Before describing bugs itself, I first explain definition of freepage.
1. pages on buddy list are counted as freepage.
2. pages on isolate migratetype buddy list are *not* counted as freepage.
3. pages on cma buddy list are counted as CMA freepage, too.
Now, I describe problems and related patch.
Patch 1: There is race conditions on getting pageblock migratetype that
it results in misplacement of freepages on buddy list, incorrect
freepage count and un-availability of freepage.
Patch 2: Freepages on pcp list could have stale cached information to
determine migratetype of buddy list to go. This causes misplacement of
freepages on buddy list and incorrect freepage count.
Patch 4: Merging between freepages on different migratetype of
pageblocks will cause freepages accouting problem. This patch fixes it.
Without patchset [3], above problem doesn't happens on my CMA allocation
test, because CMA reserved pages aren't used at all. So there is no
chance for above race.
With patchset [3], I did simple CMA allocation test and get below
result:
- Virtual machine, 4 cpus, 1024 MB memory, 256 MB CMA reservation
- run kernel build (make -j16) on background
- 30 times CMA allocation(8MB * 30 = 240MB) attempts in 5 sec interval
- Result: more than 5000 freepage count are missed
With patchset [3] and this patchset, I found that no freepage count are
missed so that I conclude that problems are solved.
On my simple memory offlining test, these problems also occur on that
environment, too.
This patch (of 4):
There are two paths to reach core free function of buddy allocator,
__free_one_page(), one is free_one_page()->__free_one_page() and the
other is free_hot_cold_page()->free_pcppages_bulk()->__free_one_page().
Each paths has race condition causing serious problems. At first, this
patch is focused on first type of freepath. And then, following patch
will solve the problem in second type of freepath.
In the first type of freepath, we got migratetype of freeing page
without holding the zone lock, so it could be racy. There are two cases
of this race.
1. pages are added to isolate buddy list after restoring orignal
migratetype
CPU1 CPU2
get migratetype => return MIGRATE_ISOLATE
call free_one_page() with MIGRATE_ISOLATE
grab the zone lock
unisolate pageblock
release the zone lock
grab the zone lock
call __free_one_page() with MIGRATE_ISOLATE
freepage go into isolate buddy list,
although pageblock is already unisolated
This may cause two problems. One is that we can't use this page anymore
until next isolation attempt of this pageblock, because freepage is on
isolate buddy list. The other is that freepage accouting could be wrong
due to merging between different buddy list. Freepages on isolate buddy
list aren't counted as freepage, but ones on normal buddy list are
counted as freepage. If merge happens, buddy freepage on normal buddy
list is inevitably moved to isolate buddy list without any consideration
of freepage accouting so it could be incorrect.
2. pages are added to normal buddy list while pageblock is isolated.
It is similar with above case.
This also may cause two problems. One is that we can't keep these
freepages from being allocated. Although this pageblock is isolated,
freepage would be added to normal buddy list so that it could be
allocated without any restriction. And the other problem is same as
case 1, that it, incorrect freepage accouting.
This race condition would be prevented by checking migratetype again
with holding the zone lock. Because it is somewhat heavy operation and
it isn't needed in common case, we want to avoid rechecking as much as
possible. So this patch introduce new variable, nr_isolate_pageblock in
struct zone to check if there is isolated pageblock. With this, we can
avoid to re-check migratetype in common case and do it only if there is
isolated pageblock or migratetype is MIGRATE_ISOLATE. This solve above
mentioned problems.
Changes from v3:
Add one more check in free_one_page() that checks whether migratetype is
MIGRATE_ISOLATE or not. Without this, abovementioned case 1 could happens.
