keep mm_users constant, take the
page table spinlock.
Signed-off-by: Anton Blanchard <anton@samba.org>
Cc: <stable@kernel.org>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
If we echo an address the hypervisor doesn't like to
/sys/devices/system/memory/probe we oops the box:
# echo 0x10000000000 > /sys/devices/system/memory/probe
kernel BUG at arch/powerpc/mm/hash_utils_64.c:541!
The backtrace is:
create_section_mapping
arch_add_memory
add_memory
memory_probe_store
sysdev_class_store
sysfs_write_file
vfs_write
SyS_write
In create_section_mapping we BUG if htab_bolt_mapping returned
an error. A better approach is to return an error which will
propagate back to userspace.
Rerunning the test with this patch applied:
# echo 0x10000000000 > /sys/devices/system/memory/probe
-bash: echo: write error: Invalid argument
Signed-off-by: Anton Blanchard <anton@samba.org>
Cc: stable@kernel.org
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
While converting code to use for_each_node_by_type I noticed a
number of coding style issues.
Signed-off-by: Anton Blanchard <anton@samba.org>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Use for_each_node_by_type instead of open coding it.
Signed-off-by: Anton Blanchard <anton@samba.org>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
During memory hotplug testing, I got the following warning:
ERROR: Bad of_node_put() on /memory@0
of_node_release
kref_put
of_node_put
of_find_node_by_type
hot_add_node_scn_to_nid
hot_add_scn_to_nid
memory_add_physaddr_to_nid
...
of_find_node_by_type() loop does the of_node_put for us so we only
need the handle the case where we terminate the loop early.
As suggested by Stephen Rothwell we can do the of_node_put
unconditionally outside of the loop since of_node_put handles a
NULL argument fine.
Signed-off-by: Anton Blanchard <anton@samba.org>
Cc: stable@kernel.org
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
We have two identical definitions of RECLAIM_DISTANCE, looks like
the patch got applied twice. Remove one.
Signed-off-by: Anton Blanchard <anton@samba.org>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
On big POWER7 boxes we see large amounts of CPU time in system
processes like workqueue and watchdog kernel threads.
We currently rebalance the entire machine each time a task goes
idle and this is very expensive on large machines. Disable newidle
balancing at the node level and rely on the scheduler tick to
rebalance across nodes.
Signed-off-by: Anton Blanchard <anton@samba.org>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
The largest POWER7 boxes have 32 nodes. SD_NODES_PER_DOMAIN groups
nodes into chunks of 16 and adds a global balancing domain
(SD_ALLNODES) above it.
If we bump SD_NODES_PER_DOMAIN to 32, then we avoid this extra
level of balancing on our largest boxes.
Signed-off-by: Anton Blanchard <anton@samba.org>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
When chasing a performance issue on ppc64, I noticed tasks
communicating via a pipe would often end up on different nodes.
It turns out SD_WAKE_AFFINE is not set in our node defition. Commit
9fcd18c9e6 (sched: re-tune balancing) enabled SD_WAKE_AFFINE
in the node definition for x86 and we need a similar change for
ppc64.
I used lmbench lat_ctx and perf bench pipe to verify this fix. Each
benchmark was run 10 times and the average taken.
lmbench lat_ctx:
before: 66565 ops/sec
after: 204700 ops/sec
3.1x faster
perf bench pipe:
before: 5.6570 usecs
after: 1.3470 usecs
4.2x faster
Signed-off-by: Anton Blanchard <anton@samba.org>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Add a new udbg driver for the PS3 gelic Ehthernet device.
This driver shares only a few stucture and constant definitions with the
gelic Ethernet device driver, so is implemented as a stand-alone driver
with no dependencies on the gelic Ethernet device driver.
Signed-off-by: Hector Martin <hector@marcansoft.com>
Signed-off-by: Andre Heider <a.heider@gmail.com>
Signed-off-by: Geoff Levand <geoff@infradead.org>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Since commit 188917e183, /proc/ppc64 is a
symlink to /proc/powerpc/. That means that creating /proc/ppc64/eeh will
end up with a unaccessible file, that is not listed under /proc/powerpc/
and, then, not listed under /proc/ppc64/.
