Validating the control action within tcf_mirred_init() proved to fix the
above issue. For the same reason, postpone the assignment of tcfa_action
and tcfm_eaction to avoid partial reconfiguration of a mirred rule when
it's replaced by another one that mirrors to a device that does not
exist. A TDC selftest is added to verify the correct behavior.
Fixes: db50514f9a9c ("net: sched: add termination action to allow goto chain") Fixes: 97763dc0f401 ("net_sched: reject unknown tcfa_action values") Signed-off-by: Davide Caratti <[email protected]> Signed-off-by: David S. Miller <[email protected]>
Davide Caratti [Wed, 20 Mar 2019 13:59:59 +0000 (14:59 +0100)]
net/sched: prepare TC actions to properly validate the control action
- pass a pointer to struct tcf_proto in each actions's init() handler,
to allow validating the control action, checking whether the chain
exists and (eventually) refcounting it.
- remove code that validates the control action after a successful call
to the action's init() handler, and replace it with a test that forbids
addition of actions having 'goto_chain' and NULL goto_chain pointer at
the same time.
- add tcf_action_check_ctrlact(), that will validate the control action
and eventually allocate the action 'goto_chain' within the init()
handler.
- add tcf_action_set_ctrlact(), that will assign the control action and
swap the current 'goto_chain' pointer with the new given one.
This disallows 'goto_chain' on actions that don't initialize it properly
in their init() handler, i.e. calling tcf_action_check_ctrlact() after
successful IDR reservation and then calling tcf_action_set_ctrlact()
to assign 'goto_chain' and 'tcf_action' consistently.
By doing this, the kernel does not leak anymore refcounts when a valid
'goto chain' handle is replaced in TC actions, causing kmemleak splats
like the following one:
Harini Katakam [Wed, 20 Mar 2019 13:42:22 +0000 (19:12 +0530)]
net: macb: Add null check for PCLK and HCLK
Both PCLK and HCLK are "required" clocks according to macb devicetree
documentation. There is a chance that devm_clk_get doesn't return a
negative error but just a NULL clock structure instead. In such a case
the driver proceeds as usual and uses pclk value 0 to calculate MDC
divisor which is incorrect. Hence fix the same in clock initialization.
Linus Torvalds [Thu, 21 Mar 2019 18:48:06 +0000 (11:48 -0700)]
Merge tag 'arm64-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/arm64/linux
Pull arm64 fixes from Catalin Marinas:
"Mostly fixes apart from the kprobe blacklist checking which was
deferred because of conflicting with a fix merged after I pinned the
arm64 for-next/core branch (f2b3d8566d81 "arm64: kprobe: Always
blacklist the KVM world-switch code").
Summary:
- Update the kprobe blacklist checking for arm64. This was supposed
to be queued during the merging window but, due to conflicts, it
was deferred post -rc1
- Extend the Fujitsu erratum 010001 workaround to A64FX v1r0
- Whitelist HiSilicon Taishan v110 CPUs as not susceptible to
Meltdown
- Export save_stack_trace_regs()
- Remove obsolete selection of MULTI_IRQ_HANDLER"
* tag 'arm64-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/arm64/linux:
arm64: remove obsolete selection of MULTI_IRQ_HANDLER
arm64: kpti: Whitelist HiSilicon Taishan v110 CPUs
arm64: Add MIDR encoding for HiSilicon Taishan CPUs
arm64/stacktrace: Export save_stack_trace_regs()
arm64: apply workaround on A64FX v1r0
arm64: kprobes: Use arch_populate_kprobe_blacklist()
arm64: kprobes: Move exception_text check in blacklist
arm64: kprobes: Remove unneeded RODATA check
arm64: kprobes: Move extable address check into arch_prepare_kprobe()
Linus Torvalds [Thu, 21 Mar 2019 17:31:55 +0000 (10:31 -0700)]
Merge tag 'fixes_for_v5.1-rc2' of git://git.kernel.org/pub/scm/linux/kernel/git/jack/linux-fs
Pull udf fixes from Jan Kara:
"Two udf error handling fixes"
* tag 'fixes_for_v5.1-rc2' of git://git.kernel.org/pub/scm/linux/kernel/git/jack/linux-fs:
udf: Propagate errors from udf_truncate_extents()
udf: Fix crash on IO error during truncate
Linus Torvalds [Thu, 21 Mar 2019 17:24:00 +0000 (10:24 -0700)]
Merge tag 'fsnotify_for_v5.1-rc2' of git://git.kernel.org/pub/scm/linux/kernel/git/jack/linux-fs
Pull fsnotify fixes from Jan Kara:
"One inotify and one fanotify fix"
* tag 'fsnotify_for_v5.1-rc2' of git://git.kernel.org/pub/scm/linux/kernel/git/jack/linux-fs:
fanotify: Allow copying of file handle to userspace
inotify: Fix fsnotify_mark refcount leak in inotify_update_existing_watch()
The following patchset contains Netfilter fixes for your net tree:
1) Remove a direct dependency with IPv6 introduced by the
sip_external_media feature, from Alin Nastac.
2) Fix bogus ENOENT when removing interval elements from set.
3) Set transport_header from br_netfilter to mimic the stack
behaviour, this partially fixes a checksum validation bug
from the SCTP connection tracking, from Xin Long.
4) Fix undefined reference to symbol in xt_TEE, due to missing
Kconfig dependencies, from Arnd Bergmann.
5) Check for NULL in skb_header_pointer() calls in ip6t_shr,
from Kangjie Lu.
6) Fix bogus EBUSY when removing an existing conntrack helper from
a transaction.
7) Fix module autoload of the redirect extension.
8) Remove duplicated transition in flowtable diagram in the existing
documentation.
9) Missing .release_ops call from error path in newrule() which
results module refcount leak, from Taehee Yoo.
====================
Erik Hugne [Thu, 21 Mar 2019 08:11:59 +0000 (09:11 +0100)]
tipc: fix cancellation of topology subscriptions
When cancelling a subscription, we have to clear the cancel bit in the
request before iterating over any established subscriptions with memcmp.
Otherwise no subscription will ever be found, and it will not be
possible to explicitly unsubscribe individual subscriptions.
Fixes: 8985ecc7c1e0 ("tipc: simplify endianness handling in topology subscriber") Signed-off-by: Erik Hugne <[email protected]> Signed-off-by: Jon Maloy <[email protected]> Signed-off-by: David S. Miller <[email protected]>
Fix the fallback definition when HAVE_LIBBPF_SUPPORT is not defined,
i.e. add the missing 'static inline' and add the __maybe_unused to the
args. Also add stdio.h since we now use FILE * in bpf-event.h.
This patch handles 3) and 4) for BPF programs loaded after 'perf
record|top'.
For timely process of these information, a dedicated event is added to
the side band evlist.
When PERF_RECORD_BPF_EVENT is received via the side band event, the
polling thread gathers 3) and 4) vis sys_bpf and store them in perf_env.
This information is saved to perf.data at the end of 'perf record'.
Committer testing:
The 'wakeup_watermark' member in 'struct perf_event_attr' is inside a
unnamed union, so can't be used in a struct designated initialization
with older gccs, get it out of that, isolating as 'attr.wakeup_watermark
= 1;' to work with all gcc versions.
We also need to add '--no-bpf-event' to the 'perf record'
perf_event_attr tests in 'perf test', as the way that that test goes is
to intercept the events being setup and looking if they match the fields
described in the control files, since now it finds first the side band
event used to catch the PERF_RECORD_BPF_EVENT, they all fail.
