The reason is that the length of the new brk is not page aligned when we
try to populate the it. There is no reason to bug on that though.
do_brk_flags already aligns the length properly so the mapping is
expanded as it should. All we need is to tell mm_populate about it.
Besides that there is absolutely no reason to to bug_on in the first
place. The worst thing that could happen is that the last page wouldn't
get populated and that is far from putting system into an inconsistent
state.
Fix the issue by moving the length sanitization code from do_brk_flags
up to vm_brk_flags. The only other caller of do_brk_flags is brk
syscall entry and it makes sure to provide the proper length so t here
is no need for sanitation and so we can use do_brk_flags without it.
Michal Hocko [Fri, 13 Jul 2018 23:59:16 +0000 (16:59 -0700)]
mm/memblock.c: do not complain about top-down allocations for !MEMORY_HOTREMOVE
Mike Rapoport is converting architectures from bootmem to nobootmem
allocator. While doing so for m68k Geert has noticed that he gets a
scary looking warning:
The warning is basically saying that a top-down allocation can break
memory hotremove because memblock allocation is not movable. But m68k
doesn't even support MEMORY_HOTREMOVE so there is no point to warn about
it.
Make the warning conditional only to configurations that care.
Oscar Salvador [Fri, 13 Jul 2018 23:59:13 +0000 (16:59 -0700)]
fs, elf: make sure to page align bss in load_elf_library
The current code does not make sure to page align bss before calling
vm_brk(), and this can lead to a VM_BUG_ON() in __mm_populate() due to
the requested lenght not being correctly aligned.
Let us make sure to align it properly.
Kees: only applicable to CONFIG_USELIB kernels: 32-bit and configured
for libc5.
Philipp Rudo [Fri, 13 Jul 2018 23:59:09 +0000 (16:59 -0700)]
x86/purgatory: add missing FORCE to Makefile target
- Build the kernel without the fix
- Add some flag to the purgatories KBUILD_CFLAGS,I used
-fno-asynchronous-unwind-tables
- Re-build the kernel
When you look at makes output you see that sha256.o is not re-build in the
last step. Also readelf -S still shows the .eh_frame section for
sha256.o.
With the fix sha256.o is rebuilt in the last step.
Without FORCE make does not detect changes only made to the command line
options. So object files might not be re-built even when they should be.
Fix this by adding FORCE where it is missing.
net/9p/client.c: put refcount of trans_mod in error case in parse_opts()
In my testing, the second mount will fail after umounting successfully.
The reason is that we put refcount of trans_mod in the correct case
rather than the error case in parse_opts() at last. That will cause the
refcount decrease to -1, and when we try to get trans_mod again in
try_module_get(), we could only increase refcount to 0 which will cause
failure as follows:
parse_opts
v9fs_get_trans_by_name
try_module_get : return NULL to caller which cause error
So we should put refcount of trans_mod in error case.
Nicholas Piggin [Fri, 13 Jul 2018 23:59:03 +0000 (16:59 -0700)]
mm: allow arch to supply p??_free_tlb functions
The mmu_gather APIs keep track of the invalidated address range
including the span covered by invalidated page table pages. Ranges
covered by page tables but not ptes (and therefore no TLBs) still need
to be invalidated because some architectures (x86) can cache
intermediate page table entries, and invalidate those with normal TLB
invalidation instructions to be almost-backward-compatible.
Architectures which don't cache intermediate page table entries, or
which invalidate these caches separately from TLB invalidation, do not
require TLB invalidation range expanded over page tables.
Allow architectures to supply their own p??_free_tlb functions, which
can avoid the __tlb_adjust_range.
Tomas Bortoli [Fri, 13 Jul 2018 23:58:59 +0000 (16:58 -0700)]
autofs: fix slab out of bounds read in getname_kernel()
The autofs subsystem does not check that the "path" parameter is present
for all cases where it is required when it is passed in via the "param"
struct.
In particular it isn't checked for the AUTOFS_DEV_IOCTL_OPENMOUNT_CMD
ioctl command.
To solve it, modify validate_dev_ioctl(function to check that a path has
been provided for ioctl commands that require it.
fs/proc/task_mmu.c: fix Locked field in /proc/pid/smaps*
Thomas reports:
"While looking around in /proc on my v4.14.52 system I noticed that all
processes got a lot of "Locked" memory in /proc/*/smaps. A lot more
memory than a regular user can usually lock with mlock().
Commit 493b0e9d945f (in v4.14-rc1) seems to have changed the behavior
of "Locked".
Before that commit the code was like this. Notice the VM_LOCKED check.
mm: do not drop unused pages when userfaultd is running
KVM guests on s390 can notify the host of unused pages. This can result
in pte_unused callbacks to be true for KVM guest memory.
If a page is unused (checked with pte_unused) we might drop this page
instead of paging it. This can have side-effects on userfaultd, when
the page in question was already migrated:
The next access of that page will trigger a fault and a user fault
instead of faulting in a new and empty zero page. As QEMU does not
expect a userfault on an already migrated page this migration will fail.
The most straightforward solution is to ignore the pte_unused hint if a
userfault context is active for this VMA.
Linus, also saw this bug on his machine, and confirmed that reverting
commit 124049decbb1 ("x86/e820: put !E820_TYPE_RAM regions into
memblock.reserved") fixes the issue.
The problem is that we incorrectly zero some struct pages after they
were setup.
The fix is to zero unavailable struct pages prior to initializing of
struct pages.
A more detailed fix should come later that would avoid double zeroing
cases: one in __init_single_page(), the other one in
zero_resv_unavail().
Fixes: 124049decbb1 ("x86/e820: put !E820_TYPE_RAM regions into memblock.reserved") Signed-off-by: Pavel Tatashin <[email protected]> Signed-off-by: Linus Torvalds <[email protected]>
Simon Ser [Mon, 9 Jul 2018 16:17:22 +0000 (11:17 -0500)]
objtool: Use '.strtab' if '.shstrtab' doesn't exist, to support ORC tables on Clang
Clang puts its section header names in the '.strtab' section instead of
'.shstrtab', which causes objtool to fail with a "can't find
.shstrtab section" warning when attempting to write ORC metadata to an
object file.
