Hi Thomas,
The 1st patch makes __irq_build_affinity_masks() more reliable, such as,
all nodes can be covered in the spread.
The 2nd patch spread vectors on node according to the ratio of this node's
CPU number to number of all remaining CPUs, then vectors assignment can
become more fair. Meantime, the warning report from Jon Derrick can be
fixed.
Please consider it for V5.4.
V5:
- remove patch 1 of V4, which is wrong
- handle vector wrapping because the 'start vector' may begin
anywhere, especially for the 2nd stage spread
- add more comment on the vector allocation algorithm
- cleanup code a bit
- run more tests to verify the change, which always get the
expected result. Covers lots of num_queues, numa topo, CPU
unpresent setting.
V4:
- provide proof why number of allocated vectors for each node is <= CPU
count of this node
V3:
- re-order the patchset
- add helper of irq_spread_vectors_on_node()
- handle vector spread correctly in case that numvecs is > ncpus
- return -ENOMEM to API's caller
V2:
- add patch3
- start to allocate vectors from node with minimized CPU number,
then every node is guaranteed to be allocated at least one vector.
- avoid cross node spread
Ming Lei (2):
genirq/affinity: Improve __irq_build_affinity_masks()
genirq/affinity: Spread vectors on node according to nr_cpu ratio
kernel/irq/affinity.c | 215 +++++++++++++++++++++++++++++++++++++-----
1 file changed, 194 insertions(+), 21 deletions(-)
Cc: Jens Axboe <[email protected]>
Cc: Christoph Hellwig <[email protected]>
Cc: Keith Busch <[email protected]>
Cc: [email protected],
Cc: Jon Derrick <[email protected]>
--
2.20.1
One invariant of __irq_build_affinity_masks() is that all CPUs in the
specified masks( cpu_mask AND node_to_cpumask for each node) should be
covered during the spread. Even though all requested vectors have been
reached, we still need to spread vectors among remained CPUs. The similar
policy has been taken in case of 'numvecs <= nodes' already:
So remove the following check inside the loop:
if (done >= numvecs)
break;
Meantime assign at least 1 vector for remained nodes if 'numvecs' vectors
have been handled already.
Also, if the specified cpumask for one numa node is empty, simply not
spread vectors on this node.
Cc: Christoph Hellwig <[email protected]>
Cc: Keith Busch <[email protected]>
Cc: [email protected],
Cc: Jon Derrick <[email protected]>
Cc: Jens Axboe <[email protected]>
Signed-off-by: Ming Lei <[email protected]>
---
kernel/irq/affinity.c | 26 ++++++++++++++++++--------
1 file changed, 18 insertions(+), 8 deletions(-)
diff --git a/kernel/irq/affinity.c b/kernel/irq/affinity.c
index 6fef48033f96..c7cca942bd8a 100644
--- a/kernel/irq/affinity.c
+++ b/kernel/irq/affinity.c
@@ -129,14 +129,26 @@ static int __irq_build_affinity_masks(unsigned int startvec,
for_each_node_mask(n, nodemsk) {
unsigned int ncpus, v, vecs_to_assign, vecs_per_node;
- /* Spread the vectors per node */
- vecs_per_node = (numvecs - (curvec - firstvec)) / nodes;
-
/* Get the cpus on this node which are in the mask */
cpumask_and(nmsk, cpu_mask, node_to_cpumask[n]);
-
- /* Calculate the number of cpus per vector */
ncpus = cpumask_weight(nmsk);
+ if (!ncpus)
+ continue;
+
+ /*
+ * Calculate the number of cpus per vector
+ *
+ * Spread the vectors evenly per node. If the requested
+ * vector number has been reached, simply allocate one
+ * vector for each remaining node so that all nodes can
+ * be covered
+ */
+ if (numvecs > done)
+ vecs_per_node = max_t(unsigned,
+ (numvecs - done) / nodes, 1);
+ else
+ vecs_per_node = 1;
+
vecs_to_assign = min(vecs_per_node, ncpus);
/* Account for rounding errors */
@@ -156,13 +168,11 @@ static int __irq_build_affinity_masks(unsigned int startvec,
}
done += v;
- if (done >= numvecs)
- break;
if (curvec >= last_affv)
curvec = firstvec;
--nodes;
}
- return done;
+ return done < numvecs ? done : numvecs;
}
/*
--
2.20.1
Now __irq_build_affinity_masks() spreads vectors evenly per node, and
all vectors may not be spread in case that each numa node has different
CPU number, then the warning in irq_build_affinity_masks() can
be triggered.
Improve current spreading algorithm by assigning vectors according to
the ratio of node's nr_cpu to nr_remaining_cpus, meantime running the
assignment from smaller nodes to bigger nodes to guarantee that every
active node gets allocated at least one vector, then we can avoid
cross-node spread in normal situation.
Meantime the reported warning can be fixed.
Another big goodness is that the spread approach becomes more fair if
node has different CPU number.
For example, on the following machine:
[root@ktest-01 ~]# lscpu
...
CPU(s): 16
On-line CPU(s) list: 0-15
Thread(s) per core: 1
Core(s) per socket: 8
Socket(s): 2
NUMA node(s): 2
...
NUMA node0 CPU(s): 0,1,3,5-9,11,13-15
NUMA node1 CPU(s): 2,4,10,12
When driver requests to allocate 8 vectors, the following spread can
be got:
irq 31, cpu list 2,4
irq 32, cpu list 10,12
irq 33, cpu list 0-1
irq 34, cpu list 3,5
irq 35, cpu list 6-7
irq 36, cpu list 8-9
irq 37, cpu list 11,13
irq 38, cpu list 14-15
Without this patch, kernel warning is triggered on above situation, and
allocation result was supposed to be 4 vectors for each node.
Cc: Christoph Hellwig <[email protected]>
Cc: Keith Busch <[email protected]>
Cc: [email protected],
Cc: Jon Derrick <[email protected]>
Cc: Jens Axboe <[email protected]>
Reported-by: Jon Derrick <[email protected]>
Signed-off-by: Ming Lei <[email protected]>
---
kernel/irq/affinity.c | 223 ++++++++++++++++++++++++++++++++++++------
1 file changed, 193 insertions(+), 30 deletions(-)
diff --git a/kernel/irq/affinity.c b/kernel/irq/affinity.c
index c7cca942bd8a..32e07e58ce81 100644
--- a/kernel/irq/affinity.c
+++ b/kernel/irq/affinity.c
@@ -7,6 +7,7 @@
#include <linux/kernel.h>
#include <linux/slab.h>
#include <linux/cpu.h>
+#include <linux/sort.h>
static void irq_spread_init_one(struct cpumask *irqmsk, struct cpumask *nmsk,
unsigned int cpus_per_vec)
@@ -94,6 +95,155 @@ static int get_nodes_in_cpumask(cpumask_var_t *node_to_cpumask,
return nodes;
}
+struct node_vectors {
+ unsigned id;
+
+ union {
+ unsigned nvectors;
+ unsigned ncpus;
+ };
+};
+
+static int ncpus_cmp_func(const void *l, const void *r)
+{
+ const struct node_vectors *ln = l;
+ const struct node_vectors *rn = r;
+
+ return ln->ncpus - rn->ncpus;
+}
+
+/*
+ * Allocate vector number for each node, so that for each node:
+ *
+ * 1) the allocated number is >= 1
+ *
+ * 2) the allocated numbver is <= active CPU number of this node
+ *
+ * The actual allocated total vectors may be less than @numvecs when
+ * active total CPU number is less than @numvecs.
+ *
+ * Active CPUs means the CPUs in '@cpu_mask AND @node_to_cpumask[]'
+ * for each node.