Creating /proc/powerpc/eeh fixes that problem and maintain the
compatibility intended with the ppc64 symlink.
Signed-off-by: Thadeu Lima de Souza Cascardo <cascardo@linux.vnet.ibm.com>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: <stable@kernel.org> [3.x]
This should fix the following warning:
LD arch/powerpc/sysdev/xics/built-in.o
WARNING: arch/powerpc/sysdev/xics/built-in.o(.text+0x1310): Section mismatch in
reference from the function .icp_native_init() to the function
.init.text:.icp_native_init_one_node()
The function .icp_native_init() references
the function __init .icp_native_init_one_node().
This is often because .icp_native_init lacks a __init
annotation or the annotation of .icp_native_init_one_node is wrong.
icp_native_init() is only referenced in `arch/powerpc/sysdev/xics/xics-common.c'
by xics_init() which is itself marked with __init.
= not built-tested =
Reported-by: Timur Tabi <timur@freescale.com>
Signed-off-by: Arnaud Lacombe <lacombar@gmail.com>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
During hotplug CPU add we get the following error:
Unexpected Error (0) returned from configure-connector
ibm,configure-connector returns 0 for configuration complete, so
catch this and avoid the error.
Signed-off-by: Anton Blanchard <anton@samba.org>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: <stable@kernel.org>
In SMP mode, the kernel would produce call trace when resumed
from hibernation. The reason is when the function destroy_context
is called to drop the resuming mm context, the mm->context.active
is 1 which is wrong and should be zero.
We pass the current->active_mm as previous mm context to function
switch_mmu_context to decrease the context.active by 1.
In UP mode, there is no effect.
Signed-off-by: Tang Yuantian <b29983@freescale.com>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
The various port_init_hw methods of ppc4xx_pciex_hwops should have been
marked __init and when I added ppc4xx_pciex_port_reset_sdr(), which is
__init. This added many section mismatch warnings like:
WARNING: arch/powerpc/sysdev/built-in.o(.text+0x5c68): Section mismatch in reference from the function ppc440spe_pciex_init_port_hw() to the function .init.text:ppc4xx_pciex_port_reset_sdr()
The function ppc440spe_pciex_init_port_hw() references
the function __init ppc4xx_pciex_port_reset_sdr().
This is often because ppc440spe_pciex_init_port_hw lacks a __init
annotation or the annotation of ppc4xx_pciex_port_reset_sdr is wrong.
Trivial patch to silence those warnings.
Reported-By: Stephen Rothwell <sfr@canb.auug.org.au>
Signed-off-by: Tony Breeds <tony@bakeyournoodle.com>
Yours Tony
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Based on a patch by Michael Ellerman <michael@ellerman.id.au>
Patch was simply forward ported upstream.
Jimi Xenidis <jimix@pobox.com>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Based on a patch by Benjamin Herrenschmidt <benh@kernel.crashing.org>
Modernized and slightly modified to not record erros into the nvram
log since we do not have that device driver just yet.
Jimi Xenidis <jimix@pobox.com>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Some config selections were applied to the platform (reference board)
when they actuall apply to the chip.
Signed-off-by: Jimi Xenidis <jimix@pobox.com>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
At this point, window has not been stored anywhere, so it has to be freed
before leaving the function.
A simplified version of the semantic match that finds this problem is as
follows: (http://coccinelle.lip6.fr/)
// <smpl>
@exists@
local idexpression x;
statement S,S1;
expression E;
identifier fl;
expression *ptr != NULL;
@@
x = \(kmalloc\|kzalloc\|kcalloc\)(...);
...
if (x == NULL) S
<... when != x
when != if (...) { <+...kfree(x)...+> }
when any
when != true x == NULL
x->fl
...>
(
if (x == NULL) S1
|
if (...) { ... when != x
when forall
(
return \(0\|<+...x...+>\|ptr\);
|
* return ...;
)
}
)
// </smpl>
Signed-off-by: Julia Lawall <julia@diku.dk>
Acked-by: Nishanth Aravamudan <nacc@us.ibm.com>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
u64 is used rather than phys_addr_t to keep things simple, as
this is called from assembly code.
