If enabled, L2TP data packets have sequence numbers which a receiver
can use to drop out of sequence frames or try to reorder them. The
first frame has sequence number 0, but the L2TP code currently expects
it to be 1. This results in the first data frame being handled as out
of sequence.
This one-line patch fixes the problem.
Signed-off-by: James Chapman <jchapman@katalix.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
When L2TP data packet reordering is enabled, packets are held in a
queue while waiting for out-of-sequence packets. If a packet gets
lost, packets will be held until the reorder timeout expires, when we
are supposed to then advance to the sequence number of the next packet
but we don't currently do so. As a result, the data channel is stuck
because we are waiting for a packet that will never arrive - all
packets age out and none are passed.
The fix is to add a flag to the session context, which is set when the
reorder timeout expires and tells the receive code to reset the next
expected sequence number to that of the next packet in the queue.
Tested in a production L2TP network with Starent and Nortel L2TP gear.
Signed-off-by: James Chapman <jchapman@katalix.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
As proposed by Eric, make the tcp_input.o thinner.
add/remove: 1/1 grow/shrink: 1/4 up/down: 868/-1329 (-461)
function old new delta
tcp_try_rmem_schedule - 864 +864
tcp_ack 4811 4815 +4
tcp_validate_incoming 817 815 -2
tcp_collapse 860 858 -2
tcp_send_rcvq 555 353 -202
tcp_data_queue 3435 3033 -402
tcp_prune_queue 721 - -721
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
Acked-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
As noted by Eric, no checks are performed on the data size we're
putting in the read queue during repair. Thus, validate the given
data size with the common rmem management routine.
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
Acked-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
It actually works on the input queue and will use its read mem
routines, thus it's better to have in in the tcp_input.c file.
Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
Acked-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
In 2009 Philip Reiser notied that a few users of netlink connector
interface needed a capability check and added the idiom
cap_raised(nsp->eff_cap, CAP_SYS_ADMIN) to a few of them, on the premise
that netlink was asynchronous.
In 2011 Patrick McHardy noticed we were being silly because netlink is
synchronous and removed eff_cap from the netlink_skb_params and changed
the idiom to cap_raised(current_cap(), CAP_SYS_ADMIN).
Looking at those spots with a fresh eye we should be calling
capable(CAP_SYS_ADMIN). The only reason I can see for not calling capable
is that it once appeared we were not in the same task as the caller which
would have made calling capable() impossible.
In the initial user_namespace the only difference between between
cap_raised(current_cap(), CAP_SYS_ADMIN) and capable(CAP_SYS_ADMIN) are a
few sanity checks and the fact that capable(CAP_SYS_ADMIN) sets
PF_SUPERPRIV if we use the capability.
Since we are going to be using root privilege setting PF_SUPERPRIV seems
the right thing to do.
The motivation for this that patch is that in a child user namespace
cap_raised(current_cap(),...) tests your capabilities with respect to that
child user namespace not capabilities in the initial user namespace and
thus will allow processes that should be unprivielged to use the kernel
services that are only protected with cap_raised(current_cap(),..).
To fix possible user_namespace issues and to just clean up the code
replace cap_raised(current_cap(), CAP_SYS_ADMIN) with
capable(CAP_SYS_ADMIN).
Signed-off-by: Eric W. Biederman <ebiederm@xmission.com>
Cc: Patrick McHardy <kaber@trash.net>
Cc: Philipp Reisner <philipp.reisner@linbit.com>
Acked-by: Serge E. Hallyn <serge.hallyn@canonical.com>
Acked-by: Andrew G. Morgan <morgan@kernel.org>
Cc: Vasiliy Kulikov <segoon@openwall.com>
Cc: David Howells <dhowells@redhat.com>
Reviewed-by: James Morris <james.l.morris@oracle.com>
Cc: David Miller <davem@davemloft.net>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
dst_check() will take care of SA (and obsolete field), hence
IPsec rekeying scenario is taken into account.
Signed-off-by: Nicolas Dichtel <nicolas.dichtel@6wind.com>
Acked-by: Vlad Yaseivch <vyasevich@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
This reverts commit 8a83a00b07.