+ */
+static void alloc_nodes_vectors(unsigned int numvecs,
+ const cpumask_var_t *node_to_cpumask,
+ const struct cpumask *cpu_mask,
+ const nodemask_t nodemsk,
+ struct cpumask *nmsk,
+ struct node_vectors *node_vectors)
+{
+ unsigned n, remaining_ncpus = 0;
+
+ for (n = 0; n < nr_node_ids; n++) {
+ node_vectors[n].id = n;
+ node_vectors[n].ncpus = UINT_MAX;
+ }
+
+ for_each_node_mask(n, nodemsk) {
+ unsigned ncpus;
+
+ cpumask_and(nmsk, cpu_mask, node_to_cpumask[n]);
+ ncpus = cpumask_weight(nmsk);
+
+ if (!ncpus)
+ continue;
+ remaining_ncpus += ncpus;
+ node_vectors[n].ncpus = ncpus;
+ }
+
+ numvecs = min_t(unsigned, remaining_ncpus, numvecs);
+
+ sort(node_vectors, nr_node_ids, sizeof(node_vectors[0]),
+ ncpus_cmp_func, NULL);
+
+ /*
+ * Allocate vectors for each node according to the ratio of this
+ * node's nr_cpus to remaining un-assigned ncpus. 'numvecs' is
+ * bigger than number of active numa nodes. Always start the
+ * allocation from the node with minimized nr_cpus.
+ *
+ * This way guarantees that each active node gets allocated at
+ * least one vector, and the theory is simple: over-allocation
+ * is only done when this node is assigned by one vector, so
+ * other nodes will be allocated >= 1 vector, since 'numvecs' is
+ * bigger than number of numa nodes.
+ *
+ * One perfect invariant is that number of allocated vectors for
+ * each node is <= CPU count of this node:
+ *
+ * 1) suppose there are two nodes: A and B
+ * ncpu(X) is CPU count of node X
+ * vecs(X) is the vector count allocated to node X via this
+ * algorithm
+ *
+ * ncpu(A) <= ncpu(B)
+ * ncpu(A) + ncpu(B) = N
+ * vecs(A) + vecs(B) = V
+ *
+ * vecs(A) = max(1, round_down(V * ncpu(A) / N))
+ * vecs(B) = V - vecs(A)
+ *
+ * both N and V are integer, and 2 <= V <= N, suppose
+ * V = N - delta, and 0 <= delta <= N - 2
+ *
+ * 2) obviously vecs(A) <= ncpu(A) because:
+ *
+ * if vecs(A) is 1, then vecs(A) <= ncpu(A) given
+ * ncpu(A) >= 1
+ *
+ * otherwise,
+ * vecs(A) <= V * ncpu(A) / N <= ncpu(A), given V <= N
+ *
+ * 3) prove how vecs(B) <= ncpu(B):
+ *
+ * if round_down(V * ncpu(A) / N) == 0, vecs(B) won't be
+ * over-allocated, so vecs(B) <= ncpu(B),
+ *
+ * otherwise:
+ *
+ * vecs(A) =
+ * round_down(V * ncpu(A) / N) =
+ * round_down((N - delta) * ncpu(A) / N) =
+ * round_down((N * ncpu(A) - delta * ncpu(A)) / N) >=
+ * round_down((N * ncpu(A) - delta * N) / N) =
+ * cpu(A) - delta
+ *
+ * then:
+ *
+ * vecs(A) - V >= ncpu(A) - delta - V
+ * =>
+ * V - vecs(A) <= V + delta - ncpu(A)
+ * =>
+ * vecs(B) <= N - ncpu(A)
+ * =>
+ * vecs(B) <= cpu(B)
+ *
+ * For nodes >= 3, it can be thought as one node and another big
+ * node given that is exactly what this algorithm is implemented,
+ * and we always re-calculate 'remaining_ncpus' & 'numvecs', and
+ * finally for each node X: vecs(X) <= ncpu(X).
+ *
+ */
+ for (n = 0; n < nr_node_ids; n++) {
+ unsigned nvectors, ncpus;
+
+ if (node_vectors[n].ncpus == UINT_MAX)
+ continue;
+
+ WARN_ON_ONCE(numvecs == 0);
+
+ ncpus = node_vectors[n].ncpus;
+ nvectors = max_t(unsigned, 1,
+ numvecs * ncpus / remaining_ncpus);
+ WARN_ON_ONCE(nvectors > ncpus);
+
+ node_vectors[n].nvectors = nvectors;
+
+ remaining_ncpus -= ncpus;
+ numvecs -= nvectors;
+ }
+}
+
static int __irq_build_affinity_masks(unsigned int startvec,
unsigned int numvecs,
unsigned int firstvec,
@@ -102,10 +252,11 @@ static int __irq_build_affinity_masks(unsigned int startvec,
struct cpumask *nmsk,
struct irq_affinity_desc *masks)
{
- unsigned int n, nodes, cpus_per_vec, extra_vecs, done = 0;
+ unsigned int i, n, nodes, cpus_per_vec, extra_vecs, done = 0;
unsigned int last_affv = firstvec + numvecs;
unsigned int curvec = startvec;
nodemask_t nodemsk = NODE_MASK_NONE;
+ struct node_vectors *node_vectors;
if (!cpumask_weight(cpu_mask))
return 0;
@@ -126,53 +277,57 @@ static int __irq_build_affinity_masks(unsigned int startvec,
return numvecs;
}
- for_each_node_mask(n, nodemsk) {
- unsigned int ncpus, v, vecs_to_assign, vecs_per_node;
+ node_vectors = kcalloc(nr_node_ids,
+ sizeof(struct node_vectors),
+ GFP_KERNEL);
+ if (!node_vectors)
+ return -ENOMEM;
+
+ /* allocate vector number for each node */
+ alloc_nodes_vectors(numvecs, node_to_cpumask, cpu_mask,
+ nodemsk, nmsk, node_vectors);
+
+ for (i = 0; i < nr_node_ids; i++) {
+ unsigned int ncpus, v;
+ struct node_vectors *nv = &node_vectors[i];
+
+ if (nv->nvectors == UINT_MAX)
+ continue;
/* Get the cpus on this node which are in the mask */
- cpumask_and(nmsk, cpu_mask, node_to_cpumask[n]);
+ cpumask_and(nmsk, cpu_mask, node_to_cpumask[nv->id]);
ncpus = cpumask_weight(nmsk);
if (!ncpus)
continue;
- /*
- * Calculate the number of cpus per vector
- *
- * Spread the vectors evenly per node. If the requested
- * vector number has been reached, simply allocate one
- * vector for each remaining node so that all nodes can
- * be covered
- */
- if (numvecs > done)
- vecs_per_node = max_t(unsigned,
- (numvecs - done) / nodes, 1);
- else
- vecs_per_node = 1;
-
- vecs_to_assign = min(vecs_per_node, ncpus);
+ WARN_ON_ONCE(nv->nvectors > ncpus);
/* Account for rounding errors */
- extra_vecs = ncpus - vecs_to_assign * (ncpus / vecs_to_assign);
+ extra_vecs = ncpus - nv->nvectors * (ncpus / nv->nvectors);
- for (v = 0; curvec < last_affv && v < vecs_to_assign;
- curvec++, v++) {
- cpus_per_vec = ncpus / vecs_to_assign;
+ /* Spread allocated vectors on CPUs of the current node */
+ for (v = 0; v < nv->nvectors; v++, curvec++) {
+ cpus_per_vec = ncpus / nv->nvectors;
/* Account for extra vectors to compensate rounding errors */
if (extra_vecs) {
cpus_per_vec++;
--extra_vecs;
}
+
+ /*
+ * wrapping has to be considered given 'startvec'
+ * may start anywhere
+ */
+ if (curvec >= last_affv)
+ curvec = firstvec;
irq_spread_init_one(&masks[curvec].mask, nmsk,
cpus_per_vec);
}
-
- done += v;
- if (curvec >= last_affv)
- curvec = firstvec;
- --nodes;
+ done += nv->nvectors;
}
- return done < numvecs ? done : numvecs;
+ kfree(node_vectors);
+ return done;
}
/*
@@ -208,6 +363,10 @@ static int irq_build_affinity_masks(unsigned int startvec, unsigned int numvecs,
nr_present = __irq_build_affinity_masks(curvec, numvecs,
firstvec, node_to_cpumask,
cpu_present_mask, nmsk, masks);
+ if (nr_present < 0) {
+ ret = nr_present;
+ goto fail_build_affinity;
+ }
/*
* Spread on non present CPUs starting from the next vector to be
@@ -223,9 +382,13 @@ static int irq_build_affinity_masks(unsigned int startvec, unsigned int numvecs,
nr_others = __irq_build_affinity_masks(curvec, numvecs,
firstvec, node_to_cpumask,
npresmsk, nmsk, masks);
+ if (nr_others < 0)
+ ret = nr_others;
+
+ fail_build_affinity:
put_online_cpus();
- if (nr_present < numvecs)
+ if (min(nr_present, nr_others) >= 0)
WARN_ON(nr_present + nr_others < numvecs);
free_node_to_cpumask(node_to_cpumask);
--
2.20.1
On Thu, Aug 15, 2019 at 07:28:49PM -0700, Ming Lei wrote:
> Now __irq_build_affinity_masks() spreads vectors evenly per node, and
> all vectors may not be spread in case that each numa node has different
> CPU number, then the warning in irq_build_affinity_masks() can
> be triggered.