If '.shstrtab' doesn't exist, use '.strtab' instead.
platform_get_resource() may fail and return NULL, so we should
better check it's return value to avoid a NULL pointer dereference
a bit later in the code.
This is detected by Coccinelle semantic patch.
@@
expression pdev, res, n, t, e, e1, e2;
@@
res = platform_get_resource(pdev, t, n);
+ if (!res)
+ return -EINVAL;
... when != res == NULL
e = devm_ioremap_nocache(e1, res->start, e2);
Fixes: cc4fa83f66e9 ("pinctrl: nsp: add pinmux driver support for Broadcom NSP SoC") Signed-off-by: Wei Yongjun <[email protected]> Reviewed-by: Ray Jui <[email protected]> Signed-off-by: Linus Walleij <[email protected]>
pinctrl: sh-pfc: r8a77970: remove SH_PFC_PIN_CFG_DRIVE_STRENGTH flag
The datasheet does not document any registers to control drive strength,
and no drive strength registers are for this reason described for this
SoC. The flags indicating that drive strength can be controlled are
however set for some pins in the driver.
This leads to a NULL pointer dereference when the sh-pfc core tries to
access the struct describing the drive strength registers, for example
when reading the sysfs file pinconf-pins.
Fix this by removing the SH_PFC_PIN_CFG_DRIVE_STRENGTH from all pins.
A kernel panic happens at dereferencing a NULL pointer: In this case, the
drvdata is still not setup properly yet when it is being accessed.
A better solution for fixing up this issue should be we should obtain the
private data from struct gpio_chip using a specific gpiochip_get_data
instead of a generic dev_get_drvdata.
Sean Wang [Fri, 22 Jun 2018 03:49:07 +0000 (11:49 +0800)]
pinctrl: mt7622: stop using the deprecated pinctrl_add_gpio_range
If the pinctrl node has the gpio-ranges property, the range will be added
by the gpio core and doesn't need to be added by the pinctrl driver.
But for keeping backward compatibility, an explicit pinctrl_add_gpio_range
is still needed to be called when there is a missing gpio-ranges in pinctrl
node in old dts files.
Sean Wang [Fri, 22 Jun 2018 03:49:06 +0000 (11:49 +0800)]
pinctrl: mt7622: fix that pinctrl_claim_hogs cannot work
To allow claiming hogs by pinctrl, we cannot enable pinctrl until all
groups and functions are being added done. Also, it's necessary that
the corresponding gpiochip is being added when the pinctrl device is
enabled.
Sean Wang [Fri, 22 Jun 2018 03:49:05 +0000 (11:49 +0800)]
pinctrl: mt7622: fix initialization sequence between eint and gpiochip
Because gpichip applied in the driver must depend on mtk eint to implement
the input data debouncing and the translation between gpio and irq, it's
better to keep logic consistent with mtk eint being built prior to gpiochip
being added.
David S. Miller [Sat, 14 Jul 2018 01:30:19 +0000 (18:30 -0700)]
Merge branch 'fix-DCTCP-delayed-ACK'
Yuchung Cheng says:
====================
fix DCTCP delayed ACK
This patch series addresses the issue that sometimes DCTCP
fail to acknowledge the latest sequence and result in sender timeout
if inflight is small.
====================
Previously, when a data segment was sent an ACK was piggybacked
on the data segment without generating a CA_EVENT_NON_DELAYED_ACK
event to notify congestion control modules. So the DCTCP
ca->delayed_ack_reserved flag could incorrectly stay set when
in fact there were no delayed ACKs being reserved. This could result
in sending a special ECN notification ACK that carries an older
ACK sequence, when in fact there was no need for such an ACK.
DCTCP keeps track of the delayed ACK status with its own separate
state ca->delayed_ack_reserved. Previously it may accidentally cancel
the delayed ACK without updating this field upon sending a special
ACK that carries a older ACK sequence. This inconsistency would
lead to DCTCP receiver never acknowledging the latest data until the
sender times out and retry in some cases.
Dan Carpenter [Thu, 12 Jul 2018 12:23:45 +0000 (15:23 +0300)]
qlogic: check kstrtoul() for errors
We accidentally left out the error handling for kstrtoul().
Fixes: a520030e326a ("qlcnic: Implement flash sysfs callback for 83xx adapter") Signed-off-by: Dan Carpenter <[email protected]> Signed-off-by: David S. Miller <[email protected]>
Merge branch 'i2c/for-current' of git://git.kernel.org/pub/scm/linux/kernel/git/wsa/linux
Pull i2c fixes from Wolfram Sang:
- I2C core bugfix regarding bus recovery
- driver bugfix for the tegra driver
- typo correction
* 'i2c/for-current' of git://git.kernel.org/pub/scm/linux/kernel/git/wsa/linux:
i2c: recovery: if possible send STOP with recovery pulses
i2c: tegra: Fix NACK error handling
i2c: stu300: use non-archaic spelling of failes
The following pull-request contains BPF updates for your *net* tree.
The main changes are:
1) Fix AF_XDP TX error reporting before final kernel release such that it
becomes consistent between copy mode and zero-copy, from Magnus.
2) Fix three different syzkaller reported issues: oob due to ld_abs
rewrite with too large offset, another oob in l3 based skb test run
and a bug leaving mangled prog in subprog JITing error path, from Daniel.
3) Fix BTF handling for bitfield extraction on big endian, from Okash.
4) Fix a missing linux/errno.h include in cgroup/BPF found by kbuild bot,
from Roman.
5) Fix xdp2skb_meta.sh sample by using just command names instead of
absolute paths for tc and ip and allow them to be redefined, from Taeung.
6) Fix availability probing for BPF seg6 helpers before final kernel ships
so they can be detected at prog load time, from Mathieu.
====================
skbuff: Unconditionally copy pfmemalloc in __skb_clone()
Commit 8b7008620b84 ("net: Don't copy pfmemalloc flag in
__copy_skb_header()") introduced a different handling for the
pfmemalloc flag in copy and clone paths.
