Merge git://git.kernel.org/pub/scm/linux/kernel/git/bart/ide-2.6
* git://git.kernel.org/pub/scm/linux/kernel/git/bart/ide-2.6: (49 commits)
ide-tape: remove tape->merge_stage
ide-tape: mv tape->merge_stage_size tape->merge_bh_size
ide-tape: mv idetape_empty_write_pipeline ide_tape_flush_merge_buffer
ide-tape: mv idetape_discard_read_pipeline ide_tape_discard_merge_buffer
ide-tape: make __idetape_discard_read_pipeline() of type void
ide: remove now unused ide_pci_create_host_proc()
ide: remove /proc/ide/ali
ide-tape: improve buffer pages freeing strategy
ide-tape: mv tape->pages_per_stage tape->pages_per_buffer
ide-tape: mv tape->stage_size tape->buffer_size
ide-tape: improve buffer allocation strategy
ide: add struct ide_io_ports (take 3)
ide: make ide_unregister() take 'ide_hwif_t *' as an argument (take 2)
ide: sanitize ide_unregister() usage
mpc8xx-ide: use ide_find_port()
ide: add "noacpi" / "acpigtf" / "acpionboot" parameters
gayle: add "doubler" parameter
ide: add "cdrom=" and "chs=" parameters
ide: add "nodma|noflush|noprobe|nowerr=" parameters
ide: remove obsoleted "hdx=autotune" kernel parameter
...
diff --git a/arch/x86/kernel/apm_32.c b/arch/x86/kernel/apm_32.c
index f0030a0..e4ea362 100644
--- a/arch/x86/kernel/apm_32.c
+++ b/arch/x86/kernel/apm_32.c
@@ -904,6 +904,7 @@
original_pm_idle();
else
default_idle();
+ local_irq_disable();
jiffies_since_last_check = jiffies - last_jiffies;
if (jiffies_since_last_check > idle_period)
goto recalc;
@@ -911,6 +912,8 @@
if (apm_idle_done)
apm_do_busy();
+
+ local_irq_enable();
}
/**
diff --git a/arch/x86/kernel/process.c b/arch/x86/kernel/process.c
index 3004d71..67e9b4a 100644
--- a/arch/x86/kernel/process.c
+++ b/arch/x86/kernel/process.c
@@ -4,6 +4,8 @@
#include <linux/smp.h>
#include <linux/slab.h>
#include <linux/sched.h>
+#include <linux/module.h>
+#include <linux/pm.h>
struct kmem_cache *task_xstate_cachep;
@@ -42,3 +44,118 @@
__alignof__(union thread_xstate),
SLAB_PANIC, NULL);
}
+
+static void do_nothing(void *unused)
+{
+}
+
+/*
+ * cpu_idle_wait - Used to ensure that all the CPUs discard old value of
+ * pm_idle and update to new pm_idle value. Required while changing pm_idle
+ * handler on SMP systems.
+ *
+ * Caller must have changed pm_idle to the new value before the call. Old
+ * pm_idle value will not be used by any CPU after the return of this function.
+ */
+void cpu_idle_wait(void)
+{
+ smp_mb();
+ /* kick all the CPUs so that they exit out of pm_idle */
+ smp_call_function(do_nothing, NULL, 0, 1);
+}
+EXPORT_SYMBOL_GPL(cpu_idle_wait);
+
+/*
+ * This uses new MONITOR/MWAIT instructions on P4 processors with PNI,
+ * which can obviate IPI to trigger checking of need_resched.
+ * We execute MONITOR against need_resched and enter optimized wait state
+ * through MWAIT. Whenever someone changes need_resched, we would be woken
+ * up from MWAIT (without an IPI).
+ *
+ * New with Core Duo processors, MWAIT can take some hints based on CPU
+ * capability.
+ */
+void mwait_idle_with_hints(unsigned long ax, unsigned long cx)
+{
+ if (!need_resched()) {
+ __monitor((void *)¤t_thread_info()->flags, 0, 0);
+ smp_mb();
+ if (!need_resched())
+ __mwait(ax, cx);
+ }
+}
+
+/* Default MONITOR/MWAIT with no hints, used for default C1 state */
+static void mwait_idle(void)
+{
+ if (!need_resched()) {
+ __monitor((void *)¤t_thread_info()->flags, 0, 0);
+ smp_mb();
+ if (!need_resched())
+ __sti_mwait(0, 0);
+ else
+ local_irq_enable();
+ } else
+ local_irq_enable();
+}
+
+
+static int __cpuinit mwait_usable(const struct cpuinfo_x86 *c)
+{
+ if (force_mwait)
+ return 1;
+ /* Any C1 states supported? */
+ return c->cpuid_level >= 5 && ((cpuid_edx(5) >> 4) & 0xf) > 0;
+}
+
+/*
+ * On SMP it's slightly faster (but much more power-consuming!)
+ * to poll the ->work.need_resched flag instead of waiting for the
+ * cross-CPU IPI to arrive. Use this option with caution.
+ */
+static void poll_idle(void)
+{
+ local_irq_enable();
+ cpu_relax();
+}
+
+void __cpuinit select_idle_routine(const struct cpuinfo_x86 *c)
+{
+ static int selected;
+
+ if (selected)
+ return;
+#ifdef CONFIG_X86_SMP
+ if (pm_idle == poll_idle && smp_num_siblings > 1) {
+ printk(KERN_WARNING "WARNING: polling idle and HT enabled,"
+ " performance may degrade.\n");
+ }
+#endif
+ if (cpu_has(c, X86_FEATURE_MWAIT) && mwait_usable(c)) {
+ /*
+ * Skip, if setup has overridden idle.
+ * One CPU supports mwait => All CPUs supports mwait
+ */
+ if (!pm_idle) {
+ printk(KERN_INFO "using mwait in idle threads.\n");
+ pm_idle = mwait_idle;
+ }
+ }
+ selected = 1;
+}
+
+static int __init idle_setup(char *str)
+{
+ if (!strcmp(str, "poll")) {
+ printk("using polling idle threads.\n");
+ pm_idle = poll_idle;
+ } else if (!strcmp(str, "mwait"))
+ force_mwait = 1;
+ else
+ return -1;
+
+ boot_option_idle_override = 1;
+ return 0;
+}
+early_param("idle", idle_setup);
+
diff --git a/arch/x86/kernel/process_32.c b/arch/x86/kernel/process_32.c
index 77de848..f8476df 100644
--- a/arch/x86/kernel/process_32.c
+++ b/arch/x86/kernel/process_32.c
@@ -111,12 +111,10 @@
*/
smp_mb();
- local_irq_disable();
- if (!need_resched()) {
+ if (!need_resched())
safe_halt(); /* enables interrupts racelessly */
- local_irq_disable();
- }
- local_irq_enable();
+ else
+ local_irq_enable();
current_thread_info()->status |= TS_POLLING;
} else {
local_irq_enable();
@@ -128,17 +126,6 @@
EXPORT_SYMBOL(default_idle);
#endif
-/*
- * On SMP it's slightly faster (but much more power-consuming!)
- * to poll the ->work.need_resched flag instead of waiting for the
- * cross-CPU IPI to arrive. Use this option with caution.
- */
-static void poll_idle(void)
-{
- local_irq_enable();
- cpu_relax();
-}
-
#ifdef CONFIG_HOTPLUG_CPU
#include <asm/nmi.h>
/* We don't actually take CPU down, just spin without interrupts. */
@@ -196,6 +183,7 @@
if (cpu_is_offline(cpu))
play_dead();
+ local_irq_disable();
__get_cpu_var(irq_stat).idle_timestamp = jiffies;
idle();
}
@@ -206,104 +194,6 @@
}
}
-static void do_nothing(void *unused)
-{
-}
-
-/*
- * cpu_idle_wait - Used to ensure that all the CPUs discard old value of
- * pm_idle and update to new pm_idle value. Required while changing pm_idle
- * handler on SMP systems.
- *
- * Caller must have changed pm_idle to the new value before the call. Old
- * pm_idle value will not be used by any CPU after the return of this function.
- */
-void cpu_idle_wait(void)
-{
- smp_mb();
- /* kick all the CPUs so that they exit out of pm_idle */
- smp_call_function(do_nothing, NULL, 0, 1);
-}
-EXPORT_SYMBOL_GPL(cpu_idle_wait);
-
-/*
- * This uses new MONITOR/MWAIT instructions on P4 processors with PNI,
- * which can obviate IPI to trigger checking of need_resched.
- * We execute MONITOR against need_resched and enter optimized wait state
- * through MWAIT. Whenever someone changes need_resched, we would be woken
- * up from MWAIT (without an IPI).
- *
- * New with Core Duo processors, MWAIT can take some hints based on CPU
- * capability.
- */
-void mwait_idle_with_hints(unsigned long ax, unsigned long cx)
-{
- if (!need_resched()) {
- __monitor((void *)¤t_thread_info()->flags, 0, 0);
- smp_mb();
- if (!need_resched())
- __sti_mwait(ax, cx);
- else
- local_irq_enable();
- } else
- local_irq_enable();
-}
-
-/* Default MONITOR/MWAIT with no hints, used for default C1 state */
-static void mwait_idle(void)
-{
- local_irq_enable();
- mwait_idle_with_hints(0, 0);
-}
-
-static int __cpuinit mwait_usable(const struct cpuinfo_x86 *c)
-{
- if (force_mwait)
- return 1;
- /* Any C1 states supported? */
- return c->cpuid_level >= 5 && ((cpuid_edx(5) >> 4) & 0xf) > 0;
-}
-
-void __cpuinit select_idle_routine(const struct cpuinfo_x86 *c)
-{
- static int selected;
-
- if (selected)
- return;
-#ifdef CONFIG_X86_SMP
- if (pm_idle == poll_idle && smp_num_siblings > 1) {
- printk(KERN_WARNING "WARNING: polling idle and HT enabled,"
- " performance may degrade.\n");
- }
-#endif
- if (cpu_has(c, X86_FEATURE_MWAIT) && mwait_usable(c)) {
- /*
- * Skip, if setup has overridden idle.
- * One CPU supports mwait => All CPUs supports mwait
- */
- if (!pm_idle) {
- printk(KERN_INFO "using mwait in idle threads.\n");
- pm_idle = mwait_idle;
- }
- }
- selected = 1;
-}
-
-static int __init idle_setup(char *str)
-{
- if (!strcmp(str, "poll")) {
- printk("using polling idle threads.\n");
- pm_idle = poll_idle;
- } else if (!strcmp(str, "mwait"))
- force_mwait = 1;
- else
- return -1;
-
- boot_option_idle_override = 1;
- return 0;
-}
-early_param("idle", idle_setup);
-
void __show_registers(struct pt_regs *regs, int all)
{
unsigned long cr0 = 0L, cr2 = 0L, cr3 = 0L, cr4 = 0L;
diff --git a/arch/x86/kernel/process_64.c b/arch/x86/kernel/process_64.c
index 131c2ee..e2319f3 100644
--- a/arch/x86/kernel/process_64.c
+++ b/arch/x86/kernel/process_64.c
@@ -106,26 +106,13 @@
* test NEED_RESCHED:
*/
smp_mb();
- local_irq_disable();
- if (!need_resched()) {
+ if (!need_resched())
safe_halt(); /* enables interrupts racelessly */
- local_irq_disable();
- }
- local_irq_enable();
+ else
+ local_irq_enable();
current_thread_info()->status |= TS_POLLING;
}
-/*
- * On SMP it's slightly faster (but much more power-consuming!)
- * to poll the ->need_resched flag instead of waiting for the
- * cross-CPU IPI to arrive. Use this option with caution.
- */
-static void poll_idle(void)
-{
- local_irq_enable();
- cpu_relax();
-}
-
#ifdef CONFIG_HOTPLUG_CPU
DECLARE_PER_CPU(int, cpu_state);
@@ -192,110 +179,6 @@
}
}
-static void do_nothing(void *unused)
-{
-}
-
-/*
- * cpu_idle_wait - Used to ensure that all the CPUs discard old value of
- * pm_idle and update to new pm_idle value. Required while changing pm_idle
- * handler on SMP systems.
- *
- * Caller must have changed pm_idle to the new value before the call. Old
- * pm_idle value will not be used by any CPU after the return of this function.
- */
-void cpu_idle_wait(void)
-{
- smp_mb();
- /* kick all the CPUs so that they exit out of pm_idle */
- smp_call_function(do_nothing, NULL, 0, 1);
-}
-EXPORT_SYMBOL_GPL(cpu_idle_wait);
-
-/*
- * This uses new MONITOR/MWAIT instructions on P4 processors with PNI,
- * which can obviate IPI to trigger checking of need_resched.