Update callers to pass a 64-bit address in r3/r4. Other unused
register assignments that were once parameters to machine_init
are dropped.
For FSL BookE, look up the physical address of the device tree from the
effective address passed in r3 by the loader. This is required for
situations where memory does not start at zero (due to AMP or IOMMU-less
virtualization), and thus the IMA doesn't start at zero, and thus the
device tree effective address does not equal the physical address.
Signed-off-by: Scott Wood <scottwood@freescale.com>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Capture more than twice as much text from the printk buffer, and
compress it to fit it in the lnx,oops-log NVRAM partition. You
can view the compressed text using the new (as of July 20) --unzip
option of the nvram command in the powerpc-utils package.
[BenH: Added select of ZLIB_DEFLATE]
Signed-off-by: Jim Keniston <jkenisto@us.ibm.com>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Currently, the build can (very rarely) fail to build because libfdt.h has
not been created or is in the process of being copied.
Signed-off-by: Matthew McClintock <msm@freescale.com>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
There is one place in the MPIC driver that assumes that the cores are numbered
from 0 to n-1. However, this is not true if the CPUs are not numbered
sequentially. This can happen on a eight-core SOC where cores two and three
are removed in the device tree. So instead of blindly looping, we iterate
over the discovered CPUs and use the SMP ID as the index.
This means that we no longer ask the MPIC how many CPUs there are, so
we also delete mpic->num_cpus.
We also catch if the number of CPUs in the SOC exceeds the number that the
MPIC supports. This should never happen, of course, but it's good to be
sure.
Signed-off-by: Timur Tabi <timur@freescale.com>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Enable hugepages on Freescale BookE processors. This allows the kernel to
use huge TLB entries to map pages, which can greatly reduce the number of
TLB misses and the amount of TLB thrashing experienced by applications with
large memory footprints. Care should be taken when using this on FSL
processors, as the number of large TLB entries supported by the core is low
(16-64) on current processors.
The supported set of hugepage sizes include 4m, 16m, 64m, 256m, and 1g.
Page sizes larger than the max zone size are called "gigantic" pages and
must be allocated on the command line (and cannot be deallocated).
This is currently only fully implemented for Freescale 32-bit BookE
processors, but there is some infrastructure in the code for
64-bit BooKE.
Signed-off-by: Becky Bruce <beckyb@kernel.crashing.org>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
This iotype is only used by the legacy_serial code in powerpc, so the
code should live there, rather than be compiled in for every 8250
driver.
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: linuxppc-dev@lists.ozlabs.org
Cc: Greg Kroah-Hartman <gregkh@suse.de>
Cc: linux-serial@vger.kernel.org
Acked-by: David Daney <david.daney@cavium.com>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
The new get_required_mask hook name is longer than many of but not all
of the prior ops. Tidy the struct initializers to align the equal signs
using the local whitespace.
Signed-off-by: Milton Miller <miltonm@bga.com>
Signed-off-by: Nishanth Aravamudan <nacc@us.ibm.com>
Cc: linuxppc-dev@lists.ozlabs.org
Cc: linux-kernel@vger.kernel.org
Cc: benh@kernel.crashing.org
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Now that the generic code has dma_map_ops set, instead of having a
messy ifdef & if block in the base dma_get_required_mask hook push
the computation into the dma ops.
If the ops fails to set the get_required_mask hook default to the
width of dma_addr_t.
This also corrects ibmbus ibmebus_dma_supported to require a 64
bit mask. I doubt anything is checking or setting the dma mask on
that bus.
