* [PATCH 09/12] ptp: Added a clock that uses the Linux system time.
From: Richard Cochran @ 2010-06-15 16:09 UTC (permalink / raw)
To: netdev; +Cc: linuxppc-dev, devicetree-discuss, Krzysztof Halasa,
linux-arm-kernel
In-Reply-To: <cover.1276615626.git.richard.cochran@omicron.at>
This PTP clock simply uses the NTP time adjustment system calls. The
driver can be used in systems that lack a special hardware PTP clock.
Signed-off-by: Richard Cochran <richard.cochran@omicron.at>
---
drivers/ptp/Kconfig | 12 ++++
drivers/ptp/Makefile | 1 +
drivers/ptp/ptp_linux.c | 136 +++++++++++++++++++++++++++++++++++++++++++++++
kernel/time/ntp.c | 2 +
4 files changed, 151 insertions(+), 0 deletions(-)
create mode 100644 drivers/ptp/ptp_linux.c
diff --git a/drivers/ptp/Kconfig b/drivers/ptp/Kconfig
index c80a25b..9390d44 100644
--- a/drivers/ptp/Kconfig
+++ b/drivers/ptp/Kconfig
@@ -23,4 +23,16 @@ config PTP_1588_CLOCK
To compile this driver as a module, choose M here: the module
will be called ptp_clock.
+config PTP_1588_CLOCK_LINUX
+ tristate "Linux system timer as PTP clock"
+ depends on PTP_1588_CLOCK
+ help
+ This driver adds support for using the standard Linux time
+ source as a PTP clock. This clock is only useful if your PTP
+ programs are using software time stamps for the PTP Ethernet
+ packets.
+
+ To compile this driver as a module, choose M here: the module
+ will be called ptp_linux.
+
endmenu
diff --git a/drivers/ptp/Makefile b/drivers/ptp/Makefile
index b86695c..1651d52 100644
--- a/drivers/ptp/Makefile
+++ b/drivers/ptp/Makefile
@@ -3,3 +3,4 @@
#
obj-$(CONFIG_PTP_1588_CLOCK) += ptp_clock.o
+obj-$(CONFIG_PTP_1588_CLOCK_LINUX) += ptp_linux.o
diff --git a/drivers/ptp/ptp_linux.c b/drivers/ptp/ptp_linux.c
new file mode 100644
index 0000000..f93ae0c
--- /dev/null
+++ b/drivers/ptp/ptp_linux.c
@@ -0,0 +1,136 @@
+/*
+ * PTP 1588 clock using the Linux system clock
+ *
+ * Copyright (C) 2010 OMICRON electronics GmbH
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
+ */
+#include <linux/device.h>
+#include <linux/err.h>
+#include <linux/hrtimer.h>
+#include <linux/init.h>
+#include <linux/kernel.h>
+#include <linux/module.h>
+#include <linux/timex.h>
+
+#include <linux/ptp_clock_kernel.h>
+
+static struct ptp_clock *linux_clock;
+
+DEFINE_SPINLOCK(adjtime_lock);
+
+static int ptp_linux_adjfreq(void *priv, s32 delta)
+{
+ struct timex txc;
+ s64 tmp = delta;
+ int err;
+ txc.freq = div_s64(tmp<<16, 1000);
+ txc.modes = ADJ_FREQUENCY;
+ err = do_adjtimex(&txc);
+ return err < 0 ? err : 0;
+}
+
+static int ptp_linux_adjtime(void *priv, struct timespec *ts)
+{
+ s64 delta;
+ ktime_t now;
+ struct timespec t2;
+ unsigned long flags;
+ int err;
+
+ delta = 1000000000LL * ts->tv_sec + ts->tv_nsec;
+
+ spin_lock_irqsave(&adjtime_lock, flags);
+
+ now = ktime_get_real();
+
+ now = delta < 0 ? ktime_sub_ns(now, -delta) : ktime_add_ns(now, delta);
+
+ t2 = ktime_to_timespec(now);
+
+ err = do_settimeofday(&t2);
+
+ spin_unlock_irqrestore(&adjtime_lock, flags);
+
+ return err;
+}
+
+static int ptp_linux_gettime(void *priv, struct timespec *ts)
+{
+ getnstimeofday(ts);
+ return 0;
+}
+
+static int ptp_linux_settime(void *priv, struct timespec *ts)
+{
+ return do_settimeofday(ts);
+}
+
+static int ptp_linux_gettimer(void *priv, int index, struct itimerspec *ts)
+{
+ /* We do not offer any ancillary features at all. */
+ return -EOPNOTSUPP;
+}
+
+static int ptp_linux_settimer(void *p, int i, int abs, struct itimerspec *ts)
+{
+ /* We do not offer any ancillary features at all. */
+ return -EOPNOTSUPP;
+}
+
+static int ptp_linux_enable(void *priv, struct ptp_clock_request *rq, int on)
+{
+ /* We do not offer any ancillary features at all. */
+ return -EOPNOTSUPP;
+}
+
+static struct ptp_clock_info ptp_linux_caps = {
+ .owner = THIS_MODULE,
+ .name = "Linux timer",
+ .max_adj = 512000,
+ .n_alarm = 0,
+ .n_ext_ts = 0,
+ .n_per_out = 0,
+ .pps = 0,
+ .priv = NULL,
+ .adjfreq = ptp_linux_adjfreq,
+ .adjtime = ptp_linux_adjtime,
+ .gettime = ptp_linux_gettime,
+ .settime = ptp_linux_settime,
+ .gettimer = ptp_linux_gettimer,
+ .settimer = ptp_linux_settimer,
+ .enable = ptp_linux_enable,
+};
+
+/* module operations */
+
+static void __exit ptp_linux_exit(void)
+{
+ ptp_clock_unregister(linux_clock);
+}
+
+static int __init ptp_linux_init(void)
+{
+ linux_clock = ptp_clock_register(&ptp_linux_caps);
+
+ return IS_ERR(linux_clock) ? PTR_ERR(linux_clock) : 0;
+}
+
+module_init(ptp_linux_init);
+module_exit(ptp_linux_exit);
+
+MODULE_AUTHOR("Richard Cochran <richard.cochran@omicron.at>");
+MODULE_DESCRIPTION("PTP clock using the Linux system timer");
+MODULE_LICENSE("GPL");
diff --git a/kernel/time/ntp.c b/kernel/time/ntp.c
index c631168..eba3bcf 100644
--- a/kernel/time/ntp.c
+++ b/kernel/time/ntp.c
@@ -14,6 +14,7 @@
#include <linux/timex.h>
#include <linux/time.h>
#include <linux/mm.h>
+#include <linux/module.h>
/*
* NTP timekeeping variables:
@@ -535,6 +536,7 @@ int do_adjtimex(struct timex *txc)
return result;
}
+EXPORT_SYMBOL(do_adjtimex);
static int __init ntp_tick_adj_setup(char *str)
{
--
1.6.3.3
^ permalink raw reply related
* [PATCH 08/12] ptp: Added a brand new class driver for ptp clocks.
From: Richard Cochran @ 2010-06-15 16:09 UTC (permalink / raw)
To: netdev; +Cc: linuxppc-dev, devicetree-discuss, Krzysztof Halasa,
linux-arm-kernel
In-Reply-To: <cover.1276615626.git.richard.cochran@omicron.at>
This patch adds an infrastructure for hardware clocks that implement
IEEE 1588, the Precision Time Protocol (PTP). A class driver offers a
registration method to particular hardware clock drivers. Each clock is
exposed to user space as a character device with ioctls that allow tuning
of the PTP clock.
Signed-off-by: Richard Cochran <richard.cochran@omicron.at>
---
Documentation/ptp/ptp.txt | 95 +++++++
Documentation/ptp/testptp.c | 269 ++++++++++++++++++++
Documentation/ptp/testptp.mk | 33 +++
drivers/Kconfig | 2 +
drivers/Makefile | 1 +
drivers/ptp/Kconfig | 26 ++
drivers/ptp/Makefile | 5 +
drivers/ptp/ptp_clock.c | 514 ++++++++++++++++++++++++++++++++++++++
include/linux/Kbuild | 1 +
include/linux/ptp_clock.h | 79 ++++++
include/linux/ptp_clock_kernel.h | 137 ++++++++++
11 files changed, 1162 insertions(+), 0 deletions(-)
create mode 100644 Documentation/ptp/ptp.txt
create mode 100644 Documentation/ptp/testptp.c
create mode 100644 Documentation/ptp/testptp.mk
create mode 100644 drivers/ptp/Kconfig
create mode 100644 drivers/ptp/Makefile
create mode 100644 drivers/ptp/ptp_clock.c
create mode 100644 include/linux/ptp_clock.h
create mode 100644 include/linux/ptp_clock_kernel.h
diff --git a/Documentation/ptp/ptp.txt b/Documentation/ptp/ptp.txt
new file mode 100644
index 0000000..46858b3
--- /dev/null
+++ b/Documentation/ptp/ptp.txt
@@ -0,0 +1,95 @@
+
+* PTP infrastructure for Linux
+
+ This patch set introduces support for IEEE 1588 PTP clocks in
+ Linux. Together with the SO_TIMESTAMPING socket options, this
+ presents a standardized method for developing PTP user space
+ programs, synchronizing Linux with external clocks, and using the
+ ancillary features of PTP hardware clocks.
+
+ A new class driver exports a kernel interface for specific clock
+ drivers and a user space interface. The infrastructure supports a
+ complete set of PTP functionality.
+
+ + Basic clock operations
+ - Set time
+ - Get time
+ - Shift the clock by a given offset atomically
+ - Adjust clock frequency
+
+ + Ancillary clock features
+ - One short or periodic alarms, with signal delivery to user program
+ - Time stamp external events
+ - Period output signals configurable from user space
+ - Synchronization of the Linux system time via the PPS subsystem
+
+** PTP kernel API
+
+ A PTP clock driver registers itself with the class driver. The
+ class driver handles all of the dealings with user space. The
+ author of a clock driver need only implement the details of
+ programming the clock hardware. The clock driver notifies the class
+ driver of asynchronous events (alarms and external time stamps) via
+ a simple message passing interface.
+
+ The class driver supports multiple PTP clock drivers. In normal use
+ cases, only one PTP clock is needed. However, for testing and
+ development, it can be useful to have more than one clock in a
+ single system, in order to allow performance comparisons.
+
+** PTP user space API
+
+ The class driver creates a character device for each registered PTP
+ clock. User space programs may control the clock using standardized
+ ioctls. A program may query, enable, configure, and disable the
+ ancillary clock features. User space can receive time stamped
+ events via blocking read() and poll(). One shot and periodic
+ signals may be configured via an ioctl API with semantics similar
+ to the POSIX timer_settime() system call.
+
+ As an real life example, the following two patches for ptpd version
+ 1.0.0 demonstrate how the API works.
+
+ https://sourceforge.net/tracker/?func=detail&aid=2992845&group_id=139814&atid=744634
+
+ https://sourceforge.net/tracker/?func=detail&aid=2992847&group_id=139814&atid=744634
+
+** Writing clock drivers
+
+ Clock drivers include include/linux/ptp_clock_kernel.h and register
+ themselves by presenting a 'struct ptp_clock_info' to the
+ registration method. Clock drivers must implement all of the
+ functions in the interface. If a clock does not offer a particular
+ ancillary feature, then the driver should just return -EOPNOTSUPP
+ from those functions.
+
+ Drivers must ensure that all of the methods in interface are
+ reentrant. Since most hardware implementations treat the time value
+ as a 64 bit integer accessed as two 32 bit registers, drivers
+ should use spin_lock_irqsave/spin_unlock_irqrestore to protect
+ against concurrent access. This locking cannot be accomplished in
+ class driver, since the lock may also be needed by the clock
+ driver's interrupt service routine.
+
+** Supported hardware
+
+ + Standard Linux system timer
+ - No special PTP features
+ - For use with software time stamping
+
+ + Freescale eTSEC gianfar
+ - 2 Time stamp external triggers, programmable polarity (opt. interrupt)
+ - 2 Alarm registers (optional interrupt)
+ - 3 Periodic signals (optional interrupt)
+
+ + National DP83640
+ - 6 GPIOs programmable as inputs or outputs
+ - 6 GPIOs with dedicated functions (LED/JTAG/clock) can also be
+ used as general inputs or outputs
+ - GPIO inputs can time stamp external triggers
+ - GPIO outputs can produce periodic signals
+ - 1 interrupt pin
+
+ + Intel IXP465
+ - Auxiliary Slave/Master Mode Snapshot (optional interrupt)
+ - Target Time (optional interrupt)
diff --git a/Documentation/ptp/testptp.c b/Documentation/ptp/testptp.c
new file mode 100644
index 0000000..e30f758
--- /dev/null
+++ b/Documentation/ptp/testptp.c
@@ -0,0 +1,269 @@
+/*
+ * PTP 1588 clock support - User space test program
+ *
+ * Copyright (C) 2010 OMICRON electronics GmbH
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
+ */
+#include <errno.h>
+#include <fcntl.h>
+#include <signal.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+#include <sys/ioctl.h>
+#include <sys/mman.h>
+#include <sys/stat.h>
+#include <sys/time.h>
+#include <sys/types.h>
+#include <time.h>
+#include <unistd.h>
+
+#include <linux/ptp_clock.h>
+
+static void handle_alarm(int s)
+{
+ printf("received signal %d\n", s);
+}
+
+static int install_handler(int signum, void (*handler)(int))
+{
+ struct sigaction action;
+ sigset_t mask;
+
+ /* Unblock the signal. */
+ sigemptyset(&mask);
+ sigaddset(&mask, signum);
+ sigprocmask(SIG_UNBLOCK, &mask, NULL);
+
+ /* Install the signal handler. */
+ action.sa_handler = handler;
+ action.sa_flags = 0;
+ sigemptyset(&action.sa_mask);
+ sigaction(signum, &action, NULL);
+
+ return 0;
+}
+
+static void usage(char *progname)
+{
+ fprintf(stderr,
+ "usage: %s [options]\n"
+ " -a val request a one-shot alarm after 'val' seconds\n"
+ " -A val request a periodic alarm every 'val' seconds\n"
+ " -c query the ptp clock's capabilities\n"
+ " -d name device to open\n"
+ " -e val read 'val' external time stamp events\n"
+ " -f val adjust the ptp clock frequency by 'val' PPB\n"
+ " -g get the ptp clock time\n"
+ " -h prints this message\n"
+ " -s set the ptp clock time from the system time\n"
+ " -t val shift the ptp clock time by 'val' seconds\n"
+ " -v query the ptp clock api version\n",
+ progname);
+}
+
+int main(int argc, char *argv[])
+{
+ struct ptp_clock_caps caps;
+ struct ptp_clock_timer timer;
+ struct ptp_extts_event event;
+ struct ptp_clock_request request;
+ struct timespec ts;
+ char *progname;
+ int c, cnt, fd, val = 0;
+
+ char *device = "/dev/ptp_clock_0";
+ int adjfreq = 0x7fffffff;
+ int adjtime = 0;
+ int capabilities = 0;
+ int extts = 0;
+ int gettime = 0;
+ int oneshot = 0;
+ int periodic = 0;
+ int settime = 0;
+ int version = 0;
+
+ progname = strrchr(argv[0], '/');
+ progname = progname ? 1+progname : argv[0];
+ while (EOF != (c = getopt(argc, argv, "a:A:cd:e:f:ghst:v"))) {
+ switch (c) {
+ case 'a':
+ oneshot = atoi(optarg);
+ break;
+ case 'A':
+ periodic = atoi(optarg);
+ break;
+ case 'c':
+ capabilities = 1;
+ break;
+ case 'd':
+ device = optarg;
+ break;
+ case 'e':
+ extts = atoi(optarg);
+ break;
+ case 'f':
+ adjfreq = atoi(optarg);
+ break;
+ case 'g':
+ gettime = 1;
+ break;
+ case 's':
+ settime = 1;
+ break;
+ case 't':
+ adjtime = atoi(optarg);
+ break;
+ case 'v':
+ version = 1;
+ break;
+ case 'h':
+ usage(progname);
+ return 0;
+ case '?':
+ default:
+ usage(progname);
+ return -1;
+ }
+ }
+
+ fd = open(device, O_RDWR);
+ if (fd < 0) {
+ fprintf(stderr, "cannot open %s: %s", device, strerror(errno));
+ return -1;
+ }
+
+ if (version) {
+ if (ioctl(fd, PTP_CLOCK_APIVERS, &val)) {
+ perror("PTP_CLOCK_APIVERS");
+ } else {
+ printf("version = 0x%08x\n", val);
+ }
+ }
+
+ if (capabilities) {
+ if (ioctl(fd, PTP_CLOCK_GETCAPS, &caps)) {
+ perror("PTP_CLOCK_GETCAPS");
+ } else {
+ printf("capabilities:\n"
+ " %d maximum frequency adjustment (PPB)\n"
+ " %d programmable alarms\n"
+ " %d external time stamp channels\n"
+ " %d programmable periodic signals\n"
+ " %d pulse per second\n",
+ caps.max_adj,
+ caps.n_alarm,
+ caps.n_ext_ts,
+ caps.n_per_out,
+ caps.pps);
+ }
+ }
+
+ if (0x7fffffff != adjfreq) {
+ if (ioctl(fd, PTP_CLOCK_ADJFREQ, adjfreq)) {
+ perror("PTP_CLOCK_ADJFREQ");
+ } else {
+ puts("frequency adjustment okay");
+ }
+ }
+
+ if (adjtime) {
+ ts.tv_sec = adjtime;
+ ts.tv_nsec = 0;
+ if (ioctl(fd, PTP_CLOCK_ADJTIME, &ts)) {
+ perror("PTP_CLOCK_ADJTIME");
+ } else {
+ puts("time shift okay");
+ }
+ }
+
+ if (gettime) {
+ if (ioctl(fd, PTP_CLOCK_GETTIME, &ts)) {
+ perror("PTP_CLOCK_GETTIME");
+ } else {
+ printf("clock time: %ld.%09ld or %s",
+ ts.tv_sec, ts.tv_nsec, ctime(&ts.tv_sec));
+ }
+ }
+
+ if (settime) {
+ clock_gettime(CLOCK_REALTIME, &ts);
+ if (ioctl(fd, PTP_CLOCK_SETTIME, &ts)) {
+ perror("PTP_CLOCK_SETTIME");
+ } else {
+ puts("set time okay");
+ }
+ }
+
+ if (extts) {
+ memset(&request, 0, sizeof(request));
+ request.type = PTP_REQUEST_EXTTS;
+ request.index = 0;
+ request.flags = PTP_ENABLE_FEATURE;
+ if (ioctl(fd, PTP_FEATURE_REQUEST, &request)) {
+ perror("PTP_FEATURE_REQUEST");
+ extts = 0;
+ } else {
+ puts("set timer okay");
+ }
+ for (; extts; extts--) {
+ cnt = read(fd, &event, sizeof(event));
+ if (cnt != sizeof(event)) {
+ perror("read");
+ break;
+ }
+ printf("event index %d at %ld.%09ld\n", event.index,
+ event.ts.tv_sec, event.ts.tv_nsec);
+ }
+ /* Disable the feature again. */
+ request.flags = 0;
+ if (ioctl(fd, PTP_FEATURE_REQUEST, &request)) {
+ perror("PTP_FEATURE_REQUEST");
+ }
+ }
+
+ if (oneshot) {
+ install_handler(SIGALRM, handle_alarm);
+ memset(&timer, 0, sizeof(timer));
+ timer.signum = SIGALRM;
+ timer.tsp.it_value.tv_sec = oneshot;
+ if (ioctl(fd, PTP_CLOCK_SETTIMER, &timer)) {
+ perror("PTP_CLOCK_SETTIMER");
+ } else {
+ puts("set timer okay");
+ }
+ pause();
+ }
+
+ if (periodic) {
+ install_handler(SIGALRM, handle_alarm);
+ memset(&timer, 0, sizeof(timer));
+ timer.signum = SIGALRM;
+ timer.tsp.it_value.tv_sec = periodic;
+ timer.tsp.it_interval.tv_sec = periodic;
+ if (ioctl(fd, PTP_CLOCK_SETTIMER, &timer)) {
+ perror("PTP_CLOCK_SETTIMER");
+ } else {
+ puts("set timer okay");
+ }
+ while (1) {
+ pause();
+ }
+ }
+
+ close(fd);
+ return 0;
+}
diff --git a/Documentation/ptp/testptp.mk b/Documentation/ptp/testptp.mk
new file mode 100644
index 0000000..4ef2d97
--- /dev/null
+++ b/Documentation/ptp/testptp.mk
@@ -0,0 +1,33 @@
+# PTP 1588 clock support - User space test program
+#
+# Copyright (C) 2010 OMICRON electronics GmbH
+#
+# This program is free software; you can redistribute it and/or modify
+# it under the terms of the GNU General Public License as published by
+# the Free Software Foundation; either version 2 of the License, or
+# (at your option) any later version.
+#
+# This program is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+# GNU General Public License for more details.
+#
+# You should have received a copy of the GNU General Public License
+# along with this program; if not, write to the Free Software
+# Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
+
+CC = $(CROSS_COMPILE)gcc
+INC = -I$(KBUILD_OUTPUT)/usr/include
+CFLAGS = -Wall $(INC)
+LDLIBS = -lrt
+PROGS = testptp
+
+all: $(PROGS)
+
+testptp: testptp.o
+
+clean:
+ rm -f testptp.o
+
+distclean: clean
+ rm -f $(PROGS)
diff --git a/drivers/Kconfig b/drivers/Kconfig
index a2b902f..774fbd7 100644
--- a/drivers/Kconfig
+++ b/drivers/Kconfig
@@ -52,6 +52,8 @@ source "drivers/spi/Kconfig"
source "drivers/pps/Kconfig"
+source "drivers/ptp/Kconfig"
+
source "drivers/gpio/Kconfig"
source "drivers/w1/Kconfig"
diff --git a/drivers/Makefile b/drivers/Makefile
index 91874e0..6d12b48 100644
--- a/drivers/Makefile
+++ b/drivers/Makefile
@@ -76,6 +76,7 @@ obj-$(CONFIG_I2O) += message/
obj-$(CONFIG_RTC_LIB) += rtc/
obj-y += i2c/ media/
obj-$(CONFIG_PPS) += pps/
+obj-$(CONFIG_PTP_1588_CLOCK) += ptp/
obj-$(CONFIG_W1) += w1/
obj-$(CONFIG_POWER_SUPPLY) += power/
obj-$(CONFIG_HWMON) += hwmon/
diff --git a/drivers/ptp/Kconfig b/drivers/ptp/Kconfig
new file mode 100644
index 0000000..c80a25b
--- /dev/null
+++ b/drivers/ptp/Kconfig
@@ -0,0 +1,26 @@
+#
+# PTP clock support configuration
+#
+
+menu "PTP clock support"
+
+config PTP_1588_CLOCK
+ tristate "PTP clock support"
+ depends on EXPERIMENTAL
+ help
+ The IEEE 1588 standard defines a method to precisely
+ synchronize distributed clocks over Ethernet networks. The
+ standard defines a Precision Time Protocol (PTP), which can
+ be used to achieve synchronization within a few dozen
+ microseconds. In addition, with the help of special hardware
+ time stamping units, it can be possible to achieve
+ synchronization to within a few hundred nanoseconds.
+
+ This driver adds support for PTP clocks as character
+ devices. If you want to use a PTP clock, then you should
+ also enable at least one clock driver as well.
+
+ To compile this driver as a module, choose M here: the module
+ will be called ptp_clock.
+
+endmenu
diff --git a/drivers/ptp/Makefile b/drivers/ptp/Makefile
new file mode 100644
index 0000000..b86695c
--- /dev/null
+++ b/drivers/ptp/Makefile
@@ -0,0 +1,5 @@
+#
+# Makefile for PTP 1588 clock support.
+#
+
+obj-$(CONFIG_PTP_1588_CLOCK) += ptp_clock.o
diff --git a/drivers/ptp/ptp_clock.c b/drivers/ptp/ptp_clock.c
new file mode 100644
index 0000000..4753bf3
--- /dev/null
+++ b/drivers/ptp/ptp_clock.c
@@ -0,0 +1,514 @@
+/*
+ * PTP 1588 clock support
+ *
+ * Partially adapted from the Linux PPS driver.