- * We execute MONITOR against need_resched and enter optimized wait state
- * through MWAIT. Whenever someone changes need_resched, we would be woken
- * up from MWAIT (without an IPI).
- *
- * New with Core Duo processors, MWAIT can take some hints based on CPU
- * capability.
- */
-void mwait_idle_with_hints(unsigned long ax, unsigned long cx)
-{
- if (!need_resched()) {
- __monitor((void *)¤t_thread_info()->flags, 0, 0);
- smp_mb();
- if (!need_resched())
- __mwait(ax, cx);
- }
-}
-
-/* Default MONITOR/MWAIT with no hints, used for default C1 state */
-static void mwait_idle(void)
-{
- if (!need_resched()) {
- __monitor((void *)¤t_thread_info()->flags, 0, 0);
- smp_mb();
- if (!need_resched())
- __sti_mwait(0, 0);
- else
- local_irq_enable();
- } else {
- local_irq_enable();
- }
-}
-
-
-static int __cpuinit mwait_usable(const struct cpuinfo_x86 *c)
-{
- if (force_mwait)
- return 1;
- /* Any C1 states supported? */
- return c->cpuid_level >= 5 && ((cpuid_edx(5) >> 4) & 0xf) > 0;
-}
-
-void __cpuinit select_idle_routine(const struct cpuinfo_x86 *c)
-{
- static int selected;
-
- if (selected)
- return;
-#ifdef CONFIG_X86_SMP
- if (pm_idle == poll_idle && smp_num_siblings > 1) {
- printk(KERN_WARNING "WARNING: polling idle and HT enabled,"
- " performance may degrade.\n");
- }
-#endif
- if (cpu_has(c, X86_FEATURE_MWAIT) && mwait_usable(c)) {
- /*
- * Skip, if setup has overridden idle.
- * One CPU supports mwait => All CPUs supports mwait
- */
- if (!pm_idle) {
- printk(KERN_INFO "using mwait in idle threads.\n");
- pm_idle = mwait_idle;
- }
- }
- selected = 1;
-}
-
-static int __init idle_setup(char *str)
-{
- if (!strcmp(str, "poll")) {
- printk("using polling idle threads.\n");
- pm_idle = poll_idle;
- } else if (!strcmp(str, "mwait"))
- force_mwait = 1;
- else
- return -1;
-
- boot_option_idle_override = 1;
- return 0;
-}
-early_param("idle", idle_setup);
-
/* Prints also some state that isn't saved in the pt_regs */
void __show_regs(struct pt_regs * regs)
{
diff --git a/drivers/acpi/processor_idle.c b/drivers/acpi/processor_idle.c
index 788da97..0d90ff5 100644
--- a/drivers/acpi/processor_idle.c
+++ b/drivers/acpi/processor_idle.c
@@ -418,13 +418,12 @@
cx = pr->power.state;
if (!cx || acpi_idle_suspend) {
- if (pm_idle_save)
- pm_idle_save();
- else
+ if (pm_idle_save) {
+ pm_idle_save(); /* enables IRQs */
+ } else {
acpi_safe_halt();
-
- if (irqs_disabled())
local_irq_enable();
+ }
return;
}
@@ -520,10 +519,12 @@
* Use the appropriate idle routine, the one that would
* be used without acpi C-states.
*/
- if (pm_idle_save)
- pm_idle_save();
- else
+ if (pm_idle_save) {
+ pm_idle_save(); /* enables IRQs */
+ } else {
acpi_safe_halt();
+ local_irq_enable();
+ }
/*
* TBD: Can't get time duration while in C1, as resumes
@@ -534,8 +535,6 @@
* skew otherwise.
*/
sleep_ticks = 0xFFFFFFFF;
- if (irqs_disabled())
- local_irq_enable();
break;
diff --git a/drivers/infiniband/hw/ehca/ehca_classes.h b/drivers/infiniband/hw/ehca/ehca_classes.h
index 0d13fe0..3d6d9461 100644
--- a/drivers/infiniband/hw/ehca/ehca_classes.h
+++ b/drivers/infiniband/hw/ehca/ehca_classes.h
@@ -160,6 +160,7 @@
};
u32 qp_type;
enum ehca_ext_qp_type ext_type;
+ enum ib_qp_state state;
struct ipz_queue ipz_squeue;
struct ipz_queue ipz_rqueue;
struct h_galpas galpas;
diff --git a/drivers/infiniband/hw/ehca/ehca_irq.c b/drivers/infiniband/hw/ehca/ehca_irq.c
index b5ca94c..ca5eb0c 100644
--- a/drivers/infiniband/hw/ehca/ehca_irq.c
+++ b/drivers/infiniband/hw/ehca/ehca_irq.c
@@ -633,7 +633,7 @@
unsigned long flags;
WARN_ON_ONCE(!in_interrupt());
- if (ehca_debug_level)
+ if (ehca_debug_level >= 3)
ehca_dmp(&cpu_online_map, sizeof(cpumask_t), "");
spin_lock_irqsave(&pool->last_cpu_lock, flags);
diff --git a/drivers/infiniband/hw/ehca/ehca_main.c b/drivers/infiniband/hw/ehca/ehca_main.c
index 65b3362..6504897 100644
--- a/drivers/infiniband/hw/ehca/ehca_main.c
+++ b/drivers/infiniband/hw/ehca/ehca_main.c
@@ -50,7 +50,7 @@
#include "ehca_tools.h"
#include "hcp_if.h"
-#define HCAD_VERSION "0025"
+#define HCAD_VERSION "0026"
MODULE_LICENSE("Dual BSD/GPL");
MODULE_AUTHOR("Christoph Raisch <raisch@de.ibm.com>");
@@ -60,7 +60,6 @@
static int ehca_open_aqp1 = 0;
static int ehca_hw_level = 0;
static int ehca_poll_all_eqs = 1;
-static int ehca_mr_largepage = 1;
int ehca_debug_level = 0;
int ehca_nr_ports = 2;
@@ -70,45 +69,40 @@
int ehca_scaling_code = 0;
int ehca_lock_hcalls = -1;
-module_param_named(open_aqp1, ehca_open_aqp1, int, S_IRUGO);
-module_param_named(debug_level, ehca_debug_level, int, S_IRUGO);
-module_param_named(hw_level, ehca_hw_level, int, S_IRUGO);
-module_param_named(nr_ports, ehca_nr_ports, int, S_IRUGO);
-module_param_named(use_hp_mr, ehca_use_hp_mr, int, S_IRUGO);
-module_param_named(port_act_time, ehca_port_act_time, int, S_IRUGO);
-module_param_named(poll_all_eqs, ehca_poll_all_eqs, int, S_IRUGO);
-module_param_named(static_rate, ehca_static_rate, int, S_IRUGO);
-module_param_named(scaling_code, ehca_scaling_code, int, S_IRUGO);
-module_param_named(mr_largepage, ehca_mr_largepage, int, S_IRUGO);
+module_param_named(open_aqp1, ehca_open_aqp1, bool, S_IRUGO);
+module_param_named(debug_level, ehca_debug_level, int, S_IRUGO);
+module_param_named(hw_level, ehca_hw_level, int, S_IRUGO);
+module_param_named(nr_ports, ehca_nr_ports, int, S_IRUGO);
+module_param_named(use_hp_mr, ehca_use_hp_mr, bool, S_IRUGO);
+module_param_named(port_act_time, ehca_port_act_time, int, S_IRUGO);
+module_param_named(poll_all_eqs, ehca_poll_all_eqs, bool, S_IRUGO);
+module_param_named(static_rate, ehca_static_rate, int, S_IRUGO);
+module_param_named(scaling_code, ehca_scaling_code, bool, S_IRUGO);
module_param_named(lock_hcalls, ehca_lock_hcalls, bool, S_IRUGO);
MODULE_PARM_DESC(open_aqp1,
- "AQP1 on startup (0: no (default), 1: yes)");
+ "Open AQP1 on startup (default: no)");
MODULE_PARM_DESC(debug_level,
- "debug level"
- " (0: no debug traces (default), 1: with debug traces)");
+ "Amount of debug output (0: none (default), 1: traces, "
+ "2: some dumps, 3: lots)");
MODULE_PARM_DESC(hw_level,
- "hardware level"
- " (0: autosensing (default), 1: v. 0.20, 2: v. 0.21)");
+ "Hardware level (0: autosensing (default), "
+ "0x10..0x14: eHCA, 0x20..0x23: eHCA2)");
MODULE_PARM_DESC(nr_ports,
"number of connected ports (-1: autodetect, 1: port one only, "
"2: two ports (default)");
MODULE_PARM_DESC(use_hp_mr,
- "high performance MRs (0: no (default), 1: yes)");
+ "Use high performance MRs (default: no)");
MODULE_PARM_DESC(port_act_time,
- "time to wait for port activation (default: 30 sec)");
+ "Time to wait for port activation (default: 30 sec)");
MODULE_PARM_DESC(poll_all_eqs,
- "polls all event queues periodically"
- " (0: no, 1: yes (default))");
+ "Poll all event queues periodically (default: yes)");
MODULE_PARM_DESC(static_rate,
- "set permanent static rate (default: disabled)");
+ "Set permanent static rate (default: no static rate)");
MODULE_PARM_DESC(scaling_code,
- "set scaling code (0: disabled/default, 1: enabled)");
-MODULE_PARM_DESC(mr_largepage,
- "use large page for MR (0: use PAGE_SIZE (default), "
- "1: use large page depending on MR size");
+ "Enable scaling code (default: no)");
MODULE_PARM_DESC(lock_hcalls,
- "serialize all hCalls made by the driver "
+ "Serialize all hCalls made by the driver "
"(default: autodetect)");
DEFINE_RWLOCK(ehca_qp_idr_lock);
@@ -275,6 +269,7 @@
u64 h_ret;
struct hipz_query_hca *rblock;
struct hipz_query_port *port;
+ const char *loc_code;
static const u32 pgsize_map[] = {
HCA_CAP_MR_PGSIZE_4K, 0x1000,
@@ -283,6 +278,12 @@
HCA_CAP_MR_PGSIZE_16M, 0x1000000,
};
+ ehca_gen_dbg("Probing adapter %s...",
+ shca->ofdev->node->full_name);
+ loc_code = of_get_property(shca->ofdev->node, "ibm,loc-code", NULL);
+ if (loc_code)
+ ehca_gen_dbg(" ... location lode=%s", loc_code);
+
rblock = ehca_alloc_fw_ctrlblock(GFP_KERNEL);
if (!rblock) {
ehca_gen_err("Cannot allocate rblock memory.");
@@ -350,11 +351,9 @@
/* translate supported MR page sizes; always support 4K */
shca->hca_cap_mr_pgsize = EHCA_PAGESIZE;
- if (ehca_mr_largepage) { /* support extra sizes only if enabled */
- for (i = 0; i < ARRAY_SIZE(pgsize_map); i += 2)
- if (rblock->memory_page_size_supported & pgsize_map[i])
- shca->hca_cap_mr_pgsize |= pgsize_map[i + 1];
- }
+ for (i = 0; i < ARRAY_SIZE(pgsize_map); i += 2)
+ if (rblock->memory_page_size_supported & pgsize_map[i])
+ shca->hca_cap_mr_pgsize |= pgsize_map[i + 1];
/* query max MTU from first port -- it's the same for all ports */
port = (struct hipz_query_port *)rblock;
@@ -567,8 +566,7 @@