It causes regressions for S390 devices, because it does an
unconditional DST drop on SKBs for vlans and the QETH device
needs the neighbour entry hung off the DST for certain things
on transmit.
Arnd can't remember exactly why he even needed this change.
Conflicts:
drivers/net/macvlan.c
net/8021q/vlan_dev.c
net/core/dev.c
Signed-off-by: David S. Miller <davem@davemloft.net>
The NEQ interrupt is only triggered when there was no previous pending
interrupt. If we request irq handling after an interrupt has occurred,
we will never get an interrupt until we call H_RESET_EVENTS.
Events seem to be cleared when we first register the NEQ. So, when we
requested irq handling right after registering it, a possible race with
an interrupt was much less likely. Now, there is a chance we may lose
this race and never get any events.
The fix here is to poll and acknowledge any events that might have
happened right after registering the irq handler.
Signed-off-by: Thadeu Lima de Souza Cascardo <cascardo@linux.vnet.ibm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Since the caller (PM resume code) is not the one holding rtnl, when taking the
'else' branch rtnl may be released at any moment, thereby defeating the whole
purpose of this code block.
Signed-off-by: Benjamin Poirier <bpoirier@suse.de>
Tested-by: Jeff Pieper <jeffrey.e.pieper@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Due to RCU lookups and RCU based release, fib_info objects can
be found during lookup which have fi->fib_dead set.
We must ignore these entries, otherwise we risk dereferencing
the parts of the entry which are being torn down.
Reported-by: Yevgen Pronenko <yevgen.pronenko@sonymobile.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Merge PA-RISC compile fixes from Rolf Eike Beer:
"Since commit d66acc39c7 ("bitops: Optimise get_order()") getorder.h
includes log2.h which leads to an include loop on PA-RISC, bringing a
bunch of other breakage to light. This patchset fixes the compilation
of the current state of 3.4 on HPPA.
Unchanged against the first version, just added an Ack by Grant."
* emailed from Rolf Eike Beer <eike-kernel@sf-tec.de>: (5 patches)
parisc: move definition of PAGE0 to asm/page.h
parisc: add missing include of asm/page.h to asm/pgtable.h
parisc: drop include of asm/pdc.h from asm/hardware.h
parisc: add missing forward declarations in asm/hardware.h
parisc: add missing includes in asm/spinlock.h
Merge misc fixes from Andrew Morton.
* emailed from Andrew Morton <akpm@linux-foundation.org>: (8 patches)
MAINTAINERS: add maintainer for LED subsystem
mm: nobootmem: fix sign extend problem in __free_pages_memory()
drivers/leds: correct __devexit annotations
memcg: free spare array to avoid memory leak
namespaces, pid_ns: fix leakage on fork() failure
hugetlb: prevent BUG_ON in hugetlb_fault() -> hugetlb_cow()
mm: fix division by 0 in percpu_pagelist_fraction()
proc/pid/pagemap: correctly report non-present ptes and holes between vmas
This was defined in asm/pdc.h which needs to include asm/page.h for
__PAGE_OFFSET. This leads to an include loop so that page.h eventually will
include pdc.h again. While this is no problem because of header guards, it is
a problem because some symbols may be undefined. Such an error is this:
In file included from include/linux/bitops.h:35:0,
from include/asm-generic/getorder.h:7,
from arch/parisc/include/asm/page.h:162,
from arch/parisc/include/asm/pdc.h:346,
from arch/parisc/include/asm/processor.h:16,
from arch/parisc/include/asm/spinlock.h:6,
from arch/parisc/include/asm/atomic.h:20,
from include/linux/atomic.h:4,
from include/linux/sysfs.h:20,
from include/linux/kobject.h:21,
from include/linux/device.h:17,
from include/linux/eisa.h:5,
from arch/parisc/kernel/pci.c:11:
arch/parisc/include/asm/bitops.h: In function ‘set_bit’:
arch/parisc/include/asm/bitops.h:82:2: error: implicit declaration of function ‘_atomic_spin_lock_irqsave’ [-Werror=implicit-function-declaration]
arch/parisc/include/asm/bitops.h:84:2: error: implicit declaration of function ‘_atomic_spin_unlock_irqrestore’ [-Werror=implicit-function-declaration]
Signed-off-by: Rolf Eike Beer <eike-kernel@sf-tec.de>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Fixes these errors:
In file included from arch/parisc/include/asm/io.h:5:0,
from include/linux/io.h:22,
from include/linux/pci.h:54,
from arch/parisc/kernel/setup.c:35:
arch/parisc/include/asm/pgtable.h:92:6: warning: "PAGE_SHIFT" is not defined [-Wundef]
arch/parisc/include/asm/pgtable.h:92:6: warning: "PAGE_SHIFT" is not defined [-Wundef]
arch/parisc/include/asm/pgtable.h:92:6: warning: "BITS_PER_PTE_ENTRY" is not defined [-Wundef]
Signed-off-by: Rolf Eike Beer <eike-kernel@sf-tec.de>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
It seems none of the symbols defined by pdc.h is needed, but it introduces an
include loop causing compile errors:
In file included from arch/parisc/include/asm/spinlock.h:4:0,
from arch/parisc/include/asm/atomic.h:20,
from include/linux/atomic.h:4,
from arch/parisc/include/asm/bitops.h:56,
from include/linux/bitops.h:35,
from include/asm-generic/getorder.h:7,
from arch/parisc/include/asm/page.h:162,
from arch/parisc/include/asm/pdc.h:346,
from arch/parisc/include/asm/hardware.h:5,
from arch/parisc/kernel/hardware.c:30:
arch/parisc/include/asm/processor.h:74:16: error: field ‘cpu_type’ has incomplete type
arch/parisc/include/asm/processor.h:77:20: error: field ‘model’ has incomplete type
arch/parisc/include/asm/processor.h: In function ‘parisc_requires_coherency’:
arch/parisc/include/asm/processor.h:349:36: error: ‘mako’ undeclared (first use in this function)
arch/parisc/include/asm/processor.h:349:36: note: each undeclared identifier is reported only once for each function it appears in
arch/parisc/include/asm/processor.h:350:30: error: ‘mako2’ undeclared (first use in this function)
Signed-off-by: Rolf Eike Beer <eike-kernel@sf-tec.de>
Acked-by: Grant Grundler <grantgrundler@gmail.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Fixes this warnings:
In file included from arch/parisc/include/asm/processor.h:15:0,
from arch/parisc/include/asm/spinlock.h:4,
from arch/parisc/include/asm/atomic.h:20,
from include/linux/atomic.h:4,
from arch/parisc/include/asm/bitops.h:11,
from include/linux/bitops.h:22,
from include/linux/kernel.h:19,
from include/linux/sched.h:55,
from arch/parisc/kernel/asm-offsets.c:31:
arch/parisc/include/asm/hardware.h:106:10: warning: ‘struct hardware_path’ declared inside parameter list [enabled by default]
arch/parisc/include/asm/hardware.h:106:10: warning: its scope is only this definition or declaration, which is probably not what you want [enabled by default]
arch/parisc/include/asm/hardware.h:116:59: warning: ‘struct hardware_path’ declared inside parameter list [enabled by default]
arch/parisc/include/asm/hardware.h:118:47: warning: ‘struct hardware_path’ declared inside parameter list [enabled by default]
arch/parisc/include/asm/hardware.h:119:57: warning: ‘struct hardware_path’ declared inside parameter list [enabled by default]
Signed-off-by: Rolf Eike Beer <eike-kernel@sf-tec.de>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
This leads to this errors:
In file included from arch/parisc/include/asm/atomic.h:20:0,
from include/linux/atomic.h:4,
from arch/parisc/include/asm/bitops.h:56,
from include/linux/bitops.h:22,
from include/linux/kernel.h:19,
from include/linux/sched.h:55,
from arch/parisc/kernel/asm-offsets.c:31:
arch/parisc/include/asm/spinlock.h: In function ‘arch_spin_is_locked’:
arch/parisc/include/asm/spinlock.h:9:2: error: implicit declaration of function ‘__ldcw_align’ [-Werror=implicit-function-declaration]
arch/parisc/include/asm/spinlock.h:9:29: warning: initialization makes pointer from integer without a cast [enabled by default]
arch/parisc/include/asm/spinlock.h: In function ‘arch_spin_lock_flags’:
arch/parisc/include/asm/spinlock.h:22:2: error: implicit declaration of function ‘mb’ [-Werror=implicit-function-declaration]
arch/parisc/include/asm/spinlock.h:23:4: warning: assignment makes pointer from integer without a cast [enabled by default]
arch/parisc/include/asm/spinlock.h:24:2: error: implicit declaration of function ‘__ldcw’ [-Werror=implicit-function-declaration]
Signed-off-by: Rolf Eike Beer <eike-kernel@sf-tec.de>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Pull a sparc fix from David Miller.