+ *
+ * Copyright (C) 2010 OMICRON electronics GmbH
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
+ */
+#include <linux/bitops.h>
+#include <linux/cdev.h>
+#include <linux/device.h>
+#include <linux/init.h>
+#include <linux/kernel.h>
+#include <linux/list.h>
+#include <linux/module.h>
+#include <linux/poll.h>
+#include <linux/sched.h>
+#include <linux/slab.h>
+#include <linux/uaccess.h>
+
+#include <linux/ptp_clock_kernel.h>
+#include <linux/ptp_clock.h>
+
+#define PTP_MAX_ALARMS 4
+#define PTP_MAX_CLOCKS BITS_PER_LONG
+#define PTP_MAX_TIMESTAMPS 128
+
+struct alarm {
+ struct pid *pid;
+ int sig;
+};
+
+struct timestamp_event_queue {
+ struct ptp_extts_event buf[PTP_MAX_TIMESTAMPS];
+ int head;
+ int tail;
+ int overflow;
+};
+
+struct ptp_clock {
+ struct list_head list;
+ struct cdev cdev;
+ struct device *dev;
+ struct ptp_clock_info *info;
+ dev_t devid;
+ int index; /* index into clocks.map, also the minor number */
+
+ struct alarm alarm[PTP_MAX_ALARMS];
+ struct mutex alarm_mux; /* one process at a time setting an alarm */
+
+ struct timestamp_event_queue tsevq; /* simple fifo for time stamps */
+ struct mutex tsevq_mux; /* one process at a time reading the fifo */
+ wait_queue_head_t tsev_wq;
+};
+
+/* private globals */
+
+static const struct file_operations ptp_fops;
+static dev_t ptp_devt;
+static struct class *ptp_class;
+
+static struct {
+ struct list_head list;
+ DECLARE_BITMAP(map, PTP_MAX_CLOCKS);
+} clocks;
+static DEFINE_SPINLOCK(clocks_lock); /* protects 'clocks' */
+
+/* time stamp event queue operations */
+
+static inline int queue_cnt(struct timestamp_event_queue *q)
+{
+ int cnt = q->tail - q->head;
+ return cnt < 0 ? PTP_MAX_TIMESTAMPS + cnt : cnt;
+}
+
+static inline int queue_free(struct timestamp_event_queue *q)
+{
+ return PTP_MAX_TIMESTAMPS - queue_cnt(q) - 1;
+}
+
+static void enqueue_external_timestamp(struct timestamp_event_queue *queue,
+ struct ptp_clock_event *src)
+{
+ struct ptp_extts_event *dst;
+ u32 remainder;
+
+ dst = &queue->buf[queue->tail];
+
+ dst->index = src->index;
+ dst->ts.tv_sec = div_u64_rem(src->timestamp, 1000000000, &remainder);
+ dst->ts.tv_nsec = remainder;
+
+ if (!queue_free(queue))
+ queue->overflow++;
+
+ queue->tail = (queue->tail + 1) % PTP_MAX_TIMESTAMPS;
+}
+
+/* public interface */
+
+struct ptp_clock *ptp_clock_register(struct ptp_clock_info *info)
+{
+ struct ptp_clock *ptp;
+ int err = 0, index, major = MAJOR(ptp_devt);
+ unsigned long flags;
+
+ if (info->n_alarm > PTP_MAX_ALARMS)
+ return ERR_PTR(-EINVAL);
+
+ /* Find a free clock slot and reserve it. */
+ err = -EBUSY;
+ spin_lock_irqsave(&clocks_lock, flags);
+ index = find_first_zero_bit(clocks.map, PTP_MAX_CLOCKS);
+ if (index < PTP_MAX_CLOCKS) {
+ set_bit(index, clocks.map);
+ spin_unlock_irqrestore(&clocks_lock, flags);
+ } else {
+ spin_unlock_irqrestore(&clocks_lock, flags);
+ goto no_clock;
+ }
+
+ /* Initialize a clock structure. */
+ err = -ENOMEM;
+ ptp = kzalloc(sizeof(struct ptp_clock), GFP_KERNEL);
+ if (ptp == NULL)
+ goto no_memory;
+
+ ptp->info = info;
+ ptp->devid = MKDEV(major, index);
+ ptp->index = index;
+ mutex_init(&ptp->alarm_mux);
+ mutex_init(&ptp->tsevq_mux);
+ init_waitqueue_head(&ptp->tsev_wq);
+
+ /* Create a new device in our class. */
+ ptp->dev = device_create(ptp_class, NULL, ptp->devid, ptp,
+ "ptp_clock_%d", ptp->index);
+ if (IS_ERR(ptp->dev))
+ goto no_device;
+
+ dev_set_drvdata(ptp->dev, ptp);
+
+ /* Register a character device. */
+ cdev_init(&ptp->cdev, &ptp_fops);
+ ptp->cdev.owner = info->owner;
+ err = cdev_add(&ptp->cdev, ptp->devid, 1);
+ if (err)
+ goto no_cdev;
+
+ /* Clock is ready, add it into the list. */
+ spin_lock_irqsave(&clocks_lock, flags);
+ list_add(&ptp->list, &clocks.list);
+ spin_unlock_irqrestore(&clocks_lock, flags);
+
+ return ptp;
+
+no_cdev:
+ device_destroy(ptp_class, ptp->devid);
+no_device:
+ mutex_destroy(&ptp->alarm_mux);
+ mutex_destroy(&ptp->tsevq_mux);
+ kfree(ptp);
+no_memory:
+ spin_lock_irqsave(&clocks_lock, flags);
+ clear_bit(index, clocks.map);
+ spin_unlock_irqrestore(&clocks_lock, flags);
+no_clock:
+ return ERR_PTR(err);
+}
+EXPORT_SYMBOL(ptp_clock_register);
+
+int ptp_clock_unregister(struct ptp_clock *ptp)
+{
+ unsigned long flags;
+
+ /* Release the clock's resources. */
+ cdev_del(&ptp->cdev);
+ device_destroy(ptp_class, ptp->devid);
+ mutex_destroy(&ptp->alarm_mux);
+ mutex_destroy(&ptp->tsevq_mux);
+
+ /* Remove the clock from the list. */
+ spin_lock_irqsave(&clocks_lock, flags);
+ list_del(&ptp->list);
+ clear_bit(ptp->index, clocks.map);
+ spin_unlock_irqrestore(&clocks_lock, flags);
+
+ kfree(ptp);
+
+ return 0;
+}
+EXPORT_SYMBOL(ptp_clock_unregister);
+
+void ptp_clock_event(struct ptp_clock *ptp, struct ptp_clock_event *event)
+{
+ switch (event->type) {
+
+ case PTP_CLOCK_ALARM:
+ kill_pid(ptp->alarm[event->index].pid,
+ ptp->alarm[event->index].sig, 1);
+ break;
+
+ case PTP_CLOCK_EXTTS:
+ enqueue_external_timestamp(&ptp->tsevq, event);
+ wake_up_interruptible(&ptp->tsev_wq);
+ break;
+
+ case PTP_CLOCK_PPS:
+ break;
+ }
+}
+EXPORT_SYMBOL(ptp_clock_event);
+
+/* character device operations */
+
+static int ptp_ioctl(struct inode *node, struct file *fp,
+ unsigned int cmd, unsigned long arg)
+{
+ struct ptp_clock_caps caps;
+ struct ptp_clock_request req;
+ struct ptp_clock_timer timer;
+ struct ptp_clock *ptp = fp->private_data;
+ struct ptp_clock_info *ops = ptp->info;
+ void *priv = ops->priv;
+ struct timespec ts;
+ int flags, index;
+ int err = 0;
+
+ switch (cmd) {
+
+ case PTP_CLOCK_APIVERS:
+ err = put_user(PTP_CLOCK_VERSION, (u32 __user *)arg);
+ break;
+
+ case PTP_CLOCK_ADJFREQ:
+ if (!capable(CAP_SYS_TIME))
+ return -EPERM;
+ err = ops->adjfreq(priv, arg);
+ break;
+
+ case PTP_CLOCK_ADJTIME:
+ if (!capable(CAP_SYS_TIME))
+ return -EPERM;
+ if (copy_from_user(&ts, (void __user *)arg, sizeof(ts)))
+ err = -EFAULT;
+ else
+ err = ops->adjtime(priv, &ts);
+ break;
+
+ case PTP_CLOCK_GETTIME:
+ err = ops->gettime(priv, &ts);
+ if (err)
+ break;
+ err = copy_to_user((void __user *)arg, &ts, sizeof(ts));
+ break;
+
+ case PTP_CLOCK_SETTIME:
+ if (!capable(CAP_SYS_TIME))
+ return -EPERM;
+ if (copy_from_user(&ts, (void __user *)arg, sizeof(ts)))
+ err = -EFAULT;
+ else
+ err = ops->settime(priv, &ts);
+ break;
+
+ case PTP_CLOCK_GETCAPS:
+ memset(&caps, 0, sizeof(caps));
+ caps.max_adj = ptp->info->max_adj;
+ caps.n_alarm = ptp->info->n_alarm;
+ caps.n_ext_ts = ptp->info->n_ext_ts;
+ caps.n_per_out = ptp->info->n_per_out;
+ caps.pps = ptp->info->pps;
+ err = copy_to_user((void __user *)arg, &caps, sizeof(caps));
+ break;
+
+ case PTP_CLOCK_GETTIMER:
+ if (copy_from_user(&timer, (void __user *)arg, sizeof(timer))) {
+ err = -EFAULT;
+ break;
+ }
+ index = timer.alarm_index;
+ if (index < 0 || index >= ptp->info->n_alarm) {
+ err = -EINVAL;
+ break;
+ }
+ err = ops->gettimer(priv, index, &timer.tsp);
+ if (err)
+ break;
+ err = copy_to_user((void __user *)arg, &timer, sizeof(timer));
+ break;
+
+ case PTP_CLOCK_SETTIMER:
+ if (copy_from_user(&timer, (void __user *)arg, sizeof(timer))) {
+ err = -EFAULT;
+ break;
+ }
+ index = timer.alarm_index;
+ if (index < 0 || index >= ptp->info->n_alarm) {
+ err = -EINVAL;
+ break;
+ }
+ if (!valid_signal(timer.signum))
+ return -EINVAL;
+ flags = timer.flags;
+ if (flags & (flags != TIMER_ABSTIME)) {
+ err = -EINVAL;
+ break;
+ }
+ if (mutex_lock_interruptible(&ptp->alarm_mux))
+ return -ERESTARTSYS;
+
+ if (ptp->alarm[index].pid)
+ put_pid(ptp->alarm[index].pid);
+
+ ptp->alarm[index].pid = get_pid(task_pid(current));
+ ptp->alarm[index].sig = timer.signum;
+ err = ops->settimer(priv, index, flags, &timer.tsp);
+
+ mutex_unlock(&ptp->alarm_mux);
+ break;
+
+ case PTP_FEATURE_REQUEST:
+ if (copy_from_user(&req, (void __user *)arg, sizeof(req))) {
+ err = -EFAULT;
+ break;
+ }
+ switch (req.type) {
+ case PTP_REQUEST_EXTTS:
+ case PTP_REQUEST_PEROUT:
+ break;
+ case PTP_REQUEST_PPS:
+ if (!capable(CAP_SYS_TIME))
+ return -EPERM;
+ break;
+ default:
+ err = -EINVAL;
+ break;
+ }
+ if (err)
+ break;
+ err = ops->enable(priv, &req,
+ req.flags & PTP_ENABLE_FEATURE ? 1 : 0);
+ break;
+
+ default:
+ err = -ENOTTY;
+ break;
+ }
+ return err;
+}
+
+static int ptp_open(struct inode *inode, struct file *fp)
+{
+ struct ptp_clock *ptp;
+ ptp = container_of(inode->i_cdev, struct ptp_clock, cdev);
+
+ fp->private_data = ptp;
+
+ return 0;
+}
+
+static unsigned int ptp_poll(struct file *fp, poll_table *wait)
+{
+ struct ptp_clock *ptp = fp->private_data;
+
+ poll_wait(fp, &ptp->tsev_wq, wait);
+
+ return queue_cnt(&ptp->tsevq) ? POLLIN : 0;
+}
+
+static ssize_t ptp_read(struct file *fp, char __user *buf,
+ size_t cnt, loff_t *off)
+{
+ struct ptp_clock *ptp = fp->private_data;
+ struct timestamp_event_queue *queue = &ptp->tsevq;
+ struct ptp_extts_event *event;
+ size_t qcnt;
+
+ if (mutex_lock_interruptible(&ptp->tsevq_mux))
+ return -ERESTARTSYS;
+
+ cnt = cnt / sizeof(struct ptp_extts_event);
+
+ if (wait_event_interruptible(ptp->tsev_wq,
+ (qcnt = queue_cnt(&ptp->tsevq)))) {
+ mutex_unlock(&ptp->tsevq_mux);
+ return -ERESTARTSYS;
+ }
+
+ if (cnt > qcnt)
+ cnt = qcnt;
+
+ event = &queue->buf[queue->head];
+
+ if (copy_to_user(buf, event, cnt * sizeof(struct ptp_extts_event))) {
+ mutex_unlock(&ptp->tsevq_mux);
+ return -EFAULT;
+ }
+ queue->head = (queue->head + cnt) % PTP_MAX_TIMESTAMPS;
+
+ mutex_unlock(&ptp->tsevq_mux);
+
+ return cnt * sizeof(struct ptp_extts_event);
+}
+
+static int ptp_release(struct inode *inode, struct file *fp)
+{
+ struct ptp_clock *ptp;
+ struct itimerspec ts = {
+ {0, 0}, {0, 0}
+ };
+ int i;
+
+ ptp = container_of(inode->i_cdev, struct ptp_clock, cdev);
+
+ for (i = 0; i < ptp->info->n_alarm; i++) {
+ if (ptp->alarm[i].pid) {
+ ptp->info->settimer(ptp->info->priv, i, 0, &ts);
+ put_pid(ptp->alarm[i].pid);
+ ptp->alarm[i].pid = NULL;
+ }
+ }
+ return 0;
+}
+
+static const struct file_operations ptp_fops = {
+ .owner = THIS_MODULE,
+ .ioctl = ptp_ioctl,
+ .open = ptp_open,
+ .poll = ptp_poll,
+ .read = ptp_read,
+ .release = ptp_release,
+};
+
+/* sysfs */
+
+static ssize_t ptp_show_status(struct device *dev,
+ struct device_attribute *attr, char *buf)
+{
+ struct ptp_clock *ptp = dev_get_drvdata(dev);
+ return sprintf(buf,
+ "maximum adjustment: %d\n"
+ "programmable alarms: %d\n"
+ "external timestamps: %d\n"
+ "periodic outputs: %d\n"
+ "has pps: %d\n"
+ "device index: %d\n",
+ ptp->info->max_adj,
+ ptp->info->n_alarm,
+ ptp->info->n_ext_ts,
+ ptp->info->n_per_out,
+ ptp->info->pps,
+ ptp->index);
+}
+
+struct device_attribute ptp_attrs[] = {
+ __ATTR(capabilities, S_IRUGO, ptp_show_status, NULL),
+ __ATTR_NULL,
+};
+
+/* module operations */
+
+static void __exit ptp_exit(void)
+{
+ class_destroy(ptp_class);
+ unregister_chrdev_region(ptp_devt, PTP_MAX_CLOCKS);
+}
+
+static int __init ptp_init(void)
+{
+ int err;
+
+ INIT_LIST_HEAD(&clocks.list);
+
+ ptp_class = class_create(THIS_MODULE, "ptp");
+ if (!ptp_class) {
+ printk(KERN_ERR "ptp: failed to allocate class\n");
+ return -ENOMEM;
+ }
+ ptp_class->dev_attrs = ptp_attrs;
+
+ err = alloc_chrdev_region(&ptp_devt, 0, PTP_MAX_CLOCKS, "ptp");
+ if (err < 0) {
+ printk(KERN_ERR "ptp: failed to allocate char device region\n");
+ goto no_region;
+ }
+
+ pr_info("PTP clock support registered\n");
+ return 0;
+
+no_region:
+ class_destroy(ptp_class);
+ return err;
+}
+
+subsys_initcall(ptp_init);
+module_exit(ptp_exit);
+
+MODULE_AUTHOR("Richard Cochran <richard.cochran@omicron.at>");
+MODULE_DESCRIPTION("PTP clocks support");
+MODULE_LICENSE("GPL");
diff --git a/include/linux/Kbuild b/include/linux/Kbuild
index 2fc8e14..9959fe4 100644
--- a/include/linux/Kbuild
+++ b/include/linux/Kbuild
@@ -140,6 +140,7 @@ header-y += pkt_sched.h
header-y += posix_types.h
header-y += ppdev.h
header-y += prctl.h
+header-y += ptp_clock.h
header-y += qnxtypes.h
header-y += qnx4_fs.h
header-y += radeonfb.h
diff --git a/include/linux/ptp_clock.h b/include/linux/ptp_clock.h
new file mode 100644
index 0000000..5a509c5
--- /dev/null
+++ b/include/linux/ptp_clock.h
@@ -0,0 +1,79 @@
+/*
+ * PTP 1588 clock support - user space interface
+ *
+ * Copyright (C) 2010 OMICRON electronics GmbH
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
+ */
+
+#ifndef _PTP_CLOCK_H_
+#define _PTP_CLOCK_H_
+
+#include <linux/ioctl.h>
+#include <linux/types.h>
+
+#define PTP_ENABLE_FEATURE (1<<0)
+#define PTP_RISING_EDGE (1<<1)
+#define PTP_FALLING_EDGE (1<<2)
+
+enum ptp_request_types {
+ PTP_REQUEST_EXTTS,
+ PTP_REQUEST_PEROUT,
+ PTP_REQUEST_PPS,
+};
+
+struct ptp_clock_caps {
+ __s32 max_adj; /* Maximum frequency adjustment, parts per billon. */
+ int n_alarm; /* Number of programmable alarms. */
+ int n_ext_ts; /* Number of external time stamp channels. */
+ int n_per_out; /* Number of programmable periodic signals. */
+ int pps; /* Whether the clock supports a PPS callback. */
+};
+
+struct ptp_clock_timer {
+ int alarm_index; /* Which alarm to query or configure. */
+ int signum; /* Requested signal. */
+ int flags; /* Zero or TIMER_ABSTIME, see TIMER_SETTIME(2) */
+ struct itimerspec tsp; /* See TIMER_SETTIME(2) */
+};
+
+struct ptp_clock_request {
+ int type; /* One of the ptp_request_types enumeration values. */
+ int index; /* Which channel to configure. */
+ struct timespec ts; /* For period signals, the desired period. */
+ int flags; /* Bit field for PTP_ENABLE_FEATURE or other flags. */
+};
+
+struct ptp_extts_event {
+ int index;
+ struct timespec ts;
+};
+
+#define PTP_CLOCK_VERSION 0x00000001
+
+#define PTP_CLK_MAGIC '='
+
+#define PTP_CLOCK_APIVERS _IOR (PTP_CLK_MAGIC, 1, __u32)
+#define PTP_CLOCK_ADJFREQ _IO (PTP_CLK_MAGIC, 2)
+#define PTP_CLOCK_ADJTIME _IOW (PTP_CLK_MAGIC, 3, struct timespec)
+#define PTP_CLOCK_GETTIME _IOR (PTP_CLK_MAGIC, 4, struct timespec)
+#define PTP_CLOCK_SETTIME _IOW (PTP_CLK_MAGIC, 5, struct timespec)
+
+#define PTP_CLOCK_GETCAPS _IOR (PTP_CLK_MAGIC, 6, struct ptp_clock_caps)
+#define PTP_CLOCK_GETTIMER _IOWR (PTP_CLK_MAGIC, 7, struct ptp_clock_timer)
+#define PTP_CLOCK_SETTIMER _IOW (PTP_CLK_MAGIC, 8, struct ptp_clock_timer)
+#define PTP_FEATURE_REQUEST _IOW (PTP_CLK_MAGIC, 9, struct ptp_clock_request)
+
+#endif
diff --git a/include/linux/ptp_clock_kernel.h b/include/linux/ptp_clock_kernel.h
new file mode 100644
index 0000000..d6cc158
--- /dev/null
+++ b/include/linux/ptp_clock_kernel.h
@@ -0,0 +1,137 @@
+/*
+ * PTP 1588 clock support
+ *
+ * Copyright (C) 2010 OMICRON electronics GmbH
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
+ */
+
+#ifndef _PTP_CLOCK_KERNEL_H_
+#define _PTP_CLOCK_KERNEL_H_
+
+#include <linux/ptp_clock.h>
+
+/**
+ * struct ptp_clock_info - decribes a PTP hardware clock
+ *
+ * @owner: The clock driver should set to THIS_MODULE.
+ * @name: A short name to identify the clock.
+ * @max_adj: The maximum possible frequency adjustment, in parts per billon.
+ * @n_alarm: The number of programmable alarms.
+ * @n_ext_ts: The number of external time stamp channels.
+ * @n_per_out: The number of programmable periodic signals.
+ * @pps: Indicates whether the clock supports a PPS callback.
+ * @priv: Passed to the clock operations, for the driver's private use.
+ *
+ * clock operations
+ *
+ * @adjfreq: Adjusts the frequency of the hardware clock.
+ * parameter delta: Desired period change in parts per billion.
+ *
+ * @adjtime: Shifts the time of the hardware clock.
+ * parameter ts: Desired change in seconds and nanoseconds.
+ *
+ * @gettime: Reads the current time from the hardware clock.
+ * parameter ts: Holds the result.
+ *
+ * @settime: Set the current time on the hardware clock.
+ * parameter ts: Time value to set.
+ *
+ * @gettimer: Reads the time remaining from the given timer.
+ * parameter index: Which alarm to query.
+ * parameter ts: Holds the result.
+ *
+ * @settimer: Arms the given timer for periodic or one shot operation.
+ * parameter index: Which alarm to set.
+ * parameter abs: TIMER_ABSTIME, or zero for relative timer.
+ * parameter ts: Alarm time and period to set.
+ *
+ * @enable: Request driver to enable or disable an ancillary feature.
+ * parameter request: Desired resource to enable or disable.
+ * parameter on: Caller passes one to enable or zero to disable.
+ *
+ * The callbacks must all return zero on success, non-zero otherwise.
+ */
+
+struct ptp_clock_info {
+ struct module *owner;
+ char name[16];
+ s32 max_adj;
+ int n_alarm;
+ int n_ext_ts;
+ int n_per_out;
+ int pps;
+ void *priv;
+ int (*adjfreq)(void *priv, s32 delta);
+ int (*adjtime)(void *priv, struct timespec *ts);
+ int (*gettime)(void *priv, struct timespec *ts);
+ int (*settime)(void *priv, struct timespec *ts);
+ int (*gettimer)(void *priv, int index, struct itimerspec *ts);
+ int (*settimer)(void *priv, int index, int abs, struct itimerspec *ts);
+ int (*enable)(void *priv, struct ptp_clock_request *request, int on);
+};
+
+struct ptp_clock;
+
+/**
+ * ptp_clock_register() - register a PTP hardware clock driver
+ *
+ * @info: Structure describing the new clock.
+ */
+
+extern struct ptp_clock *ptp_clock_register(struct ptp_clock_info *info);
+
+/**
+ * ptp_clock_unregister() - unregister a PTP hardware clock driver
+ *
+ * @ptp: The clock to remove from service.
+ */
+
+extern int ptp_clock_unregister(struct ptp_clock *ptp);
+
+
+enum ptp_clock_events {
+ PTP_CLOCK_ALARM,
+ PTP_CLOCK_EXTTS,
+ PTP_CLOCK_PPS,
+};
+
+/**
+ * struct ptp_clock_event - decribes a PTP hardware clock event
+ *
+ * @type: One of the ptp_clock_events enumeration values.
+ * @index: Identifies the source of the event.
+ * @timestamp: When the event occured.
+ */
+
+struct ptp_clock_event {
+ int type;
+ int index;
+ u64 timestamp;
+};
+
+/**
+ * ptp_clock_event() - notify the PTP layer about an event
+ *
+ * This function should only be called from interrupt context.