With these issues fixed:
Same scenario as for testing BPF programs loaded before 'perf record' or
'perf top' starts, only start the BPF programs after 'perf record|top',
so that its information get collected by the sideband threads, the rest
works as for the programs loaded before start monitoring.
Add missing 'inline' to the bpf_event__add_sb_event() when
HAVE_LIBBPF_SUPPORT is not defined, fixing the build in systems without
binutils devel files installed.
Song Liu [Tue, 12 Mar 2019 05:30:50 +0000 (22:30 -0700)]
perf evlist: Introduce side band thread
This patch introduces side band thread that captures extended
information for events like PERF_RECORD_BPF_EVENT.
This new thread uses its own evlist that uses ring buffer with very low
watermark for lower latency.
To use side band thread, we need to:
1. add side band event(s) by calling perf_evlist__add_sb_event();
2. calls perf_evlist__start_sb_thread();
3. at the end of perf run, perf_evlist__stop_sb_thread().
In the next patch, we use this thread to handle PERF_RECORD_BPF_EVENT.
Committer notes:
Add fix by Jiri Olsa for when te sb_tread can't get started and then at
the end the stop_sb_thread() segfaults when joining the (non-existing)
thread.
That can happen when running 'perf top' or 'perf record' as a normal
user, for instance.
Further checks need to be done on top of this to more graciously handle
these possible failure scenarios.
This is because it expands the macro into a non-functioning calling
sequence. The calling order must be:
outb(CX86_CCR2, 0x22);
inb(0x23);
From the comments:
* When using the old macros a line like
* setCx86(CX86_CCR2, getCx86(CX86_CCR2) | 0x88);
* gets expanded to:
* do {
* outb((CX86_CCR2), 0x22);
* outb((({
* outb((CX86_CCR2), 0x22);
* inb(0x23);
* }) | 0x88), 0x23);
* } while (0);
The new macros fix this problem, so use them instead. Tested on an
actual Geode processor.
Nick Desaulniers [Thu, 14 Mar 2019 22:14:57 +0000 (15:14 -0700)]
x86/boot: Restrict header scope to make Clang happy
The inclusion of <linux/kernel.h> was causing issue as the definition of
__arch_hweight64 from arch/x86/include/asm/arch_hweight.h eventually gets
included. The definition is problematic when compiled with -m16 (all code
in arch/x86/boot/ is) as the "D" inline assembly constraint is rejected
by both compilers when passed an argument of type long long (regardless
of signedness, anything smaller is fine).
Because GCC performs inlining before semantic analysis, and
__arch_hweight64 is dead in this translation unit, GCC does not report
any issues at compile time. Clang does the semantic analysis in the
front end, before inlining (run in the middle) can determine the code is
dead. I consider this another case of PR33587, which I think we can do
more work to solve.
It turns out that arch/x86/boot/string.c doesn't actually need
linux/kernel.h, simply linux/limits.h and linux/compiler.h.
When kzalloc() fails in push_stack(), free_verifier_state() will free
current verifier state. As push_stack() returns, dst_reg was restored
if ptr_is_dst_reg is false. However, as member of the cur_state,
dst_reg is also freed, and error occurs when dereferencing dst_reg.
Simply fix it by testing ret of push_stack() before restoring dst_reg.
Fixes: 979d63d50c0c ("bpf: prevent out of bounds speculation on pointer arithmetic") Signed-off-by: Xu Yu <[email protected]> Signed-off-by: Daniel Borkmann <[email protected]>
Jianguo Chen [Wed, 20 Mar 2019 18:54:21 +0000 (18:54 +0000)]
irqchip/mbigen: Don't clear eventid when freeing an MSI
mbigen_write_msg clears eventid bits of a mbigen register
when free a interrupt, because msi_domain_deactivate memset
struct msg to zero. Then multiple mbigen pins with zero eventid
will report the same interrupt number.
Fabien Dessenne [Thu, 7 Mar 2019 18:40:36 +0000 (19:40 +0100)]
irqchip/stm32: Don't set rising configuration registers at init
The rising configuration status register (rtsr) is not banked.
As it is shared with the co-processor, it should not be written at probe
time, else the co-processor configuration will be lost.
Fixes: f9fc1745501e ("irqchip/stm32: Add host and driver data structures") Signed-off-by: Fabien Dessenne <[email protected]> Signed-off-by: Marc Zyngier <[email protected]>
Fabien Dessenne [Thu, 7 Mar 2019 18:40:35 +0000 (19:40 +0100)]
irqchip/stm32: Don't clear rising/falling config registers at init
Falling and rising configuration and status registers are not banked.
As they are shared with M4 co-processor, they should not be cleared
at probe time, else M4 co-processor configuration will be lost.
Fixes: f9fc1745501e ("irqchip/stm32: Add host and driver data structures") Signed-off-by: Loic Pallardy <[email protected]> Signed-off-by: Fabien Dessenne <[email protected]> Signed-off-by: Marc Zyngier <[email protected]>
YueHaibing [Wed, 20 Mar 2019 14:22:20 +0000 (22:22 +0800)]
irqchip/brcmstb-l2: Make two init functions static
Fix sparse warnings:
drivers/irqchip/irq-brcmstb-l2.c:278:12: warning:
symbol 'brcmstb_l2_edge_intc_of_init' was not declared. Should it be static?
drivers/irqchip/irq-brcmstb-l2.c:285:12: warning:
symbol 'brcmstb_l2_lvl_intc_of_init' was not declared. Should it be static?
Josh Poimboeuf [Tue, 19 Mar 2019 00:09:38 +0000 (19:09 -0500)]
objtool: Move objtool_file struct off the stack
Objtool uses over 512k of stack, thanks to the hash table embedded in
the objtool_file struct. This causes an unnecessarily large stack
allocation and breaks users with low stack limits.
Bart Van Assche [Mon, 11 Mar 2019 23:02:55 +0000 (16:02 -0700)]
workqueue: Only unregister a registered lockdep key
The recent change to prevent use after free and a memory leak introduced an
unconditional call to wq_unregister_lockdep() in the error handling
path. If the lockdep key had not been registered yet, then the lockdep core
emits a warning.
Only call wq_unregister_lockdep() if wq_register_lockdep() has been
called first.
Wolfram Sang [Tue, 19 Mar 2019 10:12:59 +0000 (11:12 +0100)]
mmc: renesas_sdhi: limit block count to 16 bit for old revisions
R-Car Gen2 has two different SDHI incarnations in the same chip. The
older one does not support the recently introduced 32 bit register
access to the block count register. Make sure we use this feature only
after the first known version.
Thanks to the Renesas Testing team for this bug report!
Daniel Drake [Wed, 20 Mar 2019 06:36:53 +0000 (14:36 +0800)]
mmc: alcor: fix DMA reads
Setting max_blk_count to 1 here was causing the mmc block layer
to always use the MMC_READ_SINGLE_BLOCK command here, which the
driver does not DMA-accelerate.
Drop the max_blk_ settings here. The mmc host defaults suffice,
along with the max_segs and max_seg_size settings, which I have
now documented in more detail.