In __skb_clone(), now, the flag is set only if it was set in the
original skb, but not cleared if it wasn't. This is wrong and
might lead to socket buffers being flagged with pfmemalloc even
if the skb data wasn't allocated from pfmemalloc reserves. Copy
the flag instead of ORing it.
Reported-by: Sabrina Dubroca <[email protected]> Fixes: 8b7008620b84 ("net: Don't copy pfmemalloc flag in __copy_skb_header()") Signed-off-by: Stefano Brivio <[email protected]> Tested-by: Sabrina Dubroca <[email protected]> Signed-off-by: David S. Miller <[email protected]>
Merge branch 'timers-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip
Pull timer fixes from Ingo Molnar:
"A clocksource driver fix and a revert"
* 'timers-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
clocksource: arm_arch_timer: Set arch_mem_timer cpumask to cpu_possible_mask
Revert "tick: Prefer a lower rating device only if it's CPU local device"
Merge branch 'perf-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip
Pull perf tool fixes from Ingo Molnar:
"Misc tooling fixes: python3 related fixes, gcc8 fix, bashism fixes and
some other smaller fixes"
* 'perf-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
perf tools: Use python-config --includes rather than --cflags
perf script python: Fix dict reference counting
perf stat: Fix --interval_clear option
perf tools: Fix compilation errors on gcc8
perf test shell: Prevent temporary editor files from being considered test scripts
perf llvm-utils: Remove bashism from kernel include fetch script
perf test shell: Make perf's inet_pton test more portable
perf test shell: Replace '|&' with '2>&1 |' to work with more shells
perf scripts python: Add Python 3 support to EventClass.py
perf scripts python: Add Python 3 support to sched-migration.py
perf scripts python: Add Python 3 support to Util.py
perf scripts python: Add Python 3 support to SchedGui.py
perf scripts python: Add Python 3 support to Core.py
perf tools: Generate a Python script compatible with Python 2 and 3
Merge branch 'core-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip
Pull rseq fixes from Ingo Molnar:
"Various rseq ABI fixes and cleanups: use get_user()/put_user(),
validate parameters and use proper uapi types, etc"
* 'core-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
rseq/selftests: cleanup: Update comment above rseq_prepare_unload
rseq: Remove unused types_32_64.h uapi header
rseq: uapi: Declare rseq_cs field as union, update includes
rseq: uapi: Update uapi comments
rseq: Use get_user/put_user rather than __get_user/__put_user
rseq: Use __u64 for rseq_cs fields, validate user inputs
Merge tag 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/rdma/rdma
Pull rdma fixes from Jason Gunthorpe:
"Things have been quite slow, only 6 RC patches have been sent to the
list. Regression, user visible bugs, and crashing fixes:
- cxgb4 could wrongly fail MR creation due to a typo
- various crashes if the wrong QP type is mixed in with APIs that
expect other types
- syzkaller oops
- using ERR_PTR and NULL together cases HFI1 to crash in some cases
- mlx5 memory leak in error unwind"
* tag 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/rdma/rdma:
RDMA/mlx5: Fix memory leak in mlx5_ib_create_srq() error path
RDMA/uverbs: Don't fail in creation of multiple flows
IB/hfi1: Fix incorrect mixing of ERR_PTR and NULL return values
RDMA/uverbs: Fix slab-out-of-bounds in ib_uverbs_ex_create_flow
RDMA/uverbs: Protect from attempts to create flows on unsupported QP
iw_cxgb4: correctly enforce the max reg_mr depth
Merge tag 'kbuild-fixes-v4.18-2' of git://git.kernel.org/pub/scm/linux/kernel/git/masahiroy/linux-kbuild
Pull Kbuild fixes from Masahiro Yamada:
- update Kbuild and Kconfig documents
- sanitize -I compiler option handling
- update extract-vmlinux script to recognize LZ4 and ZSTD
- fix tools Makefiles
- update tags.sh to handle __ro_after_init
- suppress warnings in case getconf does not recognize LFS_* parameters
* tag 'kbuild-fixes-v4.18-2' of git://git.kernel.org/pub/scm/linux/kernel/git/masahiroy/linux-kbuild:
kbuild: suppress warnings from 'getconf LFS_*'
scripts/tags.sh: add __ro_after_init
tools: build: Use HOSTLDFLAGS with fixdep
tools: build: Fixup host c flags
tools build: fix # escaping in .cmd files for future Make
scripts: teach extract-vmlinux about LZ4 and ZSTD
kbuild: remove duplicated comments about PHONY
kbuild: .PHONY is not a variable, but PHONY is
kbuild: do not drop -I without parameter
kbuild: document the KBUILD_KCONFIG env. variable
kconfig: update user kconfig tools doc.
kbuild: delete INSTALL_FW_PATH from kbuild documentation
kbuild: update ARCH alias info for sparc
kbuild: update ARCH alias info for sh
Merge tag 'arm64-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/arm64/linux
Pull arm64 fixes from Will Deacon:
"Catalin's out enjoying the sunshine, so I'm sending the fixes for a
couple of weeks (although there hopefully won't be any more!).
We've got a revert of a previous fix because it broke the build with
some distro toolchains and a preemption fix when detemining whether or
not the SIMD unit is in use.
Summary:
- Revert back to the 'linux' target for LD, as 'elf' breaks some
distributions
- Fix preemption race when testing whether the vector unit is in use
or not"
* tag 'arm64-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/arm64/linux:
arm64: neon: Fix function may_use_simd() return error status
Revert "arm64: Use aarch64elf and aarch64elfb emulation mode variants"
Merge branch 'fixes' of git://git.armlinux.org.uk/~rmk/linux-arm
Pull ARM fixes from Russell King:
"A couple of small fixes this time around from Steven for an
interaction between ftrace and kernel read-only protection, and
Vladimir for nommu"
* 'fixes' of git://git.armlinux.org.uk/~rmk/linux-arm:
ARM: 8780/1: ftrace: Only set kernel memory back to read-only after boot
ARM: 8775/1: NOMMU: Use instr_sync instead of plain isb in common code
Merge tag 'trace-v4.18-rc3-3' of git://git.kernel.org/pub/scm/linux/kernel/git/rostedt/linux-trace
Pull tracing fixlet from Steven Rostedt:
"Joel Fernandes asked to add a feature in tracing that Android had its
own patch internally for. I took it back in 4.13. Now he realizes that
he had a mistake, and swapped the values from what Android had. This
means that the old Android tools will break when using a new kernel
that has the new feature on it.