+ *
+ * @ptp: The clock obtained from ptp_clock_register().
+ * @event: Message structure describing the event.
+ */
+
+extern void ptp_clock_event(struct ptp_clock *ptp,
+ struct ptp_clock_event *event);
+
+#endif
--
1.6.3.3
^ permalink raw reply related
* [PATCH 07/12] phylib: support the National Semiconductor DP83640 PHY.
From: Richard Cochran @ 2010-06-15 16:09 UTC (permalink / raw)
To: netdev; +Cc: linuxppc-dev, devicetree-discuss, Krzysztof Halasa,
linux-arm-kernel
In-Reply-To: <cover.1276615626.git.richard.cochran@omicron.at>
This patch adds a driver for the DP83640. This device is one of a new
generation of PHYs able to time stamp PTP packets.
Signed-off-by: Richard Cochran <richard.cochran@omicron.at>
---
drivers/net/phy/Kconfig | 5 +
drivers/net/phy/Makefile | 1 +
drivers/net/phy/dp83640.c | 439 +++++++++++++++++++++++++++++++++++++++++
drivers/net/phy/dp83640_reg.h | 237 ++++++++++++++++++++++
4 files changed, 682 insertions(+), 0 deletions(-)
create mode 100644 drivers/net/phy/dp83640.c
create mode 100644 drivers/net/phy/dp83640_reg.h
diff --git a/drivers/net/phy/Kconfig b/drivers/net/phy/Kconfig
index a527e37..430cab1 100644
--- a/drivers/net/phy/Kconfig
+++ b/drivers/net/phy/Kconfig
@@ -77,6 +77,11 @@ config NATIONAL_PHY
---help---
Currently supports the DP83865 PHY.
+config DP83640_PHY
+ tristate "Driver for the National Semiconductor DP83640 PHYTER"
+ ---help---
+ Supports the DP83640 PHYTER with IEEE 1588 features.
+
config STE10XP
depends on PHYLIB
tristate "Driver for STMicroelectronics STe10Xp PHYs"
diff --git a/drivers/net/phy/Makefile b/drivers/net/phy/Makefile
index 13bebab..2333215 100644
--- a/drivers/net/phy/Makefile
+++ b/drivers/net/phy/Makefile
@@ -19,6 +19,7 @@ obj-$(CONFIG_FIXED_PHY) += fixed.o
obj-$(CONFIG_MDIO_BITBANG) += mdio-bitbang.o
obj-$(CONFIG_MDIO_GPIO) += mdio-gpio.o
obj-$(CONFIG_NATIONAL_PHY) += national.o
+obj-$(CONFIG_DP83640_PHY) += dp83640.o
obj-$(CONFIG_STE10XP) += ste10Xp.o
obj-$(CONFIG_MICREL_PHY) += micrel.o
obj-$(CONFIG_MDIO_OCTEON) += mdio-octeon.o
diff --git a/drivers/net/phy/dp83640.c b/drivers/net/phy/dp83640.c
new file mode 100644
index 0000000..a3217ea
--- /dev/null
+++ b/drivers/net/phy/dp83640.c
@@ -0,0 +1,439 @@
+/*
+ * Driver for the National Semiconductor DP83640 PHYTER
+ *
+ * Copyright (C) 2010 OMICRON electronics GmbH
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
+ */
+#include <linux/ethtool.h>
+#include <linux/kernel.h>
+#include <linux/list.h>
+#include <linux/mii.h>
+#include <linux/module.h>
+#include <linux/net_tstamp.h>
+#include <linux/netdevice.h>
+#include <linux/phy.h>
+#include <linux/ptp_classify.h>
+
+#include "dp83640_reg.h"
+
+#define DP83640_PHY_ID 0x20005ce1
+#define PAGESEL 0x13
+#define LAYER4 0x02
+#define LAYER2 0x01
+#define MAX_RXTS 8
+
+struct rxts {
+ struct list_head list;
+ u64 ns;
+ unsigned long tmo;
+ u16 seqid;
+ u8 msgtype;
+ u16 hash;
+};
+
+struct dp83640_private {
+ int hwts_tx_en;
+ int hwts_rx_en;
+ int layer;
+ int version;
+ /* protects extended registers from concurrent access */
+ spinlock_t extreg_lock;
+ int page;
+ /* list of rx timestamps */
+ struct list_head rxts;
+ struct list_head pool;
+ struct rxts pool_data[MAX_RXTS];
+};
+
+/* globals */
+
+static struct sock_filter ptp_filter[] = {
+ PTP_FILTER
+};
+
+/* extended register access functions */
+
+static int ext_read(struct phy_device *phydev, int page, u32 regnum)
+{
+ struct dp83640_private *dp83640 = phydev->priv;
+ int val;
+
+ spin_lock(&dp83640->extreg_lock);
+
+ if (dp83640->page != page) {
+ phy_write(phydev, PAGESEL, page);
+ dp83640->page = page;
+ }
+ val = phy_read(phydev, regnum);
+
+ spin_unlock(&dp83640->extreg_lock);
+
+ return val;
+}
+
+static void ext_write(struct phy_device *phydev, int page, u32 regnum, u16 val)
+{
+ struct dp83640_private *dp83640 = phydev->priv;
+
+ spin_lock(&dp83640->extreg_lock);
+
+ if (dp83640->page != page) {
+ phy_write(phydev, PAGESEL, page);
+ dp83640->page = page;
+ }
+ phy_write(phydev, regnum, val);
+
+ spin_unlock(&dp83640->extreg_lock);
+}
+
+static int expired(struct rxts *rxts)
+{
+ return time_after(jiffies, rxts->tmo);
+}
+
+static int match(struct sk_buff *skb, unsigned int type, struct rxts *rxts)
+{
+ u16 *seqid;
+ u8 *msgtype, *data = skb->data;
+
+ switch (type) {
+ case PTP_CLASS_V1_IPV4:
+ msgtype = data + 42 + 32;
+ seqid = (u16 *)(data + 42 + 30);
+ break;
+ case PTP_CLASS_V1_IPV6:
+ msgtype = data + 62 + 32;
+ seqid = (u16 *)(data + 62 + 30);
+ break;
+ case PTP_CLASS_V2_IPV4:
+ msgtype = data + 42 + 0;
+ seqid = (u16 *)(data + 42 + 30);
+ break;
+ case PTP_CLASS_V2_IPV6:
+ msgtype = data + 62 + 0;
+ seqid = (u16 *)(data + 62 + 30);
+ break;
+ case PTP_CLASS_V2_L2:
+ msgtype = data + 14 + 0;
+ seqid = (u16 *)(data + 14 + 30);
+ break;
+ case PTP_CLASS_V2_VLAN:
+ msgtype = data + 18 + 0;
+ seqid = (u16 *)(data + 18 + 30);
+ break;
+ default:
+ return 0;
+ }
+
+ return ((*msgtype & 0xf) == rxts->msgtype && *seqid == rxts->seqid);
+}
+
+static int dp83640_probe(struct phy_device *phydev)
+{
+ struct dp83640_private *dp83640;
+ int i;
+
+ if (sk_chk_filter(ptp_filter, ARRAY_SIZE(ptp_filter))) {
+ pr_err("dp83640: bad ptp filter\n");
+ return -EINVAL;
+ }
+
+ dp83640 = kzalloc(sizeof(struct dp83640_private), GFP_KERNEL);
+ if (!dp83640)
+ return -ENOMEM;
+
+ spin_lock_init(&dp83640->extreg_lock);
+
+ INIT_LIST_HEAD(&dp83640->rxts);
+ INIT_LIST_HEAD(&dp83640->pool);
+
+ for (i = 0; i < MAX_RXTS; i++)
+ list_add(&dp83640->pool_data[i].list, &dp83640->pool);
+
+ phydev->priv = dp83640;
+
+ return 0;
+}
+
+static void dp83640_remove(struct phy_device *phydev)
+{
+ struct dp83640_private *dp83640 = phydev->priv;
+ kfree(dp83640);
+}
+
+static int dp83640_hwtstamp(struct phy_device *phydev, struct ifreq *ifr)
+{
+ struct dp83640_private *dp83640 = phydev->priv;
+ struct hwtstamp_config cfg;
+ u16 txcfg0, rxcfg0;
+
+ if (phydev->bus->locktype != MDIOBUS_ATOMIC_RW)
+ return -EOPNOTSUPP;
+
+ if (copy_from_user(&cfg, ifr->ifr_data, sizeof(cfg)))
+ return -EFAULT;
+
+ if (cfg.flags) /* reserved for future extensions */
+ return -EINVAL;
+
+ switch (cfg.tx_type) {
+ case HWTSTAMP_TX_OFF:
+ dp83640->hwts_tx_en = 0;
+ break;
+ case HWTSTAMP_TX_ON:
+ dp83640->hwts_tx_en = 1;
+ break;
+ default:
+ return -ERANGE;
+ }
+
+ switch (cfg.rx_filter) {
+ case HWTSTAMP_FILTER_NONE:
+ dp83640->hwts_rx_en = 0;
+ dp83640->layer = 0;
+ dp83640->version = 0;
+ break;
+ case HWTSTAMP_FILTER_PTP_V1_L4_EVENT:
+ case HWTSTAMP_FILTER_PTP_V1_L4_SYNC:
+ case HWTSTAMP_FILTER_PTP_V1_L4_DELAY_REQ:
+ dp83640->hwts_rx_en = 1;
+ dp83640->layer = LAYER4;
+ dp83640->version = 1;
+ break;
+ case HWTSTAMP_FILTER_PTP_V2_L4_EVENT:
+ case HWTSTAMP_FILTER_PTP_V2_L4_SYNC:
+ case HWTSTAMP_FILTER_PTP_V2_L4_DELAY_REQ:
+ dp83640->hwts_rx_en = 1;
+ dp83640->layer = LAYER4;
+ dp83640->version = 2;
+ break;
+ case HWTSTAMP_FILTER_PTP_V2_L2_EVENT:
+ case HWTSTAMP_FILTER_PTP_V2_L2_SYNC:
+ case HWTSTAMP_FILTER_PTP_V2_L2_DELAY_REQ:
+ dp83640->hwts_rx_en = 1;
+ dp83640->layer = LAYER2;
+ dp83640->version = 2;
+ break;
+ case HWTSTAMP_FILTER_PTP_V2_EVENT:
+ case HWTSTAMP_FILTER_PTP_V2_SYNC:
+ case HWTSTAMP_FILTER_PTP_V2_DELAY_REQ:
+ dp83640->hwts_rx_en = 1;
+ dp83640->layer = LAYER4|LAYER2;
+ dp83640->version = 2;
+ break;
+ default:
+ return -ERANGE;
+ }
+
+ txcfg0 = (dp83640->version & TX_PTP_VER_MASK) << TX_PTP_VER_SHIFT;
+ rxcfg0 = (dp83640->version & TX_PTP_VER_MASK) << TX_PTP_VER_SHIFT;
+
+ if (dp83640->layer & LAYER2) {
+ txcfg0 |= TX_L2_EN;
+ rxcfg0 |= RX_L2_EN;
+ }
+ if (dp83640->layer & LAYER4) {
+ txcfg0 |= TX_IPV6_EN | TX_IPV4_EN;
+ rxcfg0 |= RX_IPV6_EN | RX_IPV4_EN;
+ }
+
+ if (dp83640->hwts_tx_en)
+ txcfg0 |= TX_TS_EN;
+
+ if (dp83640->hwts_rx_en)
+ rxcfg0 |= RX_TS_EN;
+
+ if (dp83640->hwts_tx_en || dp83640->hwts_rx_en)
+ ext_write(phydev, PAGE4, PTP_CTL, PTP_ENABLE);
+
+ ext_write(phydev, PAGE5, PTP_TXCFG0, txcfg0);
+ ext_write(phydev, PAGE5, PTP_RXCFG0, rxcfg0);
+
+ return copy_to_user(ifr->ifr_data, &cfg, sizeof(cfg)) ? -EFAULT : 0;
+}
+
+static void read_rxts(struct phy_device *phydev, struct rxts *rxts)
+{
+ u64 ns;
+ int overflow, sec, val;
+ u16 hash, seq;
+ u8 msgtype;
+
+ val = ext_read(phydev, PAGE4, PTP_RXTS); /* ns[15:0] */
+ ns = val;
+
+ val = ext_read(phydev, PAGE4, PTP_RXTS); /* overflow[1:0], ns[29:16] */
+ overflow = val >> 14;
+ ns |= (val & 0x3fff) << 16;
+
+ val = ext_read(phydev, PAGE4, PTP_RXTS); /* sec[15:0] */
+ sec = val;
+
+ val = ext_read(phydev, PAGE4, PTP_RXTS); /* sec[31:16] */
+ sec |= val << 16;
+
+ seq = ext_read(phydev, PAGE4, PTP_RXTS); /* sequenceId[15:0] */
+ val = ext_read(phydev, PAGE4, PTP_RXTS); /* messageType[3:0], hash[11:0] */
+ msgtype = (val >> 12) & 0xf;
+ hash = val & 0x0fff;
+ ns += ((u64)sec) * 1000000000ULL;
+
+ rxts->ns = ns;
+ rxts->tmo = jiffies + HZ;
+ rxts->seqid = seq;
+ rxts->msgtype = msgtype;
+ rxts->hash = hash;
+}
+
+static int dp83640_rxtstamp(struct phy_device *phydev, struct sk_buff *skb)
+{
+ struct dp83640_private *dp83640 = phydev->priv;
+ struct skb_shared_hwtstamps *shhwtstamps;
+ struct list_head *this, *next;
+ struct rxts *rxts;
+ unsigned int type;
+ int val;
+
+ if (!dp83640->hwts_rx_en)
+ return -1;
+
+ for (;;) {
+ val = ext_read(phydev, PAGE4, PTP_STS);
+ if (!(val & RXTS_RDY))
+ break;
+
+ if (list_empty(&dp83640->pool)) {
+ pr_warning("dp83640: rx timestamp pool is empty\n");
+ break;
+ }
+
+ rxts = list_first_entry(&dp83640->pool, struct rxts, list);
+ list_del_init(&rxts->list);
+ read_rxts(phydev, rxts);
+ list_add_tail(&rxts->list, &dp83640->rxts);
+ }
+
+ type = sk_run_filter(skb, ptp_filter, ARRAY_SIZE(ptp_filter));
+
+ list_for_each_safe(this, next, &dp83640->rxts) {
+
+ rxts = list_entry(this, struct rxts, list);
+
+ /* check sequenceID, messageType, 12 bit hash of offset 20-29 */
+
+ if (match(skb, type, rxts)) {
+ shhwtstamps = skb_hwtstamps(skb);
+ memset(shhwtstamps, 0, sizeof(*shhwtstamps));
+ shhwtstamps->hwtstamp = ns_to_ktime(rxts->ns);
+ list_del_init(&rxts->list);
+ list_add(&rxts->list, &dp83640->pool);
+ return 0;
+ }
+ if (expired(rxts)) {
+ list_del_init(&rxts->list);
+ list_add(&rxts->list, &dp83640->pool);
+ }
+ }
+
+ return -1;
+}
+
+static int dp83640_txtstamp(struct phy_device *phydev, struct sk_buff *skb)
+{
+ struct dp83640_private *dp83640 = phydev->priv;
+ struct skb_shared_hwtstamps shhwtstamps;
+ union skb_shared_tx *shtx;
+ u64 ns;
+ int cnt, overflow, sec, val;
+
+ if (!dp83640->hwts_tx_en)
+ return -1;
+
+ /*
+ * This really stinks, but we have to poll for the Tx time stamp.
+ */
+ for (cnt = 0; cnt < 100; cnt++) {
+ val = ext_read(phydev, PAGE4, PTP_STS);
+ if (val & TXTS_RDY)
+ break;
+ udelay(1);
+ }
+ if (!(val & TXTS_RDY))
+ return -1;
+
+ val = ext_read(phydev, PAGE4, PTP_TXTS); /* ns[15:0] */
+ ns = val;
+
+ val = ext_read(phydev, PAGE4, PTP_TXTS); /* overflow[1:0], ns[29:16] */
+ overflow = val >> 14;
+ ns |= (val & 0x3fff) << 16;
+
+ val = ext_read(phydev, PAGE4, PTP_TXTS); /* sec[15:0] */
+ sec = val;
+
+ val = ext_read(phydev, PAGE4, PTP_TXTS); /* sec[31:16] */
+ sec |= val << 16;
+
+ ns += ((u64)sec) * 1000000000ULL;
+
+ shtx = skb_tx(skb);
+ shtx->in_progress = 1;
+ memset(&shhwtstamps, 0, sizeof(shhwtstamps));
+ shhwtstamps.hwtstamp = ns_to_ktime(ns);
+ skb_tstamp_tx(skb, &shhwtstamps);
+
+ return 0;
+}
+
+static struct phy_driver dp83640_driver = {
+ .phy_id = DP83640_PHY_ID,
+ .phy_id_mask = 0xfffffff0,
+ .name = "NatSemi DP83640",
+ .features = PHY_BASIC_FEATURES,
+ .flags = 0,
+ .probe = dp83640_probe,
+ .remove = dp83640_remove,
+ .config_aneg = genphy_config_aneg,
+ .read_status = genphy_read_status,
+ .hwtstamp = dp83640_hwtstamp,
+ .rxtstamp = dp83640_rxtstamp,
+ .txtstamp = dp83640_txtstamp,
+ .driver = {.owner = THIS_MODULE,}
+};
+
+static int __init dp83640_init(void)
+{
+ return phy_driver_register(&dp83640_driver);
+}
+
+static void __exit dp83640_exit(void)
+{
+ phy_driver_unregister(&dp83640_driver);
+}
+
+MODULE_DESCRIPTION("National Semiconductor DP83640 PHY driver");
+MODULE_AUTHOR("Richard Cochran <richard.cochran@omicron.at>");
+MODULE_LICENSE("GPL");
+
+module_init(dp83640_init);
+module_exit(dp83640_exit);
+
+static struct mdio_device_id dp83640_tbl[] = {
+ { DP83640_PHY_ID, 0xfffffff0 },
+ { }
+};
+
+MODULE_DEVICE_TABLE(mdio, dp83640_tbl);
diff --git a/drivers/net/phy/dp83640_reg.h b/drivers/net/phy/dp83640_reg.h
new file mode 100644
index 0000000..349fdbc
--- /dev/null
+++ b/drivers/net/phy/dp83640_reg.h
@@ -0,0 +1,237 @@
+/* dp83640_reg.h
+ * Generated by regen.tcl on Wed Jun 02 07:11:36 AM CEST 2010
+ */
+#ifndef HAVE_DP83640_REGISTERS
+#define HAVE_DP83640_REGISTERS
+
+#define PAGE4 0x0004
+#define PTP_CTL 0x0014 /* PTP Control Register */
+#define PTP_TDR 0x0015 /* PTP Time Data Register */
+#define PTP_STS 0x0016 /* PTP Status Register */
+#define PTP_TSTS 0x0017 /* PTP Trigger Status Register */
+#define PTP_RATEL 0x0018 /* PTP Rate Low Register */
+#define PTP_RATEH 0x0019 /* PTP Rate High Register */
+#define PTP_RDCKSUM 0x001a /* PTP Read Checksum */
+#define PTP_WRCKSUM 0x001b /* PTP Write Checksum */
+#define PTP_TXTS 0x001c /* PTP Transmit Timestamp Register, in four 16-bit reads */
+#define PTP_RXTS 0x001d /* PTP Receive Timestamp Register, in six? 16-bit reads */
+#define PTP_ESTS 0x001e /* PTP Event Status Register */
+#define PTP_EDATA 0x001f /* PTP Event Data Register */
+
+#define PAGE5 0x0005
+#define PTP_TRIG 0x0014 /* PTP Trigger Configuration Register */
+#define PTP_EVNT 0x0015 /* PTP Event Configuration Register */
+#define PTP_TXCFG0 0x0016 /* PTP Transmit Configuration Register 0 */
+#define PTP_TXCFG1 0x0017 /* PTP Transmit Configuration Register 1 */
+#define PSF_CFG0 0x0018 /* PHY Status Frame Configuration Register 0 */
+#define PTP_RXCFG0 0x0019 /* PTP Receive Configuration Register 0 */
+#define PTP_RXCFG1 0x001a /* PTP Receive Configuration Register 1 */
+#define PTP_RXCFG2 0x001b /* PTP Receive Configuration Register 2 */
+#define PTP_RXCFG3 0x001c /* PTP Receive Configuration Register 3 */
+#define PTP_RXCFG4 0x001d /* PTP Receive Configuration Register 4 */
+#define PTP_TRDL 0x001e /* PTP Temporary Rate Duration Low Register */
+#define PTP_TRDH 0x001f /* PTP Temporary Rate Duration High Register */
+
+#define PAGE6 0x0006
+#define PTP_COC 0x0014 /* PTP Clock Output Control Register */
+#define PSF_CFG1 0x0015 /* PHY Status Frame Configuration Register 1 */
+#define PSF_CFG2 0x0016 /* PHY Status Frame Configuration Register 2 */
+#define PSF_CFG3 0x0017 /* PHY Status Frame Configuration Register 3 */
+#define PSF_CFG4 0x0018 /* PHY Status Frame Configuration Register 4 */
+#define PTP_SFDCFG 0x0019 /* PTP SFD Configuration Register */
+#define PTP_INTCTL 0x001a /* PTP Interrupt Control Register */
+#define PTP_CLKSRC 0x001b /* PTP Clock Source Register */
+#define PTP_ETR 0x001c /* PTP Ethernet Type Register */
+#define PTP_OFF 0x001d /* PTP Offset Register */
+#define PTP_GPIOMON 0x001e /* PTP GPIO Monitor Register */
+#define PTP_RXHASH 0x001f /* PTP Receive Hash Register */
+
+/* Bit definitions for the PTP_CTL register */
+#define TRIG_SEL_SHIFT (10) /* PTP Trigger Select */
+#define TRIG_SEL_MASK (0x7)
+#define TRIG_DIS (1<<9) /* Disable PTP Trigger */
+#define TRIG_EN (1<<8) /* Enable PTP Trigger */
+#define TRIG_READ (1<<7) /* Read PTP Trigger */
+#define TRIG_LOAD (1<<6) /* Load PTP Trigger */
+#define PTP_RD_CLK (1<<5) /* Read PTP Clock */
+#define PTP_LOAD_CLK (1<<4) /* Load PTP Clock */
+#define PTP_STEP_CLK (1<<3) /* Step PTP Clock */
+#define PTP_ENABLE (1<<2) /* Enable PTP Clock */
+#define PTP_DISABLE (1<<1) /* Disable PTP Clock */
+#define PTP_RESET (1<<0) /* Reset PTP Clock */
+
+/* Bit definitions for the PTP_STS register */
+#define TXTS_RDY (1<<11) /* Transmit Timestamp Ready */
+#define RXTS_RDY (1<<10) /* Receive Timestamp Ready */
+#define TRIG_DONE (1<<9) /* PTP Trigger Done */
+#define EVENT_RDY (1<<8) /* PTP Event Timestamp Ready */
+#define TXTS_IE (1<<3) /* Transmit Timestamp Interrupt Enable */
+#define RXTS_IE (1<<2) /* Receive Timestamp Interrupt Enable */
+#define TRIG_IE (1<<1) /* Trigger Interrupt Enable */
+#define EVENT_IE (1<<0) /* Event Interrupt Enable */
+
+/* Bit definitions for the PTP_TSTS register */
+#define TRIG7_ERROR (1<<15) /* Trigger 7 Error */
+#define TRIG7_ACTIVE (1<<14) /* Trigger 7 Active */
+#define TRIG6_ERROR (1<<13) /* Trigger 6 Error */
+#define TRIG6_ACTIVE (1<<12) /* Trigger 6 Active */
+#define TRIG5_ERROR (1<<11) /* Trigger 5 Error */
+#define TRIG5_ACTIVE (1<<10) /* Trigger 5 Active */