* git://git.kernel.org/pub/scm/linux/kernel/git/davem/sparc:
sparc64: Do not clobber %g2 in xcall_fetch_glob_regs().
Commit 66aebce747 ("hugetlb: fix race condition in hugetlb_fault()")
added code to avoid a race condition by elevating the page refcount in
hugetlb_fault() while calling hugetlb_cow().
However, one code path in hugetlb_cow() includes an assertion that the
page count is 1, whereas it may now also have the value 2 in this path.
The consensus is that this BUG_ON has served its purpose, so rather than
extending it to cover both cases, we just remove it.
Signed-off-by: Chris Metcalf <cmetcalf@tilera.com>
Acked-by: Mel Gorman <mel@csn.ul.ie>
Acked-by: Hillf Danton <dhillf@gmail.com>
Acked-by: Hugh Dickins <hughd@google.com>
Cc: Michal Hocko <mhocko@suse.cz>
Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Cc: <stable@vger.kernel.org> [3.0.29+, 3.2.16+, 3.3.3+]
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
percpu_pagelist_fraction_sysctl_handler() has only considered -EINVAL as
a possible error from proc_dointvec_minmax().
If any other error is returned, it would proceed to divide by zero since
percpu_pagelist_fraction wasn't getting initialized at any point. For
example, writing 0 bytes into the proc file would trigger the issue.
Signed-off-by: Sasha Levin <levinsasha928@gmail.com>
Reviewed-by: Minchan Kim <minchan@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Pull percpu fixes from Tejun Heo:
"This pull request contains two patches. One is kmemleak annotation
fix which isn't critical. The other is kinda serious.
Depending on NUMA topology, percpu allocator may end up assigning
overlapping regions for the static percpu areas for different CPUs.
While critical, the bug has been there for a very long time and only
few configurations seem to be affected (NUMA configurations w/ no
memory nodes for example) - so, while it's critical, it isn't exactly
urgent."
* 'for-3.4-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/tj/percpu:
kmemleak: Fix the kmemleak tracking of the percpu areas with !SMP
percpu: pcpu_embed_first_chunk() should free unused parts after all allocs are complete
%g2 is meant to hold the CPUID number throughout this routine, since
at the very beginning, and at the very end, we use %g2 to calculate
indexes into per-cpu arrays.
However we erroneously clobber it in order to hold the %cwp register
value mid-stream.
Fix this code to use %g3 for the %cwp read and related calulcations
instead.
Reported-by: Meelis Roos <mroos@linux.ee>
Signed-off-by: David S. Miller <davem@davemloft.net>
Pull a NFS client bugfix from Trond Myklebust:
"Fix for the NFSv4 security negotiation: ensure that the security
negotiation tries all registered security flavours"
* tag 'nfs-for-3.4-5' of git://git.linux-nfs.org/projects/trondmy/linux-nfs:
auth_gss: the list of pseudoflavors not being parsed correctly
Pull sound fixes from Takashi Iwai:
"Slightly more than expected as rc7, but all are reasonablly small
fixes. A few additions of HD-audio fixup entries, a couple of other
regression fixes including a revert, and a few other trivial
oneliners."