Now each MMC command reads 4 512-byte blocks, using DMA instead of
PIO. On my SD card, this increases read performance (measured with dd)
from 167kb/sec to 4.6mb/sec.
mmc: sdhci-omap: Set caps2 to indicate no physical write protect pin
After commit 6d5cd068ee59fba ("mmc: sdhci: use WP GPIO in
sdhci_check_ro()") and commit 39ee32ce486756f ("mmc: sdhci-omap: drop
->get_ro() implementation"), sdhci-omap relied on SDHCI_PRESENT_STATE
to check if the card is read-only, if wp-gpios is not populated
in device tree. However SDHCI_PRESENT_STATE in sdhci-omap does not have
correct read-only state.
sdhci-omap can be used by platforms with both micro SD slot and standard
SD slot with physical write protect pin (using GPIO). Set caps2 to
MMC_CAP2_NO_WRITE_PROTECT based on if wp-gpios property is populated or
not.
This fix is required since existing device-tree node doesn't have
"disable-wp" property and to preserve old-dt compatibility.
Fixes: 6d5cd068ee59fba ("mmc: sdhci: use WP GPIO in sdhci_check_ro()") Fixes: 39ee32ce486756f ("mmc: sdhci-omap: drop ->get_ro() implementation") Signed-off-by: Kishon Vijay Abraham I <[email protected]> Signed-off-by: Ulf Hansson <[email protected]>
Michael Ellerman [Thu, 21 Mar 2019 04:24:33 +0000 (15:24 +1100)]
powerpc/security: Fix spectre_v2 reporting
When I updated the spectre_v2 reporting to handle software count cache
flush I got the logic wrong when there's no software count cache
enabled at all.
The result is that on systems with the software count cache flush
disabled we print:
Which correctly indicates that the count cache is disabled, but
incorrectly says the software count cache flush is enabled.
The root of the problem is that we are trying to handle all
combinations of options. But we know now that we only expect to see
the software count cache flush enabled if the other options are false.
So split the two cases, which simplifies the logic and fixes the bug.
We were also missing a space before "(hardware accelerated)".
Jian-Hong Pan [Fri, 15 Mar 2019 09:51:09 +0000 (17:51 +0800)]
ALSA: hda/realtek: Enable headset MIC of Acer AIO with ALC286
Some Acer AIO desktops like Veriton Z6860G, Z4860G and Z4660G cannot
record sound from headset MIC. This patch adds the
ALC286_FIXUP_ACER_AIO_HEADSET_MIC quirk to fix this issue.
Fixes: 9f8aefed9623 ("ALSA: hda/realtek: Fix mic issue on Acer AIO Veriton Z4660G") Fixes: b72f936f6b32 ("ALSA: hda/realtek: Fix mic issue on Acer AIO Veriton Z4860G/Z6860G") Signed-off-by: Jian-Hong Pan <[email protected]> Reviewed-by: Kailang Yang <[email protected]> Cc: <[email protected]> Signed-off-by: Takashi Iwai <[email protected]>
Martin KaFai Lau [Mon, 18 Mar 2019 17:37:13 +0000 (10:37 -0700)]
bpf: Only print ref_obj_id for refcounted reg
Naresh reported that test_align fails because of the mismatch at the
verbose printout of the register states. The reason is due to the newly
added ref_obj_id.
ref_obj_id is only useful for refcounted reg. Thus, this patch fixes it
by only printing ref_obj_id for refcounted reg. While at it, it also uses
comma instead of space to separate between "id" and "ref_obj_id".
Fixes: 1b986589680a ("bpf: Fix bpf_tcp_sock and bpf_sk_fullsock issue related to bpf_sk_release") Reported-by: Naresh Kamboju <[email protected]> Signed-off-by: Martin KaFai Lau <[email protected]> Acked-by: Andrii Nakryiko <[email protected]> Signed-off-by: Alexei Starovoitov <[email protected]>
MIXER on Exynos5 SoCs uses different synchronisation method than Exynos4
to update internal state (shadow registers).
Apparently the driver implements it incorrectly. The rule should be
as follows:
- do not request updating registers until previous request was finished,
ie. MXR_CFG_LAYER_UPDATE_COUNT must be 0.
- before setting registers synchronisation on VSYNC should be turned off,
ie. MXR_STATUS_SYNC_ENABLE should be reset,
- after finishing MXR_STATUS_SYNC_ENABLE should be set again.
The patch hopefully implements it correctly.
Below sample kernel log from page fault caused by the bug:
[ 25.670038] exynos-sysmmu 14650000.sysmmu: 14450000.mixer: PAGE FAULT occurred at 0x2247b800
[ 25.677888] ------------[ cut here ]------------
[ 25.682164] kernel BUG at ../drivers/iommu/exynos-iommu.c:450!
[ 25.687971] Internal error: Oops - BUG: 0 [#1] PREEMPT SMP ARM
[ 25.693778] Modules linked in:
[ 25.696816] CPU: 5 PID: 1553 Comm: fb-release_test Not tainted 5.0.0-rc7-01157-g5f86b1566bdd #136
[ 25.705646] Hardware name: SAMSUNG EXYNOS (Flattened Device Tree)
[ 25.711710] PC is at exynos_sysmmu_irq+0x1c0/0x264
[ 25.716470] LR is at lock_is_held_type+0x44/0x64
v2: added missing MXR_CFG_LAYER_UPDATE bit setting in mixer_enable_sync
Tyrel Datwyler [Wed, 20 Mar 2019 18:41:51 +0000 (13:41 -0500)]
scsi: ibmvscsi: Fix empty event pool access during host removal
The event pool used for queueing commands is destroyed fairly early in the
ibmvscsi_remove() code path. Since, this happens prior to the call so
scsi_remove_host() it is possible for further calls to queuecommand to be
processed which manifest as a panic due to a NULL pointer dereference as
seen here:
PANIC: "Unable to handle kernel paging request for data at address
0x00000000"
The kernel buffer log is overfilled with this log:
[11261.952732] ibmvscsi: found no event struct in pool!
This patch reorders the operations during host teardown. Start by calling
the SRP transport and Scsi_Host remove functions to flush any outstanding
work and set the host offline. LLDD teardown follows including destruction
of the event pool, freeing the Command Response Queue (CRQ), and unmapping
any persistent buffers. The event pool destruction is protected by the
scsi_host lock, and the pool is purged prior of any requests for which we
never received a response. Finally, move the removal of the scsi host from
our global list to the end so that the host is easily locatable for
debugging purposes during teardown.
Tyrel Datwyler [Wed, 20 Mar 2019 18:41:50 +0000 (13:41 -0500)]
scsi: ibmvscsi: Protect ibmvscsi_head from concurrent modificaiton
For each ibmvscsi host created during a probe or destroyed during a remove
we either add or remove that host to/from the global ibmvscsi_head
list. This runs the risk of concurrent modification.
This patch adds a simple spinlock around the list modification calls to
prevent concurrent updates as is done similarly in the ibmvfc driver and
ipr driver.