+#define TRIG4_ERROR (1<<9) /* Trigger 4 Error */
+#define TRIG4_ACTIVE (1<<8) /* Trigger 4 Active */
+#define TRIG3_ERROR (1<<7) /* Trigger 3 Error */
+#define TRIG3_ACTIVE (1<<6) /* Trigger 3 Active */
+#define TRIG2_ERROR (1<<5) /* Trigger 2 Error */
+#define TRIG2_ACTIVE (1<<4) /* Trigger 2 Active */
+#define TRIG1_ERROR (1<<3) /* Trigger 1 Error */
+#define TRIG1_ACTIVE (1<<2) /* Trigger 1 Active */
+#define TRIG0_ERROR (1<<1) /* Trigger 0 Error */
+#define TRIG0_ACTIVE (1<<0) /* Trigger 0 Active */
+
+/* Bit definitions for the PTP_RATEH register */
+#define PTP_RATE_DIR (1<<15) /* PTP Rate Direction */
+#define PTP_TMP_RATE (1<<14) /* PTP Temporary Rate */
+#define PTP_RATE_HI_SHIFT (0) /* PTP Rate High 10-bits */
+#define PTP_RATE_HI_MASK (0x3ff)
+
+/* Bit definitions for the PTP_ESTS register */
+#define EVNTS_MISSED_SHIFT (8) /* Indicates number of events missed */
+#define EVNTS_MISSED_MASK (0x7)
+#define EVNT_TS_LEN_SHIFT (6) /* Indicates length of the Timestamp field in 16-bit words minus 1 */
+#define EVNT_TS_LEN_MASK (0x3)
+#define EVNT_RF (1<<5) /* Indicates whether the event is a rise or falling event */
+#define EVNT_NUM_SHIFT (2) /* Indicates Event Timestamp Unit which detected an event */
+#define EVNT_NUM_MASK (0x7)
+#define MULT_EVNT (1<<1) /* Indicates multiple events were detected at the same time */
+#define EVENT_DET (1<<0) /* PTP Event Detected */
+
+/* Bit definitions for the PTP_EDATA register */
+#define E7_RISE (1<<15) /* Indicates direction of Event 7 */
+#define E7_DET (1<<14) /* Indicates Event 7 detected */
+#define E6_RISE (1<<13) /* Indicates direction of Event 6 */
+#define E6_DET (1<<12) /* Indicates Event 6 detected */
+#define E5_RISE (1<<11) /* Indicates direction of Event 5 */
+#define E5_DET (1<<10) /* Indicates Event 5 detected */
+#define E4_RISE (1<<9) /* Indicates direction of Event 4 */
+#define E4_DET (1<<8) /* Indicates Event 4 detected */
+#define E3_RISE (1<<7) /* Indicates direction of Event 3 */
+#define E3_DET (1<<6) /* Indicates Event 3 detected */
+#define E2_RISE (1<<5) /* Indicates direction of Event 2 */
+#define E2_DET (1<<4) /* Indicates Event 2 detected */
+#define E1_RISE (1<<3) /* Indicates direction of Event 1 */
+#define E1_DET (1<<2) /* Indicates Event 1 detected */
+#define E0_RISE (1<<1) /* Indicates direction of Event 0 */
+#define E0_DET (1<<0) /* Indicates Event 0 detected */
+
+/* Bit definitions for the PTP_TRIG register */
+#define TRIG_PULSE (1<<15) /* generate a Pulse rather than a single edge */
+#define TRIG_PER (1<<14) /* generate a periodic signal */
+#define TRIG_IF_LATE (1<<13) /* trigger immediately if already past */
+#define TRIG_NOTIFY (1<<12) /* Trigger Notification Enable */
+#define TRIG_GPIO_SHIFT (8) /* Trigger GPIO Connection, value 1-12 */
+#define TRIG_GPIO_MASK (0xf)
+#define TRIG_TOGGLE (1<<7) /* Trigger Toggle Mode Enable */
+#define TRIG_CSEL_SHIFT (1) /* Trigger Configuration Select */
+#define TRIG_CSEL_MASK (0x7)
+#define TRIG_WR (1<<0) /* Trigger Configuration Write */
+
+/* Bit definitions for the PTP_EVNT register */
+#define EVNT_RISE (1<<14) /* Event Rise Detect Enable */
+#define EVNT_FALL (1<<13) /* Event Fall Detect Enable */
+#define EVNT_SINGLE (1<<12) /* enable single event capture operation */
+#define EVNT_GPIO_SHIFT (8) /* Event GPIO Connection, value 1-12 */
+#define EVNT_GPIO_MASK (0xf)
+#define EVNT_SEL_SHIFT (1) /* Event Select */
+#define EVNT_SEL_MASK (0x7)
+#define EVNT_WR (1<<0) /* Event Configuration Write */
+
+/* Bit definitions for the PTP_TXCFG0 register */
+#define SYNC_1STEP (1<<15) /* insert timestamp into transmit Sync Messages */
+#define DR_INSERT (1<<13) /* Insert Delay_Req Timestamp in Delay_Resp (dangerous) */
+#define NTP_TS_EN (1<<12) /* Enable Timestamping of NTP Packets */
+#define IGNORE_2STEP (1<<11) /* Ignore Two_Step flag for One-Step operation */
+#define CRC_1STEP (1<<10) /* Disable checking of CRC for One-Step operation */
+#define CHK_1STEP (1<<9) /* Enable UDP Checksum correction for One-Step Operation */
+#define IP1588_EN (1<<8) /* Enable IEEE 1588 defined IP address filter */
+#define TX_L2_EN (1<<7) /* Layer2 Timestamp Enable */
+#define TX_IPV6_EN (1<<6) /* IPv6 Timestamp Enable */
+#define TX_IPV4_EN (1<<5) /* IPv4 Timestamp Enable */
+#define TX_PTP_VER_SHIFT (1) /* Enable Timestamp capture for IEEE 1588 version X */
+#define TX_PTP_VER_MASK (0xf)
+#define TX_TS_EN (1<<0) /* Transmit Timestamp Enable */
+
+/* Bit definitions for the PTP_TXCFG1 register */
+#define BYTE0_MASK_SHIFT (8) /* Bit mask to be used for matching Byte0 of the PTP Message */
+#define BYTE0_MASK_MASK (0xff)
+#define BYTE0_DATA_SHIFT (0) /* Data to be used for matching Byte0 of the PTP Message */
+#define BYTE0_DATA_MASK (0xff)
+
+/* Bit definitions for the PTP_RXCFG0 register */
+#define DOMAIN_EN (1<<15) /* Domain Match Enable */
+#define ALT_MAST_DIS (1<<14) /* Alternate Master Timestamp Disable */
+#define USER_IP_SEL (1<<13) /* Selects portion of IP address accessible thru PTP_RXCFG2 */
+#define USER_IP_EN (1<<12) /* Enable User-programmed IP address filter */
+#define RX_SLAVE (1<<11) /* Receive Slave Only */
+#define IP1588_EN_SHIFT (8) /* Enable IEEE 1588 defined IP address filters */
+#define IP1588_EN_MASK (0xf)
+#define RX_L2_EN (1<<7) /* Layer2 Timestamp Enable */
+#define RX_IPV6_EN (1<<6) /* IPv6 Timestamp Enable */
+#define RX_IPV4_EN (1<<5) /* IPv4 Timestamp Enable */
+#define RX_PTP_VER_SHIFT (1) /* Enable Timestamp capture for IEEE 1588 version X */
+#define RX_PTP_VER_MASK (0xf)
+#define RX_TS_EN (1<<0) /* Receive Timestamp Enable */
+
+/* Bit definitions for the PTP_RXCFG1 register */
+#define BYTE0_MASK_SHIFT (8) /* Bit mask to be used for matching Byte0 of the PTP Message */
+#define BYTE0_MASK_MASK (0xff)
+#define BYTE0_DATA_SHIFT (0) /* Data to be used for matching Byte0 of the PTP Message */
+#define BYTE0_DATA_MASK (0xff)
+
+/* Bit definitions for the PTP_RXCFG3 register */
+#define TS_MIN_IFG_SHIFT (12) /* Minimum Inter-frame Gap */
+#define TS_MIN_IFG_MASK (0xf)
+#define ACC_UDP (1<<11) /* Record Timestamp if UDP Checksum Error */
+#define ACC_CRC (1<<10) /* Record Timestamp if CRC Error */
+#define TS_APPEND (1<<9) /* Append Timestamp for L2 */
+#define TS_INSERT (1<<8) /* Enable Timestamp Insertion */
+#define PTP_DOMAIN_SHIFT (0) /* PTP Message domainNumber field */
+#define PTP_DOMAIN_MASK (0xff)
+
+/* Bit definitions for the PTP_RXCFG4 register */
+#define IPV4_UDP_MOD (1<<15) /* Enable IPV4 UDP Modification */
+#define TS_SEC_EN (1<<14) /* Enable Timestamp Seconds */
+#define TS_SEC_LEN_SHIFT (12) /* Inserted Timestamp Seconds Length */
+#define TS_SEC_LEN_MASK (0x3)
+#define RXTS_NS_OFF_SHIFT (6) /* Receive Timestamp Nanoseconds offset */
+#define RXTS_NS_OFF_MASK (0x3f)
+#define RXTS_SEC_OFF_SHIFT (0) /* Receive Timestamp Seconds offset */
+#define RXTS_SEC_OFF_MASK (0x3f)
+
+/* Bit definitions for the PTP_COC register */
+#define PTP_CLKOUT_EN (1<<15) /* PTP Clock Output Enable */
+#define PTP_CLKOUT_SEL (1<<14) /* PTP Clock Output Source Select */
+#define PTP_CLKOUT_SPEEDSEL (1<<13) /* PTP Clock Output I/O Speed Select */
+#define PTP_CLKDIV_SHIFT (0) /* PTP Clock Divide-by Value */
+#define PTP_CLKDIV_MASK (0xff)
+
+/* Bit definitions for the PTP_SFDCFG register */
+#define TX_SFD_GPIO_SHIFT (4) /* TX SFD GPIO Select, value 1-12 */
+#define TX_SFD_GPIO_MASK (0xf)
+#define RX_SFD_GPIO_SHIFT (0) /* RX SFD GPIO Select, value 1-12 */
+#define RX_SFD_GPIO_MASK (0xf)
+
+/* Bit definitions for the PTP_INTCTL register */
+#define PTP_INT_GPIO_SHIFT (0) /* PTP Interrupt GPIO Select */
+#define PTP_INT_GPIO_MASK (0xf)
+
+/* Bit definitions for the PTP_CLKSRC register */
+#define CLK_SRC_SHIFT (14) /* PTP Clock Source Select */
+#define CLK_SRC_MASK (0x3)
+#define CLK_SRC_PER_SHIFT (0) /* PTP Clock Source Period */
+#define CLK_SRC_PER_MASK (0x7f)
+
+/* Bit definitions for the PTP_OFF register */
+#define PTP_OFFSET_SHIFT (0) /* PTP Message offset from preceding header */
+#define PTP_OFFSET_MASK (0xff)
+
+#endif
--
1.6.3.3
^ permalink raw reply related
* [PATCH 06/12] ptp: add a BPF to help drivers detect PTP packets.
From: Richard Cochran @ 2010-06-15 16:09 UTC (permalink / raw)
To: netdev; +Cc: linuxppc-dev, devicetree-discuss, Krzysztof Halasa,
linux-arm-kernel
In-Reply-To: <cover.1276615626.git.richard.cochran@omicron.at>
Certain kinds of hardware time stamping units in both MACs and PHYs have
the limitation that they can only time stamp PTP packets. Drivers for such
hardware are left with the task of correctly matching skbs to time stamps.
This patch adds a BPF that drivers can use to classify PTP packets when
needed.
Signed-off-by: Richard Cochran <richard.cochran@omicron.at>
---
include/linux/ptp_classify.h | 118 ++++++++++++++++++++++++++++++++++++++++++
1 files changed, 118 insertions(+), 0 deletions(-)
create mode 100644 include/linux/ptp_classify.h
diff --git a/include/linux/ptp_classify.h b/include/linux/ptp_classify.h
new file mode 100644
index 0000000..e68efc3
--- /dev/null
+++ b/include/linux/ptp_classify.h
@@ -0,0 +1,118 @@
+/*
+ * PTP 1588 support
+ *
+ * This file implements a BPF that recognizes PTP event messages.
+ *
+ * Copyright (C) 2010 OMICRON electronics GmbH
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
+ */
+
+#ifndef _PTP_CLASSIFY_H_
+#define _PTP_CLASSIFY_H_
+
+#include <linux/if_ether.h>
+#include <linux/if_vlan.h>
+#include <linux/filter.h>
+#ifdef __KERNEL__
+#include <linux/in.h>
+#else
+#include <netinet/in.h>
+#endif
+
+#define PTP_CLASS_NONE 0x00 /* not a PTP event message */
+#define PTP_CLASS_V1 0x01 /* protocol version 1 */
+#define PTP_CLASS_V2 0x02 /* protocol version 2 */
+#define PTP_CLASS_VMASK 0x0f /* max protocol version is 15 */
+#define PTP_CLASS_IPV4 0x10 /* event in an IPV4 UDP packet */
+#define PTP_CLASS_IPV6 0x20 /* event in an IPV6 UDP packet */
+#define PTP_CLASS_L2 0x30 /* event in a L2 packet */
+#define PTP_CLASS_VLAN 0x40 /* event in a VLAN tagged L2 packet */
+#define PTP_CLASS_PMASK 0xf0 /* mask for the packet type field */
+
+#define PTP_CLASS_V1_IPV4 (PTP_CLASS_V1 | PTP_CLASS_IPV4)
+#define PTP_CLASS_V1_IPV6 (PTP_CLASS_V1 | PTP_CLASS_IPV6) /*probably DNE*/
+#define PTP_CLASS_V2_IPV4 (PTP_CLASS_V2 | PTP_CLASS_IPV4)
+#define PTP_CLASS_V2_IPV6 (PTP_CLASS_V2 | PTP_CLASS_IPV6)
+#define PTP_CLASS_V2_L2 (PTP_CLASS_V2 | PTP_CLASS_L2)
+#define PTP_CLASS_V2_VLAN (PTP_CLASS_V2 | PTP_CLASS_VLAN)
+
+#define PTP_EV_PORT 319
+
+#define OFF_ETYPE 12
+#define OFF_IHL 14
+#define OFF_FRAG 20
+#define OFF_PROTO4 23
+#define OFF_NEXT 6
+#define OFF_UDP_DST 2
+
+#define IP6_HLEN 40
+#define UDP_HLEN 8
+
+#define RELOFF_DST4 (ETH_HLEN + OFF_UDP_DST)
+#define OFF_DST6 (ETH_HLEN + IP6_HLEN + OFF_UDP_DST)
+#define OFF_PTP6 (ETH_HLEN + IP6_HLEN + UDP_HLEN)
+
+#define OP_AND (BPF_ALU | BPF_AND | BPF_K)
+#define OP_JEQ (BPF_JMP | BPF_JEQ | BPF_K)
+#define OP_JSET (BPF_JMP | BPF_JSET | BPF_K)
+#define OP_LDB (BPF_LD | BPF_B | BPF_ABS)
+#define OP_LDH (BPF_LD | BPF_H | BPF_ABS)
+#define OP_LDHI (BPF_LD | BPF_H | BPF_IND)
+#define OP_LDX (BPF_LDX | BPF_B | BPF_MSH)
+#define OP_OR (BPF_ALU | BPF_OR | BPF_K)
+#define OP_RETA (BPF_RET | BPF_A)
+#define OP_RETK (BPF_RET | BPF_K)
+
+#define PTP_FILTER \
+ {OP_LDH, 0, 0, OFF_ETYPE }, /* */ \
+ {OP_JEQ, 0, 12, ETH_P_IP }, /* f goto L20 */ \
+ {OP_LDB, 0, 0, OFF_PROTO4 }, /* */ \
+ {OP_JEQ, 0, 9, IPPROTO_UDP }, /* f goto L10 */ \
+ {OP_LDH, 0, 0, OFF_FRAG }, /* */ \
+ {OP_JSET, 7, 0, 0x1fff }, /* t goto L11 */ \
+ {OP_LDX, 0, 0, OFF_IHL }, /* */ \
+ {OP_LDHI, 0, 0, RELOFF_DST4 }, /* */ \
+ {OP_JEQ, 0, 4, PTP_EV_PORT }, /* f goto L12 */ \
+ {OP_LDHI, 0, 0, ETH_HLEN + UDP_HLEN }, /* */ \
+ {OP_AND, 0, 0, PTP_CLASS_VMASK }, /* */ \
+ {OP_OR, 0, 0, PTP_CLASS_IPV4 }, /* */ \
+ {OP_RETA, 0, 0, 0 }, /* */ \
+/*L1x*/ {OP_RETK, 0, 0, PTP_CLASS_NONE }, /* */ \
+/*L20*/ {OP_JEQ, 0, 9, ETH_P_IPV6 }, /* f goto L40 */ \
+ {OP_LDB, 0, 0, ETH_HLEN + OFF_NEXT }, /* */ \
+ {OP_JEQ, 0, 6, IPPROTO_UDP }, /* f goto L30 */ \
+ {OP_LDH, 0, 0, OFF_DST6 }, /* */ \
+ {OP_JEQ, 0, 4, PTP_EV_PORT }, /* f goto L31 */ \
+ {OP_LDH, 0, 0, OFF_PTP6 }, /* */ \
+ {OP_AND, 0, 0, PTP_CLASS_VMASK }, /* */ \
+ {OP_OR, 0, 0, PTP_CLASS_IPV6 }, /* */ \
+ {OP_RETA, 0, 0, 0 }, /* */ \
+/*L3x*/ {OP_RETK, 0, 0, PTP_CLASS_NONE }, /* */ \
+/*L40*/ {OP_JEQ, 0, 6, ETH_P_8021Q }, /* f goto L50 */ \
+ {OP_LDH, 0, 0, OFF_ETYPE + 4 }, /* */ \
+ {OP_JEQ, 0, 9, ETH_P_1588 }, /* f goto L60 */ \
+ {OP_LDH, 0, 0, ETH_HLEN + VLAN_HLEN }, /* */ \
+ {OP_AND, 0, 0, PTP_CLASS_VMASK }, /* */ \
+ {OP_OR, 0, 0, PTP_CLASS_VLAN }, /* */ \
+ {OP_RETA, 0, 0, 0 }, /* */ \
+/*L50*/ {OP_JEQ, 0, 4, ETH_P_1588 }, /* f goto L61 */ \
+ {OP_LDH, 0, 0, ETH_HLEN }, /* */ \
+ {OP_AND, 0, 0, PTP_CLASS_VMASK }, /* */ \
+ {OP_OR, 0, 0, PTP_CLASS_L2 }, /* */ \
+ {OP_RETA, 0, 0, 0 }, /* */ \
+/*L6x*/ {OP_RETK, 0, 0, PTP_CLASS_NONE },
+
+#endif
--
1.6.3.3
^ permalink raw reply related
* [PATCH 05/12] phylib: Allow reading and writing a mii bus from atomic context.
From: Richard Cochran @ 2010-06-15 16:08 UTC (permalink / raw)
To: netdev; +Cc: linuxppc-dev, devicetree-discuss, Krzysztof Halasa,
linux-arm-kernel
In-Reply-To: <cover.1276615626.git.richard.cochran@omicron.at>
In order to support hardware time stamping from a PHY, it is necessary to
read from the PHY while running in_interrupt(). This patch allows a mii
bus to operate in an atomic context. An mii_bus driver may declare itself
capable for this mode. Drivers which do not do this will remain with the
default that bus operations may sleep.
Signed-off-by: Richard Cochran <richard.cochran@omicron.at>
---
drivers/net/phy/mdio_bus.c | 35 ++++++++++++++++++++++++++++-------
include/linux/phy.h | 13 +++++++++++--
2 files changed, 39 insertions(+), 9 deletions(-)
diff --git a/drivers/net/phy/mdio_bus.c b/drivers/net/phy/mdio_bus.c
index 6a6b819..441be7d 100644
--- a/drivers/net/phy/mdio_bus.c
+++ b/drivers/net/phy/mdio_bus.c
@@ -36,6 +36,22 @@
#include <asm/irq.h>
#include <asm/uaccess.h>
+static inline void mdiobus_lock(struct mii_bus *bus)
+{
+ if (MDIOBUS_SLEEPS_RW == bus->locktype)
+ mutex_lock(&bus->mdio_lock.m);
+ else
+ spin_lock(&bus->mdio_lock.s);
+}
+
+static inline void mdiobus_unlock(struct mii_bus *bus)
+{
+ if (MDIOBUS_SLEEPS_RW == bus->locktype)
+ mutex_unlock(&bus->mdio_lock.m);
+ else
+ spin_unlock(&bus->mdio_lock.s);
+}
+
/**
* mdiobus_alloc - allocate a mii_bus structure
*
@@ -107,7 +123,10 @@ int mdiobus_register(struct mii_bus *bus)
return -EINVAL;
}
- mutex_init(&bus->mdio_lock);
+ if (MDIOBUS_SLEEPS_RW == bus->locktype)
+ mutex_init(&bus->mdio_lock.m);
+ else
+ spin_lock_init(&bus->mdio_lock.s);
if (bus->reset)
bus->reset(bus);
@@ -212,11 +231,12 @@ int mdiobus_read(struct mii_bus *bus, int addr, u32 regnum)
{
int retval;
- BUG_ON(in_interrupt());
+ if (MDIOBUS_SLEEPS_RW == bus->locktype)
+ BUG_ON(in_interrupt());
- mutex_lock(&bus->mdio_lock);
+ mdiobus_lock(bus);
retval = bus->read(bus, addr, regnum);
- mutex_unlock(&bus->mdio_lock);
+ mdiobus_unlock(bus);
return retval;
}
@@ -237,11 +257,12 @@ int mdiobus_write(struct mii_bus *bus, int addr, u32 regnum, u16 val)
{
int err;
- BUG_ON(in_interrupt());
+ if (MDIOBUS_SLEEPS_RW == bus->locktype)
+ BUG_ON(in_interrupt());
- mutex_lock(&bus->mdio_lock);
+ mdiobus_lock(bus);
err = bus->write(bus, addr, regnum, val);
- mutex_unlock(&bus->mdio_lock);
+ mdiobus_unlock(bus);
return err;
}
diff --git a/include/linux/phy.h b/include/linux/phy.h
index 7a8caac..352b030 100644
--- a/include/linux/phy.h
+++ b/include/linux/phy.h
@@ -98,11 +98,20 @@ struct mii_bus {
int (*write)(struct mii_bus *bus, int phy_id, int regnum, u16 val);
int (*reset)(struct mii_bus *bus);
+ /* Indicates whether bus may be used from an atomic context. */
+ enum {
+ MDIOBUS_SLEEPS_RW,
+ MDIOBUS_ATOMIC_RW
+ } locktype;
+
/*
- * A lock to ensure that only one thing can read/write
+ * A lock or mutex to ensure that only one thing can read/write
* the MDIO bus at a time
*/
- struct mutex mdio_lock;
+ union {
+ struct mutex m;
+ spinlock_t s;
+ } mdio_lock;
struct device *parent;
enum {
--
1.6.3.3
^ permalink raw reply related
* RE: [v2 Patch 1/2] s2io: add dynamic LRO disable support
From: Ramkrishna Vepa @ 2010-06-15 16:07 UTC (permalink / raw)
To: Cong Wang
Cc: Ben Hutchings, netdev@vger.kernel.org, nhorman@redhat.com,
sgruszka@redhat.com, herbert.xu@redhat.com, davem@davemloft.net
In-Reply-To: <4C173932.7010101@redhat.com>
> >> On Wed, 2010-06-09 at 06:05 -0400, Amerigo Wang wrote:
> >> [...]