* tag 'sound-3.4' of git://git.kernel.org/pub/scm/linux/kernel/git/tiwai/sound:
ASoC: sh: fix migor.c compilation
ALSA: HDA: Lessen CPU usage when waiting for chip to respond
Revert "ALSA: hda - Set codec to D3 forcibly even if not used"
ALSA: hda/realtek - Call alc_auto_parse_customize_define() always after fixup
ALSA: hdsp - Provide ioctl_compat
ALSA: hda/realtek - Add missing CD-input pin for MSI-7350 mobo
ALSA: hda/realtek - Add a fixup for Acer Aspire 5739G
ALSA: echoaudio: Remove incorrect part of assertion
compat_sys_sigprocmask reads a smaller signal mask from userspace than
sigprogmask accepts for setting. So the high word of blocked.sig[0]
will be cleared, releasing any potentially blocked RT signal.
This was discovered via userspace code that relies on get/setcontext.
glibc's i386 versions of those functions use sigprogmask instead of
rt_sigprogmask to save/restore signal mask and caused RT signal
unblocking this way.
As suggested by Linus, this replaces the sys_sigprocmask based compat
version with one that open-codes the required logic, including the merge
of the existing blocked set with the new one provided on SIG_SETMASK.
Signed-off-by: Jan Kiszka <jan.kiszka@siemens.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Initialization of irqenable, irqstatus registers is the common
operation done in this function for all OMAP platforms, viz. OMAP1,
OMAP2+. The latter _gpio_rmw()'s which supposedly got introduced
wrongly to take care of OMAP2+ platforms were overwriting initially
programmed OMAP1 value breaking functionality on OMAP1.
Somehow incorrect assumption was made that each _gpio_rmw()'s were
mutually exclusive. On close observation it is found that the first
_gpio_rmw() which is supposedly done to take care of OMAP1 platform
is generic enough and takes care of OMAP2+ platform as well.
Therefore remove the latter _gpio_rmw() to irqenable as they are
redundant now.
Writing to ctrl and debounce_en registers for OMAP2+ platforms are
modified to match the original(pre-cleanup) code where the registers
are initialized with 0. In the cleanup series since we are using
_gpio_rmw(reg, 0, 1), instead of __raw_writel(), we are just reading
and writing the same values to ctrl and debounce_en. This is not an
issue for debounce_en register because it has 0x0 as the default value.
But in the case of ctrl register the default value is 0x2 (GATINGRATIO
= 0x1) so that we end up writing 0x2 instead of intended 0 value.
Therefore changing back to __raw_writel() as this is sufficient for
this case besides simpler to understand.
Also, change irqstatus initalization logic that avoids comparison
with bool, besides making it fit in a single line.
Cc: stable@vger.kernel.org
Cc: Tony Lindgren <tony@atomide.com>
Cc: Kevin Hilman <khilman@ti.com>
Cc: Santosh Shilimkar <santosh.shilimkar@ti.com>
Cc: Grant Likely <grant.likely@secretlab.ca>
Reported-by: Janusz Krzysztofik <jkrzyszt@tis.icnet.pl>
Tested-by: Janusz Krzysztofik <jkrzyszt@tis.icnet.pl>
Signed-off-by: Tarun Kanti DebBarma <tarun.kanti@ti.com>
Signed-off-by: Kevin Hilman <khilman@ti.com>
Fix a use-after-free in the TMF path, where cmd may have been already
freed by virtscsi_complete_free when wait_for_completion restarts
executing virtscsi_tmf. Technically a race, but in practice the command
will always be freed long before the completion waiter is awoken.
The fix is to make callers specifying a completion responsible for
freeing the command in all cases.
Signed-off-by: Hu Tao <hutao@cn.fujitsu.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: James Bottomley <JBottomley@Parallels.com>
Commit 6f381fa344
Author: Lin Ming <ming.m.lin@intel.com>
[SCSI] scsi_lib: use correct DMA device in __scsi_alloc_queue
Caused a regression where we oops in every legacy mode SCSI host driver
because they supply a NULL pointer to scsi_add_host(). Fix this by checking
for the NULL in scsi_add_host_with_dma() and changing the DMA device to being
the platform_bus in that case (which replicates the original behaviour).