Fixes: 32d6e4b6e4ea ("scsi: ibmvscsi: add vscsi hosts to global list_head") Cc: <[email protected]> # v4.10+ Signed-off-by: Tyrel Datwyler <[email protected]> Signed-off-by: Martin K. Petersen <[email protected]>
Bart Van Assche [Wed, 20 Mar 2019 20:18:45 +0000 (13:18 -0700)]
blkcg: Fix kernel-doc warnings
Avoid that the following warnings are reported when building with W=1:
block/blk-cgroup.c:1755: warning: Function parameter or member 'q' not described in 'blkcg_schedule_throttle'
block/blk-cgroup.c:1755: warning: Function parameter or member 'use_memdelay' not described in 'blkcg_schedule_throttle'
block/blk-cgroup.c:1779: warning: Function parameter or member 'blkg' not described in 'blkcg_add_delay'
block/blk-cgroup.c:1779: warning: Function parameter or member 'now' not described in 'blkcg_add_delay'
block/blk-cgroup.c:1779: warning: Function parameter or member 'delta' not described in 'blkcg_add_delay'
Yufen Yu [Mon, 18 Mar 2019 14:44:41 +0000 (22:44 +0800)]
block: add BLK_MQ_POLL_CLASSIC for hybrid poll and return EINVAL for unexpected value
For q->poll_nsec == -1, means doing classic poll, not hybrid poll.
We introduce a new flag BLK_MQ_POLL_CLASSIC to replace -1, which
may make code much easier to read.
Additionally, since val is an int obtained with kstrtoint(), val can be
a negative value other than -1, so return -EINVAL for that case.
Thanks to Damien Le Moal for some good suggestion.
David S. Miller [Wed, 20 Mar 2019 19:43:52 +0000 (12:43 -0700)]
Merge branch 'ks8851-fixes'
Lukas Wunner says:
====================
ks8851 fixes & cleanups
Four fixes and two cleanups for the Microchip (formerly Micrel) KSZ8851
SPI Ethernet driver.
Some of the fixes might even pass as stable material, but I haven't marked
them as such for cautiousness: Doesn't hurt letting them bake in linux-next
for a few weeks to raise the confidence, even though we've tested them
extensively on our Revolution Pi open source PLCs.
====================
Lukas Wunner [Wed, 20 Mar 2019 14:02:00 +0000 (15:02 +0100)]
net: ks8851: Deduplicate register macros
The ks8851 chip is sold either with an SPI interface (KSZ8851SNL) or
with a so-called non-PCI interface (KSZ8851-16MLL). When the driver
for the latter was introduced with commit a55c0a0ed415 ("drivers/net:
ks8851_mll ethernet network driver"), it duplicated the register macros
introduced by the driver for the former with commit 3ba81f3ece3c ("net:
Micrel KS8851 SPI network driver").
The chips are almost identical, so the duplication seems unwarranted.
There are a handful of bits which are in use on the KSZ8851-16MLL but
reserved on the KSZ8851SNL, and vice-versa, but there are no actual
collisions.
Thus, remove the duplicate definitions from the KSZ8851-16MLL driver.
Mark all bits which differ between the two chips. Move the SPI frame
opcodes, which are specific to KSZ8851SNL, to its driver.
The KSZ8851-16MLL driver added a RXFCTR_THRESHOLD_MASK macro which is a
duplication of the RXFCTR_RXFCT_MASK macro, rename it where it's used.
Same for P1MBCR_FORCE_FDX, which duplicates the BMCR_FULLDPLX macro and
OBCR_ODS_16MA, which duplicates OBCR_ODS_16mA.
Lukas Wunner [Wed, 20 Mar 2019 14:02:00 +0000 (15:02 +0100)]
net: ks8851: Fix register macro misnomers
In the header file accompanying the ks8851 driver, the P1SCLMD register
macros are misnamed, they actually pertain to the P1CR register.
The P1CR macros in turn pertain to the P1SR register, see pages 65 to 68
of the spec:
http://www.hqchip.com/uploads/pdf/201703/47c98946d6c97a4766e14db3f24955f2.pdf
The misnomers have no negative consequences so far because the macros
aren't used by ks8851.c, but that's about to change.
Lukas Wunner [Wed, 20 Mar 2019 14:02:00 +0000 (15:02 +0100)]
net: ks8851: Set initial carrier state to down
The ks8851 chip's initial carrier state is down. A Link Change Interrupt
is signaled once interrupts are enabled if the carrier is up.
The ks8851 driver has it backwards by assuming that the initial carrier
state is up. The state is therefore misrepresented if the interface is
opened with no cable attached. Fix it.
The Link Change interrupt is sometimes not signaled unless the P1MBSR
register (which contains the Link Status bit) is read on ->ndo_open().
This might be a hardware erratum. Read the register by calling
mii_check_link(), which has the desirable side effect of setting the
carrier state to down if the cable was detached while the interface was
closed.
Lukas Wunner [Wed, 20 Mar 2019 14:02:00 +0000 (15:02 +0100)]
net: ks8851: Delay requesting IRQ until opened
The ks8851 driver currently requests the IRQ before registering the
net_device. Because the net_device name is used as IRQ name and is
still "eth%d" when the IRQ is requested, it's impossibe to tell IRQs
apart if multiple ks8851 chips are present. Most other drivers delay
requesting the IRQ until the net_device is opened. Do the same.
The driver doesn't enable interrupts on the chip before opening the
net_device and disables them when closing it, so there doesn't seem to
be a need to request the IRQ already on probe.
Lukas Wunner [Wed, 20 Mar 2019 14:02:00 +0000 (15:02 +0100)]
net: ks8851: Reassert reset pin if chip ID check fails
Commit 73fdeb82e963 ("net: ks8851: Add optional vdd_io regulator and
reset gpio") amended the ks8851 driver to briefly assert the chip's
reset pin on probe. It also amended the probe routine's error path to
reassert the reset pin if a subsequent initialization step fails.
However the commit misplaced reassertion of the reset pin in the error
path such that it is not performed if the check of the Chip ID and
Enable Register (CIDER) fails. The error path is therefore slightly
asymmetrical to the probe routine's body. Fix it.
Lukas Wunner [Wed, 20 Mar 2019 14:02:00 +0000 (15:02 +0100)]
net: ks8851: Dequeue RX packets explicitly
The ks8851 driver lets the chip auto-dequeue received packets once they
have been read in full. It achieves that by setting the ADRFE flag in
the RXQCR register ("Auto-Dequeue RXQ Frame Enable").
However if allocation of a packet's socket buffer or retrieval of the
packet over the SPI bus fails, the packet will not have been read in
full and is not auto-dequeued. Such partial retrieval of a packet
confuses the chip's RX queue management: On the next RX interrupt,
the first packet read from the queue will be the one left there
previously and this one can be retrieved without issues. But for any
newly received packets, the frame header status and byte count registers
(RXFHSR and RXFHBCR) contain bogus values, preventing their retrieval.
The chip allows explicitly dequeueing a packet from the RX queue by
setting the RRXEF flag in the RXQCR register ("Release RX Error Frame").
This could be used to dequeue the packet in case of an error, but if
that error is a failed SPI transfer, it is unknown if the packet was
transferred in full and was auto-dequeued or if it was only transferred
in part and requires an explicit dequeue. The safest approach is thus
to always dequeue packets explicitly and forgo auto-dequeueing.
Without this change, I've witnessed packet retrieval break completely
when an SPI DMA transfer fails, requiring a chip reset. Explicit
dequeueing magically fixes this and makes packet retrieval absolutely
robust for me.
The chip's documentation suggests auto-dequeuing and uses the RRXEF
flag only to dequeue error frames which the driver doesn't want to
retrieve. But that seems to be a fair-weather approach.
Song Liu [Tue, 12 Mar 2019 05:30:48 +0000 (22:30 -0700)]
perf annotate: Enable annotation of BPF programs
In symbol__disassemble(), DSO_BINARY_TYPE__BPF_PROG_INFO dso calls into
a new function symbol__disassemble_bpf(), where annotation line
information is filled based on the bpf_prog_info and btf data saved in
given perf_env.
symbol__disassemble_bpf() uses binutils's libopcodes to disassemble bpf
programs.