>
> Improve current spreading algorithm by assigning vectors according to
> the ratio of node's nr_cpu to nr_remaining_cpus, meantime running the
> assignment from smaller nodes to bigger nodes to guarantee that every
> active node gets allocated at least one vector, then we can avoid
> cross-node spread in normal situation.
>
> Meantime the reported warning can be fixed.
>
> Another big goodness is that the spread approach becomes more fair if
> node has different CPU number.
>
> For example, on the following machine:
> [root@ktest-01 ~]# lscpu
> ...
> CPU(s): 16
> On-line CPU(s) list: 0-15
> Thread(s) per core: 1
> Core(s) per socket: 8
> Socket(s): 2
> NUMA node(s): 2
> ...
> NUMA node0 CPU(s): 0,1,3,5-9,11,13-15
> NUMA node1 CPU(s): 2,4,10,12
>
> When driver requests to allocate 8 vectors, the following spread can
> be got:
> irq 31, cpu list 2,4
> irq 32, cpu list 10,12
> irq 33, cpu list 0-1
> irq 34, cpu list 3,5
> irq 35, cpu list 6-7
> irq 36, cpu list 8-9
> irq 37, cpu list 11,13
> irq 38, cpu list 14-15
>
> Without this patch, kernel warning is triggered on above situation, and
> allocation result was supposed to be 4 vectors for each node.
>
> Cc: Christoph Hellwig <[email protected]>
> Cc: Keith Busch <[email protected]>
> Cc: [email protected],
> Cc: Jon Derrick <[email protected]>
> Cc: Jens Axboe <[email protected]>
> Reported-by: Jon Derrick <[email protected]>
> Signed-off-by: Ming Lei <[email protected]>
I had every intention to thoroughly test this on imbalanced node
configurations, but that's not going to happen anytime soon. It looks
correct to me, so I'll append my review here.
I'm not sure I'd include the detailed correctness explanation in the
comments, though. It essentially boils down to "the sum of the parts
doesn't exceed the whole", and the key to carving up the parts is the
sorted iteration you've implemented.
Reviewed-by: Keith Busch <[email protected]>
> ---
> kernel/irq/affinity.c | 223 ++++++++++++++++++++++++++++++++++++------
> 1 file changed, 193 insertions(+), 30 deletions(-)
>
> diff --git a/kernel/irq/affinity.c b/kernel/irq/affinity.c
> index c7cca942bd8a..32e07e58ce81 100644
> --- a/kernel/irq/affinity.c
> +++ b/kernel/irq/affinity.c
> @@ -7,6 +7,7 @@
> #include <linux/kernel.h>
> #include <linux/slab.h>
> #include <linux/cpu.h>
> +#include <linux/sort.h>
>
> static void irq_spread_init_one(struct cpumask *irqmsk, struct cpumask *nmsk,
> unsigned int cpus_per_vec)
> @@ -94,6 +95,155 @@ static int get_nodes_in_cpumask(cpumask_var_t *node_to_cpumask,
> return nodes;
> }
>
> +struct node_vectors {
> + unsigned id;
> +
> + union {
> + unsigned nvectors;
> + unsigned ncpus;
> + };
> +};
> +
> +static int ncpus_cmp_func(const void *l, const void *r)
> +{
> + const struct node_vectors *ln = l;
> + const struct node_vectors *rn = r;
> +
> + return ln->ncpus - rn->ncpus;
> +}
> +
> +/*
> + * Allocate vector number for each node, so that for each node:
> + *
> + * 1) the allocated number is >= 1
> + *
> + * 2) the allocated numbver is <= active CPU number of this node
> + *
> + * The actual allocated total vectors may be less than @numvecs when
> + * active total CPU number is less than @numvecs.
> + *
> + * Active CPUs means the CPUs in '@cpu_mask AND @node_to_cpumask[]'
> + * for each node.
> + */
> +static void alloc_nodes_vectors(unsigned int numvecs,
> + const cpumask_var_t *node_to_cpumask,
> + const struct cpumask *cpu_mask,
> + const nodemask_t nodemsk,
> + struct cpumask *nmsk,
> + struct node_vectors *node_vectors)
> +{
> + unsigned n, remaining_ncpus = 0;
> +
> + for (n = 0; n < nr_node_ids; n++) {
> + node_vectors[n].id = n;
> + node_vectors[n].ncpus = UINT_MAX;
> + }
> +
> + for_each_node_mask(n, nodemsk) {
> + unsigned ncpus;
> +
> + cpumask_and(nmsk, cpu_mask, node_to_cpumask[n]);
> + ncpus = cpumask_weight(nmsk);
> +
> + if (!ncpus)
> + continue;
> + remaining_ncpus += ncpus;
> + node_vectors[n].ncpus = ncpus;
> + }
> +
> + numvecs = min_t(unsigned, remaining_ncpus, numvecs);
> +
> + sort(node_vectors, nr_node_ids, sizeof(node_vectors[0]),
> + ncpus_cmp_func, NULL);
> +
> + /*
> + * Allocate vectors for each node according to the ratio of this
> + * node's nr_cpus to remaining un-assigned ncpus. 'numvecs' is
> + * bigger than number of active numa nodes. Always start the
> + * allocation from the node with minimized nr_cpus.
> + *
> + * This way guarantees that each active node gets allocated at
> + * least one vector, and the theory is simple: over-allocation
> + * is only done when this node is assigned by one vector, so
> + * other nodes will be allocated >= 1 vector, since 'numvecs' is
> + * bigger than number of numa nodes.
> + *
> + * One perfect invariant is that number of allocated vectors for
> + * each node is <= CPU count of this node:
> + *
> + * 1) suppose there are two nodes: A and B
> + * ncpu(X) is CPU count of node X
> + * vecs(X) is the vector count allocated to node X via this
> + * algorithm
> + *
> + * ncpu(A) <= ncpu(B)
> + * ncpu(A) + ncpu(B) = N
> + * vecs(A) + vecs(B) = V
> + *
> + * vecs(A) = max(1, round_down(V * ncpu(A) / N))
> + * vecs(B) = V - vecs(A)
> + *
> + * both N and V are integer, and 2 <= V <= N, suppose
> + * V = N - delta, and 0 <= delta <= N - 2
> + *
> + * 2) obviously vecs(A) <= ncpu(A) because:
> + *
> + * if vecs(A) is 1, then vecs(A) <= ncpu(A) given
> + * ncpu(A) >= 1
> + *
> + * otherwise,
> + * vecs(A) <= V * ncpu(A) / N <= ncpu(A), given V <= N
> + *
> + * 3) prove how vecs(B) <= ncpu(B):
> + *
> + * if round_down(V * ncpu(A) / N) == 0, vecs(B) won't be
> + * over-allocated, so vecs(B) <= ncpu(B),
> + *
> + * otherwise:
> + *
> + * vecs(A) =
> + * round_down(V * ncpu(A) / N) =
> + * round_down((N - delta) * ncpu(A) / N) =
> + * round_down((N * ncpu(A) - delta * ncpu(A)) / N) >=
> + * round_down((N * ncpu(A) - delta * N) / N) =
> + * cpu(A) - delta
> + *
> + * then:
> + *
> + * vecs(A) - V >= ncpu(A) - delta - V
> + * =>
> + * V - vecs(A) <= V + delta - ncpu(A)
> + * =>
> + * vecs(B) <= N - ncpu(A)
> + * =>
> + * vecs(B) <= cpu(B)
> + *
> + * For nodes >= 3, it can be thought as one node and another big
> + * node given that is exactly what this algorithm is implemented,
> + * and we always re-calculate 'remaining_ncpus' & 'numvecs', and
> + * finally for each node X: vecs(X) <= ncpu(X).
