On machines without the ibm,client-architecture-support call we were missing a
newline. We may as well print the full name in all its glory too - its
ibm,client-architecture-support, not ibm,client-architecture as I mistakenly
wrote (a name only an IBM architect could love).
For my penance I will write out ibm,client-architecture-support 100 times.
Before:
Calling ibm,client-architecture...command line: root=/dev/sda6 console=hvc0 quiet
After:
Calling ibm,client-architecture-support... not implemented
command line: root=/dev/sda6 console=hvc0
Signed-off-by: Anton Blanchard <anton@samba.org>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Some System p configurations can already have more than 16 nodes so we
need to increase NODES_SHIFT. I chose 256 to give us some room to grow in the
future, although we can look at something smaller if the memory bloat is
considered too much.
Unless we clamp MAX_ACTIVE_REGIONS we end up with 300kB of extra bloat in
early_node_map in mm/page_alloc.c:
< 6144 early_node_map
> 307200 early_node_map
due to:
#if MAX_NUMNODES >= 32
/* If there can be many nodes, allow up to 50 holes per node */
#define MAX_ACTIVE_REGIONS (MAX_NUMNODES*50)
#else
/* By default, allow up to 256 distinct regions */
#define MAX_ACTIVE_REGIONS 256
Since our memory is mostly contiguous it seems reasonable to keep this
at 256 for now. I also set 32bit to 32 to save space (is there any chance
a 32bit system will have more than 32 discontiguous memory ranges?).
Even with that fixed we have a few data structures that grow:
< 896 bootmem_node_data
> 14336 bootmem_node_data
< 1280 node_devices
> 20480 node_devices
< 25088 kmalloc_caches
> 59648 kmalloc_caches
< 1632 hstates
> 21792 hstates
Signed-off-by: Anton Blanchard <anton@samba.org>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
perf_counter uses arch_vma_name() to detect a vdso region which in turn uses
current->mm->context.vdso_base. We need to initialise this before doing
the mmap or else we fail to detect the vdso.
Signed-off-by: Anton Blanchard <anton@samba.org>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
If we are using 1TB segments and we are allowed to randomise the heap, we can
put it above 1TB so it is backed by a 1TB segment. Otherwise the heap will be
in the bottom 1TB which always uses 256MB segments and this may result in a
performance penalty.
This functionality is disabled when heap randomisation is turned off:
echo 1 > /proc/sys/kernel/randomize_va_space
which may be useful when trying to allocate the maximum amount of 16M or 16G
pages.
On a microbenchmark that repeatedly touches 32GB of memory with a stride of
256MB + 4kB (designed to stress 256MB segments while still mapping nicely into
the L1 cache), we see the improvement:
Force malloc to use heap all the time:
# export MALLOC_MMAP_MAX_=0 MALLOC_TRIM_THRESHOLD_=-1
Disable heap randomization:
# echo 1 > /proc/sys/kernel/randomize_va_space
# time ./test
12.51s
Enable heap randomization:
# echo 2 > /proc/sys/kernel/randomize_va_space
# time ./test
1.70s
Signed-off-by: Anton Blanchard <anton@samba.org>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Sometimes this is used to hold a simple offset, and sometimes
it is used to hold a pointer. This patch changes it to a union containing
void * and dma_addr_t. get/set accessors are also provided, because it was
getting a bit ugly to get to the actual data.
Signed-off-by: Becky Bruce <beckyb@kernel.crashing.org>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
The former is no longer really accurate with the swiotlb case now
a possibility. I also move it into dma-mapping.h - it no longer
needs to be in dma.c, and there are about to be some more accessors
that should all end up in the same place. A comment is added to
indicate that this function is not used in configs where there is no
simple dma offset, such as the iommu case.
Signed-off-by: Becky Bruce <beckyb@kernel.crashing.org>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
When using CONFIG_RELOCATABLE, we build the kernel as a position
independent executable. The kernel then uses a little bit of relocation
code to relocate itself. That code only deals with R_PPC64_RELATIVE
relocations though. If for some reason you use assembly constructs
such as LOAD_REG_IMMEDIATE() to load the address of a symbol, you'll
generate different kinds of relocations that won't be processed properly
and bad things will happen. (We have 2 such bugs today).