static ssize_t ehca_show_debug_level(struct device_driver *ddp, char *buf)
{
- return snprintf(buf, PAGE_SIZE, "%d\n",
- ehca_debug_level);
+ return snprintf(buf, PAGE_SIZE, "%d\n", ehca_debug_level);
}
static ssize_t ehca_store_debug_level(struct device_driver *ddp,
@@ -657,14 +655,6 @@
}
static DEVICE_ATTR(adapter_handle, S_IRUGO, ehca_show_adapter_handle, NULL);
-static ssize_t ehca_show_mr_largepage(struct device *dev,
- struct device_attribute *attr,
- char *buf)
-{
- return sprintf(buf, "%d\n", ehca_mr_largepage);
-}
-static DEVICE_ATTR(mr_largepage, S_IRUGO, ehca_show_mr_largepage, NULL);
-
static struct attribute *ehca_dev_attrs[] = {
&dev_attr_adapter_handle.attr,
&dev_attr_num_ports.attr,
@@ -681,7 +671,6 @@
&dev_attr_cur_mw.attr,
&dev_attr_max_pd.attr,
&dev_attr_max_ah.attr,
- &dev_attr_mr_largepage.attr,
NULL
};
diff --git a/drivers/infiniband/hw/ehca/ehca_mrmw.c b/drivers/infiniband/hw/ehca/ehca_mrmw.c
index f26997f..46ae4eb 100644
--- a/drivers/infiniband/hw/ehca/ehca_mrmw.c
+++ b/drivers/infiniband/hw/ehca/ehca_mrmw.c
@@ -1794,8 +1794,9 @@
int t;
for (t = start_idx; t <= end_idx; t++) {
u64 pgaddr = page_to_pfn(sg_page(&page_list[t])) << PAGE_SHIFT;
- ehca_gen_dbg("chunk_page=%lx value=%016lx", pgaddr,
- *(u64 *)abs_to_virt(phys_to_abs(pgaddr)));
+ if (ehca_debug_level >= 3)
+ ehca_gen_dbg("chunk_page=%lx value=%016lx", pgaddr,
+ *(u64 *)abs_to_virt(phys_to_abs(pgaddr)));
if (pgaddr - PAGE_SIZE != *prev_pgaddr) {
ehca_gen_err("uncontiguous page found pgaddr=%lx "
"prev_pgaddr=%lx page_list_i=%x",
@@ -1862,10 +1863,13 @@
pgaddr &
~(pginfo->hwpage_size - 1));
}
- ehca_gen_dbg("kpage=%lx chunk_page=%lx "
- "value=%016lx", *kpage, pgaddr,
- *(u64 *)abs_to_virt(
- phys_to_abs(pgaddr)));
+ if (ehca_debug_level >= 3) {
+ u64 val = *(u64 *)abs_to_virt(
+ phys_to_abs(pgaddr));
+ ehca_gen_dbg("kpage=%lx chunk_page=%lx "
+ "value=%016lx",
+ *kpage, pgaddr, val);
+ }
prev_pgaddr = pgaddr;
i++;
pginfo->kpage_cnt++;
diff --git a/drivers/infiniband/hw/ehca/ehca_qp.c b/drivers/infiniband/hw/ehca/ehca_qp.c
index 3eb14a5..57bef11 100644
--- a/drivers/infiniband/hw/ehca/ehca_qp.c
+++ b/drivers/infiniband/hw/ehca/ehca_qp.c
@@ -550,6 +550,7 @@
spin_lock_init(&my_qp->spinlock_r);
my_qp->qp_type = qp_type;
my_qp->ext_type = parms.ext_type;
+ my_qp->state = IB_QPS_RESET;
if (init_attr->recv_cq)
my_qp->recv_cq =
@@ -965,7 +966,7 @@
qp_num, bad_send_wqe_p);
/* convert wqe pointer to vadr */
bad_send_wqe_v = abs_to_virt((u64)bad_send_wqe_p);
- if (ehca_debug_level)
+ if (ehca_debug_level >= 2)
ehca_dmp(bad_send_wqe_v, 32, "qp_num=%x bad_wqe", qp_num);
squeue = &my_qp->ipz_squeue;
if (ipz_queue_abs_to_offset(squeue, (u64)bad_send_wqe_p, &q_ofs)) {
@@ -978,7 +979,7 @@
wqe = (struct ehca_wqe *)ipz_qeit_calc(squeue, q_ofs);
*bad_wqe_cnt = 0;
while (wqe->optype != 0xff && wqe->wqef != 0xff) {
- if (ehca_debug_level)
+ if (ehca_debug_level >= 2)
ehca_dmp(wqe, 32, "qp_num=%x wqe", qp_num);
wqe->nr_of_data_seg = 0; /* suppress data access */
wqe->wqef = WQEF_PURGE; /* WQE to be purged */
@@ -1450,7 +1451,7 @@
/* no support for max_send/recv_sge yet */
}
- if (ehca_debug_level)
+ if (ehca_debug_level >= 2)
ehca_dmp(mqpcb, 4*70, "qp_num=%x", ibqp->qp_num);
h_ret = hipz_h_modify_qp(shca->ipz_hca_handle,
@@ -1508,6 +1509,8 @@
if (attr_mask & IB_QP_QKEY)
my_qp->qkey = attr->qkey;
+ my_qp->state = qp_new_state;
+
modify_qp_exit2:
if (squeue_locked) { /* this means: sqe -> rts */
spin_unlock_irqrestore(&my_qp->spinlock_s, flags);
@@ -1763,7 +1766,7 @@
if (qp_init_attr)
*qp_init_attr = my_qp->init_attr;
- if (ehca_debug_level)
+ if (ehca_debug_level >= 2)
ehca_dmp(qpcb, 4*70, "qp_num=%x", qp->qp_num);
query_qp_exit1:
@@ -1811,7 +1814,7 @@
goto modify_srq_exit0;
}
- if (ehca_debug_level)
+ if (ehca_debug_level >= 2)
ehca_dmp(mqpcb, 4*70, "qp_num=%x", my_qp->real_qp_num);
h_ret = hipz_h_modify_qp(shca->ipz_hca_handle, my_qp->ipz_qp_handle,
@@ -1864,7 +1867,7 @@
srq_attr->srq_limit = EHCA_BMASK_GET(
MQPCB_CURR_SRQ_LIMIT, qpcb->curr_srq_limit);
- if (ehca_debug_level)
+ if (ehca_debug_level >= 2)
ehca_dmp(qpcb, 4*70, "qp_num=%x", my_qp->real_qp_num);
query_srq_exit1:
diff --git a/drivers/infiniband/hw/ehca/ehca_reqs.c b/drivers/infiniband/hw/ehca/ehca_reqs.c
index a20bbf4..bbe0436 100644
--- a/drivers/infiniband/hw/ehca/ehca_reqs.c
+++ b/drivers/infiniband/hw/ehca/ehca_reqs.c
@@ -81,7 +81,7 @@
recv_wr->sg_list[cnt_ds].length;
}
- if (ehca_debug_level) {
+ if (ehca_debug_level >= 3) {
ehca_gen_dbg("RECEIVE WQE written into ipz_rqueue=%p",
ipz_rqueue);
ehca_dmp(wqe_p, 16*(6 + wqe_p->nr_of_data_seg), "recv wqe");
@@ -281,7 +281,7 @@
return -EINVAL;
}
- if (ehca_debug_level) {
+ if (ehca_debug_level >= 3) {
ehca_gen_dbg("SEND WQE written into queue qp=%p ", qp);
ehca_dmp( wqe_p, 16*(6 + wqe_p->nr_of_data_seg), "send wqe");
}
@@ -421,6 +421,11 @@
int ret = 0;
unsigned long flags;
+ if (unlikely(my_qp->state != IB_QPS_RTS)) {
+ ehca_err(qp->device, "QP not in RTS state qpn=%x", qp->qp_num);
+ return -EINVAL;
+ }
+
/* LOCK the QUEUE */
spin_lock_irqsave(&my_qp->spinlock_s, flags);
@@ -454,13 +459,14 @@
goto post_send_exit0;
}
wqe_cnt++;
- ehca_dbg(qp->device, "ehca_qp=%p qp_num=%x wqe_cnt=%d",
- my_qp, qp->qp_num, wqe_cnt);
} /* eof for cur_send_wr */
post_send_exit0:
iosync(); /* serialize GAL register access */
hipz_update_sqa(my_qp, wqe_cnt);
+ if (unlikely(ret || ehca_debug_level >= 2))
+ ehca_dbg(qp->device, "ehca_qp=%p qp_num=%x wqe_cnt=%d ret=%i",
+ my_qp, qp->qp_num, wqe_cnt, ret);
my_qp->message_count += wqe_cnt;
spin_unlock_irqrestore(&my_qp->spinlock_s, flags);
return ret;
@@ -520,13 +526,14 @@
goto post_recv_exit0;
}
wqe_cnt++;
- ehca_dbg(dev, "ehca_qp=%p qp_num=%x wqe_cnt=%d",
- my_qp, my_qp->real_qp_num, wqe_cnt);
} /* eof for cur_recv_wr */
post_recv_exit0:
iosync(); /* serialize GAL register access */
hipz_update_rqa(my_qp, wqe_cnt);
+ if (unlikely(ret || ehca_debug_level >= 2))
+ ehca_dbg(dev, "ehca_qp=%p qp_num=%x wqe_cnt=%d ret=%i",
+ my_qp, my_qp->real_qp_num, wqe_cnt, ret);
spin_unlock_irqrestore(&my_qp->spinlock_r, flags);
return ret;
}
@@ -570,16 +577,17 @@
struct ehca_cq *my_cq = container_of(cq, struct ehca_cq, ib_cq);
struct ehca_cqe *cqe;
struct ehca_qp *my_qp;
- int cqe_count = 0;
+ int cqe_count = 0, is_error;
poll_cq_one_read_cqe:
cqe = (struct ehca_cqe *)
ipz_qeit_get_inc_valid(&my_cq->ipz_queue);
if (!cqe) {
ret = -EAGAIN;
- ehca_dbg(cq->device, "Completion queue is empty ehca_cq=%p "
- "cq_num=%x ret=%i", my_cq, my_cq->cq_number, ret);
- goto poll_cq_one_exit0;
+ if (ehca_debug_level >= 3)
+ ehca_dbg(cq->device, "Completion queue is empty "
+ "my_cq=%p cq_num=%x", my_cq, my_cq->cq_number);
+ goto poll_cq_one_exit0;
}
/* prevents loads being reordered across this point */
@@ -609,7 +617,7 @@
ehca_dbg(cq->device,
"Got CQE with purged bit qp_num=%x src_qp=%x",
cqe->local_qp_number, cqe->remote_qp_number);
- if (ehca_debug_level)
+ if (ehca_debug_level >= 2)
ehca_dmp(cqe, 64, "qp_num=%x src_qp=%x",
cqe->local_qp_number,
cqe->remote_qp_number);
@@ -622,11 +630,13 @@
}
}
- /* tracing cqe */
- if (unlikely(ehca_debug_level)) {
+ is_error = cqe->status & WC_STATUS_ERROR_BIT;
+
+ /* trace error CQEs if debug_level >= 1, trace all CQEs if >= 3 */
+ if (unlikely(ehca_debug_level >= 3 || (ehca_debug_level && is_error))) {
ehca_dbg(cq->device,
- "Received COMPLETION ehca_cq=%p cq_num=%x -----",
- my_cq, my_cq->cq_number);
+ "Received %sCOMPLETION ehca_cq=%p cq_num=%x -----",
+ is_error ? "ERROR " : "", my_cq, my_cq->cq_number);
ehca_dmp(cqe, 64, "ehca_cq=%p cq_num=%x",
my_cq, my_cq->cq_number);
ehca_dbg(cq->device,
@@ -649,8 +659,9 @@
/* update also queue adder to throw away this entry!!! */
goto poll_cq_one_exit0;
}
+
/* eval ib_wc_status */
- if (unlikely(cqe->status & WC_STATUS_ERROR_BIT)) {
+ if (unlikely(is_error)) {
/* complete with errors */
map_ib_wc_status(cqe->status, &wc->status);
wc->vendor_err = wc->status;
@@ -671,14 +682,6 @@
wc->imm_data = cpu_to_be32(cqe->immediate_data);
wc->sl = cqe->service_level;
- if (unlikely(wc->status != IB_WC_SUCCESS))
- ehca_dbg(cq->device,
- "ehca_cq=%p cq_num=%x WARNING unsuccessful cqe "
- "OPType=%x status=%x qp_num=%x src_qp=%x wr_id=%lx "
- "cqe=%p", my_cq, my_cq->cq_number, cqe->optype,
- cqe->status, cqe->local_qp_number,
- cqe->remote_qp_number, cqe->work_request_id, cqe);
-
poll_cq_one_exit0:
if (cqe_count > 0)
hipz_update_feca(my_cq, cqe_count);
diff --git a/drivers/infiniband/hw/ehca/ehca_uverbs.c b/drivers/infiniband/hw/ehca/ehca_uverbs.c
index 1b07f2b..e43ed8f 100644
--- a/drivers/infiniband/hw/ehca/ehca_uverbs.c
+++ b/drivers/infiniband/hw/ehca/ehca_uverbs.c
@@ -211,8 +211,7 @@
break;
case 1: /* qp rqueue_addr */