> + *
> + */
> + for (n = 0; n < nr_node_ids; n++) {
> + unsigned nvectors, ncpus;
> +
> + if (node_vectors[n].ncpus == UINT_MAX)
> + continue;
> +
> + WARN_ON_ONCE(numvecs == 0);
> +
> + ncpus = node_vectors[n].ncpus;
> + nvectors = max_t(unsigned, 1,
> + numvecs * ncpus / remaining_ncpus);
> + WARN_ON_ONCE(nvectors > ncpus);
> +
> + node_vectors[n].nvectors = nvectors;
> +
> + remaining_ncpus -= ncpus;
> + numvecs -= nvectors;
> + }
> +}
> +
> static int __irq_build_affinity_masks(unsigned int startvec,
> unsigned int numvecs,
> unsigned int firstvec,
> @@ -102,10 +252,11 @@ static int __irq_build_affinity_masks(unsigned int startvec,
> struct cpumask *nmsk,
> struct irq_affinity_desc *masks)
> {
> - unsigned int n, nodes, cpus_per_vec, extra_vecs, done = 0;
> + unsigned int i, n, nodes, cpus_per_vec, extra_vecs, done = 0;
> unsigned int last_affv = firstvec + numvecs;
> unsigned int curvec = startvec;
> nodemask_t nodemsk = NODE_MASK_NONE;
> + struct node_vectors *node_vectors;
>
> if (!cpumask_weight(cpu_mask))
> return 0;
> @@ -126,53 +277,57 @@ static int __irq_build_affinity_masks(unsigned int startvec,
> return numvecs;
> }
>
> - for_each_node_mask(n, nodemsk) {
> - unsigned int ncpus, v, vecs_to_assign, vecs_per_node;
> + node_vectors = kcalloc(nr_node_ids,
> + sizeof(struct node_vectors),
> + GFP_KERNEL);
> + if (!node_vectors)
> + return -ENOMEM;
> +
> + /* allocate vector number for each node */
> + alloc_nodes_vectors(numvecs, node_to_cpumask, cpu_mask,
> + nodemsk, nmsk, node_vectors);
> +
> + for (i = 0; i < nr_node_ids; i++) {
> + unsigned int ncpus, v;
> + struct node_vectors *nv = &node_vectors[i];
> +
> + if (nv->nvectors == UINT_MAX)
> + continue;
>
> /* Get the cpus on this node which are in the mask */
> - cpumask_and(nmsk, cpu_mask, node_to_cpumask[n]);
> + cpumask_and(nmsk, cpu_mask, node_to_cpumask[nv->id]);
> ncpus = cpumask_weight(nmsk);
> if (!ncpus)
> continue;
>
> - /*
> - * Calculate the number of cpus per vector
> - *
> - * Spread the vectors evenly per node. If the requested
> - * vector number has been reached, simply allocate one
> - * vector for each remaining node so that all nodes can
> - * be covered
> - */
> - if (numvecs > done)
> - vecs_per_node = max_t(unsigned,
> - (numvecs - done) / nodes, 1);
> - else
> - vecs_per_node = 1;
> -
> - vecs_to_assign = min(vecs_per_node, ncpus);
> + WARN_ON_ONCE(nv->nvectors > ncpus);
>
> /* Account for rounding errors */
> - extra_vecs = ncpus - vecs_to_assign * (ncpus / vecs_to_assign);
> + extra_vecs = ncpus - nv->nvectors * (ncpus / nv->nvectors);
>
> - for (v = 0; curvec < last_affv && v < vecs_to_assign;
> - curvec++, v++) {
> - cpus_per_vec = ncpus / vecs_to_assign;
> + /* Spread allocated vectors on CPUs of the current node */
> + for (v = 0; v < nv->nvectors; v++, curvec++) {
> + cpus_per_vec = ncpus / nv->nvectors;
>
> /* Account for extra vectors to compensate rounding errors */
> if (extra_vecs) {
> cpus_per_vec++;
> --extra_vecs;
> }
> +
> + /*
> + * wrapping has to be considered given 'startvec'
> + * may start anywhere
> + */
> + if (curvec >= last_affv)
> + curvec = firstvec;
> irq_spread_init_one(&masks[curvec].mask, nmsk,
> cpus_per_vec);
> }
> -
> - done += v;
> - if (curvec >= last_affv)
> - curvec = firstvec;
> - --nodes;
> + done += nv->nvectors;
> }
> - return done < numvecs ? done : numvecs;
> + kfree(node_vectors);
> + return done;
> }
>
> /*
> @@ -208,6 +363,10 @@ static int irq_build_affinity_masks(unsigned int startvec, unsigned int numvecs,
> nr_present = __irq_build_affinity_masks(curvec, numvecs,
> firstvec, node_to_cpumask,
> cpu_present_mask, nmsk, masks);
> + if (nr_present < 0) {
> + ret = nr_present;
> + goto fail_build_affinity;
> + }
>
> /*
> * Spread on non present CPUs starting from the next vector to be
> @@ -223,9 +382,13 @@ static int irq_build_affinity_masks(unsigned int startvec, unsigned int numvecs,
> nr_others = __irq_build_affinity_masks(curvec, numvecs,
> firstvec, node_to_cpumask,
> npresmsk, nmsk, masks);
> + if (nr_others < 0)
> + ret = nr_others;
> +
> + fail_build_affinity:
> put_online_cpus();
>
> - if (nr_present < numvecs)
> + if (min(nr_present, nr_others) >= 0)
> WARN_ON(nr_present + nr_others < numvecs);
>
> free_node_to_cpumask(node_to_cpumask);
> --
> 2.20.1
>
On Fri, Aug 16, 2019 at 11:56 PM Keith Busch <[email protected]> wrote:
>
> On Thu, Aug 15, 2019 at 07:28:49PM -0700, Ming Lei wrote:
> > Now __irq_build_affinity_masks() spreads vectors evenly per node, and
> > all vectors may not be spread in case that each numa node has different
> > CPU number, then the warning in irq_build_affinity_masks() can
> > be triggered.
> >
> > Improve current spreading algorithm by assigning vectors according to
> > the ratio of node's nr_cpu to nr_remaining_cpus, meantime running the
> > assignment from smaller nodes to bigger nodes to guarantee that every
> > active node gets allocated at least one vector, then we can avoid
> > cross-node spread in normal situation.
> >
> > Meantime the reported warning can be fixed.
> >
> > Another big goodness is that the spread approach becomes more fair if
> > node has different CPU number.
> >
> > For example, on the following machine:
> > [root@ktest-01 ~]# lscpu
> > ...
> > CPU(s): 16
> > On-line CPU(s) list: 0-15
> > Thread(s) per core: 1
> > Core(s) per socket: 8
> > Socket(s): 2
> > NUMA node(s): 2
> > ...
> > NUMA node0 CPU(s): 0,1,3,5-9,11,13-15
> > NUMA node1 CPU(s): 2,4,10,12
> >
> > When driver requests to allocate 8 vectors, the following spread can
> > be got:
> > irq 31, cpu list 2,4
> > irq 32, cpu list 10,12
> > irq 33, cpu list 0-1
> > irq 34, cpu list 3,5
> > irq 35, cpu list 6-7
> > irq 36, cpu list 8-9
> > irq 37, cpu list 11,13
> > irq 38, cpu list 14-15
> >
> > Without this patch, kernel warning is triggered on above situation, and
> > allocation result was supposed to be 4 vectors for each node.
> >
> > Cc: Christoph Hellwig <[email protected]>
> > Cc: Keith Busch <[email protected]>
> > Cc: [email protected],
> > Cc: Jon Derrick <[email protected]>
> > Cc: Jens Axboe <[email protected]>
> > Reported-by: Jon Derrick <[email protected]>
> > Signed-off-by: Ming Lei <[email protected]>
>
> I had every intention to thoroughly test this on imbalanced node
> configurations, but that's not going to happen anytime soon. It looks
> correct to me, so I'll append my review here.
Thanks for your review!