The perl script tries to filter out "known" bad ones. It's possible
that we are missing some in the case of a weak function that nobody
implements, we'll see if we get false positive and fix it.
Signed-off-by: Tony Breeds <tony@bakeyournoodle.com>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Rename the locking free hvc_resize() function to __hvc_resize() and
provide an inline function that locks the hvc_struct and calls
__hvc_resize().
The rationale for this patch is that virtio_console calls the hvc_resize()
function without locking the hvc_struct. So it needs to call the lock
itself.
According to naming rules, the unlocked version is renamed and
prefixed with "__".
References to unlocked function calls in hvc back-ends has been updated.
Signed-off-by: Hendrik Brueckner <brueckner@linux.vnet.ibm.com>
Acked-by: Christian Borntraeger <borntraeger@de.ibm.com>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
* git://git.kernel.org/pub/scm/linux/kernel/git/rusty/linux-2.6-for-linus: (39 commits)
cpumask: Move deprecated functions to end of header.
cpumask: remove unused deprecated functions, avoid accusations of insanity
cpumask: use new-style cpumask ops in mm/quicklist.
cpumask: use mm_cpumask() wrapper: x86
cpumask: use mm_cpumask() wrapper: um
cpumask: use mm_cpumask() wrapper: mips
cpumask: use mm_cpumask() wrapper: mn10300
cpumask: use mm_cpumask() wrapper: m32r
cpumask: use mm_cpumask() wrapper: arm
cpumask: Use accessors for cpu_*_mask: um
cpumask: Use accessors for cpu_*_mask: powerpc
cpumask: Use accessors for cpu_*_mask: mips
cpumask: Use accessors for cpu_*_mask: m32r
cpumask: remove arch_send_call_function_ipi
cpumask: arch_send_call_function_ipi_mask: s390
cpumask: arch_send_call_function_ipi_mask: powerpc
cpumask: arch_send_call_function_ipi_mask: mips
cpumask: arch_send_call_function_ipi_mask: m32r
cpumask: arch_send_call_function_ipi_mask: alpha
cpumask: remove obsolete topology_core_siblings and topology_thread_siblings: ia64
...
* remove asm/atomic.h inclusion from linux/utsname.h --
not needed after kref conversion
* remove linux/utsname.h inclusion from files which do not need it
NOTE: it looks like fs/binfmt_elf.c do not need utsname.h, however
due to some personality stuff it _is_ needed -- cowardly leave ELF-related
headers and files alone.
Signed-off-by: Alexey Dobriyan <adobriyan@gmail.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
This reverts commit c02e3f361c ("kmod: fix race in usermodehelper code")
The patch is wrong. UMH_WAIT_EXEC is called with VFORK what ensures
that the child finishes prior returing back to the parent. No race.
In fact, the patch makes it even worse because it does the thing it
claims not do:
- It calls ->complete() on UMH_WAIT_EXEC
- the complete() callback may de-allocated subinfo as seen in the
following call chain:
[<c009f904>] (__link_path_walk+0x20/0xeb4) from [<c00a094c>] (path_walk+0x48/0x94)
[<c00a094c>] (path_walk+0x48/0x94) from [<c00a0a34>] (do_path_lookup+0x24/0x4c)
[<c00a0a34>] (do_path_lookup+0x24/0x4c) from [<c00a158c>] (do_filp_open+0xa4/0x83c)
[<c00a158c>] (do_filp_open+0xa4/0x83c) from [<c009ba90>] (open_exec+0x24/0xe0)
[<c009ba90>] (open_exec+0x24/0xe0) from [<c009bfa8>] (do_execve+0x7c/0x2e4)
[<c009bfa8>] (do_execve+0x7c/0x2e4) from [<c0026a80>] (kernel_execve+0x34/0x80)
[<c0026a80>] (kernel_execve+0x34/0x80) from [<c004b514>] (____call_usermodehelper+0x130/0x148)
[<c004b514>] (____call_usermodehelper+0x130/0x148) from [<c0024858>] (kernel_thread_exit+0x0/0x8)
and the path pointer was NULL. Good that ARM's kernel_execve()
doesn't check the pointer for NULL or else I wouldn't notice it.
The only race there might be is with UMH_NO_WAIT but it is too late for
me to investigate it now. UMH_WAIT_PROC could probably also use VFORK
and we could save one exec. So the only race I see is with UMH_NO_WAIT
and recent scheduler changes where the child does not always run first
might have trigger here something but as I said, it is late....