We'll see the two BPF programs that augmented_raw_syscalls.o puts in
place, one attached to the raw_syscalls:sys_enter and another to the
raw_syscalls:sys_exit tracepoints, as expected.
Now we can finally do, from the command line, annotation for one of
those two symbols, with the original BPF program source coude intermixed
with the disassembled JITed code:
Please see 'man perf-config' to see how to control what should be seen,
via ~/.perfconfig [annotate] section, for instance, one can suppress the
source code and see just the disassembly, etc.
Alternatively, use the TUI bu just using 'perf annotate', press
'/bpf_prog' to see the bpf symbols, press enter and do the interactive
annotation, which allows for dumping to a file after selecting the
the various output tunables, for instance, the above without source code
intermixed, plus showing all the instruction offsets:
Then press: 's' to hide the source code + 'O' twice to show all
instruction offsets, then 'P' to print to the
bpf_prog_819967866022f1e1_sys_enter.annotation file, which will have:
Song Liu [Tue, 12 Mar 2019 05:30:48 +0000 (22:30 -0700)]
perf build: Check what binutils's 'disassembler()' signature to use
Commit 003ca0fd2286 ("Refactor disassembler selection") in the binutils
repo, which changed the disassembler() function signature, so we must
use the feature test introduced in fb982666e380 ("tools/bpftool: fix
bpftool build with bintutils >= 2.9") to deal with that.
Committer testing:
After adding the missing function call to test-all.c, and:
Auto-detecting system features:
... dwarf: [ on ]
... dwarf_getlocations: [ on ]
... glibc: [ on ]
... gtk2: [ on ]
... libaudit: [ on ]
... libbfd: [ on ]
... libelf: [ on ]
... libnuma: [ on ]
... numa_num_possible_cpus: [ on ]
... libperl: [ on ]
... libpython: [ on ]
... libslang: [ on ]
... libcrypto: [ on ]
... libunwind: [ on ]
... libdw-dwarf-unwind: [ on ]
... zlib: [ on ]
... lzma: [ on ]
... get_cpuid: [ on ]
... bpf: [ on ]
... libaio: [ on ]
... disassembler-four-args: [ on ]
CC /tmp/build/perf/jvmti/libjvmti.o
CC /tmp/build/perf/builtin-bench.o
<SNIP>
$
$
The feature detection test-all.bin gets successfully built and linked:
$ ls -la /tmp/build/perf/feature/test-all.bin
-rwxrwxr-x. 1 acme acme 2680352 Mar 19 11:07 /tmp/build/perf/feature/test-all.bin
$ nm /tmp/build/perf/feature/test-all.bin | grep -w disassembler 0000000000061f90 T disassembler
$
Time to move on to the patches that make use of this disassembler()
routine in binutils's libopcodes.
Filipe Manana [Tue, 19 Mar 2019 17:18:13 +0000 (17:18 +0000)]
Btrfs: fix assertion failure on fsync with NO_HOLES enabled
Back in commit a89ca6f24ffe4 ("Btrfs: fix fsync after truncate when
no_holes feature is enabled") I added an assertion that is triggered when
an inline extent is found to assert that the length of the (uncompressed)
data the extent represents is the same as the i_size of the inode, since
that is true most of the time I couldn't find or didn't remembered about
any exception at that time. Later on the assertion was expanded twice to
deal with a case of a compressed inline extent representing a range that
matches the sector size followed by an expanding truncate, and another
case where fallocate can update the i_size of the inode without adding
or updating existing extents (if the fallocate range falls entirely within
the first block of the file). These two expansion/fixes of the assertion
were done by commit 7ed586d0a8241 ("Btrfs: fix assertion on fsync of
regular file when using no-holes feature") and commit 6399fb5a0b69a
("Btrfs: fix assertion failure during fsync in no-holes mode").
These however missed the case where an falloc expands the i_size of an
inode to exactly the sector size and inline extent exists, for example:
$ mkfs.btrfs -f -O no-holes /dev/sdc
$ mount /dev/sdc /mnt
$ xfs_io -f -c "pwrite -S 0xab 0 1096" /mnt/foobar
wrote 1096/1096 bytes at offset 0
1 KiB, 1 ops; 0.0002 sec (4.448 MiB/sec and 4255.3191 ops/sec)
Updating the assertion again to allow for this particular case would result
in a meaningless assertion, plus there is currently no risk of logging
content that would result in any corruption after a log replay if the size
of the data encoded in an inline extent is greater than the inode's i_size
(which is not currently possibe either with or without compression),
therefore just remove the assertion.
Luo Jiaxing [Wed, 20 Mar 2019 10:21:34 +0000 (18:21 +0800)]
scsi: hisi_sas: Add softreset in hisi_sas_I_T_nexus_reset()
We found out that for v2 hw, a SATA disk can not be written to after the
system comes up.
In commit ffb1c820b8b6 ("scsi: hisi_sas: remove the check of sas_dev status
in hisi_sas_I_T_nexus_reset()"), we introduced a path where we may issue an
internal abort for a SATA device, but without following it with a
softreset.
We need to always follow an internal abort with a software reset, as per HW
programming flow, so add this.
Fixes: ffb1c820b8b6 ("scsi: hisi_sas: remove the check of sas_dev status in hisi_sas_I_T_nexus_reset()") Signed-off-by: Luo Jiaxing <[email protected]> Signed-off-by: John Garry <[email protected]> Signed-off-by: Martin K. Petersen <[email protected]>
Rasmus Villemoes [Tue, 12 Mar 2019 17:33:46 +0000 (18:33 +0100)]
irqchip/gic-v3-its: Fix comparison logic in lpi_range_cmp
The lpi_range_list is supposed to be sorted in ascending order of
->base_id (at least if the range merging is to work), but the current
comparison function returns a positive value if rb->base_id >
ra->base_id, which means that list_sort() will put A after B in that
case - and vice versa, of course.
Xin Long [Wed, 20 Mar 2019 06:49:38 +0000 (14:49 +0800)]
sctp: use memdup_user instead of vmemdup_user
In sctp_setsockopt_bindx()/__sctp_setsockopt_connectx(), it allocates
memory with addrs_size which is passed from userspace. We used flag
GFP_USER to put some more restrictions on it in Commit cacc06215271
("sctp: use GFP_USER for user-controlled kmalloc").
However, since Commit c981f254cc82 ("sctp: use vmemdup_user() rather
than badly open-coding memdup_user()"), vmemdup_user() has been used,
which doesn't check GFP_USER flag when goes to vmalloc_*(). So when
addrs_size is a huge value, it could exhaust memory and even trigger
oom killer.
This patch is to use memdup_user() instead, in which GFP_USER would
work to limit the memory allocation with a huge addrs_size.
Note we can't fix it by limiting 'addrs_size', as there's no demand
for it from RFC.