- ehca_dbg(qp->ib_qp.device, "qp_num=%x rqueue",
- qp->ib_qp.qp_num);
+ ehca_dbg(qp->ib_qp.device, "qp_num=%x rq", qp->ib_qp.qp_num);
ret = ehca_mmap_queue(vma, &qp->ipz_rqueue,
&qp->mm_count_rqueue);
if (unlikely(ret)) {
@@ -224,8 +223,7 @@
break;
case 2: /* qp squeue_addr */
- ehca_dbg(qp->ib_qp.device, "qp_num=%x squeue",
- qp->ib_qp.qp_num);
+ ehca_dbg(qp->ib_qp.device, "qp_num=%x sq", qp->ib_qp.qp_num);
ret = ehca_mmap_queue(vma, &qp->ipz_squeue,
&qp->mm_count_squeue);
if (unlikely(ret)) {
diff --git a/drivers/infiniband/hw/ehca/hcp_if.c b/drivers/infiniband/hw/ehca/hcp_if.c
index 7029aa6..5245e13 100644
--- a/drivers/infiniband/hw/ehca/hcp_if.c
+++ b/drivers/infiniband/hw/ehca/hcp_if.c
@@ -123,8 +123,9 @@
int i, sleep_msecs;
unsigned long flags = 0;
- ehca_gen_dbg("opcode=%lx " HCALL7_REGS_FORMAT,
- opcode, arg1, arg2, arg3, arg4, arg5, arg6, arg7);
+ if (unlikely(ehca_debug_level >= 2))
+ ehca_gen_dbg("opcode=%lx " HCALL7_REGS_FORMAT,
+ opcode, arg1, arg2, arg3, arg4, arg5, arg6, arg7);
for (i = 0; i < 5; i++) {
/* serialize hCalls to work around firmware issue */
@@ -148,7 +149,8 @@
opcode, ret, arg1, arg2, arg3,
arg4, arg5, arg6, arg7);
else
- ehca_gen_dbg("opcode=%lx ret=%li", opcode, ret);
+ if (unlikely(ehca_debug_level >= 2))
+ ehca_gen_dbg("opcode=%lx ret=%li", opcode, ret);
return ret;
}
@@ -172,8 +174,10 @@
int i, sleep_msecs;
unsigned long flags = 0;
- ehca_gen_dbg("INPUT -- opcode=%lx " HCALL9_REGS_FORMAT, opcode,
- arg1, arg2, arg3, arg4, arg5, arg6, arg7, arg8, arg9);
+ if (unlikely(ehca_debug_level >= 2))
+ ehca_gen_dbg("INPUT -- opcode=%lx " HCALL9_REGS_FORMAT, opcode,
+ arg1, arg2, arg3, arg4, arg5,
+ arg6, arg7, arg8, arg9);
for (i = 0; i < 5; i++) {
/* serialize hCalls to work around firmware issue */
@@ -201,7 +205,7 @@
ret, outs[0], outs[1], outs[2], outs[3],
outs[4], outs[5], outs[6], outs[7],
outs[8]);
- } else
+ } else if (unlikely(ehca_debug_level >= 2))
ehca_gen_dbg("OUTPUT -- ret=%li " HCALL9_REGS_FORMAT,
ret, outs[0], outs[1], outs[2], outs[3],
outs[4], outs[5], outs[6], outs[7],
@@ -381,7 +385,7 @@
r_cb, /* r6 */
0, 0, 0, 0);
- if (ehca_debug_level)
+ if (ehca_debug_level >= 2)
ehca_dmp(query_port_response_block, 64, "response_block");
return ret;
@@ -731,9 +735,6 @@
u64 ret;
u64 outs[PLPAR_HCALL9_BUFSIZE];
- ehca_gen_dbg("kernel PAGE_SIZE=%x access_ctrl=%016x "
- "vaddr=%lx length=%lx",
- (u32)PAGE_SIZE, access_ctrl, vaddr, length);
ret = ehca_plpar_hcall9(H_ALLOC_RESOURCE, outs,
adapter_handle.handle, /* r4 */
5, /* r5 */
@@ -758,7 +759,7 @@
{
u64 ret;
- if (unlikely(ehca_debug_level >= 2)) {
+ if (unlikely(ehca_debug_level >= 3)) {
if (count > 1) {
u64 *kpage;
int i;
diff --git a/drivers/infiniband/hw/mlx4/cq.c b/drivers/infiniband/hw/mlx4/cq.c
index 3557e7e..5e570bb 100644
--- a/drivers/infiniband/hw/mlx4/cq.c
+++ b/drivers/infiniband/hw/mlx4/cq.c
@@ -204,7 +204,7 @@
uar = &to_mucontext(context)->uar;
} else {
- err = mlx4_ib_db_alloc(dev, &cq->db, 1);
+ err = mlx4_db_alloc(dev->dev, &cq->db, 1);
if (err)
goto err_cq;
@@ -250,7 +250,7 @@
err_db:
if (!context)
- mlx4_ib_db_free(dev, &cq->db);
+ mlx4_db_free(dev->dev, &cq->db);
err_cq:
kfree(cq);
@@ -435,7 +435,7 @@
ib_umem_release(mcq->umem);
} else {
mlx4_ib_free_cq_buf(dev, &mcq->buf, cq->cqe + 1);
- mlx4_ib_db_free(dev, &mcq->db);
+ mlx4_db_free(dev->dev, &mcq->db);
}
kfree(mcq);
diff --git a/drivers/infiniband/hw/mlx4/doorbell.c b/drivers/infiniband/hw/mlx4/doorbell.c
index 1c36087..8e342cc 100644
--- a/drivers/infiniband/hw/mlx4/doorbell.c
+++ b/drivers/infiniband/hw/mlx4/doorbell.c
@@ -34,124 +34,6 @@
#include "mlx4_ib.h"
-struct mlx4_ib_db_pgdir {
- struct list_head list;
- DECLARE_BITMAP(order0, MLX4_IB_DB_PER_PAGE);
- DECLARE_BITMAP(order1, MLX4_IB_DB_PER_PAGE / 2);
- unsigned long *bits[2];
- __be32 *db_page;
- dma_addr_t db_dma;
-};
-
-static struct mlx4_ib_db_pgdir *mlx4_ib_alloc_db_pgdir(struct mlx4_ib_dev *dev)
-{
- struct mlx4_ib_db_pgdir *pgdir;
-
- pgdir = kzalloc(sizeof *pgdir, GFP_KERNEL);
- if (!pgdir)
- return NULL;
-
- bitmap_fill(pgdir->order1, MLX4_IB_DB_PER_PAGE / 2);
- pgdir->bits[0] = pgdir->order0;
- pgdir->bits[1] = pgdir->order1;
- pgdir->db_page = dma_alloc_coherent(dev->ib_dev.dma_device,
- PAGE_SIZE, &pgdir->db_dma,
- GFP_KERNEL);
- if (!pgdir->db_page) {
- kfree(pgdir);
- return NULL;
- }
-
- return pgdir;
-}
-
-static int mlx4_ib_alloc_db_from_pgdir(struct mlx4_ib_db_pgdir *pgdir,
- struct mlx4_ib_db *db, int order)
-{
- int o;
- int i;
-
- for (o = order; o <= 1; ++o) {
- i = find_first_bit(pgdir->bits[o], MLX4_IB_DB_PER_PAGE >> o);
- if (i < MLX4_IB_DB_PER_PAGE >> o)
- goto found;
- }
-
- return -ENOMEM;
-
-found:
- clear_bit(i, pgdir->bits[o]);
-
- i <<= o;
-
- if (o > order)
- set_bit(i ^ 1, pgdir->bits[order]);
-
- db->u.pgdir = pgdir;
- db->index = i;
- db->db = pgdir->db_page + db->index;
- db->dma = pgdir->db_dma + db->index * 4;
- db->order = order;
-
- return 0;
-}
-
-int mlx4_ib_db_alloc(struct mlx4_ib_dev *dev, struct mlx4_ib_db *db, int order)
-{
- struct mlx4_ib_db_pgdir *pgdir;
- int ret = 0;
-
- mutex_lock(&dev->pgdir_mutex);
-
- list_for_each_entry(pgdir, &dev->pgdir_list, list)
- if (!mlx4_ib_alloc_db_from_pgdir(pgdir, db, order))
- goto out;
-
- pgdir = mlx4_ib_alloc_db_pgdir(dev);
- if (!pgdir) {
- ret = -ENOMEM;
- goto out;
- }
-
- list_add(&pgdir->list, &dev->pgdir_list);
-
- /* This should never fail -- we just allocated an empty page: */
- WARN_ON(mlx4_ib_alloc_db_from_pgdir(pgdir, db, order));
-
-out:
- mutex_unlock(&dev->pgdir_mutex);
-
- return ret;
-}
-
-void mlx4_ib_db_free(struct mlx4_ib_dev *dev, struct mlx4_ib_db *db)
-{
- int o;
- int i;
-
- mutex_lock(&dev->pgdir_mutex);
-
- o = db->order;
- i = db->index;
-
- if (db->order == 0 && test_bit(i ^ 1, db->u.pgdir->order0)) {
- clear_bit(i ^ 1, db->u.pgdir->order0);
- ++o;
- }
-
- i >>= o;
- set_bit(i, db->u.pgdir->bits[o]);
-
- if (bitmap_full(db->u.pgdir->order1, MLX4_IB_DB_PER_PAGE / 2)) {
- dma_free_coherent(dev->ib_dev.dma_device, PAGE_SIZE,
- db->u.pgdir->db_page, db->u.pgdir->db_dma);
- list_del(&db->u.pgdir->list);
- kfree(db->u.pgdir);
- }
-
- mutex_unlock(&dev->pgdir_mutex);
-}
-
struct mlx4_ib_user_db_page {
struct list_head list;
struct ib_umem *umem;
@@ -160,7 +42,7 @@
};
int mlx4_ib_db_map_user(struct mlx4_ib_ucontext *context, unsigned long virt,
- struct mlx4_ib_db *db)
+ struct mlx4_db *db)
{
struct mlx4_ib_user_db_page *page;
struct ib_umem_chunk *chunk;
@@ -202,7 +84,7 @@
return err;
}
-void mlx4_ib_db_unmap_user(struct mlx4_ib_ucontext *context, struct mlx4_ib_db *db)
+void mlx4_ib_db_unmap_user(struct mlx4_ib_ucontext *context, struct mlx4_db *db)
{
mutex_lock(&context->db_page_mutex);
diff --git a/drivers/infiniband/hw/mlx4/main.c b/drivers/infiniband/hw/mlx4/main.c
index 4d9b5ac..4d61e32 100644
--- a/drivers/infiniband/hw/mlx4/main.c
+++ b/drivers/infiniband/hw/mlx4/main.c
@@ -557,9 +557,6 @@
goto err_uar;
MLX4_INIT_DOORBELL_LOCK(&ibdev->uar_lock);
- INIT_LIST_HEAD(&ibdev->pgdir_list);
- mutex_init(&ibdev->pgdir_mutex);
-
ibdev->dev = dev;
strlcpy(ibdev->ib_dev.name, "mlx4_%d", IB_DEVICE_NAME_MAX);
diff --git a/drivers/infiniband/hw/mlx4/mlx4_ib.h b/drivers/infiniband/hw/mlx4/mlx4_ib.h
index 9e63732..5cf9947 100644
--- a/drivers/infiniband/hw/mlx4/mlx4_ib.h
+++ b/drivers/infiniband/hw/mlx4/mlx4_ib.h
@@ -43,24 +43,6 @@
#include <linux/mlx4/device.h>
#include <linux/mlx4/doorbell.h>
-enum {
- MLX4_IB_DB_PER_PAGE = PAGE_SIZE / 4
-};
-
-struct mlx4_ib_db_pgdir;
-struct mlx4_ib_user_db_page;
-
-struct mlx4_ib_db {
- __be32 *db;
- union {
- struct mlx4_ib_db_pgdir *pgdir;
- struct mlx4_ib_user_db_page *user_page;
- } u;
- dma_addr_t dma;
- int index;
- int order;
-};
-
struct mlx4_ib_ucontext {
struct ib_ucontext ibucontext;
struct mlx4_uar uar;
@@ -88,7 +70,7 @@
struct mlx4_cq mcq;
struct mlx4_ib_cq_buf buf;
struct mlx4_ib_cq_resize *resize_buf;
- struct mlx4_ib_db db;
+ struct mlx4_db db;
spinlock_t lock;
struct mutex resize_mutex;
struct ib_umem *umem;
@@ -127,7 +109,7 @@
struct mlx4_qp mqp;
struct mlx4_buf buf;
- struct mlx4_ib_db db;
+ struct mlx4_db db;
struct mlx4_ib_wq rq;
u32 doorbell_qpn;
@@ -154,7 +136,7 @@
struct ib_srq ibsrq;
struct mlx4_srq msrq;
struct mlx4_buf buf;
- struct mlx4_ib_db db;
+ struct mlx4_db db;
u64 *wrid;
spinlock_t lock;
int head;
@@ -175,9 +157,6 @@
struct mlx4_dev *dev;
void __iomem *uar_map;
- struct list_head pgdir_list;
- struct mutex pgdir_mutex;
-
struct mlx4_uar priv_uar;
u32 priv_pdn;
MLX4_DECLARE_DOORBELL_LOCK(uar_lock);
@@ -248,11 +227,9 @@
return container_of(ibah, struct mlx4_ib_ah, ibah);
}
-int mlx4_ib_db_alloc(struct mlx4_ib_dev *dev, struct mlx4_ib_db *db, int order);