The options are:
1. To swap it back to what Android wants.
2. Add a command line option or something to do the swap
3. Just let Android carry a patch that swaps it back
Since it requires setting a tracing option to enable this anyway, I
doubt there are other users of this than Android. Thus, I've decided
to take option 1. If someone else is actually depending on the order
that is in the kernel, then we will have to revert this change and go
to option 2 or 3"
* tag 'trace-v4.18-rc3-3' of git://git.kernel.org/pub/scm/linux/kernel/git/rostedt/linux-trace:
tracing: Reorder display of TGID to be after PID
Merge tag 'sound-4.18-rc5' of git://git.kernel.org/pub/scm/linux/kernel/git/tiwai/sound
Pull sound fixes from Takashi Iwai:
"Just a few HD-auio fixes: one fix for a possible mutex deadlock at
HDMI hotplug handling is somewhat subtle and delicate, while the rest
are usual device-specific quirks"
* tag 'sound-4.18-rc5' of git://git.kernel.org/pub/scm/linux/kernel/git/tiwai/sound:
ALSA: hda/ca0132: Update a pci quirk device name
ALSA: hda/ca0132: Add Recon3Di quirk for Gigabyte G1.Sniper Z97
ALSA: hda/realtek - two more lenovo models need fixup of MIC_LOCATION
ALSA: hda - Handle pm failure during hotplug
Merge tag 'libnvdimm-fixes-4.18-rc5' of git://git.kernel.org/pub/scm/linux/kernel/git/nvdimm/nvdimm
Pull libnvdimm fixes from Dave Jiang:
- ensure that a variable passed in by reference to acpi_nfit_ctl is
always set to a value. An incremental patch is provided due to notice
from testing in -next. The rest of the commits did not exhibit
issues.
- fix a return path in nsio_rw_bytes() that was not returning "bytes
remain" as expected for the function.
- address an issue where applications polling on scrub-completion for
the NVDIMM may falsely wakeup and read the wrong state value and
cause hang.
- change the test unit persistent capability attribute to fix up a
broken assumption in the unit test infrastructure wrt the
'write_cache' attribute
- ratelimit dev_info() in the dax device check_vma() function since
this is easily triggered from userspace
* tag 'libnvdimm-fixes-4.18-rc5' of git://git.kernel.org/pub/scm/linux/kernel/git/nvdimm/nvdimm:
nfit: fix unchecked dereference in acpi_nfit_ctl
acpi, nfit: Fix scrub idle detection
tools/testing/nvdimm: advertise a write cache for nfit_test
acpi/nfit: fix cmd_rc for acpi_nfit_ctl to always return a value
dev-dax: check_vma: ratelimit dev_info-s
libnvdimm, pmem: Fix memcpy_mcsafe() return code handling in nsio_rw_bytes()
Alex Deucher [Thu, 12 Jul 2018 13:38:09 +0000 (08:38 -0500)]
drm/amdgpu/pp/smu7: use a local variable for toc indexing
Rather than using the index variable stored in vram. If
the device fails to come back online after a resume cycle,
reads from vram will return all 1s which will cause a
segfault. Based on a patch from Thomas Martitz <[email protected]>.
This avoids the segfault, but we still need to sort out
why the GPU does not come back online after a resume.
btrfs_cmp_data_free() puts cmp's src_pages and dst_pages, but leaves
their page address intact. Now, if you hit "goto again" in
btrfs_extent_same_range() and hit some error in
btrfs_cmp_data_prepare(), you'll try to unlock/put already put pages.
This is simple fix to reset the address to avoid use-after-free.
Fixes: 67b07bd4bec5 ("Btrfs: reuse cmp workspace in EXTENT_SAME ioctl") Signed-off-by: Naohiro Aota <[email protected]> Reviewed-by: David Sterba <[email protected]> Signed-off-by: David Sterba <[email protected]>
====================
This patch set adjusts the AF_XDP TX error reporting so that it becomes
consistent between copy mode and zero-copy. First some background:
Copy-mode for TX uses the SKB path in which the action of sending the
packet is performed from process context using the sendmsg
syscall. Completions are usually done asynchronously from NAPI mode by
using a TX interrupt. In this mode, send errors can be returned back
through the syscall.
In zero-copy mode both the sending of the packet and the completions
are done asynchronously from NAPI mode for performance reasons. In
this mode, the sendmsg syscall only makes sure that the TX NAPI loop
will be run that performs both the actions of sending and
completing. In this mode it is therefore not possible to return errors
through the sendmsg syscall as the sending is done from the NAPI
loop. Note that it is possible to implement a synchronous send with
our API, but in our benchmarks that made the TX performance drop by
nearly half due to synchronization requirements and cache line
bouncing. But for some netdevs this might be preferable so let us
leave it up to the implementation to decide.
The problem is that the current code base returns some errors in
copy-mode that are not possible to return in zero-copy mode. This
patch set aligns them so that the two modes always return the same
error code. We achieve this by removing some of the errors returned by
sendmsg in copy-mode (and in one case adding an error message for
zero-copy mode) and offering alternative error detection methods that
are consistent between the two modes.
The structure of the patch set is as follows:
Patch 1: removes the ENXIO return code from copy-mode when someone has
forcefully changed the number of queues on the device so that the
queue bound to the socket is no longer available. Just silently stop
sending anything as in zero-copy mode.