Signed-off-by: Milton Miller <miltonm@bga.com>
Signed-off-by: Nishanth Aravamudan <nacc@us.ibm.com>
Cc: linuxppc-dev@lists.ozlabs.org
Cc: linux-kernel@vger.kernel.org
Cc: benh@kernel.crashing.org
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
* 'stable/bug.fixes' of git://oss.oracle.com/git/kwilk/xen:
xen/i386: follow-up to "replace order-based range checking of M2P table by linear one"
xen/irq: Alter the locking to use a mutex instead of a spinlock.
xen/e820: if there is no dom0_mem=, don't tweak extra_pages.
xen: disable PV spinlocks on HVM
On x86-64, they were just wasteful: with the explicitly added (now
unnecessary) padding, the size of the alternatives structure was 16
bytes, and an alignment of 8 bytes didn't hurt much.
However, it was still silly, since the natural size and alignment for
the structure is actually just 12 bytes, 4-byte aligned since commit
59e97e4d6f ("x86: Make alternative instruction pointers relative").
So removing the padding, and removing the extra alignment is just a good
idea.
On x86-32, the alignment of 4 bytes was correct, but was incorrectly
hardcoded as 8 bytes in <asm/alternative-asm.h>. That header file had
used to be an x86-64 only header file, but various unification efforts
have made it be used for x86-32 too (ie the unification of rwlock and
rwsem).
That in turn caused x86-32 boot failures, because the extra alignment
would result in random zero-filled words in the altinstructions section,
causing oopses early at boot when doing alternative instruction
replacement.
So just remove all the alignment noise entirely. It's wrong, and it's
unnecessary. The section itself is already properly aligned by the
linker scripts, and all additions to the section had better be of the
proper 12-byte format, keeping it aligned. So if the align directive
were to ever make a difference, that would be an indication of a serious
bug to begin with.
Reported-by: Werner Landgraf <w.landgraf@ru.r>
Acked-by: Andrew Lutomirski <luto@mit.edu>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
The numbers obtained from the hypervisor really can't ever lead to an
overflow here, only the original calculation going through the order
of the range could have. This avoids the (as Jeremy points outs)
somewhat ugly NULL-based calculation here.
Signed-off-by: Jan Beulich <jbeulich@novell.com>
Signed-off-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
Since GPIOLIB is optional on alpha, GENERIC_GPIO must not be selected by
default. If GPIOLIB is enabled, it will select GENERIC_GPIO.
See <http://bugs.debian.org/638696> for an example of what 'def_bool y'
breaks.
Reported-by: Michael Cree <mcree@orcon.net.nz>
Signed-off-by: Ben Hutchings <ben@decadent.org.uk>
Cc: Richard Henderson <rth@twiddle.net>
Cc: Ivan Kokshaysky <ink@jurassic.park.msu.ru>
Cc: Matt Turner <mattst88@gmail.com>
Cc: Michael Cree <mcree@orcon.net.nz>
Cc: Mark Brown <broonie@opensource.wolfsonmicro.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
richard@nod.at:
Fixes:
/usr/lib/gcc/x86_64-linux-gnu/4.4.5/../../../../lib/libc.a(strrchr.o): In function `rindex':
(.text+0x0): multiple definition of `strrchr'
If both STATIC_LINK and UML_NET_VDE are set to "y" libc's strrchr may
clash with the kernel implementation.
This workaround comes originally from Jeff Dike:
http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=494995#35
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Richard Weinberger <richard@nod.at>
Cc: Jeff Dike <jdike@addtoit.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
1) take subarch-specific stuff to subarch_ptrace()
2) PTRACE_{PEEK,POKE}{TEXT,DATA} is handled by ptrace_request() just fine...
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Richard Weinberger <richard@nod.at>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
It's 32bit-only, not 64bit-only... And while we are at it, it's
set_fpxregs(), not set_fpregs()...
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Richard Weinberger <richard@nod.at>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
while not doing free_irq() from irq handler is commendable, kfree() on the
data passed to said handler before free_irq() is Not Good(tm). Freeing
the stack it's being run on is also not nice... Solution: delay actually
freeing stuff.
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Richard Weinberger <richard@nod.at>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
... so set winch->fd to -1 before doing free_irq(), to avoid having
winch_interrupt() come from/during the latter and attempt to do
reactivate_fd() on something that's already gone.