> >>> +static int s2io_ethtool_set_flags(struct net_device *dev, u32 data)
> >>> +{
> >>> + struct s2io_nic *sp = netdev_priv(dev);
> >>> + int rc = 0;
> >>> + int changed = 0;
> >>> +
> >>> + if (data& ETH_FLAG_LRO) {
> >>> + if (lro_enable) {
> >>> + if (!(dev->features& NETIF_F_LRO)) {
> >>> + dev->features |= NETIF_F_LRO;
> >>> + changed = 1;
> >>> + }
> >>> + } else
> >>> + rc = -EINVAL;
> >>> + } else if (dev->features& NETIF_F_LRO) {
> >>> + dev->features&= ~NETIF_F_LRO;
> >>> + changed = 1;
> >>> + }
> >>> +
> >>> + if (changed&& netif_running(dev)) {
> >>> + s2io_stop_all_tx_queue(sp);
> >>> + s2io_card_down(sp);
> >>> + sp->lro = dev->features& NETIF_F_LRO;
> >>> + rc = s2io_card_up(sp);
> >>> + s2io_start_all_tx_queue(sp);
> >> [...]
> >>
> >> Is it safe to call s2io_start_all_tx_queue() if s2io_card_up() failed?
> > Ben,
> > Good point. If s2io_card_up() fails the chip will not be accessed, so
> it's safe but all transmit skbs will be freed without the user knowing the
> reason for failing to transmit or receive for that matter. The other
> option is to return with a failure and get the watchdog timer reset the
> adapter.
> >
>
> (Sorry for the delay, I was on vacation.)
>
> So it seems the latter option is better?
Yes, this will work.
Thanks,
Ram
The information and any attached documents contained in this message
may be confidential and/or legally privileged. The message is
intended solely for the addressee(s). If you are not the intended
recipient, you are hereby notified that any use, dissemination, or
reproduction is strictly prohibited and may be unlawful. If you are
not the intended recipient, please contact the sender immediately by
return e-mail and destroy all copies of the original message.
^ permalink raw reply
* [PATCH 04/12] phylib: add a way to make PHY time stamps possible.
From: Richard Cochran @ 2010-06-15 16:08 UTC (permalink / raw)
To: netdev; +Cc: linuxppc-dev, devicetree-discuss, Krzysztof Halasa,
linux-arm-kernel
In-Reply-To: <cover.1276615626.git.richard.cochran@omicron.at>
This patch adds a new networking option to allow hardware time stamps
from PHY devices. Using PHY time stamps will still require adding two
inline function calls to each MAC driver. The CONFIG option makes these
calls safe to add, since the calls become NOOPs when the option is
disabled.
Signed-off-by: Richard Cochran <richard.cochran@omicron.at>
---
include/linux/phy.h | 4 ++++
include/linux/skbuff.h | 32 ++++++++++++++++++++++++++++++++
net/Kconfig | 11 +++++++++++
3 files changed, 47 insertions(+), 0 deletions(-)
diff --git a/include/linux/phy.h b/include/linux/phy.h
index a5e9df1..7a8caac 100644
--- a/include/linux/phy.h
+++ b/include/linux/phy.h
@@ -234,6 +234,8 @@ enum phy_state {
PHY_RESUMING
};
+struct sk_buff;
+
/* phy_device: An instance of a PHY
*
* drv: Pointer to the driver for this PHY instance
@@ -404,6 +406,8 @@ struct phy_driver {
/* Handles SIOCSHWTSTAMP ioctl for hardware time stamping. */
int (*hwtstamp)(struct phy_device *phydev, struct ifreq *ifr);
+ int (*rxtstamp)(struct phy_device *phydev, struct sk_buff *skb);
+ int (*txtstamp)(struct phy_device *phydev, struct sk_buff *skb);
struct device_driver driver;
};
diff --git a/include/linux/skbuff.h b/include/linux/skbuff.h
index 645e78d..7b650d4 100644
--- a/include/linux/skbuff.h
+++ b/include/linux/skbuff.h
@@ -29,6 +29,7 @@
#include <linux/rcupdate.h>
#include <linux/dmaengine.h>
#include <linux/hrtimer.h>
+#include <linux/phy.h>
/* Don't change this without changing skb_csum_unnecessary! */
#define CHECKSUM_NONE 0
@@ -1942,6 +1943,37 @@ static inline ktime_t net_invalid_timestamp(void)
extern void skb_tstamp_tx(struct sk_buff *orig_skb,
struct skb_shared_hwtstamps *hwtstamps);
+#ifdef CONFIG_NETWORK_PHY_TIMESTAMPING
+
+static inline void skb_tx_timetamp(struct phy_device *phy, struct sk_buff *skb)
+{
+ union skb_shared_tx *shtx = skb_tx(skb);
+
+ if (shtx->hardware && phy && phy->drv->txtstamp)
+ phy->drv->txtstamp(phy, skb);
+
+ if (shtx->software && !shtx->in_progress)
+ skb_tstamp_tx(skb, NULL);
+}
+
+static inline void skb_rx_timetamp(struct phy_device *phy, struct sk_buff *skb)
+{
+ if (phy && phy->drv->rxtstamp)
+ phy->drv->rxtstamp(phy, skb);
+}
+
+#else /* CONFIG_NETWORK_PHY_TIMESTAMPING */
+
+static inline void skb_tx_timetamp(struct phy_device *phy, struct sk_buff *skb)
+{
+}
+
+static inline void skb_rx_timetamp(struct phy_device *phy, struct sk_buff *skb)
+{
+}
+
+#endif /* !CONFIG_NETWORK_PHY_TIMESTAMPING */
+
extern __sum16 __skb_checksum_complete_head(struct sk_buff *skb, int len);
extern __sum16 __skb_checksum_complete(struct sk_buff *skb);
diff --git a/net/Kconfig b/net/Kconfig
index 0d68b40..3fa7ae3 100644
--- a/net/Kconfig
+++ b/net/Kconfig
@@ -86,6 +86,17 @@ config NETWORK_SECMARK
to nfmark, but designated for security purposes.
If you are unsure how to answer this question, answer N.
+config NETWORK_PHY_TIMESTAMPING
+ bool "Timestamping in PHY devices"
+ depends on EXPERIMENTAL
+ help
+ This allows timestamping of network packets by PHYs with
+ hardware timestamping capabilities. This option adds some
+ overhead in the transmit and receive paths. Note that this
+ option also requires support in the MAC driver.
+
+ If you are unsure how to answer this question, answer N.
+
menuconfig NETFILTER
bool "Network packet filtering framework (Netfilter)"
---help---
--
1.6.3.3
^ permalink raw reply related
* [PATCH 03/12] phylib: add a driver method for the SIOCSHWTSTAMP ioctl.
From: Richard Cochran @ 2010-06-15 16:08 UTC (permalink / raw)
To: netdev; +Cc: linuxppc-dev, devicetree-discuss, Krzysztof Halasa,
linux-arm-kernel
In-Reply-To: <cover.1276615626.git.richard.cochran@omicron.at>
This patch adds a phy driver method for configuring hardware time stamping.
Drivers may optionally implement this function.
Signed-off-by: Richard Cochran <richard.cochran@omicron.at>
---
drivers/net/phy/phy.c | 5 +++++
include/linux/phy.h | 3 +++
2 files changed, 8 insertions(+), 0 deletions(-)
diff --git a/drivers/net/phy/phy.c b/drivers/net/phy/phy.c
index bd88d81..5130db8 100644
--- a/drivers/net/phy/phy.c
+++ b/drivers/net/phy/phy.c
@@ -361,6 +361,11 @@ int phy_mii_ioctl(struct phy_device *phydev,
}
break;
+ case SIOCSHWTSTAMP:
+ if (phydev->drv->hwtstamp)
+ return phydev->drv->hwtstamp(phydev, ifr);
+ /* fall through */
+
default:
return -EOPNOTSUPP;
}
diff --git a/include/linux/phy.h b/include/linux/phy.h
index d63736a..a5e9df1 100644
--- a/include/linux/phy.h
+++ b/include/linux/phy.h
@@ -402,6 +402,9 @@ struct phy_driver {
/* Clears up any memory if needed */
void (*remove)(struct phy_device *phydev);
+ /* Handles SIOCSHWTSTAMP ioctl for hardware time stamping. */
+ int (*hwtstamp)(struct phy_device *phydev, struct ifreq *ifr);
+
struct device_driver driver;
};
#define to_phy_driver(d) container_of(d, struct phy_driver, driver)
--
1.6.3.3
^ permalink raw reply related
* [PATCH 02/12] phylib: do not filter phy_mii_ioctl()
From: Richard Cochran @ 2010-06-15 16:07 UTC (permalink / raw)
To: netdev; +Cc: linuxppc-dev, devicetree-discuss, Krzysztof Halasa,
linux-arm-kernel
In-Reply-To: <cover.1276615626.git.richard.cochran@omicron.at>
Two maverick drivers filter the ioctl commands passed to phy_mii_ioctl().
This is unnecessary since phylib will check the command in any case.
Signed-off-by: Richard Cochran <richard.cochran@omicron.at>
---
drivers/net/cpmac.c | 5 +----
drivers/net/stmmac/stmmac_main.c | 22 ++++++++--------------
2 files changed, 9 insertions(+), 18 deletions(-)
diff --git a/drivers/net/cpmac.c b/drivers/net/cpmac.c
index d4c5ca4..0e47ca1 100644
--- a/drivers/net/cpmac.c
+++ b/drivers/net/cpmac.c
@@ -846,11 +846,8 @@ static int cpmac_ioctl(struct net_device *dev, struct ifreq *ifr, int cmd)
return -EINVAL;
if (!priv->phy)
return -EINVAL;
- if ((cmd == SIOCGMIIPHY) || (cmd == SIOCGMIIREG) ||
- (cmd == SIOCSMIIREG))
- return phy_mii_ioctl(priv->phy, ifr, cmd);
- return -EOPNOTSUPP;
+ return phy_mii_ioctl(priv->phy, ifr, cmd);
}
static int cpmac_get_settings(struct net_device *dev, struct ethtool_cmd *cmd)
diff --git a/drivers/net/stmmac/stmmac_main.c b/drivers/net/stmmac/stmmac_main.c
index 6af7471..acf0616 100644
--- a/drivers/net/stmmac/stmmac_main.c
+++ b/drivers/net/stmmac/stmmac_main.c
@@ -1437,24 +1437,18 @@ static void stmmac_poll_controller(struct net_device *dev)
static int stmmac_ioctl(struct net_device *dev, struct ifreq *rq, int cmd)
{
struct stmmac_priv *priv = netdev_priv(dev);
- int ret = -EOPNOTSUPP;
+ int ret;
if (!netif_running(dev))
return -EINVAL;
- switch (cmd) {
- case SIOCGMIIPHY:
- case SIOCGMIIREG:
- case SIOCSMIIREG:
- if (!priv->phydev)
- return -EINVAL;
-
- spin_lock(&priv->lock);
- ret = phy_mii_ioctl(priv->phydev, rq, cmd);
- spin_unlock(&priv->lock);
- default:
- break;
- }
+ if (!priv->phydev)
+ return -EINVAL;
+
+ spin_lock(&priv->lock);
+ ret = phy_mii_ioctl(priv->phydev, rq, cmd);
+ spin_unlock(&priv->lock);
+
return ret;
}
--
1.6.3.3
^ permalink raw reply related
* [PATCH 01/12] phylib: preserve ifreq parameter when calling generic phy_mii_ioctl()
From: Richard Cochran @ 2010-06-15 16:07 UTC (permalink / raw)
To: netdev; +Cc: linuxppc-dev, devicetree-discuss, Krzysztof Halasa,
linux-arm-kernel
In-Reply-To: <cover.1276615626.git.richard.cochran@omicron.at>
The phy_mii_ioctl() function unnecessarily throws away the original ifreq.
We need access to the ifreq in order to support PHYs that can perform
hardware time stamping.
Signed-off-by: Richard Cochran <richard.cochran@omicron.at>
---
drivers/net/arm/ixp4xx_eth.c | 3 ++-
drivers/net/au1000_eth.c | 2 +-
drivers/net/bcm63xx_enet.c | 2 +-
drivers/net/cpmac.c | 2 +-
drivers/net/dnet.c | 2 +-
drivers/net/ethoc.c | 2 +-
drivers/net/fec.c | 2 +-
drivers/net/fec_mpc52xx.c | 2 +-
drivers/net/fs_enet/fs_enet-main.c | 3 +--
drivers/net/gianfar.c | 2 +-
drivers/net/macb.c | 2 +-
drivers/net/mv643xx_eth.c | 2 +-
drivers/net/octeon/octeon_mgmt.c | 2 +-
drivers/net/phy/phy.c | 3 ++-
drivers/net/sb1250-mac.c | 2 +-
drivers/net/sh_eth.c | 2 +-
drivers/net/smsc911x.c | 2 +-
drivers/net/smsc9420.c | 2 +-
drivers/net/stmmac/stmmac_main.c | 2 +-
drivers/net/tc35815.c | 2 +-
drivers/net/tg3.c | 2 +-
drivers/staging/octeon/ethernet-mdio.c | 2 +-
include/linux/phy.h | 2 +-
net/dsa/slave.c | 3 +--
24 files changed, 26 insertions(+), 26 deletions(-)
diff --git a/drivers/net/arm/ixp4xx_eth.c b/drivers/net/arm/ixp4xx_eth.c
index ee2f842..4f1cc71 100644
--- a/drivers/net/arm/ixp4xx_eth.c
+++ b/drivers/net/arm/ixp4xx_eth.c
@@ -782,7 +782,8 @@ static int eth_ioctl(struct net_device *dev, struct ifreq *req, int cmd)
if (!netif_running(dev))
return -EINVAL;
- return phy_mii_ioctl(port->phydev, if_mii(req), cmd);
+
+ return phy_mii_ioctl(port->phydev, req, cmd);
}
/* ethtool support */
diff --git a/drivers/net/au1000_eth.c b/drivers/net/au1000_eth.c
index ece6128..386d4fe 100644
--- a/drivers/net/au1000_eth.c
+++ b/drivers/net/au1000_eth.c
@@ -978,7 +978,7 @@ static int au1000_ioctl(struct net_device *dev, struct ifreq *rq, int cmd)
if (!aup->phy_dev)
return -EINVAL; /* PHY not controllable */
- return phy_mii_ioctl(aup->phy_dev, if_mii(rq), cmd);
+ return phy_mii_ioctl(aup->phy_dev, rq, cmd);
}
static const struct net_device_ops au1000_netdev_ops = {
diff --git a/drivers/net/bcm63xx_enet.c b/drivers/net/bcm63xx_enet.c
index faf5add..0d2c5da 100644
--- a/drivers/net/bcm63xx_enet.c
+++ b/drivers/net/bcm63xx_enet.c
@@ -1496,7 +1496,7 @@ static int bcm_enet_ioctl(struct net_device *dev, struct ifreq *rq, int cmd)
if (priv->has_phy) {
if (!priv->phydev)
return -ENODEV;
- return phy_mii_ioctl(priv->phydev, if_mii(rq), cmd);
+ return phy_mii_ioctl(priv->phydev, rq, cmd);
} else {
struct mii_if_info mii;
diff --git a/drivers/net/cpmac.c b/drivers/net/cpmac.c
index 3c58db5..d4c5ca4 100644
--- a/drivers/net/cpmac.c
+++ b/drivers/net/cpmac.c
@@ -848,7 +848,7 @@ static int cpmac_ioctl(struct net_device *dev, struct ifreq *ifr, int cmd)
return -EINVAL;
if ((cmd == SIOCGMIIPHY) || (cmd == SIOCGMIIREG) ||
(cmd == SIOCSMIIREG))
- return phy_mii_ioctl(priv->phy, if_mii(ifr), cmd);
+ return phy_mii_ioctl(priv->phy, ifr, cmd);
return -EOPNOTSUPP;
}
diff --git a/drivers/net/dnet.c b/drivers/net/dnet.c
index 8b0f50b..4ea7141 100644
--- a/drivers/net/dnet.c
+++ b/drivers/net/dnet.c
@@ -797,7 +797,7 @@ static int dnet_ioctl(struct net_device *dev, struct ifreq *rq, int cmd)
if (!phydev)
return -ENODEV;
- return phy_mii_ioctl(phydev, if_mii(rq), cmd);
+ return phy_mii_ioctl(phydev, rq, cmd);
}
static void dnet_get_drvinfo(struct net_device *dev,
diff --git a/drivers/net/ethoc.c b/drivers/net/ethoc.c
index 6ed2df1..8102f01 100644
--- a/drivers/net/ethoc.c
+++ b/drivers/net/ethoc.c
@@ -734,7 +734,7 @@ static int ethoc_ioctl(struct net_device *dev, struct ifreq *ifr, int cmd)
phy = priv->phy;
}
- return phy_mii_ioctl(phy, mdio, cmd);
+ return phy_mii_ioctl(phy, ifr, cmd);
}
static int ethoc_config(struct net_device *dev, struct ifmap *map)
diff --git a/drivers/net/fec.c b/drivers/net/fec.c
index a3cae4e..f24f49e 100644
--- a/drivers/net/fec.c
+++ b/drivers/net/fec.c
@@ -828,7 +828,7 @@ static int fec_enet_ioctl(struct net_device *dev, struct ifreq *rq, int cmd)
if (!phydev)
return -ENODEV;
- return phy_mii_ioctl(phydev, if_mii(rq), cmd);
+ return phy_mii_ioctl(phydev, rq, cmd);
}
static void fec_enet_free_buffers(struct net_device *dev)
diff --git a/drivers/net/fec_mpc52xx.c b/drivers/net/fec_mpc52xx.c
index 25e6cc6..fdbf148 100644
--- a/drivers/net/fec_mpc52xx.c
+++ b/drivers/net/fec_mpc52xx.c
@@ -826,7 +826,7 @@ static int mpc52xx_fec_ioctl(struct net_device *dev, struct ifreq *rq, int cmd)
if (!priv->phydev)
return -ENOTSUPP;
- return phy_mii_ioctl(priv->phydev, if_mii(rq), cmd);
+ return phy_mii_ioctl(priv->phydev, rq, cmd);
}
static const struct net_device_ops mpc52xx_fec_netdev_ops = {
diff --git a/drivers/net/fs_enet/fs_enet-main.c b/drivers/net/fs_enet/fs_enet-main.c
index 309a0ea..f08cff9 100644
--- a/drivers/net/fs_enet/fs_enet-main.c
+++ b/drivers/net/fs_enet/fs_enet-main.c
@@ -963,12 +963,11 @@ static const struct ethtool_ops fs_ethtool_ops = {
static int fs_ioctl(struct net_device *dev, struct ifreq *rq, int cmd)
{
struct fs_enet_private *fep = netdev_priv(dev);
- struct mii_ioctl_data *mii = (struct mii_ioctl_data *)&rq->ifr_data;
if (!netif_running(dev))
return -EINVAL;
- return phy_mii_ioctl(fep->phydev, mii, cmd);
+ return phy_mii_ioctl(fep->phydev, rq, cmd);
}
extern int fs_mii_connect(struct net_device *dev);
diff --git a/drivers/net/gianfar.c b/drivers/net/gianfar.c
index ab54821..9d88e20 100644
--- a/drivers/net/gianfar.c
+++ b/drivers/net/gianfar.c
@@ -834,7 +834,7 @@ static int gfar_ioctl(struct net_device *dev, struct ifreq *rq, int cmd)
if (!priv->phydev)
return -ENODEV;
- return phy_mii_ioctl(priv->phydev, if_mii(rq), cmd);
+ return phy_mii_ioctl(priv->phydev, rq, cmd);
}
static unsigned int reverse_bitmap(unsigned int bit_map, unsigned int max_qs)
diff --git a/drivers/net/macb.c b/drivers/net/macb.c
index 40797fb..ff2f158 100644
--- a/drivers/net/macb.c
+++ b/drivers/net/macb.c
@@ -1082,7 +1082,7 @@ static int macb_ioctl(struct net_device *dev, struct ifreq *rq, int cmd)
if (!phydev)
return -ENODEV;
- return phy_mii_ioctl(phydev, if_mii(rq), cmd);
+ return phy_mii_ioctl(phydev, rq, cmd);
}
static const struct net_device_ops macb_netdev_ops = {
diff --git a/drivers/net/mv643xx_eth.c b/drivers/net/mv643xx_eth.c
index e345ec8..e3ab182 100644
--- a/drivers/net/mv643xx_eth.c
+++ b/drivers/net/mv643xx_eth.c
@@ -2452,7 +2452,7 @@ static int mv643xx_eth_ioctl(struct net_device *dev, struct ifreq *ifr, int cmd)
struct mv643xx_eth_private *mp = netdev_priv(dev);
if (mp->phy != NULL)
- return phy_mii_ioctl(mp->phy, if_mii(ifr), cmd);
+ return phy_mii_ioctl(mp->phy, ifr, cmd);
return -EOPNOTSUPP;
}
diff --git a/drivers/net/octeon/octeon_mgmt.c b/drivers/net/octeon/octeon_mgmt.c
index 000e792..73a2671 100644
--- a/drivers/net/octeon/octeon_mgmt.c
+++ b/drivers/net/octeon/octeon_mgmt.c
@@ -620,7 +620,7 @@ static int octeon_mgmt_ioctl(struct net_device *netdev,
if (!p->phydev)
return -EINVAL;
- return phy_mii_ioctl(p->phydev, if_mii(rq), cmd);
+ return phy_mii_ioctl(p->phydev, rq, cmd);
}
static void octeon_mgmt_adjust_link(struct net_device *netdev)
diff --git a/drivers/net/phy/phy.c b/drivers/net/phy/phy.c
index 64be466..bd88d81 100644
--- a/drivers/net/phy/phy.c
+++ b/drivers/net/phy/phy.c
@@ -309,8 +309,9 @@ EXPORT_SYMBOL(phy_ethtool_gset);
* current state. Use at own risk.