>
> I'm not sure I'd include the detailed correctness explanation in the
> comments, though. It essentially boils down to "the sum of the parts
> doesn't exceed the whole", and the key to carving up the parts is the
> sorted iteration you've implemented.
There are two invariants reached by this approach:
1) In each node, if the active CPU number isn't zero, >=1 vector is allocated
for this node
2) For each node, the allocated vector number is <= CPU number in this node.
Both two are not obviously, however, the two points are quite important wrt.
the correctness of this approach.
That is why I add the long correctness proof to the comment.
Thanks,
Ming
>
> Reviewed-by: Keith Busch <[email protected]>
>
> > ---
> > kernel/irq/affinity.c | 223 ++++++++++++++++++++++++++++++++++++------
> > 1 file changed, 193 insertions(+), 30 deletions(-)
> >
> > diff --git a/kernel/irq/affinity.c b/kernel/irq/affinity.c
> > index c7cca942bd8a..32e07e58ce81 100644
> > --- a/kernel/irq/affinity.c
> > +++ b/kernel/irq/affinity.c
> > @@ -7,6 +7,7 @@
> > #include <linux/kernel.h>
> > #include <linux/slab.h>
> > #include <linux/cpu.h>
> > +#include <linux/sort.h>
> >
> > static void irq_spread_init_one(struct cpumask *irqmsk, struct cpumask *nmsk,
> > unsigned int cpus_per_vec)
> > @@ -94,6 +95,155 @@ static int get_nodes_in_cpumask(cpumask_var_t *node_to_cpumask,
> > return nodes;
> > }
> >
> > +struct node_vectors {
> > + unsigned id;
> > +
> > + union {
> > + unsigned nvectors;
> > + unsigned ncpus;
> > + };
> > +};
> > +
> > +static int ncpus_cmp_func(const void *l, const void *r)
> > +{
> > + const struct node_vectors *ln = l;
> > + const struct node_vectors *rn = r;
> > +
> > + return ln->ncpus - rn->ncpus;
> > +}
> > +
> > +/*
> > + * Allocate vector number for each node, so that for each node:
> > + *
> > + * 1) the allocated number is >= 1
> > + *
> > + * 2) the allocated numbver is <= active CPU number of this node
> > + *
> > + * The actual allocated total vectors may be less than @numvecs when
> > + * active total CPU number is less than @numvecs.
> > + *
> > + * Active CPUs means the CPUs in '@cpu_mask AND @node_to_cpumask[]'
> > + * for each node.
> > + */
> > +static void alloc_nodes_vectors(unsigned int numvecs,
> > + const cpumask_var_t *node_to_cpumask,
> > + const struct cpumask *cpu_mask,
> > + const nodemask_t nodemsk,
> > + struct cpumask *nmsk,
> > + struct node_vectors *node_vectors)
> > +{
> > + unsigned n, remaining_ncpus = 0;
> > +
> > + for (n = 0; n < nr_node_ids; n++) {
> > + node_vectors[n].id = n;
> > + node_vectors[n].ncpus = UINT_MAX;
> > + }
> > +
> > + for_each_node_mask(n, nodemsk) {
> > + unsigned ncpus;
> > +
> > + cpumask_and(nmsk, cpu_mask, node_to_cpumask[n]);
> > + ncpus = cpumask_weight(nmsk);
> > +
> > + if (!ncpus)
> > + continue;
> > + remaining_ncpus += ncpus;
> > + node_vectors[n].ncpus = ncpus;
> > + }
> > +
> > + numvecs = min_t(unsigned, remaining_ncpus, numvecs);
> > +
> > + sort(node_vectors, nr_node_ids, sizeof(node_vectors[0]),
> > + ncpus_cmp_func, NULL);
> > +
> > + /*
> > + * Allocate vectors for each node according to the ratio of this
> > + * node's nr_cpus to remaining un-assigned ncpus. 'numvecs' is
> > + * bigger than number of active numa nodes. Always start the
> > + * allocation from the node with minimized nr_cpus.
> > + *
> > + * This way guarantees that each active node gets allocated at
> > + * least one vector, and the theory is simple: over-allocation
> > + * is only done when this node is assigned by one vector, so
> > + * other nodes will be allocated >= 1 vector, since 'numvecs' is
> > + * bigger than number of numa nodes.
> > + *
> > + * One perfect invariant is that number of allocated vectors for
> > + * each node is <= CPU count of this node:
> > + *
> > + * 1) suppose there are two nodes: A and B
> > + * ncpu(X) is CPU count of node X
> > + * vecs(X) is the vector count allocated to node X via this
> > + * algorithm
> > + *
> > + * ncpu(A) <= ncpu(B)
> > + * ncpu(A) + ncpu(B) = N
> > + * vecs(A) + vecs(B) = V
> > + *
> > + * vecs(A) = max(1, round_down(V * ncpu(A) / N))
> > + * vecs(B) = V - vecs(A)
> > + *
> > + * both N and V are integer, and 2 <= V <= N, suppose
> > + * V = N - delta, and 0 <= delta <= N - 2
> > + *
> > + * 2) obviously vecs(A) <= ncpu(A) because:
> > + *
> > + * if vecs(A) is 1, then vecs(A) <= ncpu(A) given
> > + * ncpu(A) >= 1
> > + *
> > + * otherwise,
> > + * vecs(A) <= V * ncpu(A) / N <= ncpu(A), given V <= N
> > + *
> > + * 3) prove how vecs(B) <= ncpu(B):
> > + *
> > + * if round_down(V * ncpu(A) / N) == 0, vecs(B) won't be
> > + * over-allocated, so vecs(B) <= ncpu(B),
> > + *
> > + * otherwise:
> > + *
> > + * vecs(A) =
> > + * round_down(V * ncpu(A) / N) =
> > + * round_down((N - delta) * ncpu(A) / N) =
> > + * round_down((N * ncpu(A) - delta * ncpu(A)) / N) >=
> > + * round_down((N * ncpu(A) - delta * N) / N) =
> > + * cpu(A) - delta
> > + *
> > + * then:
> > + *
> > + * vecs(A) - V >= ncpu(A) - delta - V
> > + * =>
> > + * V - vecs(A) <= V + delta - ncpu(A)
> > + * =>
> > + * vecs(B) <= N - ncpu(A)
> > + * =>
> > + * vecs(B) <= cpu(B)
> > + *
> > + * For nodes >= 3, it can be thought as one node and another big
> > + * node given that is exactly what this algorithm is implemented,
> > + * and we always re-calculate 'remaining_ncpus' & 'numvecs', and
> > + * finally for each node X: vecs(X) <= ncpu(X).