Patch 2: stop returning EAGAIN in copy mode when the completion queue
is full as zero-copy does not do this. Instead this situation can be
detected by comparing the head and tail pointers of the completion
queue in both modes. In any case, EAGAIN was not the correct error code
here since no amount of calling sendmsg will solve the problem. Only
consuming one or more messages on the completion queue will fix this.
Patch 3: Always return ENOBUFS from sendmsg if there is no TX queue
configured. This was not the case for zero-copy mode.
Patch 4: stop returning EMSGSIZE when the size of the packet is larger
than the MTU. Just send it to the device so that it will drop it as in
zero-copy mode.
Note that copy-mode can still return EAGAIN in certain circumstances,
but as these conditions cannot occur in zero-copy mode it is fine for
copy-mode to return them.
====================
Magnus Karlsson [Wed, 11 Jul 2018 08:12:52 +0000 (10:12 +0200)]
xsk: do not return EMSGSIZE in copy mode for packets larger than MTU
This patch stops returning EMSGSIZE from sendmsg in copy mode when the
size of the packet is larger than the MTU. Just send it to the device
so that it will drop it as in zero-copy mode. This makes the error
reporting consistent between copy mode and zero-copy mode.
Fixes: 35fcde7f8deb ("xsk: support for Tx") Signed-off-by: Magnus Karlsson <[email protected]> Signed-off-by: Daniel Borkmann <[email protected]>
Magnus Karlsson [Wed, 11 Jul 2018 08:12:51 +0000 (10:12 +0200)]
xsk: always return ENOBUFS from sendmsg if there is no TX queue
This patch makes sure ENOBUFS is always returned from sendmsg if there
is no TX queue configured. This was not the case for zero-copy
mode. With this patch this error reporting is consistent between copy
mode and zero-copy mode.
Fixes: ac98d8aab61b ("xsk: wire upp Tx zero-copy functions") Signed-off-by: Magnus Karlsson <[email protected]> Signed-off-by: Daniel Borkmann <[email protected]>
Magnus Karlsson [Wed, 11 Jul 2018 08:12:50 +0000 (10:12 +0200)]
xsk: do not return EAGAIN from sendmsg when completion queue is full
This patch stops returning EAGAIN in TX copy mode when the completion
queue is full as zero-copy does not do this. Instead this situation
can be detected by comparing the head and tail pointers of the
completion queue in both modes. In any case, EAGAIN was not the
correct error code here since no amount of calling sendmsg will solve
the problem. Only consuming one or more messages on the completion
queue will fix this.
With this patch, the error reporting becomes consistent between copy
mode and zero-copy mode.
Fixes: 35fcde7f8deb ("xsk: support for Tx") Signed-off-by: Magnus Karlsson <[email protected]> Signed-off-by: Daniel Borkmann <[email protected]>
Magnus Karlsson [Wed, 11 Jul 2018 08:12:49 +0000 (10:12 +0200)]
xsk: do not return ENXIO from TX copy mode
This patch removes the ENXIO return code from TX copy-mode when
someone has forcefully changed the number of queues on the device so
that the queue bound to the socket is no longer available. Just
silently stop sending anything as in zero-copy mode so the error
reporting gets consistent between the two modes.
Fixes: 35fcde7f8deb ("xsk: support for Tx") Signed-off-by: Magnus Karlsson <[email protected]> Signed-off-by: Daniel Borkmann <[email protected]>
David Sterba [Wed, 20 Jun 2018 17:51:28 +0000 (19:51 +0200)]
btrfs: restore uuid_mutex in btrfs_open_devices
Commit 542c5908abfe84f7b4c1 ("btrfs: replace uuid_mutex by
device_list_mutex in btrfs_open_devices") switched to device_list_mutex
as we need that for the device list traversal, but we also need
uuid_mutex to protect access to fs_devices::opened to be consistent with
other users of that.
Fixes: 542c5908abfe84f7b4c1 ("btrfs: replace uuid_mutex by device_list_mutex in btrfs_open_devices") Reviewed-by: Anand Jain <[email protected]> Signed-off-by: David Sterba <[email protected]>
Stephan Mueller [Sat, 7 Jul 2018 18:41:47 +0000 (20:41 +0200)]
crypto: af_alg - Initialize sg_num_bytes in error code path
The RX SGL in processing is already registered with the RX SGL tracking
list to support proper cleanup. The cleanup code path uses the
sg_num_bytes variable which must therefore be always initialized, even
in the error code path.
Willem de Bruijn [Wed, 11 Jul 2018 16:00:46 +0000 (12:00 -0400)]
selftests: in udpgso_bench do not test udp zerocopy
The udpgso benchmark compares various configurations of UDP and TCP.
Including one that is not upstream, udp zerocopy. This is a leftover
from the earlier RFC patchset.
The test is part of kselftests and run in continuous spinners. Remove
the failing case to make the test start passing.
However Android's trace visualization tools expect a slightly different
format due to an out-of-tree patch patch that was been carried for a
decade, notice that the TGID and CPU fields are reversed:
From kernel v4.13 onwards, during which TGID was introduced, tracing
with systrace on all Android kernels will break (most Android kernels
have been on 4.9 with Android patches, so this issues hasn't been seen
yet). From v4.13 onwards things will break.
The chrome browser's tracing tools also embed the systrace viewer which
uses the legacy TGID format and updates to that are known to be
difficult to make.
Considering this, I suggest we make this change to the upstream kernel
and backport it to all Android kernels. I believe this feature is merged
recently enough into the upstream kernel that it shouldn't be a problem.
Also logically, IMO it makes more sense to group the TGID with the
TASK-PID and the CPU after these.
Willem de Bruijn [Wed, 11 Jul 2018 16:00:45 +0000 (12:00 -0400)]
packet: reset network header if packet shorter than ll reserved space
If variable length link layer headers result in a packet shorter
than dev->hard_header_len, reset the network header offset. Else
skb->mac_len may exceed skb->len after skb_mac_reset_len.
packet_sendmsg_spkt already has similar logic.