Signed-off-by: Sebastian Andrzej Siewior <sebastian@breakpoint.cc>
Acked-by: Neil Horman <nhorman@tuxdriver.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
During releasepage, we try to drop any extent_state structs for the
bye offsets of the page we're releaseing. But the code was incorrectly
telling clear_extent_bit to delete the state struct unconditionallly.
Normally this would be fine because we have the page locked, but other
parts of btrfs will lock down an entire extent, the most common place
being IO completion.
releasepage was deleting the extent state without first locking the extent,
which may result in removing a state struct that another process had
locked down. The fix here is to leave the NODATASUM and EXTENT_LOCKED
bits alone in releasepage.
Signed-off-by: Chris Mason <chris.mason@oracle.com>
If test_range_bit finds an extent that goes all the way to (u64)-1, it
can incorrectly wrap the u64 instead of treaing it like the end of
the address space.
This just adds a check for the highest possible offset so we don't wrap.
Signed-off-by: Chris Mason <chris.mason@oracle.com>
Both set and clear_extent_bit allow passing a cached
state struct to reduce rbtree search times. clear_extent_bit
was improperly bypassing some of the checks around making sure
the extent state fields were correct for a given operation.
The fix used here (from Yan Zheng) is to use the hit_next
goto target instead of jumping all the way down to start clearing
bits without making sure the cached state was exactly correct
for the operation we were doing.
This also fixes up the setting of the start variable for both
ops in the case where we find an overlapping extent that
begins before the range we want to change. In both cases
we were incorrectly going backwards from the original
requested change.
Signed-off-by: Chris Mason <chris.mason@oracle.com>
Saves us one cycle of alloc-add-free if the queue was full.
Signed-off-by: Amit Shah <amit.shah@redhat.com>
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au> (modified)
Now we can tell the theoretical capacity remaining in the output
queue, virtio_net can waste entries by stopping the queue early.
It doesn't work in the case of indirect buffers and kmalloc failure,
but that's rare (we could drop the packet in that case, but other
drivers return TX_BUSY for similar reasons).
For the record, I think this patch reflects poorly on the linux
network API.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Cc: Dinesh Subhraveti <dineshs@us.ibm.com>
We put the virtio_net_hdr into the skb's cb region; turn this into a
union to clean up the code slightly and allow future expansion.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Cc: Mark McLoughlin <markmc@redhat.com>
Cc: Dinesh Subhraveti <dineshs@us.ibm.com>
The virtio_net driver is complicated by the two methods of freeing old
xmit buffers (in addition to freeing old ones at the start of the xmit
path).
The original code used a 1/10 second timer attached to xmit_free(),
reset on every xmit. Before we orphaned skbs on xmit, the
transmitting userspace could block with a full socket until the timer
fired, the skb destructor was called, and they were re-woken.
So we added the VIRTIO_F_NOTIFY_ON_EMPTY feature: supporting devices
send an interrupt (even if normally suppressed) on an empty xmit ring
which makes us schedule xmit_tasklet(). This was a benchmark win.
Unfortunately, VIRTIO_F_NOTIFY_ON_EMPTY makes quite a lot of work: a
host which is faster than the guest will fire the interrupt every xmit
packet (slowing the guest down further). Attempting mitigation in the
host adds overhead of userspace timers (possibly with the additional
pain of signals), and risks increasing latency anyway if you get it
wrong.
In practice, this effect was masked by benchmarks which take advantage
of GSO (with its inherent transmit batching), but it's still there.
Now we orphan xmitted skbs, the pressure is off: remove both paths and
no longer request VIRTIO_F_NOTIFY_ON_EMPTY. Note that the current
QEMU will notify us even if we don't negotiate this feature (legal,
but suboptimal); a patch is outstanding to improve that.
Move the skb_orphan/nf_reset to after we've done the send and notified
the other end, for a slight optimization.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Cc: Mark McLoughlin <markmc@redhat.com>
This effectively reverts 99ffc696d1
"virtio: wean net driver off NETDEV_TX_BUSY".
The complexity of queuing an skb (setting a tasklet to re-xmit) is
questionable, especially once we get rid of the other reason for the
tasklet in the next patch.