> > + *
> > + */
> > + for (n = 0; n < nr_node_ids; n++) {
> > + unsigned nvectors, ncpus;
> > +
> > + if (node_vectors[n].ncpus == UINT_MAX)
> > + continue;
> > +
> > + WARN_ON_ONCE(numvecs == 0);
> > +
> > + ncpus = node_vectors[n].ncpus;
> > + nvectors = max_t(unsigned, 1,
> > + numvecs * ncpus / remaining_ncpus);
> > + WARN_ON_ONCE(nvectors > ncpus);
> > +
> > + node_vectors[n].nvectors = nvectors;
> > +
> > + remaining_ncpus -= ncpus;
> > + numvecs -= nvectors;
> > + }
> > +}
> > +
> > static int __irq_build_affinity_masks(unsigned int startvec,
> > unsigned int numvecs,
> > unsigned int firstvec,
> > @@ -102,10 +252,11 @@ static int __irq_build_affinity_masks(unsigned int startvec,
> > struct cpumask *nmsk,
> > struct irq_affinity_desc *masks)
> > {
> > - unsigned int n, nodes, cpus_per_vec, extra_vecs, done = 0;
> > + unsigned int i, n, nodes, cpus_per_vec, extra_vecs, done = 0;
> > unsigned int last_affv = firstvec + numvecs;
> > unsigned int curvec = startvec;
> > nodemask_t nodemsk = NODE_MASK_NONE;
> > + struct node_vectors *node_vectors;
> >
> > if (!cpumask_weight(cpu_mask))
> > return 0;
> > @@ -126,53 +277,57 @@ static int __irq_build_affinity_masks(unsigned int startvec,
> > return numvecs;
> > }
> >
> > - for_each_node_mask(n, nodemsk) {
> > - unsigned int ncpus, v, vecs_to_assign, vecs_per_node;
> > + node_vectors = kcalloc(nr_node_ids,
> > + sizeof(struct node_vectors),
> > + GFP_KERNEL);
> > + if (!node_vectors)
> > + return -ENOMEM;
> > +
> > + /* allocate vector number for each node */
> > + alloc_nodes_vectors(numvecs, node_to_cpumask, cpu_mask,
> > + nodemsk, nmsk, node_vectors);
> > +
> > + for (i = 0; i < nr_node_ids; i++) {
> > + unsigned int ncpus, v;
> > + struct node_vectors *nv = &node_vectors[i];
> > +
> > + if (nv->nvectors == UINT_MAX)
> > + continue;
> >
> > /* Get the cpus on this node which are in the mask */
> > - cpumask_and(nmsk, cpu_mask, node_to_cpumask[n]);
> > + cpumask_and(nmsk, cpu_mask, node_to_cpumask[nv->id]);
> > ncpus = cpumask_weight(nmsk);
> > if (!ncpus)
> > continue;
> >
> > - /*
> > - * Calculate the number of cpus per vector
> > - *
> > - * Spread the vectors evenly per node. If the requested
> > - * vector number has been reached, simply allocate one
> > - * vector for each remaining node so that all nodes can
> > - * be covered
> > - */
> > - if (numvecs > done)
> > - vecs_per_node = max_t(unsigned,
> > - (numvecs - done) / nodes, 1);
> > - else
> > - vecs_per_node = 1;
> > -
> > - vecs_to_assign = min(vecs_per_node, ncpus);
> > + WARN_ON_ONCE(nv->nvectors > ncpus);
> >
> > /* Account for rounding errors */
> > - extra_vecs = ncpus - vecs_to_assign * (ncpus / vecs_to_assign);
> > + extra_vecs = ncpus - nv->nvectors * (ncpus / nv->nvectors);
> >
> > - for (v = 0; curvec < last_affv && v < vecs_to_assign;
> > - curvec++, v++) {
> > - cpus_per_vec = ncpus / vecs_to_assign;
> > + /* Spread allocated vectors on CPUs of the current node */
> > + for (v = 0; v < nv->nvectors; v++, curvec++) {
> > + cpus_per_vec = ncpus / nv->nvectors;
> >
> > /* Account for extra vectors to compensate rounding errors */
> > if (extra_vecs) {
> > cpus_per_vec++;
> > --extra_vecs;
> > }
> > +
> > + /*
> > + * wrapping has to be considered given 'startvec'
> > + * may start anywhere
> > + */
> > + if (curvec >= last_affv)
> > + curvec = firstvec;
> > irq_spread_init_one(&masks[curvec].mask, nmsk,
> > cpus_per_vec);
> > }
> > -
> > - done += v;
> > - if (curvec >= last_affv)
> > - curvec = firstvec;
> > - --nodes;
> > + done += nv->nvectors;
> > }
> > - return done < numvecs ? done : numvecs;
> > + kfree(node_vectors);
> > + return done;
> > }
> >
> > /*
> > @@ -208,6 +363,10 @@ static int irq_build_affinity_masks(unsigned int startvec, unsigned int numvecs,
> > nr_present = __irq_build_affinity_masks(curvec, numvecs,
> > firstvec, node_to_cpumask,
> > cpu_present_mask, nmsk, masks);
> > + if (nr_present < 0) {
> > + ret = nr_present;
> > + goto fail_build_affinity;
> > + }
> >
> > /*
> > * Spread on non present CPUs starting from the next vector to be
> > @@ -223,9 +382,13 @@ static int irq_build_affinity_masks(unsigned int startvec, unsigned int numvecs,
> > nr_others = __irq_build_affinity_masks(curvec, numvecs,
> > firstvec, node_to_cpumask,
> > npresmsk, nmsk, masks);
> > + if (nr_others < 0)
> > + ret = nr_others;
> > +
> > + fail_build_affinity:
> > put_online_cpus();
> >
> > - if (nr_present < numvecs)
> > + if (min(nr_present, nr_others) >= 0)
> > WARN_ON(nr_present + nr_others < numvecs);
> >
> > free_node_to_cpumask(node_to_cpumask);
> > --
> > 2.20.1
> >
>
> _______________________________________________
> Linux-nvme mailing list
> [email protected]
> http://lists.infradead.org/mailman/listinfo/linux-nvme
--
Ming Lei
On Fri, 2019-08-16 at 09:53 -0600, Keith Busch wrote:
> On Thu, Aug 15, 2019 at 07:28:49PM -0700, Ming Lei wrote:
> > Now __irq_build_affinity_masks() spreads vectors evenly per node, and
> > all vectors may not be spread in case that each numa node has different
> > CPU number, then the warning in irq_build_affinity_masks() can
> > be triggered.
> >
> > Improve current spreading algorithm by assigning vectors according to
> > the ratio of node's nr_cpu to nr_remaining_cpus, meantime running the
> > assignment from smaller nodes to bigger nodes to guarantee that every
> > active node gets allocated at least one vector, then we can avoid
> > cross-node spread in normal situation.
> >
> > Meantime the reported warning can be fixed.
> >
> > Another big goodness is that the spread approach becomes more fair if
> > node has different CPU number.
> >
> > For example, on the following machine:
> > [root@ktest-01 ~]# lscpu
> > ...
> > CPU(s): 16
> > On-line CPU(s) list: 0-15
> > Thread(s) per core: 1
> > Core(s) per socket: 8
> > Socket(s): 2
> > NUMA node(s): 2
> > ...
> > NUMA node0 CPU(s): 0,1,3,5-9,11,13-15
> > NUMA node1 CPU(s): 2,4,10,12
> >
> > When driver requests to allocate 8 vectors, the following spread can
> > be got:
> > irq 31, cpu list 2,4
> > irq 32, cpu list 10,12
> > irq 33, cpu list 0-1
> > irq 34, cpu list 3,5
> > irq 35, cpu list 6-7
> > irq 36, cpu list 8-9
> > irq 37, cpu list 11,13
> > irq 38, cpu list 14-15
> >
> > Without this patch, kernel warning is triggered on above situation, and
> > allocation result was supposed to be 4 vectors for each node.
> >
> > Cc: Christoph Hellwig <[email protected]>
> > Cc: Keith Busch <[email protected]>
> > Cc: [email protected],
> > Cc: Jon Derrick <[email protected]>
> > Cc: Jens Axboe <[email protected]>
> > Reported-by: Jon Derrick <[email protected]>
> > Signed-off-by: Ming Lei <[email protected]>
>
> I had every intention to thoroughly test this on imbalanced node
> configurations, but that's not going to happen anytime soon. It looks
> correct to me, so I'll append my review here.
>
I can only test this with 2 nodes but I have varied nr_cpus as well as
using different devices with fewer and more vectors than CPUs. Spread
looks good.
Thank you
Reviewed-by: Jon Derrick <[email protected]>
[snip]
The following commit has been merged into the irq/core branch of tip:
Commit-ID: 53c1788b7d7720565214a466afffdc818d8c6e5f
Gitweb: https://git.kernel.org/tip/53c1788b7d7720565214a466afffdc818d8c6e5f
Author: Ming Lei <[email protected]>
AuthorDate: Fri, 16 Aug 2019 10:28:48 +08:00
Committer: Thomas Gleixner <[email protected]>
CommitterDate: Tue, 27 Aug 2019 16:31:17 +02:00
genirq/affinity: Improve __irq_build_affinity_masks()
One invariant of __irq_build_affinity_masks() is that all CPUs in the
specified masks (cpu_mask AND node_to_cpumask for each node) should be
covered during the spread. Even though all requested vectors have been
reached, it's still required to spread vectors among remained CPUs. A
similar policy has been taken in case of 'numvecs <= nodes' already.
So remove the following check inside the loop:
if (done >= numvecs)
break;
Meantime assign at least 1 vector for remaining nodes if 'numvecs' vectors
have been handled already.
Also, if the specified cpumask for one numa node is empty, simply do not
spread vectors on this node.
Signed-off-by: Ming Lei <[email protected]>
Signed-off-by: Thomas Gleixner <[email protected]>
Link: https://lkml.kernel.org/r/[email protected]
---
kernel/irq/affinity.c | 26 ++++++++++++++++++--------
1 file changed, 18 insertions(+), 8 deletions(-)
diff --git a/kernel/irq/affinity.c b/kernel/irq/affinity.c
index 6fef480..c7cca94 100644
--- a/kernel/irq/affinity.c
+++ b/kernel/irq/affinity.c
@@ -129,14 +129,26 @@ static int __irq_build_affinity_masks(unsigned int startvec,
for_each_node_mask(n, nodemsk) {
unsigned int ncpus, v, vecs_to_assign, vecs_per_node;
- /* Spread the vectors per node */
- vecs_per_node = (numvecs - (curvec - firstvec)) / nodes;
-
/* Get the cpus on this node which are in the mask */
cpumask_and(nmsk, cpu_mask, node_to_cpumask[n]);
-
- /* Calculate the number of cpus per vector */
ncpus = cpumask_weight(nmsk);
+ if (!ncpus)
+ continue;
+
+ /*
+ * Calculate the number of cpus per vector
+ *
+ * Spread the vectors evenly per node. If the requested
+ * vector number has been reached, simply allocate one
+ * vector for each remaining node so that all nodes can
+ * be covered
+ */
+ if (numvecs > done)
+ vecs_per_node = max_t(unsigned,
+ (numvecs - done) / nodes, 1);
+ else
+ vecs_per_node = 1;
+
vecs_to_assign = min(vecs_per_node, ncpus);
/* Account for rounding errors */
@@ -156,13 +168,11 @@ static int __irq_build_affinity_masks(unsigned int startvec,
}
done += v;
- if (done >= numvecs)
- break;
if (curvec >= last_affv)
curvec = firstvec;
--nodes;
}
- return done;
+ return done < numvecs ? done : numvecs;
}
/*
The following commit has been merged into the irq/core branch of tip:
Commit-ID: b1a5a73e64e99faa5f4deef2ae96d7371a0fb5d0
Gitweb: https://git.kernel.org/tip/b1a5a73e64e99faa5f4deef2ae96d7371a0fb5d0
Author: Ming Lei <[email protected]>
AuthorDate: Fri, 16 Aug 2019 10:28:49 +08:00
Committer: Thomas Gleixner <[email protected]>
CommitterDate: Tue, 27 Aug 2019 16:31:17 +02:00
genirq/affinity: Spread vectors on node according to nr_cpu ratio
Now __irq_build_affinity_masks() spreads vectors evenly per node, but there
is a case that not all vectors have been spread when each numa node has a
different number of CPUs which triggers the warning in the spreading code.
Improve the spreading algorithm by
- assigning vectors according to the ratio of the number of CPUs on a node
to the number of remaining CPUs.
- running the assignment from smaller nodes to bigger nodes to guarantee
that every active node gets allocated at least one vector.
This ensures that all vectors are spread out. Asided of that the spread
becomes more fair if the nodes have different number of CPUs.
For example, on the following machine:
CPU(s): 16
On-line CPU(s) list: 0-15
Thread(s) per core: 1
Core(s) per socket: 8
Socket(s): 2
NUMA node(s): 2
...
NUMA node0 CPU(s): 0,1,3,5-9,11,13-15
NUMA node1 CPU(s): 2,4,10,12
When a driver requests to allocate 8 vectors, the following spread results:
irq 31, cpu list 2,4
irq 32, cpu list 10,12
irq 33, cpu list 0-1
irq 34, cpu list 3,5
irq 35, cpu list 6-7
irq 36, cpu list 8-9
irq 37, cpu list 11,13
irq 38, cpu list 14-15
So Node 0 has now 6 and Node 1 has 2 vectors assigned. The original
algorithm assigned 4 vectors on each node which was unfair versus Node 0.
[ tglx: Massaged changelog ]
Reported-by: Jon Derrick <[email protected]>
Signed-off-by: Ming Lei <[email protected]>
Signed-off-by: Thomas Gleixner <[email protected]>
Reviewed-by: Keith Busch <[email protected]>
Reviewed-by: Jon Derrick <[email protected]>
Link: https://lkml.kernel.org/r/[email protected]
---
kernel/irq/affinity.c | 239 ++++++++++++++++++++++++++++++++++-------
1 file changed, 200 insertions(+), 39 deletions(-)
diff --git a/kernel/irq/affinity.c b/kernel/irq/affinity.c
index c7cca94..d905e84 100644
--- a/kernel/irq/affinity.c
+++ b/kernel/irq/affinity.c
@@ -7,6 +7,7 @@
#include <linux/kernel.h>
#include <linux/slab.h>
#include <linux/cpu.h>
+#include <linux/sort.h>
static void irq_spread_init_one(struct cpumask *irqmsk, struct cpumask *nmsk,
unsigned int cpus_per_vec)
@@ -94,6 +95,155 @@ static int get_nodes_in_cpumask(cpumask_var_t *node_to_cpumask,
return nodes;
}
+struct node_vectors {
+ unsigned id;
+
+ union {
+ unsigned nvectors;
+ unsigned ncpus;
+ };
+};
+
+static int ncpus_cmp_func(const void *l, const void *r)
+{
+ const struct node_vectors *ln = l;
+ const struct node_vectors *rn = r;
+
+ return ln->ncpus - rn->ncpus;
+}
+
+/*
+ * Allocate vector number for each node, so that for each node:
+ *
+ * 1) the allocated number is >= 1
+ *
+ * 2) the allocated numbver is <= active CPU number of this node
+ *
+ * The actual allocated total vectors may be less than @numvecs when
+ * active total CPU number is less than @numvecs.
+ *
+ * Active CPUs means the CPUs in '@cpu_mask AND @node_to_cpumask[]'
+ * for each node.
+ */
+static void alloc_nodes_vectors(unsigned int numvecs,
+ const cpumask_var_t *node_to_cpumask,
+ const struct cpumask *cpu_mask,
+ const nodemask_t nodemsk,
+ struct cpumask *nmsk,
+ struct node_vectors *node_vectors)
+{
+ unsigned n, remaining_ncpus = 0;
+
+ for (n = 0; n < nr_node_ids; n++) {
+ node_vectors[n].id = n;
+ node_vectors[n].ncpus = UINT_MAX;
+ }
+
+ for_each_node_mask(n, nodemsk) {
+ unsigned ncpus;
+
+ cpumask_and(nmsk, cpu_mask, node_to_cpumask[n]);
+ ncpus = cpumask_weight(nmsk);
+
+ if (!ncpus)
+ continue;
+ remaining_ncpus += ncpus;
+ node_vectors[n].ncpus = ncpus;
+ }
+
+ numvecs = min_t(unsigned, remaining_ncpus, numvecs);
+
+ sort(node_vectors, nr_node_ids, sizeof(node_vectors[0]),
+ ncpus_cmp_func, NULL);
+
+ /*
+ * Allocate vectors for each node according to the ratio of this
+ * node's nr_cpus to remaining un-assigned ncpus. 'numvecs' is
+ * bigger than number of active numa nodes. Always start the
+ * allocation from the node with minimized nr_cpus.
+ *
+ * This way guarantees that each active node gets allocated at
+ * least one vector, and the theory is simple: over-allocation
+ * is only done when this node is assigned by one vector, so
+ * other nodes will be allocated >= 1 vector, since 'numvecs' is
+ * bigger than number of numa nodes.
+ *
+ * One perfect invariant is that number of allocated vectors for
+ * each node is <= CPU count of this node:
+ *
+ * 1) suppose there are two nodes: A and B
+ * ncpu(X) is CPU count of node X
+ * vecs(X) is the vector count allocated to node X via this
+ * algorithm
+ *
+ * ncpu(A) <= ncpu(B)
+ * ncpu(A) + ncpu(B) = N
+ * vecs(A) + vecs(B) = V
+ *
+ * vecs(A) = max(1, round_down(V * ncpu(A) / N))
+ * vecs(B) = V - vecs(A)
+ *
+ * both N and V are integer, and 2 <= V <= N, suppose
+ * V = N - delta, and 0 <= delta <= N - 2
+ *
+ * 2) obviously vecs(A) <= ncpu(A) because:
+ *
+ * if vecs(A) is 1, then vecs(A) <= ncpu(A) given
+ * ncpu(A) >= 1
+ *
+ * otherwise,
+ * vecs(A) <= V * ncpu(A) / N <= ncpu(A), given V <= N
+ *
+ * 3) prove how vecs(B) <= ncpu(B):
+ *
+ * if round_down(V * ncpu(A) / N) == 0, vecs(B) won't be
+ * over-allocated, so vecs(B) <= ncpu(B),
+ *
+ * otherwise:
+ *
+ * vecs(A) =
+ * round_down(V * ncpu(A) / N) =
+ * round_down((N - delta) * ncpu(A) / N) =
+ * round_down((N * ncpu(A) - delta * ncpu(A)) / N) >=
+ * round_down((N * ncpu(A) - delta * N) / N) =
+ * cpu(A) - delta
+ *
+ * then:
+ *
+ * vecs(A) - V >= ncpu(A) - delta - V
+ * =>
+ * V - vecs(A) <= V + delta - ncpu(A)
+ * =>
+ * vecs(B) <= N - ncpu(A)
+ * =>
+ * vecs(B) <= cpu(B)
+ *
+ * For nodes >= 3, it can be thought as one node and another big
+ * node given that is exactly what this algorithm is implemented,
+ * and we always re-calculate 'remaining_ncpus' & 'numvecs', and
+ * finally for each node X: vecs(X) <= ncpu(X).
+ *
+ */
+ for (n = 0; n < nr_node_ids; n++) {
+ unsigned nvectors, ncpus;
+
+ if (node_vectors[n].ncpus == UINT_MAX)
+ continue;
+
+ WARN_ON_ONCE(numvecs == 0);
+
+ ncpus = node_vectors[n].ncpus;
+ nvectors = max_t(unsigned, 1,
+ numvecs * ncpus / remaining_ncpus);
+ WARN_ON_ONCE(nvectors > ncpus);
+
+ node_vectors[n].nvectors = nvectors;
+
+ remaining_ncpus -= ncpus;
+ numvecs -= nvectors;
+ }
+}
+
static int __irq_build_affinity_masks(unsigned int startvec,
unsigned int numvecs,
unsigned int firstvec,
@@ -102,10 +252,11 @@ static int __irq_build_affinity_masks(unsigned int startvec,
struct cpumask *nmsk,
struct irq_affinity_desc *masks)
{
- unsigned int n, nodes, cpus_per_vec, extra_vecs, done = 0;
+ unsigned int i, n, nodes, cpus_per_vec, extra_vecs, done = 0;
unsigned int last_affv = firstvec + numvecs;
unsigned int curvec = startvec;
nodemask_t nodemsk = NODE_MASK_NONE;
+ struct node_vectors *node_vectors;
if (!cpumask_weight(cpu_mask))
return 0;
@@ -126,53 +277,57 @@ static int __irq_build_affinity_masks(unsigned int startvec,
return numvecs;
}
- for_each_node_mask(n, nodemsk) {
- unsigned int ncpus, v, vecs_to_assign, vecs_per_node;
+ node_vectors = kcalloc(nr_node_ids,
+ sizeof(struct node_vectors),
+ GFP_KERNEL);
+ if (!node_vectors)
+ return -ENOMEM;
+
+ /* allocate vector number for each node */
+ alloc_nodes_vectors(numvecs, node_to_cpumask, cpu_mask,
+ nodemsk, nmsk, node_vectors);
+
+ for (i = 0; i < nr_node_ids; i++) {
+ unsigned int ncpus, v;
+ struct node_vectors *nv = &node_vectors[i];
+
+ if (nv->nvectors == UINT_MAX)
+ continue;
/* Get the cpus on this node which are in the mask */
- cpumask_and(nmsk, cpu_mask, node_to_cpumask[n]);
+ cpumask_and(nmsk, cpu_mask, node_to_cpumask[nv->id]);
ncpus = cpumask_weight(nmsk);
if (!ncpus)
continue;
- /*
- * Calculate the number of cpus per vector
- *
- * Spread the vectors evenly per node. If the requested
- * vector number has been reached, simply allocate one
- * vector for each remaining node so that all nodes can
- * be covered
- */
- if (numvecs > done)
- vecs_per_node = max_t(unsigned,
- (numvecs - done) / nodes, 1);
- else
- vecs_per_node = 1;
-
- vecs_to_assign = min(vecs_per_node, ncpus);
+ WARN_ON_ONCE(nv->nvectors > ncpus);
/* Account for rounding errors */
- extra_vecs = ncpus - vecs_to_assign * (ncpus / vecs_to_assign);
+ extra_vecs = ncpus - nv->nvectors * (ncpus / nv->nvectors);
- for (v = 0; curvec < last_affv && v < vecs_to_assign;
- curvec++, v++) {
- cpus_per_vec = ncpus / vecs_to_assign;
+ /* Spread allocated vectors on CPUs of the current node */
+ for (v = 0; v < nv->nvectors; v++, curvec++) {
+ cpus_per_vec = ncpus / nv->nvectors;
/* Account for extra vectors to compensate rounding errors */
if (extra_vecs) {
cpus_per_vec++;
--extra_vecs;
}
+
+ /*
+ * wrapping has to be considered given 'startvec'
+ * may start anywhere
+ */
+ if (curvec >= last_affv)
+ curvec = firstvec;
irq_spread_init_one(&masks[curvec].mask, nmsk,
cpus_per_vec);
}
-
- done += v;
- if (curvec >= last_affv)
- curvec = firstvec;
- --nodes;
+ done += nv->nvectors;
}
- return done < numvecs ? done : numvecs;
+ kfree(node_vectors);
+ return done;
}
/*
@@ -184,7 +339,7 @@ static int irq_build_affinity_masks(unsigned int startvec, unsigned int numvecs,
unsigned int firstvec,
struct irq_affinity_desc *masks)
{
- unsigned int curvec = startvec, nr_present, nr_others;
+ unsigned int curvec = startvec, nr_present = 0, nr_others = 0;
cpumask_var_t *node_to_cpumask;
cpumask_var_t nmsk, npresmsk;
int ret = -ENOMEM;
@@ -199,15 +354,17 @@ static int irq_build_affinity_masks(unsigned int startvec, unsigned int numvecs,
if (!node_to_cpumask)
goto fail_npresmsk;
- ret = 0;
/* Stabilize the cpumasks */
get_online_cpus();
build_node_to_cpumask(node_to_cpumask);
/* Spread on present CPUs starting from affd->pre_vectors */
- nr_present = __irq_build_affinity_masks(curvec, numvecs,
- firstvec, node_to_cpumask,
- cpu_present_mask, nmsk, masks);
+ ret = __irq_build_affinity_masks(curvec, numvecs, firstvec,
+ node_to_cpumask, cpu_present_mask,
+ nmsk, masks);
+ if (ret < 0)
+ goto fail_build_affinity;
+ nr_present = ret;
/*
* Spread on non present CPUs starting from the next vector to be
@@ -220,12 +377,16 @@ static int irq_build_affinity_masks(unsigned int startvec, unsigned int numvecs,
else
curvec = firstvec + nr_present;
cpumask_andnot(npresmsk, cpu_possible_mask, cpu_present_mask);
- nr_others = __irq_build_affinity_masks(curvec, numvecs,
- firstvec, node_to_cpumask,
- npresmsk, nmsk, masks);
+ ret = __irq_build_affinity_masks(curvec, numvecs, firstvec,
+ node_to_cpumask, npresmsk, nmsk,
+ masks);
+ if (ret >= 0)
+ nr_others = ret;
+
+ fail_build_affinity:
put_online_cpus();
- if (nr_present < numvecs)
+ if (ret >= 0)
WARN_ON(nr_present + nr_others < numvecs);
free_node_to_cpumask(node_to_cpumask);
@@ -235,7 +396,7 @@ static int irq_build_affinity_masks(unsigned int startvec, unsigned int numvecs,
fail_nmsk:
free_cpumask_var(nmsk);
- return ret;
+ return ret < 0 ? ret : 0;
}
static void default_calc_sets(struct irq_affinity *affd, unsigned int affvecs)