-void mlx4_ib_db_free(struct mlx4_ib_dev *dev, struct mlx4_ib_db *db);
int mlx4_ib_db_map_user(struct mlx4_ib_ucontext *context, unsigned long virt,
- struct mlx4_ib_db *db);
-void mlx4_ib_db_unmap_user(struct mlx4_ib_ucontext *context, struct mlx4_ib_db *db);
+ struct mlx4_db *db);
+void mlx4_ib_db_unmap_user(struct mlx4_ib_ucontext *context, struct mlx4_db *db);
struct ib_mr *mlx4_ib_get_dma_mr(struct ib_pd *pd, int acc);
int mlx4_ib_umem_write_mtt(struct mlx4_ib_dev *dev, struct mlx4_mtt *mtt,
diff --git a/drivers/infiniband/hw/mlx4/qp.c b/drivers/infiniband/hw/mlx4/qp.c
index b75efae..80ea8b9 100644
--- a/drivers/infiniband/hw/mlx4/qp.c
+++ b/drivers/infiniband/hw/mlx4/qp.c
@@ -514,7 +514,7 @@
goto err;
if (!init_attr->srq) {
- err = mlx4_ib_db_alloc(dev, &qp->db, 0);
+ err = mlx4_db_alloc(dev->dev, &qp->db, 0);
if (err)
goto err;
@@ -580,7 +580,7 @@
err_db:
if (!pd->uobject && !init_attr->srq)
- mlx4_ib_db_free(dev, &qp->db);
+ mlx4_db_free(dev->dev, &qp->db);
err:
return err;
@@ -666,7 +666,7 @@
kfree(qp->rq.wrid);
mlx4_buf_free(dev->dev, qp->buf_size, &qp->buf);
if (!qp->ibqp.srq)
- mlx4_ib_db_free(dev, &qp->db);
+ mlx4_db_free(dev->dev, &qp->db);
}
}
diff --git a/drivers/infiniband/hw/mlx4/srq.c b/drivers/infiniband/hw/mlx4/srq.c
index beaa3b0..2046197 100644
--- a/drivers/infiniband/hw/mlx4/srq.c
+++ b/drivers/infiniband/hw/mlx4/srq.c
@@ -129,7 +129,7 @@
if (err)
goto err_mtt;
} else {
- err = mlx4_ib_db_alloc(dev, &srq->db, 0);
+ err = mlx4_db_alloc(dev->dev, &srq->db, 0);
if (err)
goto err_srq;
@@ -200,7 +200,7 @@
err_db:
if (!pd->uobject)
- mlx4_ib_db_free(dev, &srq->db);
+ mlx4_db_free(dev->dev, &srq->db);
err_srq:
kfree(srq);
@@ -267,7 +267,7 @@
kfree(msrq->wrid);
mlx4_buf_free(dev->dev, msrq->msrq.max << msrq->msrq.wqe_shift,
&msrq->buf);
- mlx4_ib_db_free(dev, &msrq->db);
+ mlx4_db_free(dev->dev, &msrq->db);
}
kfree(msrq);
diff --git a/drivers/infiniband/hw/nes/nes.c b/drivers/infiniband/hw/nes/nes.c
index b046262..a4e9269 100644
--- a/drivers/infiniband/hw/nes/nes.c
+++ b/drivers/infiniband/hw/nes/nes.c
@@ -139,8 +139,9 @@
addr = ntohl(ifa->ifa_address);
mask = ntohl(ifa->ifa_mask);
- nes_debug(NES_DBG_NETDEV, "nes_inetaddr_event: ip address %08X, netmask %08X.\n",
- addr, mask);
+ nes_debug(NES_DBG_NETDEV, "nes_inetaddr_event: ip address " NIPQUAD_FMT
+ ", netmask " NIPQUAD_FMT ".\n",
+ HIPQUAD(addr), HIPQUAD(mask));
list_for_each_entry(nesdev, &nes_dev_list, list) {
nes_debug(NES_DBG_NETDEV, "Nesdev list entry = 0x%p. (%s)\n",
nesdev, nesdev->netdev[0]->name);
@@ -353,13 +354,11 @@
*/
static void nes_print_macaddr(struct net_device *netdev)
{
- nes_debug(NES_DBG_INIT, "%s: MAC %02X:%02X:%02X:%02X:%02X:%02X, IRQ %u\n",
- netdev->name,
- netdev->dev_addr[0], netdev->dev_addr[1], netdev->dev_addr[2],
- netdev->dev_addr[3], netdev->dev_addr[4], netdev->dev_addr[5],
- netdev->irq);
-}
+ DECLARE_MAC_BUF(mac);
+ nes_debug(NES_DBG_INIT, "%s: %s, IRQ %u\n",
+ netdev->name, print_mac(mac, netdev->dev_addr), netdev->irq);
+}
/**
* nes_interrupt - handle interrupts
diff --git a/drivers/infiniband/hw/nes/nes_cm.c b/drivers/infiniband/hw/nes/nes_cm.c
index d073862..d940fc2 100644
--- a/drivers/infiniband/hw/nes/nes_cm.c
+++ b/drivers/infiniband/hw/nes/nes_cm.c
@@ -852,8 +852,8 @@
/* get a handle on the hte */
hte = &cm_core->connected_nodes;
- nes_debug(NES_DBG_CM, "Searching for an owner node:%x:%x from core %p->%p\n",
- loc_addr, loc_port, cm_core, hte);
+ nes_debug(NES_DBG_CM, "Searching for an owner node: " NIPQUAD_FMT ":%x from core %p->%p\n",
+ HIPQUAD(loc_addr), loc_port, cm_core, hte);
/* walk list and find cm_node associated with this session ID */
spin_lock_irqsave(&cm_core->ht_lock, flags);
@@ -902,8 +902,8 @@
}
spin_unlock_irqrestore(&cm_core->listen_list_lock, flags);
- nes_debug(NES_DBG_CM, "Unable to find listener- %x:%x\n",
- dst_addr, dst_port);
+ nes_debug(NES_DBG_CM, "Unable to find listener for " NIPQUAD_FMT ":%x\n",
+ HIPQUAD(dst_addr), dst_port);
/* no listener */
return NULL;
@@ -1054,6 +1054,7 @@
int arpindex = 0;
struct nes_device *nesdev;
struct nes_adapter *nesadapter;
+ DECLARE_MAC_BUF(mac);
/* create an hte and cm_node for this instance */
cm_node = kzalloc(sizeof(*cm_node), GFP_ATOMIC);
@@ -1066,8 +1067,9 @@
cm_node->loc_port = cm_info->loc_port;
cm_node->rem_port = cm_info->rem_port;
cm_node->send_write0 = send_first;
- nes_debug(NES_DBG_CM, "Make node addresses : loc = %x:%x, rem = %x:%x\n",
- cm_node->loc_addr, cm_node->loc_port, cm_node->rem_addr, cm_node->rem_port);
+ nes_debug(NES_DBG_CM, "Make node addresses : loc = " NIPQUAD_FMT ":%x, rem = " NIPQUAD_FMT ":%x\n",
+ HIPQUAD(cm_node->loc_addr), cm_node->loc_port,
+ HIPQUAD(cm_node->rem_addr), cm_node->rem_port);
cm_node->listener = listener;
cm_node->netdev = nesvnic->netdev;
cm_node->cm_id = cm_info->cm_id;
@@ -1116,11 +1118,8 @@
/* copy the mac addr to node context */
memcpy(cm_node->rem_mac, nesadapter->arp_table[arpindex].mac_addr, ETH_ALEN);
- nes_debug(NES_DBG_CM, "Remote mac addr from arp table:%02x,"
- " %02x, %02x, %02x, %02x, %02x\n",
- cm_node->rem_mac[0], cm_node->rem_mac[1],
- cm_node->rem_mac[2], cm_node->rem_mac[3],
- cm_node->rem_mac[4], cm_node->rem_mac[5]);
+ nes_debug(NES_DBG_CM, "Remote mac addr from arp table: %s\n",
+ print_mac(mac, cm_node->rem_mac));
add_hte_node(cm_core, cm_node);
atomic_inc(&cm_nodes_created);
@@ -1850,8 +1849,10 @@
nfo.rem_addr = ntohl(iph->saddr);
nfo.rem_port = ntohs(tcph->source);
- nes_debug(NES_DBG_CM, "Received packet: dest=0x%08X:0x%04X src=0x%08X:0x%04X\n",
- iph->daddr, tcph->dest, iph->saddr, tcph->source);
+ nes_debug(NES_DBG_CM, "Received packet: dest=" NIPQUAD_FMT
+ ":0x%04X src=" NIPQUAD_FMT ":0x%04X\n",
+ NIPQUAD(iph->daddr), tcph->dest,
+ NIPQUAD(iph->saddr), tcph->source);
/* note: this call is going to increment cm_node ref count */
cm_node = find_node(cm_core,
diff --git a/drivers/infiniband/hw/nes/nes_hw.c b/drivers/infiniband/hw/nes/nes_hw.c
index aa53aab..08964cc 100644
--- a/drivers/infiniband/hw/nes/nes_hw.c
+++ b/drivers/infiniband/hw/nes/nes_hw.c
@@ -636,6 +636,15 @@
nes_debug(NES_DBG_INIT, "Did not see full soft reset done.\n");
return 0;
}
+
+ i = 0;
+ while ((nes_read_indexed(nesdev, NES_IDX_INT_CPU_STATUS) != 0x80) && i++ < 10000)
+ mdelay(1);
+ if (i >= 10000) {
+ printk(KERN_ERR PFX "Internal CPU not ready, status = %02X\n",
+ nes_read_indexed(nesdev, NES_IDX_INT_CPU_STATUS));
+ return 0;
+ }
}
/* port reset */
@@ -684,17 +693,6 @@
}
}
-
-
- i = 0;
- while ((nes_read_indexed(nesdev, NES_IDX_INT_CPU_STATUS) != 0x80) && i++ < 10000)
- mdelay(1);
- if (i >= 10000) {
- printk(KERN_ERR PFX "Internal CPU not ready, status = %02X\n",
- nes_read_indexed(nesdev, NES_IDX_INT_CPU_STATUS));
- return 0;
- }
-
return port_count;
}
diff --git a/drivers/infiniband/hw/nes/nes_hw.h b/drivers/infiniband/hw/nes/nes_hw.h
index b7e2844..8f36e23 100644
--- a/drivers/infiniband/hw/nes/nes_hw.h
+++ b/drivers/infiniband/hw/nes/nes_hw.h
@@ -905,7 +905,7 @@
};
struct nes_hw_cq {
- struct nes_hw_cqe volatile *cq_vbase; /* PCI memory for host rings */
+ struct nes_hw_cqe *cq_vbase; /* PCI memory for host rings */
void (*ce_handler)(struct nes_device *nesdev, struct nes_hw_cq *cq);
dma_addr_t cq_pbase; /* PCI memory for host rings */
u16 cq_head;
diff --git a/drivers/infiniband/hw/nes/nes_nic.c b/drivers/infiniband/hw/nes/nes_nic.c
index 01cd0ef..e5366b0 100644
--- a/drivers/infiniband/hw/nes/nes_nic.c
+++ b/drivers/infiniband/hw/nes/nes_nic.c
@@ -787,16 +787,14 @@
int i;
u32 macaddr_low;
u16 macaddr_high;
+ DECLARE_MAC_BUF(mac);
if (!is_valid_ether_addr(mac_addr->sa_data))
return -EADDRNOTAVAIL;
memcpy(netdev->dev_addr, mac_addr->sa_data, netdev->addr_len);
- printk(PFX "%s: Address length = %d, Address = %02X%02X%02X%02X%02X%02X..\n",
- __func__, netdev->addr_len,
- mac_addr->sa_data[0], mac_addr->sa_data[1],
- mac_addr->sa_data[2], mac_addr->sa_data[3],
- mac_addr->sa_data[4], mac_addr->sa_data[5]);
+ printk(PFX "%s: Address length = %d, Address = %s\n",
+ __func__, netdev->addr_len, print_mac(mac, mac_addr->sa_data));
macaddr_high = ((u16)netdev->dev_addr[0]) << 8;
macaddr_high += (u16)netdev->dev_addr[1];
macaddr_low = ((u32)netdev->dev_addr[2]) << 24;
@@ -878,11 +876,11 @@
if (mc_nic_index < 0)
mc_nic_index = nesvnic->nic_index;
if (multicast_addr) {
- nes_debug(NES_DBG_NIC_RX, "Assigning MC Address = %02X%02X%02X%02X%02X%02X to register 0x%04X nic_idx=%d\n",
- multicast_addr->dmi_addr[0], multicast_addr->dmi_addr[1],
- multicast_addr->dmi_addr[2], multicast_addr->dmi_addr[3],
- multicast_addr->dmi_addr[4], multicast_addr->dmi_addr[5],