Linus Torvalds [Wed, 20 Mar 2019 18:01:52 +0000 (11:01 -0700)]
Merge tag 'arc-5.1-rc2' of git://git.kernel.org/pub/scm/linux/kernel/git/vgupta/arc
Pull ARC updates from Vineet Gupta:
- unaligned access support for HS cores
- Removed extra memory barrier around spinlock code
- HSDK platform updates: enable dmac, reset
- some more boot logging updates
- misc minor fixes
* tag 'arc-5.1-rc2' of git://git.kernel.org/pub/scm/linux/kernel/git/vgupta/arc:
arch: arc: Kconfig: pedantic formatting
ARCv2: spinlock: remove the extra smp_mb before lock, after unlock
ARC: unaligned: relax the check for gcc supporting -mno-unaligned-access
ARC: boot log: cut down on verbosity
ARCv2: boot log: refurbish HS core/release identification
arc: hsdk_defconfig: Enable CONFIG_BLK_DEV_RAM
ARC: u-boot args: check that magic number is correct
ARC: perf: bpok condition only exists for ARCompact
ARCv2: Add explcit unaligned access support (and ability to disable too)
ARCv2: lib: introduce memcpy optimized for unaligned access
ARC: [plat-hsdk]: Enable AXI DW DMAC support
ARC: [plat-hsdk]: Add reset controller handle to manage USB reset
ARC: DTB: [scripted] fix node name and address spelling
Christoph Paasch [Tue, 19 Mar 2019 06:14:52 +0000 (23:14 -0700)]
net/packet: Set __GFP_NOWARN upon allocation in alloc_pg_vec
Since commit fc62814d690c ("net/packet: fix 4gb buffer limit due to overflow check")
one can now allocate packet ring buffers >= UINT_MAX. However, syzkaller
found that that triggers a warning:
arm64: remove obsolete selection of MULTI_IRQ_HANDLER
The arm64 config selects MULTI_IRQ_HANDLER, which was renamed to
GENERIC_IRQ_MULTI_HANDLER by commit 4c301f9b6a94 ("ARM: Convert
to GENERIC_IRQ_MULTI_HANDLER"). The 'new' option is already
selected, so just remove the obsolete entry.
Ilya Dryomov [Wed, 20 Mar 2019 08:46:58 +0000 (09:46 +0100)]
libceph: wait for latest osdmap in ceph_monc_blacklist_add()
Because map updates are distributed lazily, an OSD may not know about
the new blacklist for quite some time after "osd blacklist add" command
is completed. This makes it possible for a blacklisted but still alive
client to overwrite a post-blacklist update, resulting in data
corruption.
Waiting for latest osdmap in ceph_monc_blacklist_add() and thus using
the post-blacklist epoch for all post-blacklist requests ensures that
all such requests "wait" for the blacklist to come into force on their
respective OSDs.
Ben Hutchings [Sun, 17 Mar 2019 01:17:56 +0000 (01:17 +0000)]
powerpc/mm: Only define MAX_PHYSMEM_BITS in SPARSEMEM configurations
MAX_PHYSMEM_BITS only needs to be defined if CONFIG_SPARSEMEM is
enabled, and that was the case before commit 4ffe713b7587
("powerpc/mm: Increase the max addressable memory to 2PB").
On 32-bit systems, where CONFIG_SPARSEMEM is not enabled, we now
define it as 46. That is larger than the real number of physical
address bits, and breaks calculations in zsmalloc:
mm/zsmalloc.c:130:49: warning: right shift count is negative
MAX(32, (ZS_MAX_PAGES_PER_ZSPAGE << PAGE_SHIFT >> OBJ_INDEX_BITS))
^~
...
mm/zsmalloc.c:253:21: error: variably modified 'size_class' at file scope
struct size_class *size_class[ZS_SIZE_CLASSES];
^~~~~~~~~~
Fixes: 4ffe713b7587 ("powerpc/mm: Increase the max addressable memory to 2PB") Cc: [email protected] # v4.20+ Signed-off-by: Ben Hutchings <[email protected]> Signed-off-by: Michael Ellerman <[email protected]>
Marco Felsch [Mon, 4 Mar 2019 10:49:40 +0000 (11:49 +0100)]
ARM: dts: pfla02: increase phy reset duration
Increase the reset duration to ensure correct phy functionality. The
reset duration is taken from barebox commit 52fdd510de ("ARM: dts:
pfla02: use long enough reset for ethernet phy"):
Use a longer reset time for ethernet phy Micrel KSZ9031RNX. Otherwise a
small percentage of modules have 'transmission timeouts' errors like
barebox@Phytec phyFLEX-i.MX6 Quad Carrier-Board:/ ifup eth0
warning: No MAC address set. Using random address 7e:94:4d:02:f8:f3
eth0: 1000Mbps full duplex link detected
eth0: transmission timeout
T eth0: transmission timeout
T eth0: transmission timeout
T eth0: transmission timeout
T eth0: transmission timeout
Taehee Yoo [Tue, 19 Mar 2019 04:22:41 +0000 (13:22 +0900)]
netfilter: nf_tables: add missing ->release_ops() in error path of newrule()
->release_ops() callback releases resources and this is used in error path.
If nf_tables_newrule() fails after ->select_ops(), it should release
resources. but it can not call ->destroy() because that should be called
after ->init().
At this point, ->release_ops() should be used for releasing resources.
Result:
Module Size Used by
xt_tcpudp 20480 2 <-- it should be 0
Fixes: b8e204006340 ("netfilter: nft_compat: use .release_ops and remove list of extension") Signed-off-by: Taehee Yoo <[email protected]> Signed-off-by: Pablo Neira Ayuso <[email protected]>
drm/vmwgfx: Don't double-free the mode stored in par->set_mode
When calling vmw_fb_set_par(), the mode stored in par->set_mode gets free'd
twice. The first free is in vmw_fb_kms_detach(), the second is near the
end of vmw_fb_set_par() under the name of 'old_mode'. The mode-setting code
only works correctly if the mode doesn't actually change. Removing
'old_mode' in favor of using par->set_mode directly fixes the problem.
This patch adds support for 6PE (RFC 4798) which uses IPv4-mapped IPv6
nexthop to connect IPv6 islands over IPv4 only MPLS network core.
Prior to this fix, to find the link-layer destination mac address, 6PE
enabled host/router was sending IPv6 ND requests for IPv4-mapped IPv6
nexthop address over the interface facing the IPv4 only core which
wouldn't success as the core is IPv6 free.
This fix changes that behavior on 6PE host to treat the nexthop as IPv4
address and send ARP requests whenever the next-hop address is an
IPv4-mapped IPv6 address.
Below topology illustrates the issue and how the patch addresses it.
R0 and R2 act as 6PE routers of IPv6 islands. R1 is IPv4 only with MPLS tunnels
between R0,R1 and R1,R2.
docker exec r0 ip -f inet6 route add abcd::2.2.2.2/128 nexthop encap mpls 100 via ::ffff:172.18.0.11 dev eth1
docker exec r2 ip -f inet6 route add abcd::1.1.1.1/128 nexthop encap mpls 200 via ::ffff:172.19.0.11 dev eth4
docker exec r1 ip -f mpls route add 100 via inet 172.19.0.12 dev eth3
docker exec r1 ip -f mpls route add 200 via inet 172.18.0.10 dev eth2
With the change, when R0 sends an IPv6 packet over MPLS tunnel to abcd::2.2.2.2,
using ::ffff:172.18.0.11 as the nexthop, it does neighbor discovery for
172.18.18.0.11.