If the skb won't fit in the tx queue, just return NETDEV_TX_BUSY.
This is frowned upon, so a followup patch uses a more complex solution.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Cc: Herbert Xu <herbert@gondor.apana.org.au>
The complex transmit free logic was introduced to avoid hangs on
removing the ip_conntrack module and also because drivers aren't
generally supposed to keep stale skbs for unbounded times.
After some debate, it was decided that while doing skb_orphan()
generally is a rat's nest, we can do it in this driver. Following
patches take advantage of this.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
The new ones have pretty kerneldoc. Move the old ones to the end to
avoid confusing people.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Cc: benh@kernel.crashing.org
We're not forcing removal of the old cpu_ functions, but we might as
well delete the now-unused ones.
Especially CPUMASK_ALLOC and friends. I actually got a phone call (!)
from a hacker who thought I had introduced them as the new cpumask
API. He seemed bewildered that I had lost all taste.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Cc: benh@kernel.crashing.org
Makes code futureproof against the impending change to mm->cpu_vm_mask (to be a pointer).
It's also a chance to use the new cpumask_ ops which take a pointer
(the older ones are deprecated, but there's no hurry for arch code).
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Makes code futureproof against the impending change to mm->cpu_vm_mask.
It's also a chance to use the new cpumask_ ops which take a pointer
(the older ones are deprecated, but there's no hurry for arch code).
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Makes code futureproof against the impending change to mm->cpu_vm_mask.
It's also a chance to use the new cpumask_ ops which take a pointer
(the older ones are deprecated, but there's no hurry for arch code).
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Makes code futureproof against the impending change to mm->cpu_vm_mask
(to be a pointer).
It's also a chance to use the new cpumask_ ops which take a pointer
(the older ones are deprecated, but there's no hurry for arch code).
Also change the actual arg name here to "mm" (which it is), not "task".
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Makes code futureproof against the impending change to mm->cpu_vm_mask.
It's also a chance to use the new cpumask_ ops which take a pointer
(the older ones are deprecated, but there's no hurry for arch code).
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Acked-by: Hirokazu Takata <takata@linux-m32r.org> (fixes)
Makes code futureproof against the impending change to mm->cpu_vm_mask.
It's also a chance to use the new cpumask_ ops which take a pointer
(the older ones are deprecated, but there's no hurry for arch code).
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Use the accessors rather than frobbing bits directly (the new versions
are const).
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Signed-off-by: Mike Travis <travis@sgi.com>
Use the accessors rather than frobbing bits directly (the new versions
are const).
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Signed-off-by: Mike Travis <travis@sgi.com>
Use the accessors rather than frobbing bits directly (the new versions
are const).
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Signed-off-by: Mike Travis <travis@sgi.com>
Use the accessors rather than frobbing bits directly (the new versions
are const).
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Signed-off-by: Mike Travis <travis@sgi.com>
We're weaning the core code off handing cpumask's around on-stack.
This introduces arch_send_call_function_ipi_mask().
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
We're weaning the core code off handing cpumask's around on-stack.
This introduces arch_send_call_function_ipi_mask(), and by defining
it, the old arch_send_call_function_ipi is defined by the core code.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
We're weaning the core code off handing cpumask's around on-stack.
This introduces arch_send_call_function_ipi_mask(), and by defining
it, the old arch_send_call_function_ipi is defined by the core code.
We also take the chance to wean the implementations off the
obsolescent for_each_cpu_mask(): making send_ipi_mask take the pointer
seemed the most natural way to ensure all implementations used
for_each_cpu.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
We're weaning the core code off handing cpumask's around on-stack.
This introduces arch_send_call_function_ipi_mask(), and by defining
it, the old arch_send_call_function_ipi is defined by the core code.
We also take the chance to wean the implementations off the
obsolescent for_each_cpu_mask(): making send_ipi_mask take the pointer
seemed the most natural way to ensure all implementations used
for_each_cpu.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
We're weaning the core code off handing cpumask's around on-stack.
This introduces arch_send_call_function_ipi_mask().
We also take the chance to wean the send_ipi_message off the
obsolescent for_each_cpu_mask(): making it take a pointer seemed the
most natural way to do this.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
smp_call_function_many is the new version: it takes a pointer. Also,
use mm accessor macro while we're changing this.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