Fixes: b84bbaf7a6c8 ("packet: in packet_snd start writing at link layer allocation") Signed-off-by: Willem de Bruijn <[email protected]> Signed-off-by: David S. Miller <[email protected]>
net: Don't copy pfmemalloc flag in __copy_skb_header()
The pfmemalloc flag indicates that the skb was allocated from
the PFMEMALLOC reserves, and the flag is currently copied on skb
copy and clone.
However, an skb copied from an skb flagged with pfmemalloc
wasn't necessarily allocated from PFMEMALLOC reserves, and on
the other hand an skb allocated that way might be copied from an
skb that wasn't.
So we should not copy the flag on skb copy, and rather decide
whether to allow an skb to be associated with sockets unrelated
to page reclaim depending only on how it was allocated.
Move the pfmemalloc flag before headers_start[0] using an
existing 1-bit hole, so that __copy_skb_header() doesn't copy
it.
When cloning, we'll now take care of this flag explicitly,
contravening to the warning comment of __skb_clone().
While at it, restore the newline usage introduced by commit b19372273164 ("net: reorganize sk_buff for faster
__copy_skb_header()") to visually separate bytes used in
bitfields after headers_start[0], that was gone after commit a9e419dc7be6 ("netfilter: merge ctinfo into nfct pointer storage
area"), and describe the pfmemalloc flag in the kernel-doc
structure comment.
This doesn't change the size of sk_buff or cacheline boundaries,
but consolidates the 15 bits hole before tc_index into a 2 bytes
hole before csum, that could now be filled more easily.
Reported-by: Patrick Talbert <[email protected]> Fixes: c93bdd0e03e8 ("netvm: allow skb allocation to use PFMEMALLOC reserves") Signed-off-by: Stefano Brivio <[email protected]> Signed-off-by: David S. Miller <[email protected]>
We should take and release the filter_sem consistently during the
reset process, in the same manner as the mac_lock and reset_lock.
For lockdep consistency we also take the filter_sem for write around
other calls to efx->type->init().
Fixes: c2bebe37c6b6 ("sfc: give ef10 its own rwsem in the filter table instead of filter_lock") Signed-off-by: Bert Kenward <[email protected]> Signed-off-by: David S. Miller <[email protected]>
In some situations we may end up calling down_read while already
holding the semaphore for write, thus hanging. This has been seen
when setting the MAC address for the interface. The hung task log
in this situation includes this stack:
down_read
efx_ef10_filter_insert
efx_ef10_filter_insert_addr_list
efx_ef10_filter_vlan_sync_rx_mode
efx_ef10_filter_add_vlan
efx_ef10_filter_table_probe
efx_ef10_set_mac_address
efx_set_mac_address
dev_set_mac_address
In addition, lockdep rightly points out that nested calling of
down_read is incorrect.
Fixes: c2bebe37c6b6 ("sfc: give ef10 its own rwsem in the filter table instead of filter_lock") Tested-by: Jarod Wilson <[email protected]> Signed-off-by: Bert Kenward <[email protected]> Signed-off-by: David S. Miller <[email protected]>
net: systemport: Fix CRC forwarding check for SYSTEMPORT Lite
SYSTEMPORT Lite reversed the logic compared to SYSTEMPORT, the
GIB_FCS_STRIP bit is set when the Ethernet FCS is stripped, and that bit
is not set by default. Fix the logic such that we properly check whether
that bit is set or not and we don't forward an extra 4 bytes to the
network stack.
Fixes: 44a4524c54af ("net: systemport: Add support for SYSTEMPORT Lite") Signed-off-by: Florian Fainelli <[email protected]> Signed-off-by: David S. Miller <[email protected]>
Wolfram Sang [Tue, 10 Jul 2018 21:42:15 +0000 (23:42 +0200)]
i2c: recovery: if possible send STOP with recovery pulses
I2C clients may misunderstand recovery pulses if they can't read SDA to
bail out early. In the worst case, as a write operation. To avoid that
and if we can write SDA, try to send STOP to avoid the
misinterpretation.
Stefan Baranoff [Tue, 10 Jul 2018 21:31:10 +0000 (17:31 -0400)]
tcp: allow user to create repair socket without window probes
Under rare conditions where repair code may be used it is possible that
window probes are either unnecessary or undesired. If the user knows that
window probes are not wanted or needed this change allows them to skip
sending them when a socket comes out of repair.
Stefan Baranoff [Tue, 10 Jul 2018 21:25:20 +0000 (17:25 -0400)]
tcp: fix sequence numbers for repaired sockets re-using TIME-WAIT sockets
This patch fixes a bug where the sequence numbers of a socket created using
TCP repair functionality are lower than set after connect is called.
This occurs when the repair socket overlaps with a TIME-WAIT socket and
triggers the re-use code. The amount lower is equal to the number of times
that a particular IP/port set is re-used and then put back into TIME-WAIT.
Re-using the first time the sequence number is 1 lower, closing that socket
and then re-opening (with repair) a new socket with the same addresses/ports
puts the sequence number 2 lower than set via setsockopt. The third time is
3 lower, etc. I have not tested what the limit of this acrewal is, if any.
The fix is, if a socket is in repair mode, to respect the already set
sequence number and timestamp when it would have already re-used the
TIME-WAIT socket.
What happens in jit_subprogs() is that kcalloc() for the subprog func
buffer is failing with NULL where we then bail out. Latter is a plain
return -ENOMEM, and this is definitely not okay since earlier in the
loop we are walking all subprogs and temporarily rewrite insn->off to
remember the subprog id as well as insn->imm to temporarily point the
call to __bpf_call_base + 1 for the initial JIT pass. Thus, bailing
out in such state and handing this over to the interpreter is troublesome
since later/subsequent e.g. find_subprog() lookups are based on wrong
insn->imm.
Therefore, once we hit this point, we need to jump to out_free path
where we undo all changes from earlier loop, so that interpreter can
work on unmodified insn->{off,imm}.
Another point is that should find_subprog() fail in jit_subprogs() due
to a verifier bug, then we also should not simply defer the program to
the interpreter since also here we did partial modifications. Instead
we should just bail out entirely and return an error to the user who is
trying to load the program.