*/
int phy_mii_ioctl(struct phy_device *phydev,
- struct mii_ioctl_data *mii_data, int cmd)
+ struct ifreq *ifr, int cmd)
{
+ struct mii_ioctl_data *mii_data = if_mii(ifr);
u16 val = mii_data->val_in;
switch (cmd) {
diff --git a/drivers/net/sb1250-mac.c b/drivers/net/sb1250-mac.c
index 1f3acc3..e585c3f 100644
--- a/drivers/net/sb1250-mac.c
+++ b/drivers/net/sb1250-mac.c
@@ -2532,7 +2532,7 @@ static int sbmac_mii_ioctl(struct net_device *dev, struct ifreq *rq, int cmd)
if (!netif_running(dev) || !sc->phy_dev)
return -EINVAL;
- return phy_mii_ioctl(sc->phy_dev, if_mii(rq), cmd);
+ return phy_mii_ioctl(sc->phy_dev, rq, cmd);
}
static int sbmac_close(struct net_device *dev)
diff --git a/drivers/net/sh_eth.c b/drivers/net/sh_eth.c
index 501a55f..8279f8e 100644
--- a/drivers/net/sh_eth.c
+++ b/drivers/net/sh_eth.c
@@ -1233,7 +1233,7 @@ static int sh_eth_do_ioctl(struct net_device *ndev, struct ifreq *rq,
if (!phydev)
return -ENODEV;
- return phy_mii_ioctl(phydev, if_mii(rq), cmd);
+ return phy_mii_ioctl(phydev, rq, cmd);
}
#if defined(SH_ETH_HAS_TSU)
diff --git a/drivers/net/smsc911x.c b/drivers/net/smsc911x.c
index cc55974..56dc2ff 100644
--- a/drivers/net/smsc911x.c
+++ b/drivers/net/smsc911x.c
@@ -1538,7 +1538,7 @@ static int smsc911x_do_ioctl(struct net_device *dev, struct ifreq *ifr, int cmd)
if (!netif_running(dev) || !pdata->phy_dev)
return -EINVAL;
- return phy_mii_ioctl(pdata->phy_dev, if_mii(ifr), cmd);
+ return phy_mii_ioctl(pdata->phy_dev, ifr, cmd);
}
static int
diff --git a/drivers/net/smsc9420.c b/drivers/net/smsc9420.c
index 6cdee6a..b09ee1c 100644
--- a/drivers/net/smsc9420.c
+++ b/drivers/net/smsc9420.c
@@ -245,7 +245,7 @@ static int smsc9420_do_ioctl(struct net_device *dev, struct ifreq *ifr, int cmd)
if (!netif_running(dev) || !pd->phy_dev)
return -EINVAL;
- return phy_mii_ioctl(pd->phy_dev, if_mii(ifr), cmd);
+ return phy_mii_ioctl(pd->phy_dev, ifr, cmd);
}
static int smsc9420_ethtool_get_settings(struct net_device *dev,
diff --git a/drivers/net/stmmac/stmmac_main.c b/drivers/net/stmmac/stmmac_main.c
index a31d580..6af7471 100644
--- a/drivers/net/stmmac/stmmac_main.c
+++ b/drivers/net/stmmac/stmmac_main.c
@@ -1450,7 +1450,7 @@ static int stmmac_ioctl(struct net_device *dev, struct ifreq *rq, int cmd)
return -EINVAL;
spin_lock(&priv->lock);
- ret = phy_mii_ioctl(priv->phydev, if_mii(rq), cmd);
+ ret = phy_mii_ioctl(priv->phydev, rq, cmd);
spin_unlock(&priv->lock);
default:
break;
diff --git a/drivers/net/tc35815.c b/drivers/net/tc35815.c
index be08b75..99e423a 100644
--- a/drivers/net/tc35815.c
+++ b/drivers/net/tc35815.c
@@ -2066,7 +2066,7 @@ static int tc35815_ioctl(struct net_device *dev, struct ifreq *rq, int cmd)
return -EINVAL;
if (!lp->phy_dev)
return -ENODEV;
- return phy_mii_ioctl(lp->phy_dev, if_mii(rq), cmd);
+ return phy_mii_ioctl(lp->phy_dev, rq, cmd);
}
static void tc35815_chip_reset(struct net_device *dev)
diff --git a/drivers/net/tg3.c b/drivers/net/tg3.c
index 289cdc5..d4163f2 100644
--- a/drivers/net/tg3.c
+++ b/drivers/net/tg3.c
@@ -10954,7 +10954,7 @@ static int tg3_ioctl(struct net_device *dev, struct ifreq *ifr, int cmd)
if (!(tp->tg3_flags3 & TG3_FLG3_PHY_CONNECTED))
return -EAGAIN;
phydev = tp->mdio_bus->phy_map[TG3_PHY_MII_ADDR];
- return phy_mii_ioctl(phydev, data, cmd);
+ return phy_mii_ioctl(phydev, ifr, cmd);
}
switch (cmd) {
diff --git a/drivers/staging/octeon/ethernet-mdio.c b/drivers/staging/octeon/ethernet-mdio.c
index 7e0be8d..10a82ef 100644
--- a/drivers/staging/octeon/ethernet-mdio.c
+++ b/drivers/staging/octeon/ethernet-mdio.c
@@ -113,7 +113,7 @@ int cvm_oct_ioctl(struct net_device *dev, struct ifreq *rq, int cmd)
if (!priv->phydev)
return -EINVAL;
- return phy_mii_ioctl(priv->phydev, if_mii(rq), cmd);
+ return phy_mii_ioctl(priv->phydev, rq, cmd);
}
static void cvm_oct_adjust_link(struct net_device *dev)
diff --git a/include/linux/phy.h b/include/linux/phy.h
index 987e111..d63736a 100644
--- a/include/linux/phy.h
+++ b/include/linux/phy.h
@@ -498,7 +498,7 @@ void phy_stop_machine(struct phy_device *phydev);
int phy_ethtool_sset(struct phy_device *phydev, struct ethtool_cmd *cmd);
int phy_ethtool_gset(struct phy_device *phydev, struct ethtool_cmd *cmd);
int phy_mii_ioctl(struct phy_device *phydev,
- struct mii_ioctl_data *mii_data, int cmd);
+ struct ifreq *ifr, int cmd);
int phy_start_interrupts(struct phy_device *phydev);
void phy_print_status(struct phy_device *phydev);
struct phy_device* phy_device_create(struct mii_bus *bus, int addr, int phy_id);
diff --git a/net/dsa/slave.c b/net/dsa/slave.c
index 8fdca56..64ca2a6 100644
--- a/net/dsa/slave.c
+++ b/net/dsa/slave.c
@@ -164,10 +164,9 @@ out:
static int dsa_slave_ioctl(struct net_device *dev, struct ifreq *ifr, int cmd)
{
struct dsa_slave_priv *p = netdev_priv(dev);
- struct mii_ioctl_data *mii_data = if_mii(ifr);
if (p->phy != NULL)
- return phy_mii_ioctl(p->phy, mii_data, cmd);
+ return phy_mii_ioctl(p->phy, ifr, cmd);
return -EOPNOTSUPP;
}
--
1.6.3.3
^ permalink raw reply related
* [PATCH v4 00/12] ptp: IEEE 1588 clock support
From: Richard Cochran @ 2010-06-15 16:06 UTC (permalink / raw)
To: netdev; +Cc: linuxppc-dev, devicetree-discuss, Krzysztof Halasa,
linux-arm-kernel
Now and again there has been talk on the netdev list of adding PTP
support into Linux. One part of the picture is already in place, the
SO_TIMESTAMPING API for hardware time stamping. This patch set offers
the missing second part needed for complete IEEE 1588 support.
The only feature still to be implemented is the hook into the PPS
subsystem, to synchronize the Linux clock to the PTP clock.
The first seven patches concern phylib. A new generation of PHYs with
timestamping capabilities is appearing on the market. In order to
support hardware timestamping on these devices in Linux, a number of
adjustments will have to be made in the PHY subsystem. The reasons for
the changes are explained inline at the top of the patches.
The last five patches provide the PTP subsystem itself.
* Why all the CCs?
1. The patches probably should go through netdev.
2. One driver is for PowerPC, and adds device tree stuff.
3. One driver is for the ARM Xscale IXP465
* Open Issues:
** DP83640
In order to make this work, three lines must be added into the MAC
driver. Since this PHY can be connected to almost any MAC, I did
not do anything about that. If you have the DP83640 and want to try
the driver, you need to add three lines to your MAC driver.
1. Before mdio_register, add
bus->locktype = MDIOBUS_ATOMIC_RW;
2. In the .ndo_start_xmit function, add
skb_tx_timestamp()
3. In the NAPI poll function, add
skb_rx_timestamp()
** IXP465
I do not know how to correctly choose the timestamp "channel" based
on the port identifier:
#define PORT2CHANNEL(p) 1
/*
* PHYSICAL_ID(p->id) ?
* TODO - Figure out correct mapping.
*/
* Patch ChangeLog
** v4
*** general
- Added a clock driver for the National Semiconductor PHYTER.
- Added a clock driver for the Intel IXP465.
- Made more stylish according to checkstyle.pl.
*** gianfar
- Replace device_type and model with compatible string ("fsl,etsec-ptp")
- Register only one interrupt, since others are superfluous.
- Combine ptp clock instance with private variable structure.
- ISR now returns NONE or HANDLED properly.
- Print error message if something is missing from the device nodes.
** v3
*** general
- Added documentation on writing clock drivers.
- Added the ioctls for the ancillary clock features.
- Changed wrong subsys_initcall() to module_init() in clock drivers.
- Removed the (too coarse) character device mutex.
- Setting the clock now requires CAP_SYS_TIME.
*** gianfar
- Added alarm feature.
- Added device tree node binding description.
- Added fine grain locking of the clock registers.
- Added the external time stamp feature.
- Added white space for better style.
- Coverted base+offset to structure pointers for register access.
- When removing the driver, we now disable all PTP functions.
** v2
- Changed clock list from a static array into a dynamic list. Also,
use a bitmap to manage the clock's minor numbers.
- Replaced character device semaphore with a mutex.
- Drop .ko from module names in Kbuild help.
- Replace deprecated unifdef-y with header-y for user space header file.
- Added links to both of the ptpd patches on sourceforge.
- Gianfar driver now gets parameters from device tree.
- Added API documentation to Documentation/ptp/ptp.txt
Looking forward to your feedback,
Richard
Richard Cochran (12):
phylib: preserve ifreq parameter when calling generic phy_mii_ioctl()
phylib: do not filter phy_mii_ioctl()
phylib: add a driver method for the SIOCSHWTSTAMP ioctl.
phylib: add a way to make PHY time stamps possible.
phylib: Allow reading and writing a mii bus from atomic context.
ptp: add a BPF to help drivers detect PTP packets.
phylib: support the National Semiconductor DP83640 PHY.
ptp: Added a brand new class driver for ptp clocks.
ptp: Added a clock that uses the Linux system time.
ptp: Added a clock that uses the eTSEC found on the MPC85xx.
ptp: Added a clock driver for the IXP46x.
ptp: Added a clock driver for the National Semiconductor PHYTER.
Documentation/powerpc/dts-bindings/fsl/tsec.txt | 54 ++
Documentation/ptp/ptp.txt | 95 ++++
Documentation/ptp/testptp.c | 269 ++++++++++
Documentation/ptp/testptp.mk | 33 ++
arch/arm/mach-ixp4xx/include/mach/ixp46x_ts.h | 67 +++
arch/powerpc/boot/dts/mpc8313erdb.dts | 13 +
arch/powerpc/boot/dts/p2020ds.dts | 13 +
arch/powerpc/boot/dts/p2020rdb.dts | 13 +
drivers/Kconfig | 2 +
drivers/Makefile | 1 +
drivers/net/Makefile | 1 +
drivers/net/arm/ixp4xx_eth.c | 197 ++++++++-
drivers/net/au1000_eth.c | 2 +-
drivers/net/bcm63xx_enet.c | 2 +-
drivers/net/cpmac.c | 5 +-
drivers/net/dnet.c | 2 +-
drivers/net/ethoc.c | 2 +-
drivers/net/fec.c | 2 +-
drivers/net/fec_mpc52xx.c | 2 +-
drivers/net/fs_enet/fs_enet-main.c | 3 +-
drivers/net/gianfar.c | 2 +-
drivers/net/gianfar_ptp.c | 518 ++++++++++++++++++++
drivers/net/gianfar_ptp_reg.h | 113 +++++
drivers/net/macb.c | 2 +-
drivers/net/mv643xx_eth.c | 2 +-
drivers/net/octeon/octeon_mgmt.c | 2 +-
drivers/net/phy/Kconfig | 16 +
drivers/net/phy/Makefile | 1 +
drivers/net/phy/dp83640.c | 595 +++++++++++++++++++++++
drivers/net/phy/dp83640_reg.h | 237 +++++++++
drivers/net/phy/mdio_bus.c | 35 ++-
drivers/net/phy/phy.c | 8 +-
drivers/net/sb1250-mac.c | 2 +-
drivers/net/sh_eth.c | 2 +-
drivers/net/smsc911x.c | 2 +-
drivers/net/smsc9420.c | 2 +-
drivers/net/stmmac/stmmac_main.c | 22 +-
drivers/net/tc35815.c | 2 +-
drivers/net/tg3.c | 2 +-
drivers/ptp/Kconfig | 64 +++
drivers/ptp/Makefile | 7 +
drivers/ptp/ptp_clock.c | 514 ++++++++++++++++++++
drivers/ptp/ptp_ixp46x.c | 231 +++++++++
drivers/ptp/ptp_linux.c | 136 ++++++
drivers/staging/octeon/ethernet-mdio.c | 2 +-
include/linux/Kbuild | 1 +
include/linux/phy.h | 22 +-
include/linux/ptp_classify.h | 118 +++++
include/linux/ptp_clock.h | 79 +++
include/linux/ptp_clock_kernel.h | 137 ++++++
include/linux/skbuff.h | 32 ++
kernel/time/ntp.c | 2 +
net/Kconfig | 11 +
net/dsa/slave.c | 3 +-
54 files changed, 3651 insertions(+), 51 deletions(-)
create mode 100644 Documentation/ptp/ptp.txt
create mode 100644 Documentation/ptp/testptp.c
create mode 100644 Documentation/ptp/testptp.mk
create mode 100644 arch/arm/mach-ixp4xx/include/mach/ixp46x_ts.h
create mode 100644 drivers/net/gianfar_ptp.c
create mode 100644 drivers/net/gianfar_ptp_reg.h
create mode 100644 drivers/net/phy/dp83640.c
create mode 100644 drivers/net/phy/dp83640_reg.h
create mode 100644 drivers/ptp/Kconfig
create mode 100644 drivers/ptp/Makefile
create mode 100644 drivers/ptp/ptp_clock.c
create mode 100644 drivers/ptp/ptp_ixp46x.c
create mode 100644 drivers/ptp/ptp_linux.c
create mode 100644 include/linux/ptp_classify.h
create mode 100644 include/linux/ptp_clock.h
create mode 100644 include/linux/ptp_clock_kernel.h
^ permalink raw reply
* [PATCH 12/17] netfilter: nfnetlink_log: RCU conversion, part 2
From: kaber @ 2010-06-15 15:38 UTC (permalink / raw)
To: davem; +Cc: netfilter-devel, netdev
In-Reply-To: <1276616323-20880-1-git-send-email-kaber@trash.net>
From: Eric Dumazet <eric.dumazet@gmail.com>
- must use atomic_inc_not_zero() in instance_lookup_get()
- must use hlist_add_head_rcu() instead of hlist_add_head()
- must use hlist_del_rcu() instead of hlist_del()
- Introduce NFULNL_COPY_DISABLED to stop lockless reader from using an
instance, before we do final instance_put() on it.
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: Patrick McHardy <kaber@trash.net>
---
include/linux/netfilter/nfnetlink_log.h | 1 +
net/netfilter/nfnetlink_log.c | 18 ++++++++++++------
2 files changed, 13 insertions(+), 6 deletions(-)
diff --git a/include/linux/netfilter/nfnetlink_log.h b/include/linux/netfilter/nfnetlink_log.h
index d3bab7a..1d0b84a 100644
--- a/include/linux/netfilter/nfnetlink_log.h
+++ b/include/linux/netfilter/nfnetlink_log.h
@@ -89,6 +89,7 @@ enum nfulnl_attr_config {
#define NFULNL_COPY_NONE 0x00
#define NFULNL_COPY_META 0x01
#define NFULNL_COPY_PACKET 0x02
+#define NFULNL_COPY_DISABLED 0x03
#define NFULNL_CFG_F_SEQ 0x0001
#define NFULNL_CFG_F_SEQ_GLOBAL 0x0002
diff --git a/net/netfilter/nfnetlink_log.c b/net/netfilter/nfnetlink_log.c
index 8ec23ec..fb86a51 100644
--- a/net/netfilter/nfnetlink_log.c
+++ b/net/netfilter/nfnetlink_log.c
@@ -109,8 +109,8 @@ instance_lookup_get(u_int16_t group_num)
rcu_read_lock_bh();
inst = __instance_lookup(group_num);
- if (inst)
- instance_get(inst);
+ if (inst && !atomic_inc_not_zero(&inst->use))
+ inst = NULL;
rcu_read_unlock_bh();
return inst;
@@ -171,7 +171,7 @@ instance_create(u_int16_t group_num, int pid)
inst->copy_mode = NFULNL_COPY_PACKET;
inst->copy_range = NFULNL_COPY_RANGE_MAX;
- hlist_add_head(&inst->hlist,
+ hlist_add_head_rcu(&inst->hlist,
&instance_table[instance_hashfn(group_num)]);
spin_unlock_bh(&instances_lock);
@@ -185,18 +185,23 @@ out_unlock:
static void __nfulnl_flush(struct nfulnl_instance *inst);
+/* called with BH disabled */
static void
__instance_destroy(struct nfulnl_instance *inst)
{
/* first pull it out of the global list */
- hlist_del(&inst->hlist);
+ hlist_del_rcu(&inst->hlist);
/* then flush all pending packets from skb */
- spin_lock_bh(&inst->lock);
+ spin_lock(&inst->lock);
+
+ /* lockless readers wont be able to use us */
+ inst->copy_mode = NFULNL_COPY_DISABLED;
+
if (inst->skb)
__nfulnl_flush(inst);
- spin_unlock_bh(&inst->lock);
+ spin_unlock(&inst->lock);
/* and finally put the refcount */
instance_put(inst);
@@ -624,6 +629,7 @@ nfulnl_log_packet(u_int8_t pf,
size += nla_total_size(data_len);
break;
+ case NFULNL_COPY_DISABLED:
default:
goto unlock_and_release;
}
--
1.7.0.4
^ permalink raw reply related
* [PATCH 10/17] netfilter: nfnetlink_queue: some optimizations
From: kaber @ 2010-06-15 15:38 UTC (permalink / raw)
To: davem; +Cc: netfilter-devel, netdev
In-Reply-To: <1276616323-20880-1-git-send-email-kaber@trash.net>
From: Eric Dumazet <eric.dumazet@gmail.com>
- Use an atomic_t for id_sequence to avoid a spin_lock/spin_unlock pair
- Group highly modified struct nfqnl_instance fields together
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: Patrick McHardy <kaber@trash.net>
---
net/netfilter/nfnetlink_queue.c | 33 ++++++++++++++-------------------
1 files changed, 14 insertions(+), 19 deletions(-)
diff --git a/net/netfilter/nfnetlink_queue.c b/net/netfilter/nfnetlink_queue.c
index 12e1ab3..d05605b 100644
--- a/net/netfilter/nfnetlink_queue.c
+++ b/net/netfilter/nfnetlink_queue.c
@@ -46,17 +46,19 @@ struct nfqnl_instance {
int peer_pid;
unsigned int queue_maxlen;
unsigned int copy_range;
- unsigned int queue_total;
unsigned int queue_dropped;
unsigned int queue_user_dropped;
- unsigned int id_sequence; /* 'sequence' of pkt ids */
u_int16_t queue_num; /* number of this queue */
u_int8_t copy_mode;
-
- spinlock_t lock;
-
+/*
+ * Following fields are dirtied for each queued packet,
+ * keep them in same cache line if possible.
+ */
+ spinlock_t lock;
+ unsigned int queue_total;
+ atomic_t id_sequence; /* 'sequence' of pkt ids */
struct list_head queue_list; /* packets in queue */
};
@@ -238,32 +240,24 @@ nfqnl_build_packet_message(struct nfqnl_instance *queue,
outdev = entry->outdev;
- spin_lock_bh(&queue->lock);
-
- switch ((enum nfqnl_config_mode)queue->copy_mode) {
+ switch ((enum nfqnl_config_mode)ACCESS_ONCE(queue->copy_mode)) {
case NFQNL_COPY_META:
case NFQNL_COPY_NONE:
break;
case NFQNL_COPY_PACKET:
if (entskb->ip_summed == CHECKSUM_PARTIAL &&
- skb_checksum_help(entskb)) {
- spin_unlock_bh(&queue->lock);
+ skb_checksum_help(entskb))
return NULL;
- }
- if (queue->copy_range == 0
- || queue->copy_range > entskb->len)
+
+ data_len = ACCESS_ONCE(queue->copy_range);
+ if (data_len == 0 || data_len > entskb->len)
data_len = entskb->len;
- else
- data_len = queue->copy_range;
size += nla_total_size(data_len);
break;
}
- entry->id = queue->id_sequence++;
-
- spin_unlock_bh(&queue->lock);
skb = alloc_skb(size, GFP_ATOMIC);
if (!skb)
@@ -278,6 +272,7 @@ nfqnl_build_packet_message(struct nfqnl_instance *queue,
nfmsg->version = NFNETLINK_V0;
nfmsg->res_id = htons(queue->queue_num);
+ entry->id = atomic_inc_return(&queue->id_sequence);
pmsg.packet_id = htonl(entry->id);
pmsg.hw_protocol = entskb->protocol;
pmsg.hook = entry->hook;
@@ -866,7 +861,7 @@ static int seq_show(struct seq_file *s, void *v)
inst->peer_pid, inst->queue_total,
inst->copy_mode, inst->copy_range,
inst->queue_dropped, inst->queue_user_dropped,
- inst->id_sequence, 1);
+ atomic_read(&inst->id_sequence), 1);
}
static const struct seq_operations nfqnl_seq_ops = {
--
1.7.0.4
^ permalink raw reply related
* [PATCH 09/17] netfilter: ip6_queue: rwlock to spinlock conversion
From: kaber @ 2010-06-15 15:38 UTC (permalink / raw)
To: davem; +Cc: netfilter-devel, netdev
In-Reply-To: <1276616323-20880-1-git-send-email-kaber@trash.net>
From: Eric Dumazet <eric.dumazet@gmail.com>
Converts queue_lock rwlock to a spinlock.
(readlocked part can be changed by reads of integer values)
One atomic operation instead of four per ipq_enqueue_packet() call.
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: Patrick McHardy <kaber@trash.net>
---
net/ipv6/netfilter/ip6_queue.c | 57 +++++++++++++++++----------------------
1 files changed, 25 insertions(+), 32 deletions(-)
diff --git a/net/ipv6/netfilter/ip6_queue.c b/net/ipv6/netfilter/ip6_queue.c
index 8c20174..413ab07 100644
--- a/net/ipv6/netfilter/ip6_queue.c
+++ b/net/ipv6/netfilter/ip6_queue.c
@@ -43,7 +43,7 @@ typedef int (*ipq_cmpfn)(struct nf_queue_entry *, unsigned long);
static unsigned char copy_mode __read_mostly = IPQ_COPY_NONE;
static unsigned int queue_maxlen __read_mostly = IPQ_QMAX_DEFAULT;
-static DEFINE_RWLOCK(queue_lock);
+static DEFINE_SPINLOCK(queue_lock);
static int peer_pid __read_mostly;
static unsigned int copy_range __read_mostly;
static unsigned int queue_total;
@@ -73,10 +73,10 @@ __ipq_set_mode(unsigned char mode, unsigned int range)
break;
case IPQ_COPY_PACKET:
- copy_mode = mode;
+ if (range > 0xFFFF)
+ range = 0xFFFF;
copy_range = range;
- if (copy_range > 0xFFFF)
- copy_range = 0xFFFF;
+ copy_mode = mode;
break;
default:
@@ -102,7 +102,7 @@ ipq_find_dequeue_entry(unsigned long id)
{
struct nf_queue_entry *entry = NULL, *i;
- write_lock_bh(&queue_lock);
+ spin_lock_bh(&queue_lock);
list_for_each_entry(i, &queue_list, list) {
if ((unsigned long)i == id) {
@@ -116,7 +116,7 @@ ipq_find_dequeue_entry(unsigned long id)
queue_total--;
}
- write_unlock_bh(&queue_lock);
+ spin_unlock_bh(&queue_lock);
return entry;
}
@@ -137,9 +137,9 @@ __ipq_flush(ipq_cmpfn cmpfn, unsigned long data)
static void
ipq_flush(ipq_cmpfn cmpfn, unsigned long data)
{
- write_lock_bh(&queue_lock);
+ spin_lock_bh(&queue_lock);
__ipq_flush(cmpfn, data);
- write_unlock_bh(&queue_lock);
+ spin_unlock_bh(&queue_lock);
}
static struct sk_buff *
@@ -153,9 +153,7 @@ ipq_build_packet_message(struct nf_queue_entry *entry, int *errp)
struct nlmsghdr *nlh;
struct timeval tv;
- read_lock_bh(&queue_lock);
-
- switch (copy_mode) {
+ switch (ACCESS_ONCE(copy_mode)) {
case IPQ_COPY_META:
case IPQ_COPY_NONE:
size = NLMSG_SPACE(sizeof(*pmsg));
@@ -163,26 +161,21 @@ ipq_build_packet_message(struct nf_queue_entry *entry, int *errp)
case IPQ_COPY_PACKET:
if (entry->skb->ip_summed == CHECKSUM_PARTIAL &&
- (*errp = skb_checksum_help(entry->skb))) {
- read_unlock_bh(&queue_lock);
+ (*errp = skb_checksum_help(entry->skb)))
return NULL;
- }
- if (copy_range == 0 || copy_range > entry->skb->len)
+
+ data_len = ACCESS_ONCE(copy_range);
+ if (data_len == 0 || data_len > entry->skb->len)
data_len = entry->skb->len;
- else
- data_len = copy_range;
size = NLMSG_SPACE(sizeof(*pmsg) + data_len);
break;
default:
*errp = -EINVAL;
- read_unlock_bh(&queue_lock);
return NULL;
}
- read_unlock_bh(&queue_lock);
-
skb = alloc_skb(size, GFP_ATOMIC);
if (!skb)
goto nlmsg_failure;
@@ -242,7 +235,7 @@ ipq_enqueue_packet(struct nf_queue_entry *entry, unsigned int queuenum)
if (nskb == NULL)
return status;
- write_lock_bh(&queue_lock);
+ spin_lock_bh(&queue_lock);
if (!peer_pid)
goto err_out_free_nskb;
@@ -266,14 +259,14 @@ ipq_enqueue_packet(struct nf_queue_entry *entry, unsigned int queuenum)
__ipq_enqueue_entry(entry);
- write_unlock_bh(&queue_lock);
+ spin_unlock_bh(&queue_lock);
return status;
err_out_free_nskb:
kfree_skb(nskb);
err_out_unlock:
- write_unlock_bh(&queue_lock);
+ spin_unlock_bh(&queue_lock);
return status;
}
@@ -342,9 +335,9 @@ ipq_set_mode(unsigned char mode, unsigned int range)
{
int status;
- write_lock_bh(&queue_lock);
+ spin_lock_bh(&queue_lock);
status = __ipq_set_mode(mode, range);
- write_unlock_bh(&queue_lock);
+ spin_unlock_bh(&queue_lock);
return status;
}
@@ -441,11 +434,11 @@ __ipq_rcv_skb(struct sk_buff *skb)
if (security_netlink_recv(skb, CAP_NET_ADMIN))
RCV_SKB_FAIL(-EPERM);
- write_lock_bh(&queue_lock);
+ spin_lock_bh(&queue_lock);
if (peer_pid) {
if (peer_pid != pid) {
- write_unlock_bh(&queue_lock);
+ spin_unlock_bh(&queue_lock);
RCV_SKB_FAIL(-EBUSY);
}
} else {
@@ -453,7 +446,7 @@ __ipq_rcv_skb(struct sk_buff *skb)
peer_pid = pid;
}
- write_unlock_bh(&queue_lock);
+ spin_unlock_bh(&queue_lock);
status = ipq_receive_peer(NLMSG_DATA(nlh), type,
nlmsglen - NLMSG_LENGTH(0));
@@ -498,10 +491,10 @@ ipq_rcv_nl_event(struct notifier_block *this,
struct netlink_notify *n = ptr;
if (event == NETLINK_URELEASE && n->protocol == NETLINK_IP6_FW) {
- write_lock_bh(&queue_lock);
+ spin_lock_bh(&queue_lock);
if ((net_eq(n->net, &init_net)) && (n->pid == peer_pid))
__ipq_reset();
- write_unlock_bh(&queue_lock);
+ spin_unlock_bh(&queue_lock);
}
return NOTIFY_DONE;
}
@@ -528,7 +521,7 @@ static ctl_table ipq_table[] = {
#ifdef CONFIG_PROC_FS
static int ip6_queue_show(struct seq_file *m, void *v)
{
- read_lock_bh(&queue_lock);
+ spin_lock_bh(&queue_lock);
seq_printf(m,
"Peer PID : %d\n"
@@ -546,7 +539,7 @@ static int ip6_queue_show(struct seq_file *m, void *v)
queue_dropped,
queue_user_dropped);
- read_unlock_bh(&queue_lock);
+ spin_unlock_bh(&queue_lock);
return 0;
}
--
1.7.0.4
^ permalink raw reply related
* [PATCH 04/17] netfilter: xt_rateest: Better struct xt_rateest layout
From: kaber @ 2010-06-15 15:38 UTC (permalink / raw)
To: davem; +Cc: netfilter-devel, netdev
In-Reply-To: <1276616323-20880-1-git-send-email-kaber@trash.net>
From: Eric Dumazet <eric.dumazet@gmail.com>
We currently dirty two cache lines in struct xt_rateest, this hurts SMP
performance.
