When CONFIG_GPIOLIB is unset, its stubs will return -ENOSYS. That means
when the mmc core parses DT for CD/WP GPIOs via mmc_of_parse(), -ENOSYS
becomes propagated to the caller. Typically this means that the mmc host
driver fails to probe.
As the CD/WP GPIOs are already treated as optional, let's extend that to
cover the case when CONFIG_GPIOLIB is unset.
Reported-by: Michal Simek <michal.simek@xilinx.com> Fixes: 16b23787fc70 ("mmc: sdhci-of-arasan: Call OF parsing for MMC") Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org> Tested-by: Michal Simek <michal.simek@xilinx.com> Acked-by: Venu Byravarasu <vbyravarasu@nvidia.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Currently one mrq->data maybe execute dma_map_sg() twice
when mmc subsystem prepare over one new request, and the
following log show up:
sdhci[sdhci_pre_dma_transfer] invalid cookie: 24, next-cookie 25
In this condition, mrq->date map a dma-memory(1) in sdhci_pre_req
for the first time, and map another dma-memory(2) in sdhci_prepare_data
for the second time. But driver only unmap the dma-memory(2), and
dma-memory(1) never unmapped, which cause the dma memory leak issue.
This patch use another method to map the dma memory for the mrq->data
which can fix this dma memory leak issue.
UBI: attaching mtd1 to ubi0
UBI: scanning is finished
UBI error: init_volumes: not enough PEBs, required 706, available 686
UBI error: ubi_wl_init: no enough physical eraseblocks (-20, need 1)
UBI error: ubi_attach_mtd_dev: failed to attach mtd1, error -12 <= NOT ENOMEM
UBI error: ubi_init: cannot attach mtd1
If available PEBs are not enough when initializing volumes, return -ENOSPC
directly. If available PEBs are not enough when initializing WL, return
-ENOSPC instead of -ENOMEM.
Signed-off-by: Sheng Yong <shengyong1@huawei.com> Signed-off-by: Richard Weinberger <richard@nod.at> Reviewed-by: David Gstir <david@sigma-star.at> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Make sure that data_size is less than LEB size.
Otherwise a handcrafted UBI image is able to trigger
an out of bounds memory access in ubi_compare_lebs().
Signed-off-by: Richard Weinberger <richard@nod.at> Reviewed-by: David Gstir <david@sigma-star.at> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
While the lockdep splat is a false positive, becuase path_openat holds i_mutex
of the parent directory and ubifs_init_security() tries to acquire i_mutex
of a new inode, it reveals that taking i_mutex in ubifs_init_security() is
in vain because it is only being called in the inode allocation path
and therefore nobody else can see the inode yet.
Reported-and-tested-by: Boris Brezillon <boris.brezillon@free-electrons.com> Reviewed-and-tested-by: Dongsheng Yang <yangds.fnst@cn.fujitsu.com> Signed-off-by: Richard Weinberger <richard@nod.at> Signed-off-by: dedekind1@gmail.com Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
When replacing del_timer() with del_timer_sync(), I introduced
a deadlock condition :
reqsk_queue_unlink() is called from inet_csk_reqsk_queue_drop()
inet_csk_reqsk_queue_drop() can be called from many contexts,
one being the timer handler itself (reqsk_timer_handler()).
In this case, del_timer_sync() loops forever.
Simple fix is to test if timer is pending.
Fixes: 2235f2ac75fd ("inet: fix races with reqsk timers") Signed-off-by: Eric Dumazet <edumazet@google.com> Signed-off-by: David S. Miller <davem@davemloft.net> Cc: Holger Hoffstätte <holger.hoffstaette@googlemail.com> Cc: Andre Tomt <andre@tomt.net> Cc: Chris Caputo <ccaputo@alt.net> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Commit 5d5cd85ff441 ("rsi: Fix failure to load firmware after memory
leak fix and fix the leak") also added a check on the allocation of
DMA-accessible memory that may directly return. In that case the
already allocated firmware data is leaked. Make sure the data is
always freed correctly. Detected by Coverity CID 1316519.
Fixes: 5d5cd85ff441 ("rsi: Fix failure to load firmware after memory leak fix and fix the leak") Signed-off-by: Christian Engelmayer <cengelma@gmx.at> Signed-off-by: Kalle Valo <kvalo@codeaurora.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
This fixes a race which can result in the same virtual IRQ number
being assigned to two different MSI interrupts. The most visible
consequence of that is usually a warning and stack trace from the
sysfs code about an attempt to create a duplicate entry in sysfs.
The race happens when one CPU (say CPU 0) is disposing of an MSI
while another CPU (say CPU 1) is setting up an MSI. CPU 0 calls
(for example) pnv_teardown_msi_irqs(), which calls
msi_bitmap_free_hwirqs() to indicate that the MSI (i.e. its
hardware IRQ number) is no longer in use. Then, before CPU 0 gets
to calling irq_dispose_mapping() to free up the virtal IRQ number,
CPU 1 comes in and calls msi_bitmap_alloc_hwirqs() to allocate an
MSI, and gets the same hardware IRQ number that CPU 0 just freed.
CPU 1 then calls irq_create_mapping() to get a virtual IRQ number,
which sees that there is currently a mapping for that hardware IRQ
number and returns the corresponding virtual IRQ number (which is
the same virtual IRQ number that CPU 0 was using). CPU 0 then
calls irq_dispose_mapping() and frees that virtual IRQ number.
Now, if another CPU comes along and calls irq_create_mapping(), it
is likely to get the virtual IRQ number that was just freed,
resulting in the same virtual IRQ number apparently being used for
two different hardware interrupts.
To fix this race, we just move the call to msi_bitmap_free_hwirqs()
to after the call to irq_dispose_mapping(). Since virq_to_hw()
doesn't work for the virtual IRQ number after irq_dispose_mapping()
has been called, we need to call it before irq_dispose_mapping() and
remember the result for the msi_bitmap_free_hwirqs() call.
The pattern of calling msi_bitmap_free_hwirqs() before
irq_dispose_mapping() appears in 5 places under arch/powerpc, and
appears to have originated in commit 05af7bd2d75e ("[POWERPC] MPIC
U3/U4 MSI backend") from 2007.