Fixes: 1c2a088a6626 ("bpf: x64: add JIT support for multi-function programs") Reported-by: [email protected] Signed-off-by: Daniel Borkmann <[email protected]> Signed-off-by: Alexei Starovoitov <[email protected]>
Jacob Keller [Tue, 10 Jul 2018 21:22:27 +0000 (14:22 -0700)]
sch_fq_codel: zero q->flows_cnt when fq_codel_init fails
When fq_codel_init fails, qdisc_create_dflt will cleanup by using
qdisc_destroy. This function calls the ->reset() op prior to calling the
->destroy() op.
Unfortunately, during the failure flow for sch_fq_codel, the ->flows
parameter is not initialized, so the fq_codel_reset function will null
pointer dereference.
This is caused because flows_cnt is non-zero, but flows hasn't been
initialized. fq_codel_init has left the private data in a partially
initialized state.
To fix this, reset flows_cnt to 0 when we fail to initialize.
Additionally, to make the state more consistent, also cleanup the flows
pointer when the allocation of backlogs fails.
This fixes the NULL pointer dereference, since both the for-loop and
memset in fq_codel_reset will be no-ops when flow_cnt is zero.
Suppress warnings for systems that do not recognize LFS_*.
getconf: no such configuration parameter `LFS_CFLAGS'
getconf: no such configuration parameter `LFS_LDFLAGS'
getconf: no such configuration parameter `LFS_LIBS'
Fixes: d7f14c66c273 ("kbuild: Enable Large File Support for hostprogs") Reported-by: Chen Feng <[email protected]> Signed-off-by: Masahiro Yamada <[email protected]> Acked-by: Uwe Kleine-König <[email protected]>
Laura Abbott [Tue, 10 Jul 2018 00:45:56 +0000 (17:45 -0700)]
tools: build: Fixup host c flags
Commit 0c3b7e42616f ("tools build: Add support for host programs format")
introduced host_c_flags which referenced CHOSTFLAGS. The actual name of the
variable is HOSTCFLAGS. Fix this up.
Fixes: 0c3b7e42616f ("tools build: Add support for host programs format") Signed-off-by: Laura Abbott <[email protected]> Acked-by: Jiri Olsa <[email protected]> Signed-off-by: Masahiro Yamada <[email protected]>
Documentation/networking/e1000.rst:83: ERROR: Unexpected indentation.
Documentation/networking/e1000.rst:84: WARNING: Block quote ends without a blank line; unexpected unindent.
Documentation/networking/e1000.rst:173: WARNING: Definition list ends without a blank line; unexpected unindent.
Documentation/networking/e1000.rst:236: WARNING: Definition list ends without a blank line; unexpected unindent.
While here, fix highlights and mark a table as such.
David Francis [Thu, 12 Jul 2018 14:07:49 +0000 (10:07 -0400)]
amd/dc/dce100: On dce100, set clocks to 0 on suspend
[Why]
When a dce100 asic was suspended, the clocks were not set to 0.
Upon resume, the new clock was compared to the existing clock,
they were found to be the same, and so the clock was not set.
This resulted in a pernicious blackscreen.
[How]
In atomic commit, check to see if there are any active pipes.
If no, set clocks to 0
Harry Wentland [Thu, 12 Jul 2018 14:23:21 +0000 (10:23 -0400)]
drm/amd/display: Convert 10kHz clks from PPLib into kHz for Vega
The driver is expecting clock frequency in kHz, while SMU returns
the values in 10kHz, which causes the bandwidth validation to fail
4.18 has the faulty clock assignment in pp_to_dc_clock_levels_with_latency
only, which is only used by Vega. Make sure we multiply these values
by 10 here, as we do for other ASICs as powerplay assigned them
wrong. 4.19 has the proper fix in powerplay.
v2: Add Fixes tag
v3: Fixes -> Bugzilla, with simplified link
Alexander Duyck [Mon, 18 Jun 2018 16:02:00 +0000 (12:02 -0400)]
ixgbe: Be more careful when modifying MAC filters
This change makes it so that we are much more explicit about the ordering
of updates to the receive address register (RAR) table. Prior to this patch
I believe we may have been updating the table while entries were still
active, or possibly allowing for reordering of things since we weren't
explicitly flushing writes to either the lower or upper portion of the
register prior to accessing the other half.
Adam Ford [Wed, 11 Jul 2018 17:54:54 +0000 (12:54 -0500)]
ARM: dts: am3517.dtsi: Disable reference to OMAP3 OTG controller
The AM3517 has a different OTG controller location than the OMAP3,
which is included from omap3.dtsi. This results in a hwmod error.
Since the AM3517 has a different OTG controller address, this patch
disabes one that is isn't available.
ARM: DRA7/OMAP5: Enable ACTLR[0] (Enable invalidates of BTB) for secondary cores
Call secure services to enable ACTLR[0] (Enable invalidates of BTB with
ICIALLU) when branch hardening is enabled for kernel.
On GP devices OMAP5/DRA7, there is no possibility to update secure
side since "secure world" is ROM and there are no override mechanisms
possible. On HS devices, appropriate PPA should do the workarounds as
well.
However, the configuration is only done for secondary core, since it is
expected that firmware/bootloader will have enabled the required
configuration for the primary boot core (note: bootloaders typically
will NOT enable secondary processors, since it has no need to do so).
xen: remove global bit from __default_kernel_pte_mask for pv guests
When removing the global bit from __supported_pte_mask do the same for
__default_kernel_pte_mask in order to avoid the WARN_ONCE() in
check_pgprot() when setting a kernel pte before having called
init_mem_mapping().
David S. Miller [Thu, 12 Jul 2018 07:19:06 +0000 (00:19 -0700)]
Merge branch 'ieee802154-for-davem-2018-07-11' of git://git.kernel.org/pub/scm/linux/kernel/git/sschmidt/wpan
Stefan Schmidt says:
====================
pull-request: ieee802154 for net 2018-07-11
An update from ieee802154 for your *net* tree.
Build system fix for a missing include from Arnd Bergmann.
Setting the IFLA_LINK for the lowpan parent from Lubomir Rintel.
Fixes for some RX corner cases in adf7242 driver by Michael Hennerich.
And some small patches to cleanup our BUG_ON vs WARN_ON usage.
====================
Keith Busch [Wed, 11 Jul 2018 22:44:44 +0000 (16:44 -0600)]
nvme-pci: fix memory leak on probe failure
The nvme driver specific structures need to be initialized prior to
enabling the generic controller so we can unwind on failure with out
using the reference counting callbacks so that 'probe' and 'remove'
can be symmetric.
The newly added iod_mempool is the only resource that was being
allocated out of order, and a failure there would leak the generic
controller memory. This patch just moves that allocation above the
controller initialization.
Fixes: 943e942e6266f ("nvme-pci: limit max IO size and segments to avoid high order allocations") Reported-by: Weiping Zhang <[email protected]> Signed-off-by: Keith Busch <[email protected]> Signed-off-by: Christoph Hellwig <[email protected]>
Russell King [Tue, 10 Jul 2018 11:05:36 +0000 (12:05 +0100)]
sfp: fix module initialisation with netdev already up
It was been observed that with a particular order of initialisation,
the netdev can be up, but the SFP module still has its TX_DISABLE
signal asserted. This occurs when the network device brought up before
the SFP kernel module has been inserted by userspace.
This occurs because sfp-bus layer does not hear about the change in
network device state, and so assumes that it is still down. Set
netdev->sfp when the upstream is registered to work around this problem.
Russell King [Tue, 10 Jul 2018 11:05:31 +0000 (12:05 +0100)]
sfp: ensure we clean up properly on bus registration failure
We fail to correctly clean up after a bus registration failure, which
can lead to an incorrect assumption about the registration state of
the upstream or sfp cage.
After scrubbing the BPF prog a bit from the noise, turns out it called
bpf_skb_change_head() for the lwt_xmit prog with headroom of 2. Nothing
wrong in that, however, this was run with repeat >> 0 in bpf_prog_test_run_skb()
and the same skb thus keeps changing until the pskb_expand_head() called
from skb_cow() keeps bailing out in atomic alloc context with -ENOMEM.
So upon return we'll basically have 0 headroom left yet blindly do the
__skb_push() of 14 bytes and keep copying data from there in bpf_test_finish()
out of bounds. Fix to check if we have enough headroom and if pskb_expand_head()
fails, bail out with error.
Another bug independent of this fix (but related in triggering above) is
that BPF_PROG_TEST_RUN should be reworked to reset the skb/xdp buffer to
it's original state from input as otherwise repeating the same test in a
loop won't work for benchmarking when underlying input buffer is getting
changed by the prog each time and reused for the next run leading to
unexpected results.
ARM: 8780/1: ftrace: Only set kernel memory back to read-only after boot
Dynamic ftrace requires modifying the code segments that are usually
set to read-only. To do this, a per arch function is called both before
and after the ftrace modifications are performed. The "before" function
will set kernel code text to read-write to allow for ftrace to make the
modifications, and the "after" function will set the kernel code text
back to "read-only" to keep the kernel code text protected.
The issue happens when dynamic ftrace is tested at boot up. The test is
done before the kernel code text has been set to read-only. But the
"before" and "after" calls are still performed. The "after" call will
change the kernel code text to read-only prematurely, and other boot
code that expects this code to be read-write will fail.
The solution is to add a variable that is set when the kernel code text
is expected to be converted to read-only, and make the ftrace "before"
and "after" calls do nothing if that variable is not yet set. This is
similar to the x86 solution from commit 162396309745 ("ftrace, x86:
make kernel text writable only for conversions").
When extracting bitfield from a number, btf_int_bits_seq_show() builds
a mask and accesses least significant byte of the number in a way
specific to little-endian. This patch fixes that by checking endianness
of the machine and then shifting left and right the unneeded bits.
Thanks to Martin Lau for the help in navigating potential pitfalls when
dealing with endianess and for the final solution.
Fixes: b00b8daec828 ("bpf: btf: Add pretty print capability for data with BTF type info") Signed-off-by: Okash Khawaja <[email protected]> Acked-by: Martin KaFai Lau <[email protected]> Signed-off-by: Daniel Borkmann <[email protected]>
bpf_lwt_seg6_* helpers require CONFIG_IPV6_SEG6_BPF, and currently
return -EOPNOTSUPP to indicate unavailability. This patch forces the
BPF verifier to reject programs using these helpers when
!CONFIG_IPV6_SEG6_BPF, allowing users to more easily probe if they are
available or not.
Merge branch 'for-4.18-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/tj/libata
Pull libata fixes from Tejun Heo:
- Jens's patches to expand the usable command depth from 31 to 32 broke
sata_fsl due to a subtle command iteration bug. Fixed by introducing
explicit iteration helpers and using the correct variant.
- On some laptops, enabling LPM by default reportedly led to occasional
hard hangs. Blacklist the affected cases.
- Other misc fixes / changes.
* 'for-4.18-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/tj/libata:
ata: Remove depends on HAS_DMA in case of platform dependency
ata: Fix ZBC_OUT all bit handling
ata: Fix ZBC_OUT command block check
ahci: Add Intel Ice Lake LP PCI ID
ahci: Disable LPM on Lenovo 50 series laptops with a too old BIOS
sata_nv: remove redundant pointers sdev0 and sdev1
sata_fsl: remove dead code in tag retrieval
sata_fsl: convert to command iterator
libata: convert eh to command iterators
libata: add command iterator helpers
ata: ahci_mvebu: ahci_mvebu_stop_engine() can be static
libahci: Fix possible Spectre-v1 pmp indexing in ahci_led_store()
sample: vfio-mdev: avoid deadlock in mdev_access()
mdev_access() calls mbochs_get_page() with mdev_state->ops_lock held,
while mbochs_get_page() locks the mutex by itself.
It leads to unavoidable deadlock.
Found by Linux Driver Verification project (linuxtesting.org).
projects / linux.git / log