- perfect_filter_register_address+(mc_index * 8), mc_nic_index);
+ DECLARE_MAC_BUF(mac);
+ nes_debug(NES_DBG_NIC_RX, "Assigning MC Address %s to register 0x%04X nic_idx=%d\n",
+ print_mac(mac, multicast_addr->dmi_addr),
+ perfect_filter_register_address+(mc_index * 8),
+ mc_nic_index);
macaddr_high = ((u16)multicast_addr->dmi_addr[0]) << 8;
macaddr_high += (u16)multicast_addr->dmi_addr[1];
macaddr_low = ((u32)multicast_addr->dmi_addr[2]) << 24;
diff --git a/drivers/infiniband/hw/nes/nes_utils.c b/drivers/infiniband/hw/nes/nes_utils.c
index f9db07c..c6d5631 100644
--- a/drivers/infiniband/hw/nes/nes_utils.c
+++ b/drivers/infiniband/hw/nes/nes_utils.c
@@ -660,7 +660,9 @@
/* DELETE or RESOLVE */
if (arp_index == nesadapter->arp_table_size) {
- nes_debug(NES_DBG_NETDEV, "mac address not in ARP table - cannot delete or resolve\n");
+ nes_debug(NES_DBG_NETDEV, "MAC for " NIPQUAD_FMT " not in ARP table - cannot %s\n",
+ HIPQUAD(ip_addr),
+ action == NES_ARP_RESOLVE ? "resolve" : "delete");
return -1;
}
diff --git a/drivers/infiniband/hw/nes/nes_verbs.c b/drivers/infiniband/hw/nes/nes_verbs.c
index f9a5d43..ee74f7c 100644
--- a/drivers/infiniband/hw/nes/nes_verbs.c
+++ b/drivers/infiniband/hw/nes/nes_verbs.c
@@ -1976,7 +1976,7 @@
if (nescq->cq_mem_size)
pci_free_consistent(nesdev->pcidev, nescq->cq_mem_size,
- (void *)nescq->hw_cq.cq_vbase, nescq->hw_cq.cq_pbase);
+ nescq->hw_cq.cq_vbase, nescq->hw_cq.cq_pbase);
kfree(nescq);
return ret;
@@ -3610,6 +3610,12 @@
while (cqe_count < num_entries) {
if (le32_to_cpu(nescq->hw_cq.cq_vbase[head].cqe_words[NES_CQE_OPCODE_IDX]) &
NES_CQE_VALID) {
+ /*
+ * Make sure we read CQ entry contents *after*
+ * we've checked the valid bit.
+ */
+ rmb();
+
cqe = nescq->hw_cq.cq_vbase[head];
nescq->hw_cq.cq_vbase[head].cqe_words[NES_CQE_OPCODE_IDX] = 0;
u32temp = le32_to_cpu(cqe.cqe_words[NES_CQE_COMP_COMP_CTX_LOW_IDX]);
diff --git a/drivers/infiniband/ulp/ipoib/ipoib.h b/drivers/infiniband/ulp/ipoib/ipoib.h
index 73b2b17..f1f142d 100644
--- a/drivers/infiniband/ulp/ipoib/ipoib.h
+++ b/drivers/infiniband/ulp/ipoib/ipoib.h
@@ -56,11 +56,11 @@
/* constants */
enum {
- IPOIB_PACKET_SIZE = 2048,
- IPOIB_BUF_SIZE = IPOIB_PACKET_SIZE + IB_GRH_BYTES,
-
IPOIB_ENCAP_LEN = 4,
+ IPOIB_UD_HEAD_SIZE = IB_GRH_BYTES + IPOIB_ENCAP_LEN,
+ IPOIB_UD_RX_SG = 2, /* max buffer needed for 4K mtu */
+
IPOIB_CM_MTU = 0x10000 - 0x10, /* padding to align header to 16 */
IPOIB_CM_BUF_SIZE = IPOIB_CM_MTU + IPOIB_ENCAP_LEN,
IPOIB_CM_HEAD_SIZE = IPOIB_CM_BUF_SIZE % PAGE_SIZE,
@@ -139,7 +139,7 @@
struct ipoib_rx_buf {
struct sk_buff *skb;
- u64 mapping;
+ u64 mapping[IPOIB_UD_RX_SG];
};
struct ipoib_tx_buf {
@@ -294,6 +294,7 @@
unsigned int admin_mtu;
unsigned int mcast_mtu;
+ unsigned int max_ib_mtu;
struct ipoib_rx_buf *rx_ring;
@@ -305,6 +306,9 @@
struct ib_send_wr tx_wr;
unsigned tx_outstanding;
+ struct ib_recv_wr rx_wr;
+ struct ib_sge rx_sge[IPOIB_UD_RX_SG];
+
struct ib_wc ibwc[IPOIB_NUM_WC];
struct list_head dead_ahs;
@@ -366,6 +370,14 @@
struct list_head list;
};
+#define IPOIB_UD_MTU(ib_mtu) (ib_mtu - IPOIB_ENCAP_LEN)
+#define IPOIB_UD_BUF_SIZE(ib_mtu) (ib_mtu + IB_GRH_BYTES)
+
+static inline int ipoib_ud_need_sg(unsigned int ib_mtu)
+{
+ return IPOIB_UD_BUF_SIZE(ib_mtu) > PAGE_SIZE;
+}
+
/*
* We stash a pointer to our private neighbour information after our
* hardware address in neigh->ha. The ALIGN() expression here makes
diff --git a/drivers/infiniband/ulp/ipoib/ipoib_ib.c b/drivers/infiniband/ulp/ipoib/ipoib_ib.c
index 0205eb7..7cf1fa7 100644
--- a/drivers/infiniband/ulp/ipoib/ipoib_ib.c
+++ b/drivers/infiniband/ulp/ipoib/ipoib_ib.c
@@ -89,28 +89,59 @@
spin_unlock_irqrestore(&priv->lock, flags);
}
+static void ipoib_ud_dma_unmap_rx(struct ipoib_dev_priv *priv,
+ u64 mapping[IPOIB_UD_RX_SG])
+{
+ if (ipoib_ud_need_sg(priv->max_ib_mtu)) {
+ ib_dma_unmap_single(priv->ca, mapping[0], IPOIB_UD_HEAD_SIZE,
+ DMA_FROM_DEVICE);
+ ib_dma_unmap_page(priv->ca, mapping[1], PAGE_SIZE,
+ DMA_FROM_DEVICE);
+ } else
+ ib_dma_unmap_single(priv->ca, mapping[0],
+ IPOIB_UD_BUF_SIZE(priv->max_ib_mtu),
+ DMA_FROM_DEVICE);
+}
+
+static void ipoib_ud_skb_put_frags(struct ipoib_dev_priv *priv,
+ struct sk_buff *skb,
+ unsigned int length)
+{
+ if (ipoib_ud_need_sg(priv->max_ib_mtu)) {
+ skb_frag_t *frag = &skb_shinfo(skb)->frags[0];
+ unsigned int size;
+ /*
+ * There is only two buffers needed for max_payload = 4K,
+ * first buf size is IPOIB_UD_HEAD_SIZE
+ */
+ skb->tail += IPOIB_UD_HEAD_SIZE;
+ skb->len += length;
+
+ size = length - IPOIB_UD_HEAD_SIZE;
+
+ frag->size = size;
+ skb->data_len += size;
+ skb->truesize += size;
+ } else
+ skb_put(skb, length);
+
+}
+
static int ipoib_ib_post_receive(struct net_device *dev, int id)
{
struct ipoib_dev_priv *priv = netdev_priv(dev);
- struct ib_sge list;
- struct ib_recv_wr param;
struct ib_recv_wr *bad_wr;
int ret;
- list.addr = priv->rx_ring[id].mapping;
- list.length = IPOIB_BUF_SIZE;
- list.lkey = priv->mr->lkey;
+ priv->rx_wr.wr_id = id | IPOIB_OP_RECV;
+ priv->rx_sge[0].addr = priv->rx_ring[id].mapping[0];
+ priv->rx_sge[1].addr = priv->rx_ring[id].mapping[1];
- param.next = NULL;
- param.wr_id = id | IPOIB_OP_RECV;
- param.sg_list = &list;
- param.num_sge = 1;
- ret = ib_post_recv(priv->qp, ¶m, &bad_wr);
+ ret = ib_post_recv(priv->qp, &priv->rx_wr, &bad_wr);
if (unlikely(ret)) {
ipoib_warn(priv, "receive failed for buf %d (%d)\n", id, ret);
- ib_dma_unmap_single(priv->ca, priv->rx_ring[id].mapping,
- IPOIB_BUF_SIZE, DMA_FROM_DEVICE);
+ ipoib_ud_dma_unmap_rx(priv, priv->rx_ring[id].mapping);
dev_kfree_skb_any(priv->rx_ring[id].skb);
priv->rx_ring[id].skb = NULL;
}
@@ -118,15 +149,21 @@
return ret;
}
-static int ipoib_alloc_rx_skb(struct net_device *dev, int id)
+static struct sk_buff *ipoib_alloc_rx_skb(struct net_device *dev, int id)
{
struct ipoib_dev_priv *priv = netdev_priv(dev);
struct sk_buff *skb;
- u64 addr;
+ int buf_size;
+ u64 *mapping;
- skb = dev_alloc_skb(IPOIB_BUF_SIZE + 4);
- if (!skb)
- return -ENOMEM;
+ if (ipoib_ud_need_sg(priv->max_ib_mtu))
+ buf_size = IPOIB_UD_HEAD_SIZE;
+ else
+ buf_size = IPOIB_UD_BUF_SIZE(priv->max_ib_mtu);
+
+ skb = dev_alloc_skb(buf_size + 4);
+ if (unlikely(!skb))
+ return NULL;
/*
* IB will leave a 40 byte gap for a GRH and IPoIB adds a 4 byte
@@ -135,17 +172,32 @@
*/
skb_reserve(skb, 4);
- addr = ib_dma_map_single(priv->ca, skb->data, IPOIB_BUF_SIZE,
- DMA_FROM_DEVICE);
- if (unlikely(ib_dma_mapping_error(priv->ca, addr))) {
- dev_kfree_skb_any(skb);
- return -EIO;
+ mapping = priv->rx_ring[id].mapping;
+ mapping[0] = ib_dma_map_single(priv->ca, skb->data, buf_size,
+ DMA_FROM_DEVICE);
+ if (unlikely(ib_dma_mapping_error(priv->ca, mapping[0])))
+ goto error;
+
+ if (ipoib_ud_need_sg(priv->max_ib_mtu)) {
+ struct page *page = alloc_page(GFP_ATOMIC);
+ if (!page)
+ goto partial_error;
+ skb_fill_page_desc(skb, 0, page, 0, PAGE_SIZE);
+ mapping[1] =
+ ib_dma_map_page(priv->ca, skb_shinfo(skb)->frags[0].page,
+ 0, PAGE_SIZE, DMA_FROM_DEVICE);
+ if (unlikely(ib_dma_mapping_error(priv->ca, mapping[1])))
+ goto partial_error;
}
- priv->rx_ring[id].skb = skb;
- priv->rx_ring[id].mapping = addr;
+ priv->rx_ring[id].skb = skb;
+ return skb;
- return 0;
+partial_error:
+ ib_dma_unmap_single(priv->ca, mapping[0], buf_size, DMA_FROM_DEVICE);
+error:
+ dev_kfree_skb_any(skb);
+ return NULL;
}
static int ipoib_ib_post_receives(struct net_device *dev)
@@ -154,7 +206,7 @@
int i;
for (i = 0; i < ipoib_recvq_size; ++i) {
- if (ipoib_alloc_rx_skb(dev, i)) {
+ if (!ipoib_alloc_rx_skb(dev, i)) {
ipoib_warn(priv, "failed to allocate receive buffer %d\n", i);
return -ENOMEM;
}
@@ -172,7 +224,7 @@
struct ipoib_dev_priv *priv = netdev_priv(dev);
unsigned int wr_id = wc->wr_id & ~IPOIB_OP_RECV;
struct sk_buff *skb;
- u64 addr;
+ u64 mapping[IPOIB_UD_RX_SG];
ipoib_dbg_data(priv, "recv completion: id %d, status: %d\n",
wr_id, wc->status);
@@ -184,15 +236,13 @@
}
skb = priv->rx_ring[wr_id].skb;
- addr = priv->rx_ring[wr_id].mapping;
if (unlikely(wc->status != IB_WC_SUCCESS)) {
if (wc->status != IB_WC_WR_FLUSH_ERR)
ipoib_warn(priv, "failed recv event "
"(status=%d, wrid=%d vend_err %x)\n",
wc->status, wr_id, wc->vendor_err);
- ib_dma_unmap_single(priv->ca, addr,
- IPOIB_BUF_SIZE, DMA_FROM_DEVICE);
+ ipoib_ud_dma_unmap_rx(priv, priv->rx_ring[wr_id].mapping);
dev_kfree_skb_any(skb);
priv->rx_ring[wr_id].skb = NULL;
return;
@@ -205,11 +255,14 @@
if (wc->slid == priv->local_lid && wc->src_qp == priv->qp->qp_num)
goto repost;
+ memcpy(mapping, priv->rx_ring[wr_id].mapping,
+ IPOIB_UD_RX_SG * sizeof *mapping);
+
/*
* If we can't allocate a new RX buffer, dump
* this packet and reuse the old buffer.