So far the mcast tvlv handler did not anticipate the processing of
multiple incoming OGMs from the same originator at the same time. This
can lead to various issues:
* Broken refcounting: For instance two mcast handlers might both assume
that an originator just got multicast capabilities and will together
wrongly decrease mcast.num_disabled by two, potentially leading to
an integer underflow.
* Potential kernel panic on hlist_del_rcu(): Two mcast handlers might
one after another try to do an
hlist_del_rcu(&orig->mcast_want_all_*_node). The second one will
cause memory corruption / crashes.
(Reported by: Sven Eckelmann <sven@narfation.org>)
Right in the beginning the code path makes assumptions about the current
multicast related state of an originator and bases all updates on that. The
easiest and least error prune way to fix the issues in this case is to
serialize multiple mcast handler invocations with a spinlock.
Fixes: 60432d756cf0 ("batman-adv: Announce new capability via multicast TVLV") Signed-off-by: Linus Lüssing <linus.luessing@c0d3.blue> Signed-off-by: Marek Lindner <mareklindner@neomailbox.ch> Signed-off-by: Antonio Quartulli <antonio@meshcoding.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Bitwise OR/AND assignments in C aren't guaranteed to be atomic. One
OGM handler might undo the set/clear of a specific bit from another
handler run in between.
Fix this by using the atomic set_bit()/clear_bit()/test_bit() functions.
Fixes: 60432d756cf0 ("batman-adv: Announce new capability via multicast TVLV") Signed-off-by: Linus Lüssing <linus.luessing@c0d3.blue> Signed-off-by: Marek Lindner <mareklindner@neomailbox.ch> Signed-off-by: Antonio Quartulli <antonio@meshcoding.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Bitwise OR/AND assignments in C aren't guaranteed to be atomic. One
OGM handler might undo the set/clear of a specific bit from another
handler run in between.
Fix this by using the atomic set_bit()/clear_bit()/test_bit() functions.
Bitwise OR/AND assignments in C aren't guaranteed to be atomic. One
OGM handler might undo the set/clear of a specific bit from another
handler run in between.
Fix this by using the atomic set_bit()/clear_bit()/test_bit() functions.
If there is a DMA zone (usually 24bit = 16MB I believe), but no DMA32
zone, as is the case for some 32-bit kernels, then massage_gfp_flags()
will cause DMA memory allocated for devices with a 32..63-bit
coherent_dma_mask to fall back to using __GFP_DMA, even though there may
only be 32-bits of physical address available anyway.
Correct that case to compare against a mask the size of phys_addr_t
instead of always using a 64-bit mask.
Tolerance applies on both sides of the target voltage, i.e. both min and
max sides. But while checking if a voltage is supported by the regulator
or not, we haven't taken care of tolerance on the lower side. Fix that.
Cc: Lucas Stach <l.stach@pengutronix.de> Fixes: 045ee45c4ff2 ("cpufreq: cpufreq-dt: disable unsupported OPPs") Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org> Reviewed-by: Lucas Stach <l.stach@pengutronix.de> Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
cache_shared_cpu_map_setup() did check sibling CPUs cacheinfo descriptor
while the respective teardown path cache_shared_cpu_map_remove() didn't.
Fix that.
>From tglx's version: to be on the safe side, move the cacheinfo
descriptor check to free_cache_attributes(), thus cleaning up the
hotplug path a little and making this even more robust.
Add a device quirk for the Logitech PTZ Pro Camera and its sibling the
ConferenceCam CC3000e Camera.
This fixes the failed camera enumeration on some boot, particularly on
machines with fast CPU.
Tested by connecting a Logitech PTZ Pro Camera to a machine with a
Haswell Core i7-4600U CPU @ 2.10GHz, and doing thousands of reboot cycles
while recording the kernel logs and taking camera picture after each boot.
Before the patch, more than 7% of the boots show some enumeration transfer
failures and in a few of them, the kernel is giving up before actually
enumerating the webcam. After the patch, the enumeration has been correct
on every reboot.
Signed-off-by: Vincent Palatin <vpalatin@chromium.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
since commit 33c300cb90a6 ("usb: musb: dsps:
don't fake of_node to musb core") we have been
preventing CPPI 4.1 from probing due to NULL
of_node. We can't revert said commit otherwise
a different regression would show up, so the fix
is to look for the parent device's (glue layer's)
of_node instead, since that's the thing which
is actually described in DTS.
Signed-off-by: Felipe Balbi <balbi@ti.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The gpio-desc migration done in v4.0 caused a regression
with legacy boots due to reversed reset logic.
e.g. omap3-beagle USB host breaks on legacy boot.
Request the reset GPIO with GPIOF_ACTIVE_LOW flag so that
it matches the driver logic and pin behaviour.
Fixes: e9f2cefb0cdc ("usb: phy: generic: migrate to gpio_desc") Tested-by: Fabio Estevam <fabio.estevam@freescale.com> Signed-off-by: Roger Quadros <rogerq@ti.com> Signed-off-by: Felipe Balbi <balbi@ti.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
According to spec, there are functional and protocol stalls.
For functional stall, it is for bulk and interrupt endpoints,
below are cases for it:
- Host sends SET_FEATURE request for Set-Halt, the udc driver
needs to set stall, and return true unconditionally.
- The gadget driver may call usb_ep_set_halt to stall certain
endpoints, if there is a transfer in pending, the udc driver
should not set stall, and return -EAGAIN accordingly.
These two kinds of stall need to be cleared by host using CLEAR_FEATURE
request (Clear-Halt).
For protocol stall, it is for control endpoint, this stall will
be set if the control request has failed. This stall will be
cleared by next setup request (hardware will do it).
It fixed usbtest (drivers/usb/misc/usbtest.c) Test 13 "set/clear halt"
test failure, meanwhile, this change has been verified by
USB2 CV Compliance Test and MSC Tests.
Cc: Alan Stern <stern@rowland.harvard.edu> Cc: Felipe Balbi <balbi@ti.com> Signed-off-by: Peter Chen <peter.chen@freescale.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Fix the regression caused by commit ad78c918602 ("usb: musb: dsps: just
start polling already") which causes polling the ID pin status even in
device-only mode.
Fixes: ad78c918602c ("usb: musb: dsps: just start polling already") Signed-off-by: Bin Liu <b-liu@ti.com> Signed-off-by: Felipe Balbi <balbi@ti.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The in kernel snprintf() will conveniently return the actual length of
the printed string even if not given an output beffer at all so just do
that rather than relying on the user to pass in a suitable buffer,
ensuring that we don't need to worry if the buffer was truncated due to
the size of the buffer passed in.