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Richard Weinberger <richard@nod.at>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
tty->count is decremented only after ->close() had been called and
several tasks can hit it in parallel. As the result, using tty->count
to check if you are the last one is broken. We end up leaving line->tty
not reset to NULL and the next IRQ on that sucker will blow up trying to
dereference pointers from kfree'd struct tty.
Fix is obvious: we need to use a counter of our own.
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Richard Weinberger <richard@nod.at>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Some time ago Jeff prepared 42daba3165 ("uml: stop saving process FP
state") for UML to stop saving the process FP state between task
switches. The assumption was that since with SKAS0 every guest process
runs inside a host process context the host OS will take care of keeping
the proper FP state.
Unfortunately this is not true for multi-threaded applications, where
all guest threads share a single host process context yet all may use
the FPU on their own. Although I haven't verified it I suspect things
to be even worse in SKAS3 mode where all guest processes run inside a
single host process.
The patch reintroduces the saving and restoring of the FP context
between task switches.
[richard@nod.at: Ingo posted this patch in 2009, sadly it was never applied
and got lost. Now in 2011 the problem was reported by Gunnar.]
Signed-off-by: Ingo van Lil <inguin@gmx.de>
Signed-off-by: Richard Weinberger <richard@nod.at>
Reported-by: <gunnarlindroth@hotmail.com>
Tested-by: <gunnarlindroth@hotmail.com>
Cc: Stanislav Meduna <stano@meduna.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
I could use out_close1, but that seems to be the code path to close the fd
returned by os_create_unix_socket, and using it to close the fd returned
by mkstemp might lead to some confusion, so I don't do it.
Signed-off-by: Jonathan Neuschäfer <j.neuschaefer@gmx.net>
Signed-off-by: Richard Weinberger <richard@nod.at>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Commit b789ef518b ("slub: Add cmpxchg_double_slab()") tests for
cmpxchg_double support in the SLUB code and it breaks UML builds with
SLUB. Since UML does not support checking for CPU features, disable
CMPXCHG_DOUBLE just like CMPXCHG_LOCAL is disabled for UML.
Signed-off-by: Thadeu Lima de Souza Cascardo <cascardo@holoscopio.com>
Reviewed-by: Christoph Lameter <cl@linux.com>
Cc: Jeff Dike <jdike@addtoit.com>
Cc: Richard Weinberger <richard@nod.at>
Signed-off-by: Richard Weinberger <richard@nod.at>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
The patch "xen: use maximum reservation to limit amount of usable RAM"
(d312ae878b) breaks machines that
do not use 'dom0_mem=' argument with:
reserve RAM buffer: 000000133f2e2000 - 000000133fffffff
(XEN) mm.c:4976:d0 Global bit is set to kernel page fffff8117e
(XEN) domain_crash_sync called from entry.S
(XEN) Domain 0 (vcpu#0) crashed on cpu#0:
...
The reason being that the last E820 entry is created using the
'extra_pages' (which is based on how many pages have been freed).
The mentioned git commit sets the initial value of 'extra_pages'
using a hypercall which returns the number of pages (if dom0_mem
has been used) or -1 otherwise. If the later we return with
MAX_DOMAIN_PAGES as basis for calculation:
return min(max_pages, MAX_DOMAIN_PAGES);
and use it:
extra_limit = xen_get_max_pages();
if (extra_limit >= max_pfn)
extra_pages = extra_limit - max_pfn;
else
extra_pages = 0;
which means we end up with extra_pages = 128GB in PFNs (33554432)
- 8GB in PFNs (2097152, on this specific box, can be larger or smaller),
and then we add that value to the E820 making it:
Xen: 00000000ff000000 - 0000000100000000 (reserved)
Xen: 0000000100000000 - 000000133f2e2000 (usable)
which is clearly wrong. It should look as so:
Xen: 00000000ff000000 - 0000000100000000 (reserved)
Xen: 0000000100000000 - 000000027fbda000 (usable)
Naturally this problem does not present itself if dom0_mem=max:X
is used.
CC: stable@kernel.org
Signed-off-by: David Vrabel <david.vrabel@citrix.com>
Signed-off-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