Reported-by: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: James Bottomley <JBottomley@Parallels.com>
In case of firmmware detected under-run condition and scsi status of
task_set_full or busy_condition, return that to the mid layer for proper error
handling instead of DID_ERROR (which causes error handler activation and a
full retry).
Signed-off-by: Giridhar Malavali <giridhar.malavali@qlogic.com>
Signed-off-by: Chad Dupuis <chad.dupuis@qlogic.com>
Signed-off-by: James Bottomley <JBottomley@Parallels.com>
When the hwmon code was initially added it was with the assumption that a
sysfs patch would be also coming soon. Since that isn't the case some
clean up needs to be done. This patch does that.
Signed-off-by: Don Skidmore <donald.c.skidmore@intel.com>
Tested-by: Stephen Ko <stephen.s.ko@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Kernel software timestamping requires that the driver calls skb_tx_timestamp
just before passing the skb to the MAC, in order to provide the best software
timestamps. This patch adds this call for that support.
Signed-off-by: Jacob Keller <jacob.e.keller@intel.com>
Tested-by: Phil Schmitt <phillip.j.schmitt@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
This patch adds support for the ethtool get_ts_info operation, which enables
access of available timestamp/timesync support for that device. It can query
which ptp clock device is associated with the particular port.
Signed-off-by: Jacob Keller <jacob.e.keller@intel.com>
Tested-by: Stephen Ko <stephen.s.ko@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
The current value of the udelay timeout for ixgbe_disable_rx_buff is too
short. This causes the security path to not not be properly disabled during
the section that is meant to have it turned off. The end result causes a race
condition that results in RX issues.
Signed-off-by: Jacob Keller <jacob.e.keller@intel.com>
Tested-by: Phil Schmitt <phillip.j.schmitt@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
This patch enables the PPS system in the PHC framework, by enabling
the clock-out feature on the X540 device. Causes the SDP0 to be set as
a 1Hz clock. Also configures the timesync interrupt cause in order to
report each pulse to the PPS via the PHC framework, which can be used
for general system clock synchronization. (This allows a stable method
for tuning the general system time via the on-board SYSTIM register
based clock.)
Signed-off-by: Jacob E Keller <jacob.e.keller@intel.com>
Tested-by: Stephen Ko <stephen.s.ko@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
This patch enables hardware timestamping for use with PTP software by
extracting a ns counter from an arbitrary fixed point cycles counter.
The hardware generates SYSTIME registers using the DMA tick which
changes based on the current link speed. These SYSTIME registers are
converted to ns using the cyclecounter and timecounter structures
provided by the kernel. Using the SO_TIMESTAMPING api, software can
enable and access timestamps for PTP packets.
The SO_TIMESTAMPING API has space for 3 different kinds of timestamps,
SYS, RAW, and SOF. SYS hardware timestamps are hardware ns values that
are then scaled to the software clock. RAW hardware timestamps are the
direct raw value of the ns counter. SOF software timestamps are the
software timestamp calculated as close as possible to the software
transmit, but are not offloaded to the hardware. This patch only
supports the RAW hardware timestamps due to inefficiency of the SYS
design.
This patch also enables the PHC subsystem features for atomically
adjusting the cycle register, and adjusting the clock frequency in
parts per billion. This frequency adjustment works by slightly
adjusting the value added to the cycle registers each DMA tick. This
causes the hardware registers to overflow rapidly (approximately once
every 34 seconds, when at 10gig link). To solve this, the timecounter
structure is used, along with a timer set for every 25 seconds. This
allows for detecting register overflow and converting the cycle
counter registers into ns values needed for providing useful
timestamps to the network stack.
Only the basic required clock functions are supported at this time,
although the hardware supports some ancillary features and these could
easily be enabled in the future.
Note that use of this hardware timestamping requires modifying daemon
software to use the SO_TIMESTAMPING API for timestamps, and the
ptp_clock PHC framework for accessing the clock. The timestamps have
no relation to the system time at all, so software must use the posix
clock generated by the PHC framework instead.
Signed-off-by: Jacob E Keller <jacob.e.keller@intel.com>
Tested-by: Stephen Ko <stephen.s.ko@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