Reported-by: Rasmus Villemoes <linux@rasmusvillemoes.dk> Signed-off-by: Mark Brown <broonie@kernel.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
If a read is attempted which is smaller than the line length then we may
underflow the subtraction we're doing with the unsigned size_t type so
move some of the calculation to be additions on the right hand side
instead in order to avoid this.
Reported-by: Rasmus Villemoes <linux@rasmusvillemoes.dk> Signed-off-by: Mark Brown <broonie@kernel.org> Cc: stable@vger.kernel.org Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Remove unnecessary check that disabled SIS pipe commands for SIS-32
devices. This change was sufficient to enable raw mode and send SIS
pipe commands for a 57B3 device.
Fixes: f8ee25d7d239 ("ipr: AF DASD raw mode implementation in ipr driver") Signed-off-by: Gabriel Krisman Bertazi <krisman@linux.vnet.ibm.com> Reviewed-by: Wen Xiong <wenxiong@linux.vnet.ibm.com> Acked-by: Brian King <brking@linux.vnet.ibm.com> Signed-off-by: James Bottomley <JBottomley@Odin.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Fix the lack of clk_put() in sa11xx_base.c's error cleanup paths by
converting the driver to the devm_* API.
Fixes: 86d88bfca475 ("ARM: 8247/2: pcmcia: sa1100: make use of device clock") Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Even though there's a WMI enum for fragmentation
threshold no known firmware actually implements
it. Moreover it is not possible to rely frame
fragmentation to mac80211 because firmware clears
the "more fragments" bit in frame control making
it impossible for remote devices to reassemble
frames.
Hence implement a dummy callback just to say
fragmentation isn't supported. This effectively
prevents mac80211 from doing frame fragmentation
in software.
This fixes Tx becoming broken after setting
fragmentation threshold.
Fixes: 1010ba4c5d1c ("ath10k: unregister and remove frag_threshold callback") Signed-off-by: Michal Kazior <michal.kazior@tieto.com> Signed-off-by: Kalle Valo <kvalo@qca.qualcomm.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
In device_add_property_set() we check pset parameter for a NULL, but few lines
later we do a pointer arithmetic without check that will crash kernel in the
set_secondary_fwnode().
Here we check if pset parameter is NULL and return immediately.
Fixes: 16ba08d5c9ec (device property: Introduce firmware node type for platform data) Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com> Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The rockchip io-domain driver currently only depends on ARCH_ROCKCHIP
itself. This makes it possible to select the power-domain driver, but
not the POWER_AVS class and results in the iodomain-driver not getting
build in this case.
So add the additional dependency, which also results in the driver
config option now being placed nicely into the AVS submenu.
Fixes: 662a958638bd ("PM / AVS: rockchip-io: add driver handling Rockchip io domains") Signed-off-by: Heiko Stuebner <heiko@sntech.de> Acked-by: Kevin Hilman <khilman@linaro.org> Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The USER_DATA register cannot be accessed using byte accessors on A13
SoCs, thus triggering a bug when using memcpy_toio on this register.
Declare an helper macros to convert an OOB buffer into a suitable
USER_DATA value and vice-versa.
This patch also fixes an error in the oob_required logic (some OOB data
are not written even if the user required it) by removing the
oob_required condition, which is perfectly valid since the core already
fill ->oob_poi with FFs when oob_required is false.
When keeping the configuration set by the bootloader (by using
the marvell,nand-keep-config property), the pxa3xx_nand_detect_config()
function is called and set the chunk size to 512 as a default value if
NDCR_PAGE_SZ is not set.
In the other case, when not keeping the bootloader configuration, no
chunk size is set. Fix this by adding a default chunk size of 512.
When the ITS is configured for non-cacheable transactions, make sure
that the allocated, zeroed memory is flushed to the Point of
Coherency, allowing the ITS to observe the zeros instead of random
garbage (or even get its own data overwritten by zeros being evicted
from the cache...).
Fixes: 241a386c7dbb "irqchip: gicv3-its: Use non-cacheable accesses when no shareability" Reported-and-tested-by: Stuart Yoder <stuart.yoder@freescale.com> Signed-off-by: Marc Zyngier <marc.zyngier@arm.com> Cc: linux-arm-kernel@lists.infradead.org Cc: Pavel Fedin <p.fedin@samsung.com> Cc: Jason Cooper <jason@lakedaemon.net> Link: http://lkml.kernel.org/r/1442142873-20213-3-git-send-email-marc.zyngier@arm.com Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
When masking/unmasking interrupts, mask_cache is updated and used later
for suspend/resume. Unfortunately, it always was the mask_cache
associated with the first irq chip which was updated. So when performing
resume, only irqs 0-31 could be enabled.
Linux cifs mount with ntlmssp against an Mac OS X (Yosemite
10.10.5) share fails in case the clocks differ more than +/-2h:
digest-service: digest-request: od failed with 2 proto=ntlmv2
digest-service: digest-request: kdc failed with -1561745592 proto=ntlmv2
Fix this by (re-)using the given server timestamp for the
ntlmv2 authentication (as Windows 7 does).
A related problem was also reported earlier by Namjae Jaen (see below):
Windows machine has extended security feature which refuse to allow
authentication when there is time difference between server time and
client time when ntlmv2 negotiation is used. This problem is prevalent
in embedded enviornment where system time is set to default 1970.
Modern servers send the server timestamp in the TargetInfo Av_Pair
structure in the challenge message [see MS-NLMP 2.2.2.1]
In [MS-NLMP 3.1.5.1.2] it is explicitly mentioned that the client must
use the server provided timestamp if present OR current time if it is
not
Reported-by: Namjae Jeon <namjae.jeon@samsung.com> Signed-off-by: Peter Seiderer <ps.report@gmx.net> Signed-off-by: Steve French <smfrench@gmail.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
cd-gpios polarity should be changed to GPIO_ACTIVE_LOW and wp-gpios
should be changed to GPIO_ACTIVE_HIGH.
Otherwise, the SD may not work properly due to wrong polarity inversion
specified in DT after switch to common parsing function mmc_of_parse().
cd-gpios polarity should be changed to GPIO_ACTIVE_LOW and wp-gpios
should be changed to GPIO_ACTIVE_HIGH.
Otherwise, the SD may not work properly due to wrong polarity inversion
specified in DT after switch to common parsing function mmc_of_parse().
cd-gpios polarity should be changed to GPIO_ACTIVE_LOW and wp-gpios
should be changed to GPIO_ACTIVE_HIGH.
Otherwise, the SD may not work properly due to wrong polarity inversion
specified in DT after switch to common parsing function mmc_of_parse().
Current card detect probe process is that when driver finds a valid
ESDHC_CD_GPIO, it will clear the quirk SDHCI_QUIRK_BROKEN_CARD_DETECTION
which is set by default for all esdhc/usdhc controllers.
Then host driver will know there's a valid card detect function.
Commit 8d86e4fcccf6 ("mmc: sdhci-esdhc-imx: Call mmc_of_parse()")
breaks GPIO CD function for dt platform that it will return directly
when find ESDHC_CD_GPIO for dt platform which result in the later wrongly
to keep SDHCI_QUIRK_BROKEN_CARD_DETECTION for all dt platforms.
Then MMC_CAP_NEEDS_POLL will be used instead even there's a valid
GPIO card detect.
This patch adds back this function and follows the original approach to
clear the quirk if find an valid CD GPIO for dt platforms.
The only user of this driver that has not been converted to fully
device tree is the i.MX35 SoC.
There is a i.MX35-based board (mach-pcm043.c) that uses platform data
to pass wp_gpio and cd_gpio information.
Commit 8d86e4fcccf61ba ("mmc: sdhci-esdhc-imx: Call mmc_of_parse()")
broke the platform data case by removing mmc_gpio_request_ro() and
mmc_gpio_request_cd(), so restore the functionality for the non-dt
case.
Also, restore the check for ESDHC_CD_CONTROLLER so that we can still
support the "fsl,cd-controller" property.
As per DW MobileStorage databook "each descriptor can transfer up to 4kB
of data in chained mode", moreover buffer size that is put in "des1" is
limited to 13 bits, i.e. for example on attempt to
IDMAC_SET_BUFFER1_SIZE(desc, 8192) size value that's effectively written
will be 0.
On the platform with 8kB PAGE_SIZE I see dw_mmc gets data blocks in
SG-list of 8kB size and that leads to unpredictable behavior of the
SD/MMC controller.
In particular on write to FAT partition of SD-card the controller will
stuck in the middle of DMA transaction.
Solution to the problem is simple - we need to pass large (> 4kB) data
buffers to the controller via multiple descriptors. And that's what
that change does.
What's interesting I did try original driver on same platform but
configured with 4kB PAGE_SIZE and may confirm that data blocks passed
in SG-list to dw_mmc never exeed 4kB limit - that explains why nobody
ever faced a problem I did.
Bitwise OR/AND assignments in C aren't guaranteed to be atomic. One
OGM handler might undo the set/clear of a specific bit from another
handler run in between.
Fix this by using the atomic set_bit()/clear_bit()/test_bit() functions.
The tt_local_entry deletion performed in batadv_tt_local_remove() was neither
protecting against simultaneous deletes nor checking whether the element was
still part of the list before calling hlist_del_rcu().
Replacing the hlist_del_rcu() call with batadv_hash_remove() provides adequate
protection via hash spinlocks as well as an is-element-still-in-hash check to
avoid 'blind' hash removal.
Commit 11c32d7b6274cb0f554943d65bd4a126c4a86dcd
"video: move Versatile CLCD helpers" missed the fact
that the Integrator/CP is also using the helper, and
as a result the platform got only stubs and no graphics.
Add this as a default selection to Kconfig so we have
graphics again.
It is possible that we bind against a local socket in early_demux when we
are actually going to want to forward it. In this case, the socket serves
no purpose and only serves to confuse things (particularly functions which
implicitly expect sk_fullsock to be true, like ip_local_out).
Additionally, skb_set_owner_w is totally broken for non full-socks.
Signed-off-by: Alex Gartrell <agartrell@fb.com> Fixes: 41063e9dd119 ("ipv4: Early TCP socket demux.") Acked-by: Julian Anastasov <ja@ssi.bg> Signed-off-by: Simon Horman <horms@verge.net.au> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
I overlooked the svc->sched_data usage from schedulers
when the services were converted to RCU in 3.10. Now
the rare ipvsadm -E command can change the scheduler
but due to the reverse order of ip_vs_bind_scheduler
and ip_vs_unbind_scheduler we provide new sched_data
to the old scheduler resulting in a crash.
To fix it without changing the scheduler methods we
have to use synchronize_rcu() only for the editing case.
It means all svc->scheduler readers should expect a
NULL value. To avoid breakage for the service listing
and ipvsadm -R we can use the "none" name to indicate
that scheduler is not assigned, a state when we drop
new connections.
Reported-by: Alexander Vasiliev <a.vasylev@404-group.com> Fixes: ceec4c381681 ("ipvs: convert services to rcu") Signed-off-by: Julian Anastasov <ja@ssi.bg> Signed-off-by: Simon Horman <horms@verge.net.au> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Michael Vallaly reports about wrong source address used
in rare cases for tunneled traffic. Looks like
__ip_vs_get_out_rt in 3.10+ is providing uninitialized
dest_dst->dst_saddr.ip because ip_vs_dest_dst_alloc uses
kmalloc. While we retry after seeing EINVAL from routing
for data that does not look like valid local address, it
still succeeded when this memory was previously used from
other dests and with different local addresses. As result,
we can use valid local address that is not suitable for
our real server.
Fix it by providing 0.0.0.0 every time our cache is refreshed.
By this way we will get preferred source address from routing.
Reported-by: Michael Vallaly <lvs@nolatency.com> Fixes: 026ace060dfe ("ipvs: optimize dst usage for real server") Signed-off-by: Julian Anastasov <ja@ssi.bg> Signed-off-by: Simon Horman <horms@verge.net.au> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Dave Martin [Fri, 24 Jul 2015 16:39:21 +0000 (17:39 +0100)]
serial/amba-pl011: Disable interrupts around TX softirq
pl011_tx_softirq() currently uses spin_{,un}lock(), which are not
sufficient to inhibit pl011_int() from being triggered by a local
IRQ and trying to re-take the same lock. This can lead to
deadlocks.
This patch uses the _irq() locking variants instead to ensure that
pl011_int() handling for a given port is deferred until any
pl011_tx_softirq() work for that port is complete.
Notes for stable:
This patch fixes an issue that is fixed by the following upstream
commit, which is a more substantial rewrite of the affected code,
fixing multiple, mostly more minor issues:
The upstream patch was rejected for stable on the reasonable grounds
that it was too big and complex a patch. The original buggy code was
merged in v4.1, and the rewrite was merged in v4.2, leaving only v4.1
affected.
The optimized task selection logic optimistically selects a new task
to run without first doing a full put_prev_task(). This is so that we
can avoid a put/set on the common ancestors of the old and new task.
Similarly, we should only call check_cfs_rq_runtime() to throttle
eligible groups if they're part of the common ancestry, otherwise it
is possible to end up with no eligible task in the simple task
selection.
Imagine:
/root
/prev /next
/A /B
If our optimistic selection ends up throttling /next, we goto simple
and our put_prev_task() ends up throttling /prev, after which we're
going to bug out in set_next_entity() because there aren't any tasks
left.
Avoid this scenario by only throttling common ancestors.
Reported-by: Mohammed Naser <mnaser@vexxhost.com> Reported-by: Konstantin Khlebnikov <khlebnikov@yandex-team.ru> Signed-off-by: Ben Segall <bsegall@google.com>
[ munged Changelog ] Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: H. Peter Anvin <hpa@zytor.com> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Roman Gushchin <klamm@yandex-team.ru> Cc: Thomas Gleixner <tglx@linutronix.de> Cc: pjt@google.com Fixes: 678d5718d8d0 ("sched/fair: Optimize cgroup pick_next_task_fair()") Link: http://lkml.kernel.org/r/xm26wq1oswoq.fsf@sword-of-the-dawn.mtv.corp.google.com Signed-off-by: Ingo Molnar <mingo@kernel.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
As reported by Dmitry Vyukov, we really shouldn't do ipc_addid() before
having initialized the IPC object state. Yes, we initialize the IPC
object in a locked state, but with all the lockless RCU lookup work,
that IPC object lock no longer means that the state cannot be seen.
We already did this for the IPC semaphore code (see commit e8577d1f0329:
"ipc/sem.c: fully initialize sem_array before making it visible") but we
clearly forgot about msg and shm.
This commit checks for the URB_ZERO_PACKET flag and creates an extra
zero-length td if the urb transfer length is a multiple of the endpoint's
max packet length.
Don't check if timer is running with a timer_pending() before
deleting it with del_timer_sync(), this defies the whole point of
the sync part and can cause a possible race.
Instead we just want to make sure the timer is initialized early enough
before we have a chance to delete it.
Some changes between xhci 0.96 and xhci 1.0 specifications forced us to
check the hci version in code, some of these checks were implemented as
hci_version == 1.0, which will not work with new xhci 1.1 controllers.
xhci 1.1 behaves similar to xhci 1.0 in these cases, so change these
checks to hci_version >= 1.0
During quick plug/removal of OTG adapter during dual-role testing
it can happen that xhci_alloc_device() is called for the newly
detected device after the DRD library has called xhci_stop to
remove the HCD.
If that is the case, just fail early to prevent the following warning.
[ 154.732649] hub 4-0:1.0: USB hub found
[ 154.742204] hub 4-0:1.0: 1 port detected
[ 154.824458] hub 3-0:1.0: state 7 ports 1 chg 0002 evt 0000
[ 154.854609] hub 4-0:1.0: state 7 ports 1 chg 0000 evt 0000
[ 154.944430] usb 3-1: new high-speed USB device number 2 using xhci-hcd
[ 154.951009] xhci-hcd xhci-hcd.0.auto: xhci_setup_device
[ 155.038191] xhci-hcd xhci-hcd.0.auto: remove, state 4
[ 155.043315] usb usb4: USB disconnect, device number 1
[ 155.055270] xhci-hcd xhci-hcd.0.auto: xhci_stop
[ 155.060094] xhci-hcd xhci-hcd.0.auto: USB bus 4 deregistered
[ 155.066576] xhci-hcd xhci-hcd.0.auto: remove, state 1
[ 155.071710] usb usb3: USB disconnect, device number 1
[ 155.077124] xhci-hcd xhci-hcd.0.auto: xhci_setup_device
[ 155.082389] ------------[ cut here ]------------
[ 155.087690] WARNING: CPU: 0 PID: 72 at drivers/usb/host/xhci.c:3800 xhci_setup_device+0x410/0x484 [xhci_hcd]()
[ 155.097861] Modules linked in: sd_mod usb_storage scsi_mod usb_f_ss_lb g_zero libcomposite ipv6 xhci_plat_hcd xhci_hcd usbcore dwc3 udc_core evdev ti_am335x_adc joydev kfifo_buf industrialio snd_soc_simple_cc
[ 155.146734] CPU: 0 PID: 72 Comm: kworker/0:3 Tainted: G W 4.1.4-00834-gcd9380b-dirty #50
[ 155.156073] Hardware name: Generic AM43 (Flattened Device Tree)
[ 155.162117] Workqueue: usb_hub_wq hub_event [usbcore]
[ 155.167249] Backtrace:
[ 155.169751] [<c0012af0>] (dump_backtrace) from [<c0012c8c>] (show_stack+0x18/0x1c)
[ 155.177390] r6:c089d4a4 r5:ffffffff r4:00000000 r3:ee46c000
[ 155.183137] [<c0012c74>] (show_stack) from [<c05f7c14>] (dump_stack+0x84/0xd0)
[ 155.190446] [<c05f7b90>] (dump_stack) from [<c00439ac>] (warn_slowpath_common+0x80/0xbc)
[ 155.198605] r7:00000009 r6:00000ed8 r5:bf27eb70 r4:00000000
[ 155.204348] [<c004392c>] (warn_slowpath_common) from [<c0043a0c>] (warn_slowpath_null+0x24/0x2c)
[ 155.213202] r8:ee49f000 r7:ee7c0004 r6:00000000 r5:ee7c0158 r4:ee7c0000
[ 155.220051] [<c00439e8>] (warn_slowpath_null) from [<bf27eb70>] (xhci_setup_device+0x410/0x484 [xhci_hcd])
[ 155.229816] [<bf27e760>] (xhci_setup_device [xhci_hcd]) from [<bf27ec10>] (xhci_address_device+0x14/0x18 [xhci_hcd])
[ 155.240415] r10:ee598200 r9:00000001 r8:00000002 r7:00000001 r6:00000003 r5:00000002
[ 155.248363] r4:ee49f000
[ 155.250978] [<bf27ebfc>] (xhci_address_device [xhci_hcd]) from [<bf20cb94>] (hub_port_init+0x1b8/0xa9c [usbcore])
[ 155.261403] [<bf20c9dc>] (hub_port_init [usbcore]) from [<bf2101e0>] (hub_event+0x738/0x1020 [usbcore])
[ 155.270874] r10:ee598200 r9:ee7c0000 r8:ee7c0038 r7:ee518800 r6:ee49f000 r5:00000001
[ 155.278822] r4:00000000
[ 155.281426] [<bf20faa8>] (hub_event [usbcore]) from [<c005754c>] (process_one_work+0x128/0x340)
[ 155.290196] r10:00000000 r9:00000003 r8:00000000 r7:fedfa000 r6:eeec5400 r5:ee598314
[ 155.298151] r4:ee434380
[ 155.300718] [<c0057424>] (process_one_work) from [<c00578f8>] (worker_thread+0x158/0x49c)
[ 155.308963] r10:ee434380 r9:00000003 r8:eeec5400 r7:00000008 r6:ee434398 r5:eeec5400
[ 155.316913] r4:eeec5414
[ 155.319482] [<c00577a0>] (worker_thread) from [<c005cc40>] (kthread+0xdc/0xf8)
[ 155.326765] r10:00000000 r9:00000000 r8:00000000 r7:c00577a0 r6:ee434380 r5:ee4441c0
[ 155.334713] r4:00000000 r3:00000000
[ 155.338341] [<c005cb64>] (kthread) from [<c000fc08>] (ret_from_fork+0x14/0x2c)
[ 155.345626] r7:00000000 r6:00000000 r5:c005cb64 r4:ee4441c0
[ 155.356108] ---[ end trace a58d34c223b190e6 ]---
[ 155.360783] xhci-hcd xhci-hcd.0.auto: Virt dev invalid for slot_id 0x1!
[ 155.574404] xhci-hcd xhci-hcd.0.auto: xhci_setup_device
[ 155.579667] ------------[ cut here ]------------
Fix potential null-pointer dereference at probe by making sure that the
required endpoints are present.
The whiteheat driver assumes there are at least five pairs of bulk
endpoints, of which the final pair is used for the "command port". An
attempt to bind to an interface with fewer bulk endpoints would
currently lead to an oops.
we validate the mstb structs in the work function, and doing
that takes a reference. So we should never get here with the
work function running using the mstb device, only if the work
function hasn't run yet or is running for another mstb.
So we don't need to sync the work here, this was causing
lockdep spew as below.
[ +0.000160] =============================================
[ +0.000001] [ INFO: possible recursive locking detected ]
[ +0.000002] 3.10.0-320.el7.rhel72.stable.backport.3.x86_64.debug #1 Tainted: G W ------------
[ +0.000001] ---------------------------------------------
[ +0.000001] kworker/4:2/1262 is trying to acquire lock:
[ +0.000001] ((&mgr->work)){+.+.+.}, at: [<ffffffff810b29a5>] flush_work+0x5/0x2e0
[ +0.000007]
but task is already holding lock:
[ +0.000001] ((&mgr->work)){+.+.+.}, at: [<ffffffff810b57e4>] process_one_work+0x1b4/0x710
[ +0.000004]
other info that might help us debug this:
[ +0.000001] Possible unsafe locking scenario:
Instead of only enabling the backlight (which seems to set it to max
brightness), just re-set the current backlight level, which also takes
care of enabling the backlight if necessary.
Only the radeon_atom_encoder_dpms_dig part tested on a Kaveri laptop,
the radeon_atom_encoder_dpms_avivo part is only compile tested.
Signed-off-by: Michel Dänzer <michel.daenzer@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
I've done some extensive history digging across libdrm, mesa and
xf86-video-{intel,nouveau,ati}. The only potential user of this with
kms drivers I could find was ttmtest, which once used drmGetLock
still. But that mistake was quickly fixed up. Even the intel xvmc
library (which otherwise was really good with using dri1 stuff in kms
mode) managed to never take the hw lock for dri2 (and hence kms).
Hence it should be save to unconditionally disallow this.
Cc: Peter Antoine <peter.antoine@intel.com> Reviewed-by: Peter Antoine <peter.antoine@intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@intel.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The VBT MIPI Sequence Block version 3 has forward incompatible changes:
First, the block size in the header has been specified reserved, and the
actual size is a separate 32-bit value within the block. The current
find_section() function to will only look at the size in the block
header, and, depending on what's in that now reserved size field,
continue looking for other sections in the wrong place.
Fix this by taking the new block size field into account. This will
ensure that the lookups for other sections will work properly, as long
as the new 32-bit size does not go beyond the opregion VBT mailbox size.
Second, the contents of the block have been completely
changed. Gracefully refuse parsing the yet unknown data version.
Cc: Deepak M <m.deepak@intel.com> Reviewed-by: Deepak M <m.deepak@intel.com> Signed-off-by: Jani Nikula <jani.nikula@intel.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The error paths in set_file_size for cifs and smb3 are incorrect.
In the unlikely event that a server did not support set file info
of the file size, the code incorrectly falls back to trying SMBWriteX
(note that only the original core SMB Write, used for example by DOS,
can set the file size this way - this actually does not work for the more
recent SMBWriteX). The idea was since the old DOS SMB Write could set
the file size if you write zero bytes at that offset then use that if
server rejects the normal set file info call.
Fortunately the SMBWriteX will never be sent on the wire (except when
file size is zero) since the length and offset fields were reversed
in the two places in this function that call SMBWriteX causing
the fall back path to return an error. It is also important to never call
an SMB request from an SMB2/sMB3 session (which theoretically would
be possible, and can cause a brief session drop, although the client
recovers) so this should be fixed. In practice this path does not happen
with modern servers but the error fall back to SMBWriteX is clearly wrong.
Removing the calls to SMBWriteX in the error paths in cifs_set_file_size
Pointed out by PaX/grsecurity team
Signed-off-by: Steve French <steve.french@primarydata.com> Reported-by: PaX Team <pageexec@freemail.hu> CC: Emese Revfy <re.emese@gmail.com> CC: Brad Spengler <spender@grsecurity.net> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Kerberos, which is very important for security, was only enabled for
CIFS not SMB2/SMB3 mounts (e.g. vers=3.0)
Patch based on the information detailed in
http://thread.gmane.org/gmane.linux.kernel.cifs/10081/focus=10307
to enable Kerberized SMB2/SMB3
a) SMB2_negotiate: enable/use decode_negTokenInit in SMB2_negotiate
b) SMB2_sess_setup: handle Kerberos sectype and replicate Kerberos
SMB1 processing done in sess_auth_kerberos
Signed-off-by: Noel Power <noel.power@suse.com> Signed-off-by: Jim McDonough <jmcd@samba.org> Signed-off-by: Steve French <steve.french@primarydata.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
If all other conditions in nfs_can_extend_write() are met, and there
are no locks, then we should be able to assume close-to-open semantics
and the ability to extend our write to cover the whole page.
With this patch, the xfstests generic/074 test completes in 242s instead
of >1400s on my test rig.
Fixes: bd61e0a9c852 ("locks: convert posix locks to file_lock_context") Cc: Jeff Layton <jlayton@primarydata.com> Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
We really want sizeof(struct page *) instead. Otherwise we limit
maximum IO size to 64 pages rather than 512 pages on a 64bit system.
Fixes 2e11f829(nfs: cap request size to fit a kmalloced page array).
Cc: Christoph Hellwig <hch@lst.de> Signed-off-by: Peng Tao <tao.peng@primarydata.com> Fixes: 2e11f8296d22 ("nfs: cap request size to fit a kmalloced page array") Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
There are several actions that smp_conn_security() might make that do
not require a valid SMP context (conn->smp pointer). One of these
actions is to encrypt the link with an existing LTK. If the SMP
context wasn't initialized properly we should still allow the
independent actions to be done, i.e. the check for the context should
only be done at the last possible moment.
like nf_log_unset, nf_log_unregister must not reset the list of loggers.
Otherwise, a call to nf_log_unregister() will render loggers of other nf
protocols unusable:
iptables -A INPUT -j LOG
modprobe nf_log_arp ; rmmod nf_log_arp
iptables -A INPUT -j LOG
iptables: No chain/target/match by that name
Fixes: 30e0c6a6be ("netfilter: nf_log: prepare net namespace support for loggers") Signed-off-by: Florian Westphal <fw@strlen.de> Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Fix lookup of existing match/target structures in the corresponding list
by skipping the family check if NFPROTO_UNSPEC is used.
This is resulting in the allocation and insertion of one match/target
structure for each use of them. So this not only bloats memory
consumption but also severely affects the time to reload the ruleset
from the iptables-compat utility.
After this patch, iptables-compat-restore and iptables-compat take
almost the same time to reload large rulesets.
Fixes: 0ca743a55991 ("netfilter: nf_tables: add compatibility layer for x_tables") Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The nf_log_unregister() function needs to call synchronize_rcu() to make sure
that the objects are not dereferenced anymore on module removal.
Fixes: 5962815a6a56 ("netfilter: nf_log: use an array of loggers instead of list") Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Currenlty nf_tables chains added in one network namespace are being
run in all network namespace. The issues are myriad with the simplest
being an unprivileged user can cause any network packets to be dropped.
Address this by simply not running nf_tables chains in the wrong
network namespace.
Signed-off-by: "Eric W. Biederman" <ebiederm@xmission.com> Acked-by: Pablo Neira Ayuso <pablo@netfilter.org> Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Add code to nf_unregister_hook to flush the nf_queue when a hook is
unregistered. This guarantees that the pointer that the nf_queue code
retains into the nf_hook list will remain valid while a packet is
queued.
I tested what would happen if we do not flush queued packets and was
trivially able to obtain the oops below. All that was required was
to stop the nf_queue listening process, to delete all of the nf_tables,
and to awaken the nf_queue listening process.
Signed-off-by: "Eric W. Biederman" <ebiederm@xmission.com> Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
When zones were originally introduced, the expectation functions were
all extended to perform lookup using the zone. However, insertion was
not modified to check the zone. This means that two expectations which
are intended to apply for different connections that have the same tuple
but exist in different zones cannot both be tracked.
Fixes: 5d0aa2ccd4 (netfilter: nf_conntrack: add support for "conntrack zones") Signed-off-by: Joe Stringer <joestringer@nicira.com> Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The convention in nfnetlink is to use network byte order in every header field
as well as in the attribute payload. The initial version of the batching
infrastructure assumes that res_id comes in host byte order though.
The only client of the batching infrastructure is nf_tables, so let's add a
workaround to address this inconsistency. We currently have 11 nfnetlink
subsystems according to NFNL_SUBSYS_COUNT, so we can assume that the subsystem
2560, ie. htons(10), will not be allocated anytime soon, so it can be an alias
of nf_tables from the nfnetlink batching path when interpreting the res_id
field.
Commit 3a0f9aaee028 ("dm raid: round region_size to power of two")
intended to make sure that the default region size is a power of two.
However, the logic in that commit is incorrect and sets the variable
region_size to 0 or 1, depending on whether min_region_size is a power
of two.
Fix this logic, using roundup_pow_of_two(), so that region_size is
properly rounded up to the next power of two.
Signed-off-by: Mikulas Patocka <mpatocka@redhat.com> Fixes: 3a0f9aaee028 ("dm raid: round region_size to power of two") Signed-off-by: Mike Snitzer <snitzer@redhat.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Calling e.g. blk_queue_max_hw_sectors() after calls to
disk_stack_limits() discards the settings determined by
disk_stack_limits().
So we need to make those calls first.
When a (e.g.) RAID5 array is reshaped to RAID0, the updating
of queue parameters (e.g. max number of sectors per bio) is
done in the wrong place.
It should be part of ->run, but it is actually part of ->takeover.
This means it happens before level_store() calls:
blk_set_stacking_limits(&mddev->queue->limits);
and so it ineffective. This can lead to errors from underlying
devices.
So move all the relevant settings out of create_stripe_zones()
and into raid0_run().
As this can lead to a bug-on it is suitable for any -stable
kernel which supports reshape to RAID0. So 2.6.35 or later.
As the bug has been present for five years there is no urgency,
so no need to rush into -stable.
Fixes: 9af204cf720c ("md: Add support for Raid5->Raid0 and Raid10->Raid0 takeover") Reported-by: Yi Zhang <yizhan@redhat.com> Signed-off-by: NeilBrown <neilb@suse.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Fixes: 83271f626 ("ion: hold reference to handle...") Signed-off-by: Shawn Lin <shawn.lin@rock-chips.com> Reviewed-by: Laura Abbott <labbott@redhat.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
When using nested btrees, the top leaves of the top levels contain
block addresses for the root of the next tree down. If we shadow a
shared leaf node the leaf values (sub tree roots) should be incremented
accordingly.
This is only an issue if there is metadata sharing in the top levels.
Which only occurs if metadata snapshots are being used (as is possible
with dm-thinp). And could result in a block from the thinp metadata
snap being reused early, thus corrupting the thinp metadata snap.
Signed-off-by: Joe Thornber <ejt@redhat.com> Signed-off-by: Mike Snitzer <snitzer@redhat.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The Linux NFS server returns garbage in the data payload of inline
NFS/RDMA READ replies. These are READs of under 1000 bytes or so
where the client has not provided either a reply chunk or a write
list.
The NFS server delivers the data payload for an NFS READ reply to
the transport in an xdr_buf page list. If the NFS client did not
provide a reply chunk or a write list, send_reply() is supposed to
set up a separate sge for the page containing the READ data, and
another sge for XDR padding if needed, then post all of the sges via
a single SEND Work Request.
The problem is send_reply() does not advance through the xdr_buf
when setting up scatter/gather entries for SEND WR. It always calls
dma_map_xdr with xdr_off set to zero. When there's more than one
sge, dma_map_xdr() sets up the SEND sge's so they all point to the
xdr_buf's head.
The current Linux NFS/RDMA client always provides a reply chunk or
a write list when performing an NFS READ over RDMA. Therefore, it
does not exercise this particular case. The Linux server has never
had to use more than one extra sge for building RPC/RDMA replies
with a Linux client.
However, an NFS/RDMA client _is_ allowed to send small NFS READs
without setting up a write list or reply chunk. The NFS READ reply
fits entirely within the inline reply buffer in this case. This is
perhaps a more efficient way of performing NFS READs that the Linux
NFS/RDMA client may some day adopt.
Fixes: b432e6b3d9c1 ('svcrdma: Change DMA mapping logic to . . .') BugLink: https://bugzilla.linux-nfs.org/show_bug.cgi?id=285 Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: J. Bruce Fields <bfields@redhat.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Setting the dm-crypt device's max_segment_size to PAGE_SIZE is an
unfortunate constraint that is required to avoid the potential for
exceeding dm-crypt's underlying device's max_segments limits -- due to
crypt_alloc_buffer() possibly allocating pages for the encryption bio
that are not as physically contiguous as the original bio.
It is interesting to note that this problem was already fixed back in
2007 via commit 91e106259 ("dm crypt: use bio_add_page"). But Linux 4.0
commit cf2f1abfb ("dm crypt: don't allocate pages for a partial
request") regressed dm-crypt back to _not_ using bio_add_page(). But
given dm-crypt's cpu parallelization changes all depend on commit cf2f1abfb's abandoning of the more complex io fragments processing that
dm-crypt previously had we cannot easily go back to using
bio_add_page().
So all said the cleanest way to resolve this issue is to fix dm-crypt to
properly constrain the original bios entering dm-crypt so the encryption
bios that dm-crypt generates from the original bios are always
compatible with the underlying device's max_segments queue limits.
It should be noted that technically Linux 4.3 does _not_ need this fix
because of the block core's new late bio-splitting capability. But, it
is reasoned, there is little to be gained by having the block core split
the encrypted bio that is composed of PAGE_SIZE segments. That said, in
the future we may revert this change.
Fixes: cf2f1abfb ("dm crypt: don't allocate pages for a partial request") Fixes: https://bugzilla.kernel.org/show_bug.cgi?id=104421 Suggested-by: Jeff Moyer <jmoyer@redhat.com> Signed-off-by: Mike Snitzer <snitzer@redhat.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
c770cb4cb505 ("PCI: Mark invalid BARs as unassigned") sets IORESOURCE_UNSET
if we fail to claim a resource. If we tried to claim a bridge window,
failed, clipped the window, and tried to claim the clipped window, we
failed again because of IORESOURCE_UNSET:
pci_bus 0000:00: root bus resource [mem 0xc0000000-0xffffffff window]
pci 0000:00:01.0: can't claim BAR 15 [mem 0xbdf00000-0xddefffff 64bit pref]: no compatible bridge window
pci 0000:00:01.0: [mem size 0x20000000 64bit pref] clipped to [mem size 0x1df00000 64bit pref]
pci 0000:00:01.0: bridge window [mem size 0x1df00000 64bit pref]
pci 0000:00:01.0: can't claim BAR 15 [mem size 0x1df00000 64bit pref]: no address assigned
The 00:01.0 window started as [mem 0xbdf00000-0xddefffff 64bit pref]. That
starts before the host bridge window [mem 0xc0000000-0xffffffff window], so
we clipped the 00:01.0 window to [mem 0xc0000000-0xddefffff 64bit pref].
But we left it marked IORESOURCE_UNSET, so the second claim failed when it
should have succeeded.
This means downstream devices will also fail for lack of resources, e.g.,
in the bugzilla below,
radeon 0000:01:00.0: Fatal error during GPU init
Clear IORESOURCE_UNSET when we clip a bridge window. Also clear
IORESOURCE_UNSET in our copy of the unclipped window so we can see exactly
what the original window was and how it now fits inside the upstream
window.
Fixes: c770cb4cb505 ("PCI: Mark invalid BARs as unassigned") Link: https://bugzilla.kernel.org/show_bug.cgi?id=85491#c47 Based-on-patch-by: Lorenzo Pieralisi <lorenzo.pieralisi@arm.com> Based-on-patch-by: Yinghai Lu <yinghai@kernel.org> Tested-by: Lorenzo Pieralisi <lorenzo.pieralisi@arm.com> Signed-off-by: Bjorn Helgaas <bhelgaas@google.com> Reviewed-by: Lorenzo Pieralisi <lorenzo.pieralisi@arm.com> Acked-by: Yinghai Lu <yinghai@kernel.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
projects / users / jedix / linux-maple.git / log