Arnd Bergmann [Tue, 19 Mar 2019 21:15:58 +0000 (22:15 +0100)]
3c515: fix integer overflow warning
clang points out a harmless signed integer overflow:
drivers/net/ethernet/3com/3c515.c:1530:66: error: implicit conversion from 'int' to 'short' changes value from 32783 to -32753 [-Werror,-Wconstant-conversion]
new_mode = SetRxFilter | RxStation | RxMulticast | RxBroadcast | RxProm;
~ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~^~~~~~~~
drivers/net/ethernet/3com/3c515.c:1532:52: error: implicit conversion from 'int' to 'short' changes value from 32775 to -32761 [-Werror,-Wconstant-conversion]
new_mode = SetRxFilter | RxStation | RxMulticast | RxBroadcast;
~ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~^~~~~~~~~~~~~
drivers/net/ethernet/3com/3c515.c:1534:38: error: implicit conversion from 'int' to 'short' changes value from 32773 to -32763 [-Werror,-Wconstant-conversion]
new_mode = SetRxFilter | RxStation | RxBroadcast;
~ ~~~~~~~~~~~~~~~~~~~~~~~~^~~~~~~~~~~~~
Jiada Wang [Tue, 12 Mar 2019 06:51:28 +0000 (15:51 +0900)]
PM / Domains: Avoid a potential deadlock
Lockdep warns that prepare_lock and genpd->mlock can cause a deadlock
the deadlock scenario is like following:
First thread is probing cs2000
cs2000_probe()
clk_register()
__clk_core_init()
clk_prepare_lock() ----> acquires prepare_lock
cs2000_recalc_rate()
i2c_smbus_read_byte_data()
rcar_i2c_master_xfer()
dma_request_chan()
rcar_dmac_of_xlate()
rcar_dmac_alloc_chan_resources()
pm_runtime_get_sync()
__pm_runtime_resume()
rpm_resume()
rpm_callback()
genpd_runtime_resume() ----> acquires genpd->mlock
Second thread is attaching any device to the same PM domain
genpd_add_device()
genpd_lock() ----> acquires genpd->mlock
cpg_mssr_attach_dev()
of_clk_get_from_provider()
__of_clk_get_from_provider()
__clk_create_clk()
clk_prepare_lock() ----> acquires prepare_lock
Since currently no PM provider access genpd's critical section
in .attach_dev, and .detach_dev callbacks, so there is no need to protect
these two callbacks with genpd->mlock.
This patch avoids a potential deadlock by moving out .attach_dev and .detach_dev
from genpd->mlock, so that genpd->mlock won't be held when prepare_lock is acquired
in .attach_dev and .detach_dev
Andy Shevchenko [Mon, 18 Mar 2019 18:47:09 +0000 (21:47 +0300)]
ACPI / utils: Drop reference in test for device presence
When commit 8661423eea1a ("ACPI / utils: Add new acpi_dev_present
helper") introduced acpi_dev_present(), it missed the fact that
bus_find_device() took a reference on the device found by it and
the callers of acpi_dev_present() don't drop that reference.
Drop the reference on the device in acpi_dev_present().
Fixes: 8661423eea1a ("ACPI / utils: Add new acpi_dev_present helper") Signed-off-by: Andy Shevchenko <[email protected]> Reviewed-by: Hans de Goede <[email protected]> Signed-off-by: Rafael J. Wysocki <[email protected]>
Eric Dumazet [Tue, 19 Mar 2019 12:45:35 +0000 (05:45 -0700)]
tcp: do not use ipv6 header for ipv4 flow
When a dual stack tcp listener accepts an ipv4 flow,
it should not attempt to use an ipv6 header or tcp_v6_iif() helper.
Fixes: 1397ed35f22d ("ipv6: add flowinfo for tcp6 pkt_options for all cases") Fixes: df3687ffc665 ("ipv6: add the IPV6_FL_F_REFLECT flag to IPV6_FL_A_GET") Fixes: 1da177e4c3f4 ("Linux-2.6.12-rc2") Signed-off-by: Eric Dumazet <[email protected]> Signed-off-by: David S. Miller <[email protected]>
YueHaibing [Tue, 19 Mar 2019 02:16:53 +0000 (10:16 +0800)]
net-sysfs: call dev_hold if kobject_init_and_add success
In netdev_queue_add_kobject and rx_queue_add_kobject,
if sysfs_create_group failed, kobject_put will call
netdev_queue_release to decrease dev refcont, however
dev_hold has not be called. So we will see this while
unregistering dev:
unregister_netdevice: waiting for bcsh0 to become free. Usage count = -1
Reported-by: Hulk Robot <[email protected]> Fixes: d0d668371679 ("net: don't decrement kobj reference count on init failure") Signed-off-by: YueHaibing <[email protected]> Signed-off-by: David S. Miller <[email protected]>
Aaro Koskinen [Mon, 18 Mar 2019 21:36:08 +0000 (23:36 +0200)]
net: stmmac: fix memory corruption with large MTUs
When using 16K DMA buffers and ring mode, the DES3 refill is not working
correctly as the function is using a bogus pointer for checking the
private data. As a result stale pointers will remain in the RX descriptor
ring, so DMA will now likely overwrite/corrupt some already freed memory.
As simple reproducer, just receive some UDP traffic:
# ifconfig eth0 down; ifconfig eth0 mtu 9000; ifconfig eth0 up
# iperf3 -c 192.168.253.40 -u -b 0 -R
If you didn't crash by now check the RX descriptors to find non-contiguous
RX buffers:
# ping -s 8200 192.168.253.40
PING 192.168.253.40 (192.168.253.40) 8200(8228) bytes of data.
8208 bytes from 192.168.253.40: icmp_seq=1 ttl=64 time=1.00 ms
wrong data byte #8144 should be 0xd0 but was 0x88
Fix the wrong pointer. Also we must refill DES3 only if the DMA buffer
size is 16K.
Fixes: 54139cf3bb33 ("net: stmmac: adding multiple buffers for rx") Signed-off-by: Aaro Koskinen <[email protected]> Acked-by: Jose Abreu <[email protected]> Signed-off-by: David S. Miller <[email protected]>
A recently added function in mlxsw triggers a harmless compiler warning:
In file included from drivers/net/ethernet/mellanox/mlxsw/core.h:17,
from drivers/net/ethernet/mellanox/mlxsw/core_env.c:7:
drivers/net/ethernet/mellanox/mlxsw/core_env.c: In function 'mlxsw_env_module_temp_thresholds_get':
drivers/net/ethernet/mellanox/mlxsw/reg.h:8015:45: error: '*' in boolean context, suggest '&&' instead [-Werror=int-in-bool-context]
#define MLXSW_REG_MTMP_TEMP_TO_MC(val) (val * 125)
~~~~~^~~~~~
drivers/net/ethernet/mellanox/mlxsw/core_env.c:116:8: note: in expansion of macro 'MLXSW_REG_MTMP_TEMP_TO_MC'
if (!MLXSW_REG_MTMP_TEMP_TO_MC(module_temp)) {
^~~~~~~~~~~~~~~~~~~~~~~~~
The warning is normally disabled, but it would be nice to enable
it to find real bugs, and there are no other known instances at
the moment.
Replace the negation with a zero-comparison, which also matches
the comment above it.
Fixes: d93c19a1d95c ("mlxsw: core: Add API for QSFP module temperature thresholds reading") Signed-off-by: Arnd Bergmann <[email protected]> Acked-by: Jiri Pirko <[email protected]> Signed-off-by: David S. Miller <[email protected]>
Song Liu [Tue, 12 Mar 2019 05:30:49 +0000 (22:30 -0700)]
perf bpf: Process PERF_BPF_EVENT_PROG_LOAD for annotation
This patch adds processing of PERF_BPF_EVENT_PROG_LOAD, which sets
proper DSO type/id/etc of memory regions mapped to BPF programs to
DSO_BINARY_TYPE__BPF_PROG_INFO.