*/
- if (unlikely(ipoib_alloc_rx_skb(dev, wr_id))) {
+ if (unlikely(!ipoib_alloc_rx_skb(dev, wr_id))) {
++dev->stats.rx_dropped;
goto repost;
}
@@ -217,9 +270,9 @@
ipoib_dbg_data(priv, "received %d bytes, SLID 0x%04x\n",
wc->byte_len, wc->slid);
- ib_dma_unmap_single(priv->ca, addr, IPOIB_BUF_SIZE, DMA_FROM_DEVICE);
+ ipoib_ud_dma_unmap_rx(priv, mapping);
+ ipoib_ud_skb_put_frags(priv, skb, wc->byte_len);
- skb_put(skb, wc->byte_len);
skb_pull(skb, IB_GRH_BYTES);
skb->protocol = ((struct ipoib_header *) skb->data)->proto;
@@ -733,10 +786,8 @@
rx_req = &priv->rx_ring[i];
if (!rx_req->skb)
continue;
- ib_dma_unmap_single(priv->ca,
- rx_req->mapping,
- IPOIB_BUF_SIZE,
- DMA_FROM_DEVICE);
+ ipoib_ud_dma_unmap_rx(priv,
+ priv->rx_ring[i].mapping);
dev_kfree_skb_any(rx_req->skb);
rx_req->skb = NULL;
}
diff --git a/drivers/infiniband/ulp/ipoib/ipoib_main.c b/drivers/infiniband/ulp/ipoib/ipoib_main.c
index bd07f02..7a4ed9d 100644
--- a/drivers/infiniband/ulp/ipoib/ipoib_main.c
+++ b/drivers/infiniband/ulp/ipoib/ipoib_main.c
@@ -195,7 +195,7 @@
return 0;
}
- if (new_mtu > IPOIB_PACKET_SIZE - IPOIB_ENCAP_LEN)
+ if (new_mtu > IPOIB_UD_MTU(priv->max_ib_mtu))
return -EINVAL;
priv->admin_mtu = new_mtu;
@@ -971,10 +971,6 @@
NETIF_F_LLTX |
NETIF_F_HIGHDMA);
- /* MTU will be reset when mcast join happens */
- dev->mtu = IPOIB_PACKET_SIZE - IPOIB_ENCAP_LEN;
- priv->mcast_mtu = priv->admin_mtu = dev->mtu;
-
memcpy(dev->broadcast, ipv4_bcast_addr, INFINIBAND_ALEN);
netif_carrier_off(dev);
@@ -1107,6 +1103,7 @@
{
struct ipoib_dev_priv *priv;
struct ib_device_attr *device_attr;
+ struct ib_port_attr attr;
int result = -ENOMEM;
priv = ipoib_intf_alloc(format);
@@ -1115,6 +1112,18 @@
SET_NETDEV_DEV(priv->dev, hca->dma_device);
+ if (!ib_query_port(hca, port, &attr))
+ priv->max_ib_mtu = ib_mtu_enum_to_int(attr.max_mtu);
+ else {
+ printk(KERN_WARNING "%s: ib_query_port %d failed\n",
+ hca->name, port);
+ goto device_init_failed;
+ }
+
+ /* MTU will be reset when mcast join happens */
+ priv->dev->mtu = IPOIB_UD_MTU(priv->max_ib_mtu);
+ priv->mcast_mtu = priv->admin_mtu = priv->dev->mtu;
+
result = ib_query_pkey(hca, port, 0, &priv->pkey);
if (result) {
printk(KERN_WARNING "%s: ib_query_pkey port %d failed (ret = %d)\n",
diff --git a/drivers/infiniband/ulp/ipoib/ipoib_multicast.c b/drivers/infiniband/ulp/ipoib/ipoib_multicast.c
index 31a53c5..d00a2c1 100644
--- a/drivers/infiniband/ulp/ipoib/ipoib_multicast.c
+++ b/drivers/infiniband/ulp/ipoib/ipoib_multicast.c
@@ -567,8 +567,7 @@
return;
}
- priv->mcast_mtu = ib_mtu_enum_to_int(priv->broadcast->mcmember.mtu) -
- IPOIB_ENCAP_LEN;
+ priv->mcast_mtu = IPOIB_UD_MTU(ib_mtu_enum_to_int(priv->broadcast->mcmember.mtu));
if (!ipoib_cm_admin_enabled(dev))
dev->mtu = min(priv->mcast_mtu, priv->admin_mtu);
diff --git a/drivers/infiniband/ulp/ipoib/ipoib_verbs.c b/drivers/infiniband/ulp/ipoib/ipoib_verbs.c
index 8a20e37..07c03f1 100644
--- a/drivers/infiniband/ulp/ipoib/ipoib_verbs.c
+++ b/drivers/infiniband/ulp/ipoib/ipoib_verbs.c
@@ -150,7 +150,7 @@
.max_send_wr = ipoib_sendq_size,
.max_recv_wr = ipoib_recvq_size,
.max_send_sge = 1,
- .max_recv_sge = 1
+ .max_recv_sge = IPOIB_UD_RX_SG
},
.sq_sig_type = IB_SIGNAL_ALL_WR,
.qp_type = IB_QPT_UD
@@ -215,6 +215,19 @@
priv->tx_wr.sg_list = priv->tx_sge;
priv->tx_wr.send_flags = IB_SEND_SIGNALED;
+ priv->rx_sge[0].lkey = priv->mr->lkey;
+ if (ipoib_ud_need_sg(priv->max_ib_mtu)) {
+ priv->rx_sge[0].length = IPOIB_UD_HEAD_SIZE;
+ priv->rx_sge[1].length = PAGE_SIZE;
+ priv->rx_sge[1].lkey = priv->mr->lkey;
+ priv->rx_wr.num_sge = IPOIB_UD_RX_SG;
+ } else {
+ priv->rx_sge[0].length = IPOIB_UD_BUF_SIZE(priv->max_ib_mtu);
+ priv->rx_wr.num_sge = 1;
+ }
+ priv->rx_wr.next = NULL;
+ priv->rx_wr.sg_list = priv->rx_sge;
+
return 0;
out_free_cq:
diff --git a/drivers/infiniband/ulp/ipoib/ipoib_vlan.c b/drivers/infiniband/ulp/ipoib/ipoib_vlan.c
index 293f5b8..431fdea 100644
--- a/drivers/infiniband/ulp/ipoib/ipoib_vlan.c
+++ b/drivers/infiniband/ulp/ipoib/ipoib_vlan.c
@@ -89,6 +89,7 @@
goto err;
}
+ priv->max_ib_mtu = ppriv->max_ib_mtu;
set_bit(IPOIB_FLAG_SUBINTERFACE, &priv->flags);
priv->pkey = pkey;
diff --git a/drivers/input/joystick/xpad.c b/drivers/input/joystick/xpad.c
index 4b07bda..b29e3af 100644
--- a/drivers/input/joystick/xpad.c
+++ b/drivers/input/joystick/xpad.c
@@ -444,6 +444,23 @@
__FUNCTION__, retval);
}
+static void xpad_bulk_out(struct urb *urb)
+{
+ switch (urb->status) {
+ case 0:
+ /* success */
+ break;
+ case -ECONNRESET:
+ case -ENOENT:
+ case -ESHUTDOWN:
+ /* this urb is terminated, clean up */
+ dbg("%s - urb shutting down with status: %d", __FUNCTION__, urb->status);
+ break;
+ default:
+ dbg("%s - nonzero urb status received: %d", __FUNCTION__, urb->status);
+ }
+}
+
#if defined(CONFIG_JOYSTICK_XPAD_FF) || defined(CONFIG_JOYSTICK_XPAD_LEDS)
static void xpad_irq_out(struct urb *urb)
{
@@ -475,23 +492,6 @@
__FUNCTION__, retval);
}
-static void xpad_bulk_out(struct urb *urb)
-{
- switch (urb->status) {
- case 0:
- /* success */
- break;
- case -ECONNRESET:
- case -ENOENT:
- case -ESHUTDOWN:
- /* this urb is terminated, clean up */
- dbg("%s - urb shutting down with status: %d", __FUNCTION__, urb->status);
- break;
- default:
- dbg("%s - nonzero urb status received: %d", __FUNCTION__, urb->status);
- }
-}
-
static int xpad_init_output(struct usb_interface *intf, struct usb_xpad *xpad)
{
struct usb_endpoint_descriptor *ep_irq_out;
diff --git a/drivers/net/Kconfig b/drivers/net/Kconfig
index 0697aa8..8082c1d 100644
--- a/drivers/net/Kconfig
+++ b/drivers/net/Kconfig
@@ -2011,7 +2011,7 @@
config E1000E
tristate "Intel(R) PRO/1000 PCI-Express Gigabit Ethernet support"
- depends on PCI
+ depends on PCI && (!SPARC32 || BROKEN)
---help---
This driver supports the PCI-Express Intel(R) PRO/1000 gigabit
ethernet family of adapters. For PCI or PCI-X e1000 adapters,
diff --git a/drivers/net/mlx4/alloc.c b/drivers/net/mlx4/alloc.c
index 75ef9d0..f9d6b4d 100644
--- a/drivers/net/mlx4/alloc.c
+++ b/drivers/net/mlx4/alloc.c
@@ -196,3 +196,160 @@
}
}
EXPORT_SYMBOL_GPL(mlx4_buf_free);
+
+static struct mlx4_db_pgdir *mlx4_alloc_db_pgdir(struct device *dma_device)
+{
+ struct mlx4_db_pgdir *pgdir;
+
+ pgdir = kzalloc(sizeof *pgdir, GFP_KERNEL);
+ if (!pgdir)
+ return NULL;
+
+ bitmap_fill(pgdir->order1, MLX4_DB_PER_PAGE / 2);
+ pgdir->bits[0] = pgdir->order0;
+ pgdir->bits[1] = pgdir->order1;
+ pgdir->db_page = dma_alloc_coherent(dma_device, PAGE_SIZE,
+ &pgdir->db_dma, GFP_KERNEL);
+ if (!pgdir->db_page) {
+ kfree(pgdir);
+ return NULL;
+ }
+
+ return pgdir;
+}
+
+static int mlx4_alloc_db_from_pgdir(struct mlx4_db_pgdir *pgdir,
+ struct mlx4_db *db, int order)
+{
+ int o;
+ int i;
+
+ for (o = order; o <= 1; ++o) {
+ i = find_first_bit(pgdir->bits[o], MLX4_DB_PER_PAGE >> o);
+ if (i < MLX4_DB_PER_PAGE >> o)
+ goto found;
+ }
+
+ return -ENOMEM;
+
+found:
+ clear_bit(i, pgdir->bits[o]);
+
+ i <<= o;
+
+ if (o > order)
+ set_bit(i ^ 1, pgdir->bits[order]);
+
+ db->u.pgdir = pgdir;
+ db->index = i;
+ db->db = pgdir->db_page + db->index;
+ db->dma = pgdir->db_dma + db->index * 4;
+ db->order = order;
+
+ return 0;
+}
+
+int mlx4_db_alloc(struct mlx4_dev *dev, struct mlx4_db *db, int order)
+{
+ struct mlx4_priv *priv = mlx4_priv(dev);
+ struct mlx4_db_pgdir *pgdir;
+ int ret = 0;
+