Signed-off-by: Joonsoo Kim <iamjoonsoo.kim@lge.com>
Acked-by: Minchan Kim <minchan@kernel.org>
Acked-by: Michal Nazarewicz <mina86@mina86.com>
Acked-by: Vlastimil Babka <vbabka@suse.cz>
Cc: "Kirill A. Shutemov" <kirill@shutemov.name>
Cc: Mel Gorman <mgorman@suse.de>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: Yasuaki Ishimatsu <isimatu.yasuaki@jp.fujitsu.com>
Cc: Zhang Yanfei <zhangyanfei@cn.fujitsu.com>
Cc: Tang Chen <tangchen@cn.fujitsu.com>
Cc: Naoya Horiguchi <n-horiguchi@ah.jp.nec.com>
Cc: Bartlomiej Zolnierkiewicz <b.zolnierkie@samsung.com>
Cc: Wen Congyang <wency@cn.fujitsu.com>
Cc: Marek Szyprowski <m.szyprowski@samsung.com>
Cc: Laura Abbott <lauraa@codeaurora.org>
Cc: Heesub Shin <heesub.shin@samsung.com>
Cc: "Aneesh Kumar K.V" <aneesh.kumar@linux.vnet.ibm.com>
Cc: Ritesh Harjani <ritesh.list@gmail.com>
Cc: Gioh Kim <gioh.kim@lge.com>
Cc: <stable@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Commit 7d49d88683 ("mm, compaction: reduce zone checking frequency in
the migration scanner") has a side-effect that changes the iteration
range calculation. Before the change, block_end_pfn is calculated using
start_pfn, but now it blindly adds pageblock_nr_pages to the previous
value.
This causes the problem that isolation_start_pfn is larger than
block_end_pfn when we isolate the page with more than pageblock order.
In this case, isolation would fail due to an invalid range parameter.
To prevent this, this patch implements skipping the range until a proper
target pageblock is met. Without this patch, CMA with more than
pageblock order always fails but with this patch it will succeed.
Signed-off-by: Joonsoo Kim <iamjoonsoo.kim@lge.com>
Cc: Vlastimil Babka <vbabka@suse.cz>
Cc: Minchan Kim <minchan@kernel.org>
Cc: Michal Nazarewicz <mina86@mina86.com>
Cc: Naoya Horiguchi <n-horiguchi@ah.jp.nec.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
zram could kunmap_atomic() a NULL pointer in a rare situation: a zram
page becomes a full-zeroed page after a partial write io. The current
code doesn't handle this case and performs kunmap_atomic() on a NULL
pointer, which panics the kernel.
This patch fixes this issue.
Signed-off-by: Weijie Yang <weijie.yang@samsung.com>
Cc: Sergey Senozhatsky <sergey.senozhatsky@gmail.com>
Cc: Dan Streetman <ddstreet@ieee.org>
Cc: Nitin Gupta <ngupta@vflare.org>
Cc: Weijie Yang <weijie.yang.kh@gmail.com>
Acked-by: Jerome Marchand <jmarchan@redhat.com>
Cc: <stable@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Use the global current_stack_pointer to get the value of the stack pointer.
This change supports being able to compile the kernel with both gcc and clang.
Signed-off-by: Behan Webster <behanw@converseincode.com>
Reviewed-by: Mark Charlebois <charlebm@gmail.com>
Reviewed-by: Jan-Simon Möller <dl9pf@gmx.de>
Acked-by: Will Deacon <will.deacon@arm.com>
Acked-by: Nicolas Pitre <nico@linaro.org>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
Use the global current_stack_pointer to calculate the end of the stack for
current_pt_regs()
Signed-off-by: Behan Webster <behanw@converseincode.com>
Reviewed-by: Mark Charlebois <charlebm@gmail.com>
Reviewed-by: Jan-Simon Möller <dl9pf@gmx.de>
Acked-by: Will Deacon <will.deacon@arm.com>
Acked-by: Nicolas Pitre <nico@linaro.org>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
Using global current_stack_pointer works on both clang and gcc.
current_stack_pointer is an unsigned long and needs to be cast
as a pointer to dereference.
Signed-off-by: Mark Charlebois <charlebm@gmail.com>
Signed-off-by: Behan Webster <behanw@converseincode.com>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
Use the global current_stack_pointer to get the value of the stack pointer.
This change supports being able to compile the kernel with both gcc and clang.
Signed-off-by: Behan Webster <behanw@converseincode.com>
Reviewed-by: Mark Charlebois <charlebm@gmail.com>
Reviewed-by: Jan-Simon Möller <dl9pf@gmx.de>
Acked-by: Will Deacon <will.deacon@arm.com>
Acked-by: Nicolas Pitre <nico@linaro.org>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
Use the global current_stack_pointer to get the value of the stack pointer.