Introduce a new dso type DSO_BINARY_TYPE__BPF_PROG_INFO for BPF programs. In
symbol__disassemble(), DSO_BINARY_TYPE__BPF_PROG_INFO dso will call into a new
function symbol__disassemble_bpf() in an upcoming patch, where annotation line
information is filled based bpf_prog_info and btf saved in given perf_env.
Committer notes:
Removed the unnamed union with 'bpf_prog' and 'cache' in 'struct dso',
to fix this bug when exiting 'perf top':
Song Liu [Tue, 12 Mar 2019 05:30:47 +0000 (22:30 -0700)]
perf feature detection: Add -lopcodes to feature-libbfd
Both libbfd and libopcodes are distributed with binutil-dev/devel. When
libbfd is present, it is OK to assume that libopcodes also present. This
has been a safe assumption for bpftool.
This patch adds -lopcodes to perf/Makefile.config. libopcodes will be
used in the next commit for BPF annotation.
Just to compile and load a BPF program that attaches to the
raw_syscalls:sys_{enter,exit} tracepoints to trace the syscalls ending
in "msg" (recvmsg, sendmsg, recvmmsg, sendmmsg, etc).
Make sure you have a recent enough clang, say version 9, to get the
BTF ELF sections needed for this testing:
Song Liu [Tue, 12 Mar 2019 05:30:43 +0000 (22:30 -0700)]
perf bpf: Save bpf_prog_info information as headers to perf.data
This patch enables perf-record to save bpf_prog_info information as
headers to perf.data. A new header type HEADER_BPF_PROG_INFO is
introduced for this data.
Committer testing:
As root, being on the kernel sources top level directory, run:
Just to compile and load a BPF program that attaches to the
raw_syscalls:sys_{enter,exit} tracepoints to trace the syscalls ending
in "msg" (recvmsg, sendmsg, recvmmsg, sendmmsg, etc).
Then do a systemwide perf record session for a few seconds:
# perf record -a sleep 2s
Then look at:
# perf report --header-only | grep -i bpf
# bpf_prog_info of id 13
# bpf_prog_info of id 14
# bpf_prog_info of id 15
# bpf_prog_info of id 16
# bpf_prog_info of id 17
# bpf_prog_info of id 18
# bpf_prog_info of id 21
# bpf_prog_info of id 22
# bpf_prog_info of id 208
# bpf_prog_info of id 209
#
We need to show more info about these programs, like bpftool does for
the ones running on the system, i.e. 'perf record/perf report' become a
way of saving the BPF state in a machine to then analyse on another,
together with all the other information that is already saved in the
perf.data header:
# perf report --header-only
# ========
# captured on : Tue Mar 12 11:42:13 2019
# header version : 1
# data offset : 296
# data size : 16294184
# feat offset : 16294480
# hostname : quaco
# os release : 5.0.0+
# perf version : 5.0.gd783c8
# arch : x86_64
# nrcpus online : 8
# nrcpus avail : 8
# cpudesc : Intel(R) Core(TM) i7-8650U CPU @ 1.90GHz
# cpuid : GenuineIntel,6,142,10
# total memory : 24555720 kB
# cmdline : /home/acme/bin/perf (deleted) record -a
# event : name = cycles:ppp, , id = { 3190123, 3190124, 3190125, 3190126, 3190127, 3190128, 3190129, 3190130 }, size = 112, { sample_period, sample_freq } = 4000, sample_type = IP|TID|TIME|CPU|PERIOD, read_format = ID, disabled = 1, inherit = 1, mmap = 1, comm = 1, freq = 1, task = 1, precise_ip = 3, sample_id_all = 1, exclude_guest = 1, mmap2 = 1, comm_exec = 1
# CPU_TOPOLOGY info available, use -I to display
# NUMA_TOPOLOGY info available, use -I to display
# pmu mappings: intel_pt = 8, software = 1, power = 11, uprobe = 7, uncore_imc = 12, cpu = 4, cstate_core = 18, uncore_cbox_2 = 15, breakpoint = 5, uncore_cbox_0 = 13, tracepoint = 2, cstate_pkg = 19, uncore_arb = 17, kprobe = 6, i915 = 10, msr = 9, uncore_cbox_3 = 16, uncore_cbox_1 = 14
# CACHE info available, use -I to display
# time of first sample : 116392.441701
# time of last sample : 116400.932584
# sample duration : 8490.883 ms
# MEM_TOPOLOGY info available, use -I to display
# bpf_prog_info of id 13
# bpf_prog_info of id 14
# bpf_prog_info of id 15
# bpf_prog_info of id 16
# bpf_prog_info of id 17
# bpf_prog_info of id 18
# bpf_prog_info of id 21
# bpf_prog_info of id 22
# bpf_prog_info of id 208
# bpf_prog_info of id 209
# missing features: TRACING_DATA BRANCH_STACK GROUP_DESC AUXTRACE STAT CLOCKID DIR_FORMAT
# ========
#
Committer notes:
We can't use the libbpf unconditionally, as the build may have been with
NO_LIBBPF, when we end up with linking errors, so provide dummy
{process,write}_bpf_prog_info() wrapped by HAVE_LIBBPF_SUPPORT for that
case.
Printing are not affected by this, so can continue as is.
Before this set, 1) and 2) in the list are already saved to perf.data
file. For BPF programs that are already loaded before perf run, 1) and 2)
are synthesized by perf_event__synthesize_bpf_events(). For short living
BPF programs, 1) and 2) are generated by kernel.
This set handles 3) and 4) from the list. Again, it is necessary to handle
existing BPF program and short living program separately.
This patch handles 3) for exising BPF programs while synthesizing 1) and
2) in perf_event__synthesize_bpf_events(). These data are stored in
perf_env. The next patch saves these data from perf_env to perf.data as
headers.
Similarly, the two patches after the next saves 4) of existing BPF
programs to perf_env and perf.data.
Another patch later will handle 3) and 4) for short living BPF programs
by monitoring 1) and 2) in a dedicate thread.
Song Liu [Tue, 12 Mar 2019 05:30:41 +0000 (22:30 -0700)]
perf bpf: Make synthesize_bpf_events() receive perf_session pointer instead of perf_tool
This patch changes the arguments of perf_event__synthesize_bpf_events()
to include perf_session* instead of perf_tool*. perf_session will be
used in the next patch.
util/bpf-event.c: In function 'perf_event__synthesize_one_bpf_prog':
util/bpf-event.c:143:35: error: cast to pointer from integer of different size [-Werror=int-to-pointer-cast]
u8 (*prog_tags)[BPF_TAG_SIZE] = (void *)(info->prog_tags);
^
util/bpf-event.c:144:22: error: cast to pointer from integer of different size [-Werror=int-to-pointer-cast]
__u32 *prog_lens = (__u32 *)(info->jited_func_lens);
^
util/bpf-event.c:145:23: error: cast to pointer from integer of different size [-Werror=int-to-pointer-cast]
__u64 *prog_addrs = (__u64 *)(info->jited_ksyms);
^
util/bpf-event.c:146:22: error: cast to pointer from integer of different size [-Werror=int-to-pointer-cast]
void *func_infos = (void *)(info->func_info);
^
cc1: all warnings being treated as errors
projects / linux.git / log