+ mutex_lock(&priv->pgdir_mutex);
+
+ list_for_each_entry(pgdir, &priv->pgdir_list, list)
+ if (!mlx4_alloc_db_from_pgdir(pgdir, db, order))
+ goto out;
+
+ pgdir = mlx4_alloc_db_pgdir(&(dev->pdev->dev));
+ if (!pgdir) {
+ ret = -ENOMEM;
+ goto out;
+ }
+
+ list_add(&pgdir->list, &priv->pgdir_list);
+
+ /* This should never fail -- we just allocated an empty page: */
+ WARN_ON(mlx4_alloc_db_from_pgdir(pgdir, db, order));
+
+out:
+ mutex_unlock(&priv->pgdir_mutex);
+
+ return ret;
+}
+EXPORT_SYMBOL_GPL(mlx4_db_alloc);
+
+void mlx4_db_free(struct mlx4_dev *dev, struct mlx4_db *db)
+{
+ struct mlx4_priv *priv = mlx4_priv(dev);
+ int o;
+ int i;
+
+ mutex_lock(&priv->pgdir_mutex);
+
+ o = db->order;
+ i = db->index;
+
+ if (db->order == 0 && test_bit(i ^ 1, db->u.pgdir->order0)) {
+ clear_bit(i ^ 1, db->u.pgdir->order0);
+ ++o;
+ }
+ i >>= o;
+ set_bit(i, db->u.pgdir->bits[o]);
+
+ if (bitmap_full(db->u.pgdir->order1, MLX4_DB_PER_PAGE / 2)) {
+ dma_free_coherent(&(dev->pdev->dev), PAGE_SIZE,
+ db->u.pgdir->db_page, db->u.pgdir->db_dma);
+ list_del(&db->u.pgdir->list);
+ kfree(db->u.pgdir);
+ }
+
+ mutex_unlock(&priv->pgdir_mutex);
+}
+EXPORT_SYMBOL_GPL(mlx4_db_free);
+
+int mlx4_alloc_hwq_res(struct mlx4_dev *dev, struct mlx4_hwq_resources *wqres,
+ int size, int max_direct)
+{
+ int err;
+
+ err = mlx4_db_alloc(dev, &wqres->db, 1);
+ if (err)
+ return err;
+
+ *wqres->db.db = 0;
+
+ err = mlx4_buf_alloc(dev, size, max_direct, &wqres->buf);
+ if (err)
+ goto err_db;
+
+ err = mlx4_mtt_init(dev, wqres->buf.npages, wqres->buf.page_shift,
+ &wqres->mtt);
+ if (err)
+ goto err_buf;
+
+ err = mlx4_buf_write_mtt(dev, &wqres->mtt, &wqres->buf);
+ if (err)
+ goto err_mtt;
+
+ return 0;
+
+err_mtt:
+ mlx4_mtt_cleanup(dev, &wqres->mtt);
+err_buf:
+ mlx4_buf_free(dev, size, &wqres->buf);
+err_db:
+ mlx4_db_free(dev, &wqres->db);
+
+ return err;
+}
+EXPORT_SYMBOL_GPL(mlx4_alloc_hwq_res);
+
+void mlx4_free_hwq_res(struct mlx4_dev *dev, struct mlx4_hwq_resources *wqres,
+ int size)
+{
+ mlx4_mtt_cleanup(dev, &wqres->mtt);
+ mlx4_buf_free(dev, size, &wqres->buf);
+ mlx4_db_free(dev, &wqres->db);
+}
+EXPORT_SYMBOL_GPL(mlx4_free_hwq_res);
diff --git a/drivers/net/mlx4/cq.c b/drivers/net/mlx4/cq.c
index caa5bcf..6fda0af 100644
--- a/drivers/net/mlx4/cq.c
+++ b/drivers/net/mlx4/cq.c
@@ -180,7 +180,7 @@
cq_context->mtt_base_addr_h = mtt_addr >> 32;
cq_context->mtt_base_addr_l = cpu_to_be32(mtt_addr & 0xffffffff);
- err = mlx4_MODIFY_CQ(dev, mailbox, cq->cqn, 1);
+ err = mlx4_MODIFY_CQ(dev, mailbox, cq->cqn, 0);
mlx4_free_cmd_mailbox(dev, mailbox);
return err;
diff --git a/drivers/net/mlx4/main.c b/drivers/net/mlx4/main.c
index 49a4aca..a6aa49f 100644
--- a/drivers/net/mlx4/main.c
+++ b/drivers/net/mlx4/main.c
@@ -798,6 +798,9 @@
INIT_LIST_HEAD(&priv->ctx_list);
spin_lock_init(&priv->ctx_lock);
+ INIT_LIST_HEAD(&priv->pgdir_list);
+ mutex_init(&priv->pgdir_mutex);
+
/*
* Now reset the HCA before we touch the PCI capabilities or
* attempt a firmware command, since a boot ROM may have left
diff --git a/drivers/net/mlx4/mlx4.h b/drivers/net/mlx4/mlx4.h
index 7333681..a4023c2 100644
--- a/drivers/net/mlx4/mlx4.h
+++ b/drivers/net/mlx4/mlx4.h
@@ -257,6 +257,9 @@
struct list_head ctx_list;
spinlock_t ctx_lock;
+ struct list_head pgdir_list;
+ struct mutex pgdir_mutex;
+
struct mlx4_fw fw;
struct mlx4_cmd cmd;
diff --git a/drivers/net/mlx4/qp.c b/drivers/net/mlx4/qp.c
index fa24e65..ee5484c 100644
--- a/drivers/net/mlx4/qp.c
+++ b/drivers/net/mlx4/qp.c
@@ -299,3 +299,34 @@
}
EXPORT_SYMBOL_GPL(mlx4_qp_query);
+int mlx4_qp_to_ready(struct mlx4_dev *dev, struct mlx4_mtt *mtt,
+ struct mlx4_qp_context *context,
+ struct mlx4_qp *qp, enum mlx4_qp_state *qp_state)
+{
+ int err;
+ int i;
+ enum mlx4_qp_state states[] = {
+ MLX4_QP_STATE_RST,
+ MLX4_QP_STATE_INIT,
+ MLX4_QP_STATE_RTR,
+ MLX4_QP_STATE_RTS
+ };
+
+ for (i = 0; i < ARRAY_SIZE(states) - 1; i++) {
+ context->flags &= cpu_to_be32(~(0xf << 28));
+ context->flags |= cpu_to_be32(states[i + 1] << 28);
+ err = mlx4_qp_modify(dev, mtt, states[i], states[i + 1],
+ context, 0, 0, qp);
+ if (err) {
+ mlx4_err(dev, "Failed to bring QP to state: "
+ "%d with error: %d\n",
+ states[i + 1], err);
+ return err;
+ }
+
+ *qp_state = states[i + 1];
+ }
+
+ return 0;
+}
+EXPORT_SYMBOL_GPL(mlx4_qp_to_ready);
diff --git a/include/asm-x86/processor.h b/include/asm-x86/processor.h
index 117343b..2e7974e 100644
--- a/include/asm-x86/processor.h
+++ b/include/asm-x86/processor.h
@@ -722,6 +722,7 @@
static inline void __sti_mwait(unsigned long eax, unsigned long ecx)
{
+ trace_hardirqs_on();
/* "mwait %eax, %ecx;" */
asm volatile("sti; .byte 0x0f, 0x01, 0xc9;"
:: "a" (eax), "c" (ecx));
diff --git a/include/linux/mlx4/device.h b/include/linux/mlx4/device.h
index ff7df1a..9fa1a80 100644
--- a/include/linux/mlx4/device.h
+++ b/include/linux/mlx4/device.h
@@ -208,6 +208,38 @@
int page_shift;
};
+enum {
+ MLX4_DB_PER_PAGE = PAGE_SIZE / 4
+};
+
+struct mlx4_db_pgdir {
+ struct list_head list;
+ DECLARE_BITMAP(order0, MLX4_DB_PER_PAGE);
+ DECLARE_BITMAP(order1, MLX4_DB_PER_PAGE / 2);
+ unsigned long *bits[2];
+ __be32 *db_page;
+ dma_addr_t db_dma;
+};
+
+struct mlx4_ib_user_db_page;
+
+struct mlx4_db {
+ __be32 *db;
+ union {
+ struct mlx4_db_pgdir *pgdir;
+ struct mlx4_ib_user_db_page *user_page;
+ } u;
+ dma_addr_t dma;
+ int index;
+ int order;
+};
+
+struct mlx4_hwq_resources {
+ struct mlx4_db db;
+ struct mlx4_mtt mtt;
+ struct mlx4_buf buf;
+};
+
struct mlx4_mr {
struct mlx4_mtt mtt;
u64 iova;
@@ -341,6 +373,14 @@
int mlx4_buf_write_mtt(struct mlx4_dev *dev, struct mlx4_mtt *mtt,
struct mlx4_buf *buf);
+int mlx4_db_alloc(struct mlx4_dev *dev, struct mlx4_db *db, int order);
+void mlx4_db_free(struct mlx4_dev *dev, struct mlx4_db *db);
+
+int mlx4_alloc_hwq_res(struct mlx4_dev *dev, struct mlx4_hwq_resources *wqres,
+ int size, int max_direct);
+void mlx4_free_hwq_res(struct mlx4_dev *mdev, struct mlx4_hwq_resources *wqres,
+ int size);
+
int mlx4_cq_alloc(struct mlx4_dev *dev, int nent, struct mlx4_mtt *mtt,
struct mlx4_uar *uar, u64 db_rec, struct mlx4_cq *cq);
void mlx4_cq_free(struct mlx4_dev *dev, struct mlx4_cq *cq);
diff --git a/include/linux/mlx4/qp.h b/include/linux/mlx4/qp.h
index a5e43fe..7f128b2 100644
--- a/include/linux/mlx4/qp.h
+++ b/include/linux/mlx4/qp.h
@@ -296,6 +296,10 @@
int mlx4_qp_query(struct mlx4_dev *dev, struct mlx4_qp *qp,
struct mlx4_qp_context *context);
+int mlx4_qp_to_ready(struct mlx4_dev *dev, struct mlx4_mtt *mtt,
+ struct mlx4_qp_context *context,
+ struct mlx4_qp *qp, enum mlx4_qp_state *qp_state);
+
static inline struct mlx4_qp *__mlx4_qp_lookup(struct mlx4_dev *dev, u32 qpn)
{
return radix_tree_lookup(&dev->qp_table_tree, qpn & (dev->caps.num_qps - 1));
diff --git a/net/mac80211/mesh.h b/net/mac80211/mesh.h
index 742003d..9ee3aff 100644
--- a/net/mac80211/mesh.h
+++ b/net/mac80211/mesh.h
@@ -13,6 +13,7 @@
#include <linux/types.h>
#include <linux/jhash.h>
+#include <asm/unaligned.h>
#include "ieee80211_i.h"
diff --git a/net/mac80211/mesh_hwmp.c b/net/mac80211/mesh_hwmp.c
index 02de8f1..3df8092 100644
--- a/net/mac80211/mesh_hwmp.c
+++ b/net/mac80211/mesh_hwmp.c
@@ -7,7 +7,6 @@
* published by the Free Software Foundation.
*/
-#include <asm/unaligned.h>
#include "mesh.h"
#define TEST_FRAME_LEN 8192