This patch moves lock/bstats/rstats at beginning of structure so that
they share a single cache line.
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: Patrick McHardy <kaber@trash.net>
---
include/net/netfilter/xt_rateest.h | 10 +++++++---
1 files changed, 7 insertions(+), 3 deletions(-)
diff --git a/include/net/netfilter/xt_rateest.h b/include/net/netfilter/xt_rateest.h
index ddbf37e..b1d780e 100644
--- a/include/net/netfilter/xt_rateest.h
+++ b/include/net/netfilter/xt_rateest.h
@@ -2,13 +2,17 @@
#define _XT_RATEEST_H
struct xt_rateest {
+ /* keep lock and bstats on same cache line to speedup xt_rateest_tg() */
+ struct gnet_stats_basic_packed bstats;
+ spinlock_t lock;
+ /* keep rstats and lock on same cache line to speedup xt_rateest_mt() */
+ struct gnet_stats_rate_est rstats;
+
+ /* following fields not accessed in hot path */
struct hlist_node list;
char name[IFNAMSIZ];
unsigned int refcnt;
- spinlock_t lock;
struct gnet_estimator params;
- struct gnet_stats_rate_est rstats;
- struct gnet_stats_basic_packed bstats;
};
extern struct xt_rateest *xt_rateest_lookup(const char *name);
--
1.7.0.4
^ permalink raw reply related
* [PATCH 03/17] netfilter: vmalloc_node cleanup
From: kaber @ 2010-06-15 15:38 UTC (permalink / raw)
To: davem; +Cc: netfilter-devel, netdev
In-Reply-To: <1276616323-20880-1-git-send-email-kaber@trash.net>
From: Eric Dumazet <eric.dumazet@gmail.com>
Using vmalloc_node(size, numa_node_id()) for temporary storage is not
needed. vmalloc(size) is more respectful of user NUMA policy.
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: Patrick McHardy <kaber@trash.net>
---
net/ipv4/netfilter/arp_tables.c | 7 +++----
net/ipv4/netfilter/ip_tables.c | 4 ++--
net/ipv6/netfilter/ip6_tables.c | 7 +++----
3 files changed, 8 insertions(+), 10 deletions(-)
diff --git a/net/ipv4/netfilter/arp_tables.c b/net/ipv4/netfilter/arp_tables.c
index 1ac01b1..16c0ba0 100644
--- a/net/ipv4/netfilter/arp_tables.c
+++ b/net/ipv4/netfilter/arp_tables.c
@@ -758,7 +758,7 @@ static struct xt_counters *alloc_counters(const struct xt_table *table)
* about).
*/
countersize = sizeof(struct xt_counters) * private->number;
- counters = vmalloc_node(countersize, numa_node_id());
+ counters = vmalloc(countersize);
if (counters == NULL)
return ERR_PTR(-ENOMEM);
@@ -1005,8 +1005,7 @@ static int __do_replace(struct net *net, const char *name,
struct arpt_entry *iter;
ret = 0;
- counters = vmalloc_node(num_counters * sizeof(struct xt_counters),
- numa_node_id());
+ counters = vmalloc(num_counters * sizeof(struct xt_counters));
if (!counters) {
ret = -ENOMEM;
goto out;
@@ -1159,7 +1158,7 @@ static int do_add_counters(struct net *net, const void __user *user,
if (len != size + num_counters * sizeof(struct xt_counters))
return -EINVAL;
- paddc = vmalloc_node(len - size, numa_node_id());
+ paddc = vmalloc(len - size);
if (!paddc)
return -ENOMEM;
diff --git a/net/ipv4/netfilter/ip_tables.c b/net/ipv4/netfilter/ip_tables.c
index 63958f3..7c0b8ad 100644
--- a/net/ipv4/netfilter/ip_tables.c
+++ b/net/ipv4/netfilter/ip_tables.c
@@ -928,7 +928,7 @@ static struct xt_counters *alloc_counters(const struct xt_table *table)
(other than comefrom, which userspace doesn't care
about). */
countersize = sizeof(struct xt_counters) * private->number;
- counters = vmalloc_node(countersize, numa_node_id());
+ counters = vmalloc(countersize);
if (counters == NULL)
return ERR_PTR(-ENOMEM);
@@ -1352,7 +1352,7 @@ do_add_counters(struct net *net, const void __user *user,
if (len != size + num_counters * sizeof(struct xt_counters))
return -EINVAL;
- paddc = vmalloc_node(len - size, numa_node_id());
+ paddc = vmalloc(len - size);
if (!paddc)
return -ENOMEM;
diff --git a/net/ipv6/netfilter/ip6_tables.c b/net/ipv6/netfilter/ip6_tables.c
index 6f517bd..82945ef 100644
--- a/net/ipv6/netfilter/ip6_tables.c
+++ b/net/ipv6/netfilter/ip6_tables.c
@@ -943,7 +943,7 @@ static struct xt_counters *alloc_counters(const struct xt_table *table)
(other than comefrom, which userspace doesn't care
about). */
countersize = sizeof(struct xt_counters) * private->number;
- counters = vmalloc_node(countersize, numa_node_id());
+ counters = vmalloc(countersize);
if (counters == NULL)
return ERR_PTR(-ENOMEM);
@@ -1213,8 +1213,7 @@ __do_replace(struct net *net, const char *name, unsigned int valid_hooks,
struct ip6t_entry *iter;
ret = 0;
- counters = vmalloc_node(num_counters * sizeof(struct xt_counters),
- numa_node_id());
+ counters = vmalloc(num_counters * sizeof(struct xt_counters));
if (!counters) {
ret = -ENOMEM;
goto out;
@@ -1368,7 +1367,7 @@ do_add_counters(struct net *net, const void __user *user, unsigned int len,
if (len != size + num_counters * sizeof(struct xt_counters))
return -EINVAL;
- paddc = vmalloc_node(len - size, numa_node_id());
+ paddc = vmalloc(len - size);
if (!paddc)
return -ENOMEM;
--
1.7.0.4
^ permalink raw reply related
* [PATCH 01/17] netfilter: br_netfilter: use skb_set_noref()
From: kaber @ 2010-06-15 15:38 UTC (permalink / raw)
To: davem; +Cc: netfilter-devel, netdev
In-Reply-To: <1276616323-20880-1-git-send-email-kaber@trash.net>
From: Eric Dumazet <eric.dumazet@gmail.com>
Avoid dirtying bridge_parent_rtable refcount, using new dst noref
infrastructure.
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: Patrick McHardy <kaber@trash.net>
---
net/bridge/br_netfilter.c | 6 ++----
1 files changed, 2 insertions(+), 4 deletions(-)
diff --git a/net/bridge/br_netfilter.c b/net/bridge/br_netfilter.c
index 4442099..cbea5af 100644
--- a/net/bridge/br_netfilter.c
+++ b/net/bridge/br_netfilter.c
@@ -244,8 +244,7 @@ static int br_nf_pre_routing_finish_ipv6(struct sk_buff *skb)
kfree_skb(skb);
return 0;
}
- dst_hold(&rt->u.dst);
- skb_dst_set(skb, &rt->u.dst);
+ skb_dst_set_noref(skb, &rt->u.dst);
skb->dev = nf_bridge->physindev;
nf_bridge_update_protocol(skb);
@@ -396,8 +395,7 @@ bridged_dnat:
kfree_skb(skb);
return 0;
}
- dst_hold(&rt->u.dst);
- skb_dst_set(skb, &rt->u.dst);
+ skb_dst_set_noref(skb, &rt->u.dst);
}
skb->dev = nf_bridge->physindev;
--
1.7.0.4
^ permalink raw reply related
* [PATCH 2/2] ehea: Fix kernel deadlock in DLPAR-mem processing
From: Jan-Bernd Themann @ 2010-06-15 15:35 UTC (permalink / raw)
To: David Miller
Cc: netdev, linuxppc-dev, linux-kernel, tklein, Andre Detsch, themann
Port reset operations and memory add/remove operations need to
be serialized to avoid a kernel deadlock. The deadlock is caused
by calling the napi_disable() function twice.
Therefore we have to employ the dlpar_mem_lock in the ehea_reset_port
function as well
Signed-off-by: Jan-Bernd Themann <themann@de.ibm.com>
---
created against kernel-2.6.35-rc3
drivers/net/ehea/ehea.h | 2 +-
drivers/net/ehea/ehea_main.c | 8 +++-----
2 files changed, 4 insertions(+), 6 deletions(-)
diff --git a/drivers/net/ehea/ehea.h b/drivers/net/ehea/ehea.h
index 0630980..0060e42 100644
--- a/drivers/net/ehea/ehea.h
+++ b/drivers/net/ehea/ehea.h
@@ -40,7 +40,7 @@
#include <asm/io.h>
#define DRV_NAME "ehea"
-#define DRV_VERSION "EHEA_0103"
+#define DRV_VERSION "EHEA_0105"
/* eHEA capability flags */
#define DLPAR_PORT_ADD_REM 1
diff --git a/drivers/net/ehea/ehea_main.c b/drivers/net/ehea/ehea_main.c
index fd890fa..8b92acb 100644
--- a/drivers/net/ehea/ehea_main.c
+++ b/drivers/net/ehea/ehea_main.c
@@ -2860,6 +2860,7 @@ static void ehea_reset_port(struct work_struct *work)
container_of(work, struct ehea_port, reset_task);
struct net_device *dev = port->netdev;
+ mutex_lock(&dlpar_mem_lock);
port->resets++;
mutex_lock(&port->port_lock);
netif_stop_queue(dev);
@@ -2882,6 +2883,7 @@ static void ehea_reset_port(struct work_struct *work)
netif_wake_queue(dev);
out:
mutex_unlock(&port->port_lock);
+ mutex_unlock(&dlpar_mem_lock);
}
static void ehea_rereg_mrs(struct work_struct *work)
@@ -3543,10 +3545,7 @@ static int ehea_mem_notifier(struct notifier_block *nb,
int ret = NOTIFY_BAD;
struct memory_notify *arg = data;
- if (!mutex_trylock(&dlpar_mem_lock)) {
- ehea_info("ehea_mem_notifier must not be called parallelized");
- goto out;
- }
+ mutex_lock(&dlpar_mem_lock);
switch (action) {
case MEM_CANCEL_OFFLINE:
@@ -3575,7 +3574,6 @@ static int ehea_mem_notifier(struct notifier_block *nb,
out_unlock:
mutex_unlock(&dlpar_mem_lock);
-out:
return ret;
}
--
1.7.0
^ permalink raw reply related
* [PATCH 1/2] ehea: fix delayed packet processing
From: Jan-Bernd Themann @ 2010-06-15 15:35 UTC (permalink / raw)
To: David Miller
Cc: netdev, linuxppc-dev, linux-kernel, tklein, Andre Detsch, themann
In the eHEA poll function an rmb() is required. Without that some packets
on the receive queue are not seen and thus delayed until the next interrupt
is handled for the same receive queue.
Signed-off-by: Jan-Bernd Themann <themann@de.ibm.com>
---
Patch created against 2.6.35-rc3
drivers/net/ehea/ehea_main.c | 1 +
1 files changed, 1 insertions(+), 0 deletions(-)
diff --git a/drivers/net/ehea/ehea_main.c b/drivers/net/ehea/ehea_main.c
index f547894..fd890fa 100644
--- a/drivers/net/ehea/ehea_main.c
+++ b/drivers/net/ehea/ehea_main.c
@@ -867,6 +867,7 @@ static int ehea_poll(struct napi_struct *napi, int budget)
ehea_reset_cq_ep(pr->send_cq);
ehea_reset_cq_n1(pr->recv_cq);
ehea_reset_cq_n1(pr->send_cq);
+ rmb();
cqe = ehea_poll_rq1(pr->qp, &wqe_index);
cqe_skb = ehea_poll_cq(pr->send_cq);
--
1.7.0
^ permalink raw reply related
* [PATCH 17/17] netfilter: xtables: idletimer target implementation
From: kaber @ 2010-06-15 15:38 UTC (permalink / raw)
To: davem; +Cc: netfilter-devel, netdev
In-Reply-To: <1276616323-20880-1-git-send-email-kaber@trash.net>
From: Luciano Coelho <luciano.coelho@nokia.com>
This patch implements an idletimer Xtables target that can be used to
identify when interfaces have been idle for a certain period of time.
Timers are identified by labels and are created when a rule is set with a new
label. The rules also take a timeout value (in seconds) as an option. If
more than one rule uses the same timer label, the timer will be restarted
whenever any of the rules get a hit.
One entry for each timer is created in sysfs. This attribute contains the
timer remaining for the timer to expire. The attributes are located under
the xt_idletimer class:
/sys/class/xt_idletimer/timers/<label>
When the timer expires, the target module sends a sysfs notification to the
userspace, which can then decide what to do (eg. disconnect to save power).
Cc: Timo Teras <timo.teras@iki.fi>
Signed-off-by: Luciano Coelho <luciano.coelho@nokia.com>
Signed-off-by: Patrick McHardy <kaber@trash.net>
---
include/linux/netfilter/Kbuild | 1 +
include/linux/netfilter/xt_IDLETIMER.h | 45 +++++
net/netfilter/Kconfig | 12 ++
net/netfilter/Makefile | 1 +
net/netfilter/xt_IDLETIMER.c | 314 ++++++++++++++++++++++++++++++++
5 files changed, 373 insertions(+), 0 deletions(-)
create mode 100644 include/linux/netfilter/xt_IDLETIMER.h
create mode 100644 net/netfilter/xt_IDLETIMER.c
diff --git a/include/linux/netfilter/Kbuild b/include/linux/netfilter/Kbuild
index 48767cd..bb103f4 100644
--- a/include/linux/netfilter/Kbuild
+++ b/include/linux/netfilter/Kbuild
@@ -8,6 +8,7 @@ header-y += xt_CONNMARK.h
header-y += xt_CONNSECMARK.h
header-y += xt_CT.h
header-y += xt_DSCP.h
+header-y += xt_IDLETIMER.h
header-y += xt_LED.h
header-y += xt_MARK.h
header-y += xt_NFLOG.h
diff --git a/include/linux/netfilter/xt_IDLETIMER.h b/include/linux/netfilter/xt_IDLETIMER.h
new file mode 100644
index 0000000..3e1aa1b
--- /dev/null
+++ b/include/linux/netfilter/xt_IDLETIMER.h
@@ -0,0 +1,45 @@
+/*
+ * linux/include/linux/netfilter/xt_IDLETIMER.h
+ *
+ * Header file for Xtables timer target module.
+ *
+ * Copyright (C) 2004, 2010 Nokia Corporation
+ * Written by Timo Teras <ext-timo.teras@nokia.com>
+ *
+ * Converted to x_tables and forward-ported to 2.6.34
+ * by Luciano Coelho <luciano.coelho@nokia.com>
+ *
+ * Contact: Luciano Coelho <luciano.coelho@nokia.com>
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License
+ * version 2 as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful, but
+ * WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ * General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA
+ * 02110-1301 USA
+ */
+
+#ifndef _XT_IDLETIMER_H
+#define _XT_IDLETIMER_H
+
+#include <linux/types.h>
+
+#define MAX_IDLETIMER_LABEL_SIZE 28
+
+struct idletimer_tg_info {
+ __u32 timeout;
+
+ char label[MAX_IDLETIMER_LABEL_SIZE];
+
+ /* for kernel module internal use only */
+ struct idletimer_tg *timer __attribute((aligned(8)));
+};
+
+#endif
diff --git a/net/netfilter/Kconfig b/net/netfilter/Kconfig
index 8593a77..413ed24 100644
--- a/net/netfilter/Kconfig
+++ b/net/netfilter/Kconfig
@@ -424,6 +424,18 @@ config NETFILTER_XT_TARGET_HL
since you can easily create immortal packets that loop
forever on the network.
+config NETFILTER_XT_TARGET_IDLETIMER
+ tristate "IDLETIMER target support"
+ depends on NETFILTER_ADVANCED
+ help
+
+ This option adds the `IDLETIMER' target. Each matching packet
+ resets the timer associated with label specified when the rule is
+ added. When the timer expires, it triggers a sysfs notification.
+ The remaining time for expiration can be read via sysfs.
+
+ To compile it as a module, choose M here. If unsure, say N.
+
config NETFILTER_XT_TARGET_LED
tristate '"LED" target support'
depends on LEDS_CLASS && LEDS_TRIGGERS
diff --git a/net/netfilter/Makefile b/net/netfilter/Makefile
index 14e3a8f..e28420a 100644
--- a/net/netfilter/Makefile
+++ b/net/netfilter/Makefile
@@ -61,6 +61,7 @@ obj-$(CONFIG_NETFILTER_XT_TARGET_TCPMSS) += xt_TCPMSS.o
obj-$(CONFIG_NETFILTER_XT_TARGET_TCPOPTSTRIP) += xt_TCPOPTSTRIP.o
obj-$(CONFIG_NETFILTER_XT_TARGET_TEE) += xt_TEE.o
obj-$(CONFIG_NETFILTER_XT_TARGET_TRACE) += xt_TRACE.o
+obj-$(CONFIG_NETFILTER_XT_TARGET_IDLETIMER) += xt_IDLETIMER.o
# matches
obj-$(CONFIG_NETFILTER_XT_MATCH_CLUSTER) += xt_cluster.o
diff --git a/net/netfilter/xt_IDLETIMER.c b/net/netfilter/xt_IDLETIMER.c
new file mode 100644
index 0000000..e11090a
--- /dev/null
+++ b/net/netfilter/xt_IDLETIMER.c
@@ -0,0 +1,314 @@
+/*
+ * linux/net/netfilter/xt_IDLETIMER.c
+ *
+ * Netfilter module to trigger a timer when packet matches.
+ * After timer expires a kevent will be sent.
+ *
+ * Copyright (C) 2004, 2010 Nokia Corporation
+ * Written by Timo Teras <ext-timo.teras@nokia.com>
+ *
+ * Converted to x_tables and reworked for upstream inclusion
+ * by Luciano Coelho <luciano.coelho@nokia.com>
+ *
+ * Contact: Luciano Coelho <luciano.coelho@nokia.com>
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License
+ * version 2 as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful, but
+ * WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ * General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA
+ * 02110-1301 USA
+ */
+
+#define pr_fmt(fmt) KBUILD_MODNAME ": " fmt
+
+#include <linux/module.h>
+#include <linux/timer.h>
+#include <linux/list.h>
+#include <linux/mutex.h>
+#include <linux/netfilter.h>
+#include <linux/netfilter/x_tables.h>
+#include <linux/netfilter/xt_IDLETIMER.h>
+#include <linux/kobject.h>
+#include <linux/workqueue.h>
+#include <linux/sysfs.h>
+
+struct idletimer_tg_attr {
+ struct attribute attr;
+ ssize_t (*show)(struct kobject *kobj,
+ struct attribute *attr, char *buf);
+};
+
+struct idletimer_tg {
+ struct list_head entry;
+ struct timer_list timer;
+ struct work_struct work;
+
+ struct kobject *kobj;
+ struct idletimer_tg_attr attr;
+
+ unsigned int refcnt;
+};
+
+static LIST_HEAD(idletimer_tg_list);
+static DEFINE_MUTEX(list_mutex);
+
+static struct kobject *idletimer_tg_kobj;
+
+static
+struct idletimer_tg *__idletimer_tg_find_by_label(const char *label)
+{
+ struct idletimer_tg *entry;
+
+ BUG_ON(!label);
+
+ list_for_each_entry(entry, &idletimer_tg_list, entry) {
+ if (!strcmp(label, entry->attr.attr.name))
+ return entry;
+ }
+
+ return NULL;
+}
+
+static ssize_t idletimer_tg_show(struct kobject *kobj, struct attribute *attr,
+ char *buf)
+{
+ struct idletimer_tg *timer;
+ unsigned long expires = 0;
+
+ mutex_lock(&list_mutex);
+
+ timer = __idletimer_tg_find_by_label(attr->name);
+ if (timer)
+ expires = timer->timer.expires;
+
+ mutex_unlock(&list_mutex);
+
+ if (time_after(expires, jiffies))
+ return sprintf(buf, "%u\n",
+ jiffies_to_msecs(expires - jiffies) / 1000);
+
+ return sprintf(buf, "0\n");
+}
+
+static void idletimer_tg_work(struct work_struct *work)
+{
+ struct idletimer_tg *timer = container_of(work, struct idletimer_tg,
+ work);
+
+ sysfs_notify(idletimer_tg_kobj, NULL, timer->attr.attr.name);
+}
+
+static void idletimer_tg_expired(unsigned long data)
+{
+ struct idletimer_tg *timer = (struct idletimer_tg *) data;
+
+ pr_debug("timer %s expired\n", timer->attr.attr.name);
+
+ schedule_work(&timer->work);
+}
+
+static int idletimer_tg_create(struct idletimer_tg_info *info)
+{
+ int ret;
+
+ info->timer = kmalloc(sizeof(*info->timer), GFP_KERNEL);
+ if (!info->timer) {
+ pr_debug("couldn't alloc timer\n");
+ ret = -ENOMEM;
+ goto out;
+ }
+
+ info->timer->attr.attr.name = kstrdup(info->label, GFP_KERNEL);
+ if (!info->timer->attr.attr.name) {
+ pr_debug("couldn't alloc attribute name\n");
+ ret = -ENOMEM;
+ goto out_free_timer;
+ }
+ info->timer->attr.attr.mode = S_IRUGO;
+ info->timer->attr.show = idletimer_tg_show;
+
+ ret = sysfs_create_file(idletimer_tg_kobj, &info->timer->attr.attr);
+ if (ret < 0) {
+ pr_debug("couldn't add file to sysfs");
+ goto out_free_attr;
+ }
+
+ list_add(&info->timer->entry, &idletimer_tg_list);
+
+ setup_timer(&info->timer->timer, idletimer_tg_expired,
+ (unsigned long) info->timer);
+ info->timer->refcnt = 1;
+
+ mod_timer(&info->timer->timer,
+ msecs_to_jiffies(info->timeout * 1000) + jiffies);
+
+ INIT_WORK(&info->timer->work, idletimer_tg_work);
+
+ return 0;
+
+out_free_attr:
+ kfree(info->timer->attr.attr.name);
+out_free_timer:
+ kfree(info->timer);
+out:
+ return ret;
+}
+
+/*
+ * The actual xt_tables plugin.