This change supports being able to compile the kernel with both gcc and clang.
Signed-off-by: Behan Webster <behanw@converseincode.com>
Reviewed-by: Mark Charlebois <charlebm@gmail.com>
Reviewed-by: Jan-Simon Möller <dl9pf@gmx.de>
Acked-by: Will Deacon <will.deacon@arm.com>
Acked-by: Nicolas Pitre <nico@linaro.org>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
Use the global current_stack_pointer to get the value of the stack pointer.
This change supports being able to compile the kernel with both gcc and Clang.
Signed-off-by: Behan Webster <behanw@converseincode.com>
Reviewed-by: Mark Charlebois <charlebm@gmail.com>
Reviewed-by: Jan-Simon Möller <dl9pf@gmx.de>
Acked-by: Will Deacon <will.deacon@arm.com>
Acked-by: Nicolas Pitre <nico@linaro.org>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
Define a global named register for current_stack_pointer. The use of this new
variable guarantees that both gcc and clang can access this register in C code.
Signed-off-by: Behan Webster <behanw@converseincode.com>
Reviewed-by: Jan-Simon Möller <dl9pf@gmx.de>
Reviewed-by: Mark Charlebois <charlebm@gmail.com>
Acked-by: Will Deacon <will.deacon@arm.com>
Acked-by: Nicolas Pitre <nico@linaro.org>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
The kuser helpers page is not set up on non-MMU systems, so it does
not make sense to allow CONFIG_KUSER_HELPERS to be enabled when
CONFIG_MMU=n. Allowing it to be set on !MMU results in an oops in
set_tls (used in execve and the arm_syscall trap handler):
Unhandled exception: IPSR = 00000005 LR = fffffff1
CPU: 0 PID: 1 Comm: swapper Not tainted 3.18.0-rc1-00041-ga30465a #216
task: 8b838000 ti: 8b82a000 task.ti: 8b82a000
PC is at flush_thread+0x32/0x40
LR is at flush_thread+0x21/0x40
pc : [<8f00157a>] lr : [<8f001569>] psr: 4100000b
sp : 8b82be20 ip : 00000000 fp : 8b83c000
r10: 00000001 r9 : 88018c84 r8 : 8bb85000
r7 : 8b838000 r6 : 00000000 r5 : 8bb77400 r4 : 8b82a000
r3 : ffff0ff0 r2 : 8b82a000 r1 : 00000000 r0 : 88020354
xPSR: 4100000b
CPU: 0 PID: 1 Comm: swapper Not tainted 3.18.0-rc1-00041-ga30465a #216
[<8f002bc1>] (unwind_backtrace) from [<8f002033>] (show_stack+0xb/0xc)
[<8f002033>] (show_stack) from [<8f00265b>] (__invalid_entry+0x4b/0x4c)
As best I can tell this issue existed for the set_tls ARM syscall
before commit fbfb872f5f "ARM: 8148/1: flush TLS and thumbee
register state during exec" consolidated the TLS manipulation code
into the set_tls helper function, but now that we're using it to flush
register state during execve, !MMU users encounter the oops at the
first exec.
Prevent CONFIG_MMU=n configurations from enabling
CONFIG_KUSER_HELPERS.
Fixes: fbfb872f5f (ARM: 8148/1: flush TLS and thumbee register state during exec)
Signed-off-by: Nathan Lynch <nathan_lynch@mentor.com>
Reported-by: Stefan Agner <stefan@agner.ch>
Acked-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Cc: stable@vger.kernel.org
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
To speed up decompression, the decompressor sets up a flat, cacheable
mapping of memory. However, when there is insufficient space to hold
the page tables for this mapping, we don't bother to enable the caches
and subsequently skip all the cache maintenance hooks.
Skipping the cache maintenance before jumping to the relocated code
allows the processor to predict the branch and populate the I-cache
with stale data before the relocation loop has completed (since a
bootloader may have SCTLR.I set, which permits normal, cacheable
instruction fetches regardless of SCTLR.M).
This patch moves the cache maintenance check into the maintenance
routines themselves, allowing the v6/v7 versions to invalidate the
I-cache regardless of the MMU state.
Cc: <stable@vger.kernel.org>
Reported-by: Marc Carino <marc.ceeeee@gmail.com>
Tested-by: Julien Grall <julien.grall@linaro.org>
Signed-off-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
The pcc-cpufreq driver is not automatically loaded on systems where
the platform's power management setting requires this driver.
Instead, on those systems no CPU frequency driver is registered and
active.
Make the autoloading matching criteria for loading the pcc-cpufreq
driver the same as done in acpi-cpufreq by commit c655affbd5
("ACPI / cpufreq: Add ACPI processor device IDs to acpi-cpufreq").
x86 CPU frequency drivers are now typically autoloaded by specifying
MODULE_DEVICE_TABLE entries and x86cpu model specific matching.
But pcc-cpufreq was omitted when acpi-cpufreq and other drivers were
changed to use this approach.
Both acpi-cpufreq and pcc-cpufreq depend on a distinct and mutually
exclusive set of ACPI methods which are not directly tied to specific
processor model numbers. Both of these drivers have init routines
which look for their required ACPI methods. As a result, only the
appropriate driver registers as the cpu frequency driver and the other
one ends up being unloaded.
Tested on various systems where acpi-cpufreq, intel_pstate, and
pcc-cpufreq are the expected cpu frequency drivers.
Signed-off-by: Lenny Szubowicz <lszubowi@redhat.com>
Signed-off-by: Joseph Szczypek <joseph.szczypek@hp.com>
Reported-by: Trinh Dao <trinh.dao@hp.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
PM domains are powered on/off from various places. Some callers do
latency measurements, others don't. Consolidate using two helper
functions, which always measure the latencies, and update the stored
latencies when needed.
Other minor changes:
- Use pr_warn() instead of pr_warning(),
- There's no need to check genpd->name, %s handles NULL pointers fine,
- Make the warning format strings identical, to save memory.
Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be>
Reviewed-by: Ulf Hansson <ulf.hansson@linaro.org>
Reviewed-by: Kevin Hilman <khilman@linaro.org>
Acked-by: Pavel Machek <pavel@ucw.cz>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
The PM domain pointed to by the genpd parameter is never modified.
Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be>
Acked-by: Kevin Hilman <khilman@linaro.org>
Acked-by: Pavel Machek <pavel@ucw.cz>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Previously we applied _HPX type 2 record Link Control register settings
only to bridges with a subordinate bus. But it's better to apply them to
all devices with a link because if the subordinate bus has not been
allocated yet, we won't apply settings to the device.
Use pcie_cap_has_lnkctl() to determine whether the device has a Link
Control register instead of looking at dev->subordinate.
[bhelgaas: changelog]
Fixes: 6cd33649fa ("PCI: Add pci_configure_device() during enumeration")
Signed-off-by: Yinghai Lu <yinghai@kernel.org>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
41e5c0f81d ("of/pci: Add pci_get_new_domain_nr() and
of_get_pci_domain_nr()") added parsing of the "linux,pci-domain" property,
but didn't add the binding documentation.
Since this property will be supported by a number of host bridge drivers,
add it to the common PCI binding doc.
Fixes: 41e5c0f81d ("of/pci: Add pci_get_new_domain_nr() and of_get_pci_domain_nr()")
Signed-off-by: Lucas Stach <l.stach@pengutronix.de>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Acked-by: Liviu Dudau <Liviu.Dudau@arm.com>
Acked-by: Rob Herring <robh@kernel.org>
This is a single patch that fixes the VBLANK machinery after:
7ffd7a6851 drm: Always reject drm_vblank_get() after drm_vblank_off()
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2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=oKTG
-----END PGP SIGNATURE-----
Merge tag 'drm/tegra/for-3.18-rc5' of git://people.freedesktop.org/~tagr/linux into drm-fixes
drm/tegra: Fixes for v3.18-rc5
This is a single patch that fixes the VBLANK machinery after:
7ffd7a6851 drm: Always reject drm_vblank_get() after drm_vblank_off()
* tag 'drm/tegra/for-3.18-rc5' of git://people.freedesktop.org/~tagr/linux:
drm/tegra: dc: Add missing call to drm_vblank_on()
One modesetting, one gk20a fix.
* 'linux-3.18' of git://anongit.freedesktop.org/git/nouveau/linux-2.6:
drm/nouveau/nv50/disp: Fix modeset on G94
drm/gk20a/fb: fix setting of large page size bit
In DC world, GSO packets initially cooked by tcp_sendmsg() are usually
big, as sk_pacing_rate is high.
When network is congested, cwnd can be smaller than the GSO packets
found in socket write queue. tcp_write_xmit() splits GSO packets
using the available cwnd, and we end up sending a single GSO packet,
consuming all available cwnd.
With GRO aggregation on the receiver, we might handle a single GRO
packet, sending back a single ACK.
1) This single ACK might be lost
TLP or RTO are forced to attempt a retransmit.
2) This ACK releases a full cwnd, sender sends another big GSO packet,
in a ping pong mode.
This behavior does not fill the pipes in the best way, because of
scheduling artifacts.
Make sure we always have at least two GSO packets in flight.
This allows us to safely increase GRO efficiency without risking
spurious retransmits.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Acked-by: Neal Cardwell <ncardwell@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
one regression fix.
* tag 'drm-intel-fixes-2014-11-13' of git://anongit.freedesktop.org/drm-intel:
drm/i915: Fix obj->map_and_fenceable across tiling changes
Currently, we only match against local port number in order to reuse
socket. But if this new vxlan wants an IPv6 socket and a IPv4 one bound
to that port, vxlan will reuse an IPv4 socket as IPv6 and a panic will
follow. The following steps reproduce it:
# ip link add vxlan6 type vxlan id 42 group 229.10.10.10 \
srcport 5000 6000 dev eth0
# ip link add vxlan7 type vxlan id 43 group ff0e::110 \
srcport 5000 6000 dev eth0
# ip link set vxlan6 up
# ip link set vxlan7 up
<panic>
[ 4.187481] BUG: unable to handle kernel NULL pointer dereference at 0000000000000058
...
[ 4.188076] Call Trace:
[ 4.188085] [<ffffffff81667c4a>] ? ipv6_sock_mc_join+0x3a/0x630
[ 4.188098] [<ffffffffa05a6ad6>] vxlan_igmp_join+0x66/0xd0 [vxlan]
[ 4.188113] [<ffffffff810a3430>] process_one_work+0x220/0x710
[ 4.188125] [<ffffffff810a33c4>] ? process_one_work+0x1b4/0x710
[ 4.188138] [<ffffffff810a3a3b>] worker_thread+0x11b/0x3a0
[ 4.188149] [<ffffffff810a3920>] ? process_one_work+0x710/0x710
So address family must also match in order to reuse a socket.
Reported-by: Jean-Tsung Hsiao <jhsiao@redhat.com>
Signed-off-by: Marcelo Ricardo Leitner <mleitner@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Reallocation is only required for shrinking and expanding and both rely
on a mutex for synchronization and callers of rhashtable_init() are in
non atomic context. Therefore, no reason to continue passing allocation
hints through the API.
Instead, use GFP_KERNEL and add __GFP_NOWARN | __GFP_NORETRY to allow
for silent fall back to vzalloc() without the OOM killer jumping in as
pointed out by Eric Dumazet and Eric W. Biederman.
Signed-off-by: Thomas Graf <tgraf@suug.ch>
Signed-off-by: David S. Miller <davem@davemloft.net>
Or Gerlitz says:
====================
mlx4: Flexible (asymmetric) allocation of EQs and MSI-X vectors
This series from Matan Barak is built as follows:
The 1st two patches fix small bugs w.r.t firmware spec. Next
are two patches which do more re-factoring of the init/fini flow
and a patch that adds support for the QUERY_FUNC firmware command,
these are all pre-steps for the major patch of the series. In this
patch (#6) we change the order of talking/querying the firmware
and enabling SRIOV. This allows to remote worst-case assumption
w.r.t the number of available MSI-X vectors and EQs per function.
The last patch easily enjoys this ordering change, to enable
supports > 64 VFs over a firmware that allows for that.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
We now allow up to 126 VFs. Note though that certain firmware
versions only allow up to 80 VFs. Moreover, old HCAs only support 64 VFs.
In these cases, we limit the maximum number of VFs to 64.
Signed-off-by: Matan Barak <matanb@mellanox.com>
Signed-off-by: Or Gerlitz <ogerlitz@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Previously, the driver queried the firmware in order to get the number
of supported EQs. Under SRIOV, since this was done before the driver
notified the firmware how many VFs it actually needs, the firmware had
to take into account a worst case scenario and always allocated four EQs
per VF, where one was used for events while the others were used for completions.
Now, when the firmware supports the asymmetric allocation scheme, denoted
by exposing num_sys_eqs > 0 (--> MLX4_DEV_CAP_FLAG2_SYS_EQS), we use the
QUERY_FUNC command to query the firmware before enabling SRIOV. Thus we
can get more EQs and MSI-X vectors per function.
Moreover, when running in the new firmware/driver mode, the limitation
that the number of EQs should be a power of two is lifted.
Signed-off-by: Jack Morgenstein <jackm@dev.mellanox.co.il>
Signed-off-by: Matan Barak <matanb@mellanox.com>
Signed-off-by: Or Gerlitz <ogerlitz@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
QUERY_FUNC firmware command could be used in order to query the
number of EQs, reserved EQs, etc for a specific function.
Signed-off-by: Matan Barak <matanb@mellanox.com>
Signed-off-by: Or Gerlitz <ogerlitz@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Refactor mlx4_load_one, as a preparation step for a new and
more complicated load function. The goal is to support both
newer firmware that required init_hca to be done before
enable_sriov and legacy firmwares that requires things to
be done the other way around.
Signed-off-by: Matan Barak <matanb@mellanox.com>
Signed-off-by: Or Gerlitz <ogerlitz@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Refactoring mlx4_cmd_init and mlx4_cmd_cleanup such that partial init
and cleanup are possible. After this refactoring, calling mlx4_cmd_init
several times is safe.
This is necessary in the VF init flow when mlx4_init_hca returns -EACCESS,
we need to issue cleanup and re-attempt to call it with the slave flag.
Signed-off-by: Matan Barak <matanb@mellanox.com>
Signed-off-by: Or Gerlitz <ogerlitz@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
We've used an incorrect type for the loop counter and the
mlx4_QUERY_FUNC_CAP function. The current input modifier
is either a port or a boolean.
Since the number of ports is always a positive value < 255,
we should use u8 instead of an integer with casting.
Signed-off-by: Matan Barak <matanb@mellanox.com>
Signed-off-by: Or Gerlitz <ogerlitz@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
We mistakenly read the reserved_eqs field as a standard
numeric value rather than a log2 value.
Signed-off-by: Matan Barak <matanb@mellanox.com>
Signed-off-by: Or Gerlitz <ogerlitz@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Herbert Xu says:
====================
rhashtable: Allow local locks to be used and tested
This series moves mutex_is_held entirely under PROVE_LOCKING so
there is zero foot print when we're not debugging. More importantly
it adds a parrent argument to mutex_is_held so that we can test
local locks rather than global ones (e.g., per-namespace locks).
====================
Acked-by: Thomas Graf <tgraf@suug.ch>
Signed-off-by: David S. Miller <davem@davemloft.net>
Currently mutex_is_held can only test locks in the that are global
since it takes no arguments. This prevents rhashtable from being
used in places where locks are lock, e.g., per-namespace locks.
This patch adds a parent field to mutex_is_held and rhashtable_params
so that local locks can be used (and tested).
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: David S. Miller <davem@davemloft.net>
The rhashtable function mutex_is_held is only used when PROVE_LOCKING
is enabled. This patch makes the mutex_is_held field in rhashtable
optional depending on PROVE_LOCKING.
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: David S. Miller <davem@davemloft.net>
The rhashtable function mutex_is_held is only used when PROVE_LOCKING
is enabled. This patch modifies netfilter so that we can rhashtable.h
itself can later make mutex_is_held optional depending on PROVE_LOCKING.
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: David S. Miller <davem@davemloft.net>
The rhashtable function mutex_is_held is only used when PROVE_LOCKING
is enabled. This patch modifies netlink so that we can rhashtable.h
itself can later make mutex_is_held optional depending on PROVE_LOCKING.
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: David S. Miller <davem@davemloft.net>
With commit be9dad1f9f ("net: phy: suspend phydev when going
to HALTED"), the PHY device will be put in a low-power mode using
BMCR_PDOWN if the the interface is set down. The smsc911x driver does
a software_reset opening the device driver (ndo_open). In such case,
the PHY must be powered-up before access to any register and before
calling the software_reset function. Otherwise, as the PHY is powered
down the software reset fails and the interface can not be enabled
again.
This patch fixes this scenario that is easy to reproduce setting down
the network interface and setting up again.
$ ifconfig eth0 down
$ ifconfig eth0 up
ifconfig: SIOCSIFFLAGS: Input/output error
Signed-off-by: Enric Balletbo i Serra <eballetbo@iseebcn.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The device tree probing for R-Car M2N (r8a7793) is added.
Signed-off-by: Hisashi Nakamura <hisashi.nakamura.ak@renesas.com>
Signed-off-by: Yoshihiro Kaneko <ykaneko0929@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
When using RMMI mode, it is necessary to change in probe.
Signed-off-by: Hisashi Nakamura <hisashi.nakamura.ak@renesas.com>
Signed-off-by: Yoshihiro Kaneko <ykaneko0929@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Large receive offloading is known to cause problems if received packets
are passed to other host. Therefore the kernel disables it by calling
dev_disable_lro() whenever a network device is enslaved in a bridge or
forwarding is enabled for it (or globally). For virtual devices we need
to disable LRO on the underlying physical device (which is actually
receiving the packets).
Current dev_disable_lro() code handles this propagation for a vlan
(including 802.1ad nested vlan), macvlan or a vlan on top of a macvlan.
It doesn't handle other stacked devices and their combinations, in
particular propagation from a bond to its slaves which often causes
problems in virtualization setups.
As we now have generic data structures describing the upper-lower device
relationship, dev_disable_lro() can be generalized to disable LRO also
for all lower devices (if any) once it is disabled for the device
itself.
For bonding and teaming devices, it is necessary to disable LRO not only
on current slaves at the moment when dev_disable_lro() is called but
also on any slave (port) added later.
v2: use lower device links for all devices (including vlan and macvlan)
Signed-off-by: Michal Kubecek <mkubecek@suse.cz>
Acked-by: Veaceslav Falico <vfalico@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
There was a missing break statement so we set everything to
PKT_HASH_TYPE_L3 even when we intended to use PKT_HASH_TYPE_L4.
Fixes: 5b9dfe299e ('amd-xgbe: Provide support for receive side scaling')
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
My editor spewed garbage that looked like memory corruption on
my screen. It turns out that a number of occurences of "fi" got
turned into a ligature.
This patch replaces these ligatures with the ASCII letters "fi".
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Cheers,
Acked-by: Thomas Graf <tgraf@suug.ch>
Signed-off-by: David S. Miller <davem@davemloft.net>
Increased delay in the smsc911x_phy_disable_energy_detect (from 1ms to 2ms).
Dropped delays in the smsc911x_phy_enable_energy_detect (100ms and 1ms).
The patch affect SMSC LAN generation 4 chips with integrated PHY (LAN9221).
I saw problems with soft reset due to wrong udelay timings.
After I fixed udelay, I measured the time needed to bring integrated PHY
from power-down to operational mode (the time beetween clearing EDPWRDOWN
bit and soft reset complete event). I got 1ms (measured using ktime_get).
The value is equal to the current value (1ms) used in the
smsc911x_phy_disable_energy_detect. It is near the upper bound and in order
to avoid rare soft reset faults it is doubled (2ms).
I don't know official timing for bringing up integrated PHY as specs doesn't
clarify this (or may be I didn't found).
It looks safe to drop delays before and after setting EDPWRDOWN bit
(enable PHY power-down mode). I didn't saw any regressions with the patch.
The patch was reviewed by Steve Glendinning and Microchip Team.
Signed-off-by: Alexander Kochetkov <al.kochet@gmail.com>
Acked-by: Steve Glendinning <steve.glendinning@shawell.net>
Signed-off-by: David S. Miller <davem@davemloft.net>
The patch affect SMSC LAN generation 4 chips with integrated PHY (LAN9221).
It is possible that PHY could enter power-down mode (ENERGYON clear),
between ENERGYON bit check in smsc911x_phy_disable_energy_detect and SRST
bit set in smsc911x_soft_reset. This could happen, for example, if someone
disconnect ethernet cable between the checks. The PHY in a power-down mode
would prevent the MAC portion of chip to be software reseted.
Initially found by code review, confirmed later using test case.
This is low probability issue, and in order to reproduce it you have to
run the script:
while true; do
ifconfig eth0 down
ifconfig eth0 up || break
done
While the script is running you have to plug/unplug ethernet cable many
times (using gpio controlled ethernet switch, for example) until get:
[ 4516.477783] ADDRCONF(NETDEV_UP): eth0: link is not ready
[ 4516.512207] smsc911x smsc911x.0: eth0: SMSC911x/921x identified at 0xce006000, IRQ: 336
[ 4516.524658] ADDRCONF(NETDEV_UP): eth0: link is not ready
[ 4516.559082] smsc911x smsc911x.0: eth0: SMSC911x/921x identified at 0xce006000, IRQ: 336
[ 4516.571990] ADDRCONF(NETDEV_UP): eth0: link is not ready
ifconfig: SIOCSIFFLAGS: Input/output error
The patch was reviewed by Steve Glendinning and Microchip Team.
Signed-off-by: Alexander Kochetkov <al.kochet@gmail.com>
Acked-by: Steve Glendinning <steve.glendinning@shawell.net>
Signed-off-by: David S. Miller <davem@davemloft.net>
Refactored all macros used in cxgb4i as part of previously started cxgb4 macro
names cleanup. Makes them more uniform and avoids namespace collision.
Minor changes in other drivers where required as some of these macros are used
by multiple drivers, affected drivers are iw_cxgb4, cxgb4(vf) & csiostor
Signed-off-by: Anish Bhatt <anish@chelsio.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
This patch fixes two issues after using iovec iterators:
- vlan_offset should be initialized to zero, otherwise unexpected offset
will be used in skb_copy_datagram_iter()
- advance iovec iterator when vnet_hdr_sz is greater than sizeof(gso), this
is the case when mergeable rx buffer were enabled for a virt guest.
Fixes e0b46d0ee9 ("tun: Use iovec iterators")
Cc: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: Jason Wang <jasowang@redhat.com>
Acked-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: David S. Miller <davem@davemloft.net>
net/ipv4/fou.c: In function ‘ip_tunnel_encap_del_fou_ops’:
net/ipv4/fou.c:861:1: warning: no return statement in function returning non-void [-Wreturn-type]
Fixes: a8c5f90fb5 ("ip_tunnel: Ops registration for secondary encap (fou, gue)")
Signed-off-by: Thomas Graf <tgraf@suug.ch>
Signed-off-by: David S. Miller <davem@davemloft.net>
kick_requests() can put linger requests on the notarget list. This
means we need to clear the much-overloaded req->r_req_lru_item in
__unregister_linger_request() as well, or we get an assertion failure
in ceph_osdc_release_request() - !list_empty(&req->r_req_lru_item).
AFAICT the assumption was that registered linger requests cannot be on
any of req->r_req_lru_item lists, but that's clearly not the case.
Signed-off-by: Ilya Dryomov <idryomov@redhat.com>
Reviewed-by: Alex Elder <elder@linaro.org>
Requests have to be unlinked from both osd->o_requests (normal
requests) and osd->o_linger_requests (linger requests) lists when
clearing req->r_osd. Otherwise __unregister_linger_request() gets
confused and we trip over a !list_empty(&osd->o_linger_requests)
assert in __remove_osd().
MON=1 OSD=1:
# cat remove-osd.sh
#!/bin/bash
rbd create --size 1 test
DEV=$(rbd map test)
ceph osd out 0
sleep 3
rbd map dne/dne # obtain a new osdmap as a side effect
rbd unmap $DEV & # will block
sleep 3
ceph osd in 0
Signed-off-by: Ilya Dryomov <idryomov@redhat.com>
Reviewed-by: Alex Elder <elder@linaro.org>