+ */
+static unsigned int idletimer_tg_target(struct sk_buff *skb,
+ const struct xt_action_param *par)
+{
+ const struct idletimer_tg_info *info = par->targinfo;
+
+ pr_debug("resetting timer %s, timeout period %u\n",
+ info->label, info->timeout);
+
+ BUG_ON(!info->timer);
+
+ mod_timer(&info->timer->timer,
+ msecs_to_jiffies(info->timeout * 1000) + jiffies);
+
+ return XT_CONTINUE;
+}
+
+static int idletimer_tg_checkentry(const struct xt_tgchk_param *par)
+{
+ struct idletimer_tg_info *info = par->targinfo;
+ int ret;
+
+ pr_debug("checkentry targinfo%s\n", info->label);
+
+ if (info->timeout == 0) {
+ pr_debug("timeout value is zero\n");
+ return -EINVAL;
+ }
+
+ if (info->label[0] == '\0' ||
+ strnlen(info->label,
+ MAX_IDLETIMER_LABEL_SIZE) == MAX_IDLETIMER_LABEL_SIZE) {
+ pr_debug("label is empty or not nul-terminated\n");
+ return -EINVAL;
+ }
+
+ mutex_lock(&list_mutex);
+
+ info->timer = __idletimer_tg_find_by_label(info->label);
+ if (info->timer) {
+ info->timer->refcnt++;
+ mod_timer(&info->timer->timer,
+ msecs_to_jiffies(info->timeout * 1000) + jiffies);
+
+ pr_debug("increased refcnt of timer %s to %u\n",
+ info->label, info->timer->refcnt);
+ } else {
+ ret = idletimer_tg_create(info);
+ if (ret < 0) {
+ pr_debug("failed to create timer\n");
+ mutex_unlock(&list_mutex);
+ return ret;
+ }
+ }
+
+ mutex_unlock(&list_mutex);
+ return 0;
+}
+
+static void idletimer_tg_destroy(const struct xt_tgdtor_param *par)
+{
+ const struct idletimer_tg_info *info = par->targinfo;
+
+ pr_debug("destroy targinfo %s\n", info->label);
+
+ mutex_lock(&list_mutex);
+
+ if (--info->timer->refcnt == 0) {
+ pr_debug("deleting timer %s\n", info->label);
+
+ list_del(&info->timer->entry);
+ del_timer_sync(&info->timer->timer);
+ sysfs_remove_file(idletimer_tg_kobj, &info->timer->attr.attr);
+ kfree(info->timer->attr.attr.name);
+ kfree(info->timer);
+ } else {
+ pr_debug("decreased refcnt of timer %s to %u\n",
+ info->label, info->timer->refcnt);
+ }
+
+ mutex_unlock(&list_mutex);
+}
+
+static struct xt_target idletimer_tg __read_mostly = {
+ .name = "IDLETIMER",
+ .family = NFPROTO_UNSPEC,
+ .target = idletimer_tg_target,
+ .targetsize = sizeof(struct idletimer_tg_info),
+ .checkentry = idletimer_tg_checkentry,
+ .destroy = idletimer_tg_destroy,
+ .me = THIS_MODULE,
+};
+
+static struct class *idletimer_tg_class;
+
+static struct device *idletimer_tg_device;
+
+static int __init idletimer_tg_init(void)
+{
+ int err;
+
+ idletimer_tg_class = class_create(THIS_MODULE, "xt_idletimer");
+ err = PTR_ERR(idletimer_tg_class);
+ if (IS_ERR(idletimer_tg_class)) {
+ pr_debug("couldn't register device class\n");
+ goto out;
+ }
+
+ idletimer_tg_device = device_create(idletimer_tg_class, NULL,
+ MKDEV(0, 0), NULL, "timers");
+ err = PTR_ERR(idletimer_tg_device);
+ if (IS_ERR(idletimer_tg_device)) {
+ pr_debug("couldn't register system device\n");
+ goto out_class;
+ }
+
+ idletimer_tg_kobj = &idletimer_tg_device->kobj;
+
+ err = xt_register_target(&idletimer_tg);
+ if (err < 0) {
+ pr_debug("couldn't register xt target\n");
+ goto out_dev;
+ }
+
+ return 0;
+out_dev:
+ device_destroy(idletimer_tg_class, MKDEV(0, 0));
+out_class:
+ class_destroy(idletimer_tg_class);
+out:
+ return err;
+}
+
+static void __exit idletimer_tg_exit(void)
+{
+ xt_unregister_target(&idletimer_tg);
+
+ device_destroy(idletimer_tg_class, MKDEV(0, 0));
+ class_destroy(idletimer_tg_class);
+}
+
+module_init(idletimer_tg_init);
+module_exit(idletimer_tg_exit);
+
+MODULE_AUTHOR("Timo Teras <ext-timo.teras@nokia.com>");
+MODULE_AUTHOR("Luciano Coelho <luciano.coelho@nokia.com>");
+MODULE_DESCRIPTION("Xtables: idle time monitor");
+MODULE_LICENSE("GPL v2");
--
1.7.0.4
^ permalink raw reply related
* [PATCH 16/17] netfilter: CLUSTERIP: RCU conversion
From: kaber @ 2010-06-15 15:38 UTC (permalink / raw)
To: davem; +Cc: netfilter-devel, netdev
In-Reply-To: <1276616323-20880-1-git-send-email-kaber@trash.net>
From: Eric Dumazet <eric.dumazet@gmail.com>
- clusterip_lock becomes a spinlock
- lockless lookups
- kfree() deferred after RCU grace period
- rcu_barrier_bh() inserted in clusterip_tg_exit()
v2)
- As Patrick pointed out, we use atomic_inc_not_zero() in
clusterip_config_find_get().
- list_add_rcu() and list_del_rcu() variants are used.
- atomic_dec_and_lock() used in clusterip_config_entry_put()
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: Patrick McHardy <kaber@trash.net>
---
net/ipv4/netfilter/ipt_CLUSTERIP.c | 48 +++++++++++++++++++++--------------
1 files changed, 29 insertions(+), 19 deletions(-)
diff --git a/net/ipv4/netfilter/ipt_CLUSTERIP.c b/net/ipv4/netfilter/ipt_CLUSTERIP.c
index f91c94b..64d0875 100644
--- a/net/ipv4/netfilter/ipt_CLUSTERIP.c
+++ b/net/ipv4/netfilter/ipt_CLUSTERIP.c
@@ -53,12 +53,13 @@ struct clusterip_config {
#endif
enum clusterip_hashmode hash_mode; /* which hashing mode */
u_int32_t hash_initval; /* hash initialization */
+ struct rcu_head rcu;
};
static LIST_HEAD(clusterip_configs);
/* clusterip_lock protects the clusterip_configs list */
-static DEFINE_RWLOCK(clusterip_lock);
+static DEFINE_SPINLOCK(clusterip_lock);
#ifdef CONFIG_PROC_FS
static const struct file_operations clusterip_proc_fops;
@@ -71,11 +72,17 @@ clusterip_config_get(struct clusterip_config *c)
atomic_inc(&c->refcount);
}
+
+static void clusterip_config_rcu_free(struct rcu_head *head)
+{
+ kfree(container_of(head, struct clusterip_config, rcu));
+}
+
static inline void
clusterip_config_put(struct clusterip_config *c)
{
if (atomic_dec_and_test(&c->refcount))
- kfree(c);
+ call_rcu_bh(&c->rcu, clusterip_config_rcu_free);
}
/* decrease the count of entries using/referencing this config. If last
@@ -84,10 +91,11 @@ clusterip_config_put(struct clusterip_config *c)
static inline void
clusterip_config_entry_put(struct clusterip_config *c)
{
- write_lock_bh(&clusterip_lock);
- if (atomic_dec_and_test(&c->entries)) {
- list_del(&c->list);
- write_unlock_bh(&clusterip_lock);
+ local_bh_disable();
+ if (atomic_dec_and_lock(&c->entries, &clusterip_lock)) {
+ list_del_rcu(&c->list);
+ spin_unlock(&clusterip_lock);
+ local_bh_enable();
dev_mc_del(c->dev, c->clustermac);
dev_put(c->dev);
@@ -100,7 +108,7 @@ clusterip_config_entry_put(struct clusterip_config *c)
#endif
return;
}
- write_unlock_bh(&clusterip_lock);
+ local_bh_enable();
}
static struct clusterip_config *
@@ -108,7 +116,7 @@ __clusterip_config_find(__be32 clusterip)
{
struct clusterip_config *c;
- list_for_each_entry(c, &clusterip_configs, list) {
+ list_for_each_entry_rcu(c, &clusterip_configs, list) {
if (c->clusterip == clusterip)
return c;
}
@@ -121,16 +129,15 @@ clusterip_config_find_get(__be32 clusterip, int entry)
{
struct clusterip_config *c;
- read_lock_bh(&clusterip_lock);
+ rcu_read_lock_bh();
c = __clusterip_config_find(clusterip);
- if (!c) {
- read_unlock_bh(&clusterip_lock);
- return NULL;
+ if (c) {
+ if (unlikely(!atomic_inc_not_zero(&c->refcount)))
+ c = NULL;
+ else if (entry)
+ atomic_inc(&c->entries);
}
- atomic_inc(&c->refcount);
- if (entry)
- atomic_inc(&c->entries);
- read_unlock_bh(&clusterip_lock);
+ rcu_read_unlock_bh();
return c;
}
@@ -181,9 +188,9 @@ clusterip_config_init(const struct ipt_clusterip_tgt_info *i, __be32 ip,
}
#endif
- write_lock_bh(&clusterip_lock);
- list_add(&c->list, &clusterip_configs);
- write_unlock_bh(&clusterip_lock);
+ spin_lock_bh(&clusterip_lock);
+ list_add_rcu(&c->list, &clusterip_configs);
+ spin_unlock_bh(&clusterip_lock);
return c;
}
@@ -733,6 +740,9 @@ static void __exit clusterip_tg_exit(void)
#endif
nf_unregister_hook(&cip_arp_ops);
xt_unregister_target(&clusterip_tg_reg);
+
+ /* Wait for completion of call_rcu_bh()'s (clusterip_config_rcu_free) */
+ rcu_barrier_bh();
}
module_init(clusterip_tg_init);
--
1.7.0.4
^ permalink raw reply related
* [PATCH 11/17] netfilter: nfnetlink_log: RCU conversion
From: kaber @ 2010-06-15 15:38 UTC (permalink / raw)
To: davem; +Cc: netfilter-devel, netdev
In-Reply-To: <1276616323-20880-1-git-send-email-kaber@trash.net>
From: Eric Dumazet <eric.dumazet@gmail.com>
- instances_lock becomes a spinlock
- lockless lookups
While nfnetlink_log probably not performance critical, using less
rwlocks in our code is always welcomed...
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: Patrick McHardy <kaber@trash.net>
---
net/netfilter/nfnetlink_log.c | 49 ++++++++++++++++++++++------------------
1 files changed, 27 insertions(+), 22 deletions(-)
diff --git a/net/netfilter/nfnetlink_log.c b/net/netfilter/nfnetlink_log.c
index fc9a211..8ec23ec 100644
--- a/net/netfilter/nfnetlink_log.c
+++ b/net/netfilter/nfnetlink_log.c
@@ -66,9 +66,10 @@ struct nfulnl_instance {
u_int16_t group_num; /* number of this queue */
u_int16_t flags;
u_int8_t copy_mode;
+ struct rcu_head rcu;
};
-static DEFINE_RWLOCK(instances_lock);
+static DEFINE_SPINLOCK(instances_lock);
static atomic_t global_seq;
#define INSTANCE_BUCKETS 16
@@ -88,7 +89,7 @@ __instance_lookup(u_int16_t group_num)
struct nfulnl_instance *inst;
head = &instance_table[instance_hashfn(group_num)];
- hlist_for_each_entry(inst, pos, head, hlist) {
+ hlist_for_each_entry_rcu(inst, pos, head, hlist) {
if (inst->group_num == group_num)
return inst;
}
@@ -106,22 +107,26 @@ instance_lookup_get(u_int16_t group_num)
{
struct nfulnl_instance *inst;
- read_lock_bh(&instances_lock);
+ rcu_read_lock_bh();
inst = __instance_lookup(group_num);
if (inst)
instance_get(inst);
- read_unlock_bh(&instances_lock);
+ rcu_read_unlock_bh();
return inst;
}
+static void nfulnl_instance_free_rcu(struct rcu_head *head)
+{
+ kfree(container_of(head, struct nfulnl_instance, rcu));
+ module_put(THIS_MODULE);
+}
+
static void
instance_put(struct nfulnl_instance *inst)
{
- if (inst && atomic_dec_and_test(&inst->use)) {
- kfree(inst);
- module_put(THIS_MODULE);
- }
+ if (inst && atomic_dec_and_test(&inst->use))
+ call_rcu_bh(&inst->rcu, nfulnl_instance_free_rcu);
}
static void nfulnl_timer(unsigned long data);
@@ -132,7 +137,7 @@ instance_create(u_int16_t group_num, int pid)
struct nfulnl_instance *inst;
int err;
- write_lock_bh(&instances_lock);
+ spin_lock_bh(&instances_lock);
if (__instance_lookup(group_num)) {
err = -EEXIST;
goto out_unlock;
@@ -169,12 +174,12 @@ instance_create(u_int16_t group_num, int pid)
hlist_add_head(&inst->hlist,
&instance_table[instance_hashfn(group_num)]);
- write_unlock_bh(&instances_lock);
+ spin_unlock_bh(&instances_lock);
return inst;
out_unlock:
- write_unlock_bh(&instances_lock);
+ spin_unlock_bh(&instances_lock);
return ERR_PTR(err);
}
@@ -200,9 +205,9 @@ __instance_destroy(struct nfulnl_instance *inst)
static inline void
instance_destroy(struct nfulnl_instance *inst)
{
- write_lock_bh(&instances_lock);
+ spin_lock_bh(&instances_lock);
__instance_destroy(inst);
- write_unlock_bh(&instances_lock);
+ spin_unlock_bh(&instances_lock);
}
static int
@@ -672,7 +677,7 @@ nfulnl_rcv_nl_event(struct notifier_block *this,
int i;
/* destroy all instances for this pid */
- write_lock_bh(&instances_lock);
+ spin_lock_bh(&instances_lock);
for (i = 0; i < INSTANCE_BUCKETS; i++) {
struct hlist_node *tmp, *t2;
struct nfulnl_instance *inst;
@@ -684,7 +689,7 @@ nfulnl_rcv_nl_event(struct notifier_block *this,
__instance_destroy(inst);
}
}
- write_unlock_bh(&instances_lock);
+ spin_unlock_bh(&instances_lock);
}
return NOTIFY_DONE;
}
@@ -861,19 +866,19 @@ static struct hlist_node *get_first(struct iter_state *st)
for (st->bucket = 0; st->bucket < INSTANCE_BUCKETS; st->bucket++) {
if (!hlist_empty(&instance_table[st->bucket]))
- return instance_table[st->bucket].first;
+ return rcu_dereference_bh(instance_table[st->bucket].first);
}
return NULL;
}
static struct hlist_node *get_next(struct iter_state *st, struct hlist_node *h)
{
- h = h->next;
+ h = rcu_dereference_bh(h->next);
while (!h) {
if (++st->bucket >= INSTANCE_BUCKETS)
return NULL;
- h = instance_table[st->bucket].first;
+ h = rcu_dereference_bh(instance_table[st->bucket].first);
}
return h;
}
@@ -890,9 +895,9 @@ static struct hlist_node *get_idx(struct iter_state *st, loff_t pos)
}
static void *seq_start(struct seq_file *seq, loff_t *pos)
- __acquires(instances_lock)
+ __acquires(rcu_bh)
{
- read_lock_bh(&instances_lock);
+ rcu_read_lock_bh();
return get_idx(seq->private, *pos);
}
@@ -903,9 +908,9 @@ static void *seq_next(struct seq_file *s, void *v, loff_t *pos)
}
static void seq_stop(struct seq_file *s, void *v)
- __releases(instances_lock)
+ __releases(rcu_bh)
{
- read_unlock_bh(&instances_lock);
+ rcu_read_unlock_bh();
}
static int seq_show(struct seq_file *s, void *v)
--
1.7.0.4
^ permalink raw reply related
* [PATCH 15/17] netfilter: defrag: kill unused work parameter of frag_kfree_skb()
From: kaber @ 2010-06-15 15:38 UTC (permalink / raw)
To: davem; +Cc: netfilter-devel, netdev
In-Reply-To: <1276616323-20880-1-git-send-email-kaber@trash.net>
From: Shan Wei <shanwei@cn.fujitsu.com>
The parameter (work) is unused, remove it.
Reported from Eric Dumazet.
Signed-off-by: Shan Wei <shanwei@cn.fujitsu.com>
Acked-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: Patrick McHardy <kaber@trash.net>
---
net/ipv6/netfilter/nf_conntrack_reasm.c | 6 ++----
1 files changed, 2 insertions(+), 4 deletions(-)
diff --git a/net/ipv6/netfilter/nf_conntrack_reasm.c b/net/ipv6/netfilter/nf_conntrack_reasm.c
index bc5b86d..9254008 100644
--- a/net/ipv6/netfilter/nf_conntrack_reasm.c
+++ b/net/ipv6/netfilter/nf_conntrack_reasm.c
@@ -114,10 +114,8 @@ static void nf_skb_free(struct sk_buff *skb)
}
/* Memory Tracking Functions. */
-static inline void frag_kfree_skb(struct sk_buff *skb, unsigned int *work)
+static void frag_kfree_skb(struct sk_buff *skb)
{
- if (work)
- *work -= skb->truesize;
atomic_sub(skb->truesize, &nf_init_frags.mem);
nf_skb_free(skb);
kfree_skb(skb);
@@ -335,7 +333,7 @@ static int nf_ct_frag6_queue(struct nf_ct_frag6_queue *fq, struct sk_buff *skb,
fq->q.fragments = next;
fq->q.meat -= free_it->len;
- frag_kfree_skb(free_it, NULL);
+ frag_kfree_skb(free_it);
}
}
--
1.7.0.4
^ permalink raw reply related
* [PATCH 14/17] netfilter: defrag: remove one redundant atomic ops
From: kaber @ 2010-06-15 15:38 UTC (permalink / raw)
To: davem; +Cc: netfilter-devel, netdev
In-Reply-To: <1276616323-20880-1-git-send-email-kaber@trash.net>
From: Shan Wei <shanwei@cn.fujitsu.com>
Instead of doing one atomic operation per frag, we can factorize them.
Reported from Eric Dumazet.
Signed-off-by: Shan Wei <shanwei@cn.fujitsu.com>
Acked-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: Patrick McHardy <kaber@trash.net>
---
net/ipv6/netfilter/nf_conntrack_reasm.c | 3 +--
1 files changed, 1 insertions(+), 2 deletions(-)
diff --git a/net/ipv6/netfilter/nf_conntrack_reasm.c b/net/ipv6/netfilter/nf_conntrack_reasm.c
index 6fb8901..bc5b86d 100644
--- a/net/ipv6/netfilter/nf_conntrack_reasm.c
+++ b/net/ipv6/netfilter/nf_conntrack_reasm.c
@@ -442,7 +442,6 @@ nf_ct_frag6_reasm(struct nf_ct_frag6_queue *fq, struct net_device *dev)
skb_shinfo(head)->frag_list = head->next;
skb_reset_transport_header(head);
skb_push(head, head->data - skb_network_header(head));
- atomic_sub(head->truesize, &nf_init_frags.mem);
for (fp=head->next; fp; fp = fp->next) {
head->data_len += fp->len;
@@ -452,8 +451,8 @@ nf_ct_frag6_reasm(struct nf_ct_frag6_queue *fq, struct net_device *dev)
else if (head->ip_summed == CHECKSUM_COMPLETE)
head->csum = csum_add(head->csum, fp->csum);
head->truesize += fp->truesize;
- atomic_sub(fp->truesize, &nf_init_frags.mem);
}
+ atomic_sub(head->truesize, &nf_init_frags.mem);
head->next = NULL;
head->dev = dev;
--
1.7.0.4
^ permalink raw reply related
* [PATCH 13/17] netfilter: kill redundant check code in which setting ip_summed value
From: kaber @ 2010-06-15 15:38 UTC (permalink / raw)
To: davem; +Cc: netfilter-devel, netdev
In-Reply-To: <1276616323-20880-1-git-send-email-kaber@trash.net>
From: Shan Wei <shanwei@cn.fujitsu.com>
If the returned csum value is 0, We has set ip_summed with
CHECKSUM_UNNECESSARY flag in __skb_checksum_complete_head().
So this patch kills the check and changes to return to upper
caller directly.
Signed-off-by: Shan Wei <shanwei@cn.fujitsu.com>
Signed-off-by: Patrick McHardy <kaber@trash.net>
---
net/ipv4/netfilter.c | 4 +---
net/ipv6/netfilter.c | 4 +---
2 files changed, 2 insertions(+), 6 deletions(-)
diff --git a/net/ipv4/netfilter.c b/net/ipv4/netfilter.c
index 07de855..acd1ea8 100644
--- a/net/ipv4/netfilter.c
+++ b/net/ipv4/netfilter.c
@@ -212,9 +212,7 @@ static __sum16 nf_ip_checksum_partial(struct sk_buff *skb, unsigned int hook,
skb->csum = csum_tcpudp_nofold(iph->saddr, iph->daddr, protocol,
skb->len - dataoff, 0);
skb->ip_summed = CHECKSUM_NONE;
- csum = __skb_checksum_complete_head(skb, dataoff + len);
- if (!csum)
- skb->ip_summed = CHECKSUM_UNNECESSARY;
+ return __skb_checksum_complete_head(skb, dataoff + len);
}
return csum;
}
diff --git a/net/ipv6/netfilter.c b/net/ipv6/netfilter.c
index a74951c..7155b24 100644
--- a/net/ipv6/netfilter.c
+++ b/net/ipv6/netfilter.c
@@ -151,9 +151,7 @@ static __sum16 nf_ip6_checksum_partial(struct sk_buff *skb, unsigned int hook,
protocol,
csum_sub(0, hsum)));
skb->ip_summed = CHECKSUM_NONE;
- csum = __skb_checksum_complete_head(skb, dataoff + len);
- if (!csum)
- skb->ip_summed = CHECKSUM_UNNECESSARY;
+ return __skb_checksum_complete_head(skb, dataoff + len);
}
return csum;
};
--
1.7.0.4
^ permalink raw reply related
page: next (older) | prev (newer) | latest
- recent:[subjects (threaded)|topics (new)|topics (active)]
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox