From: John Garry <john.garry@huawei.com>
To: <will@kernel.org>, <robin.murphy@arm.com>
Cc: <joro@8bytes.org>, <trivial@kernel.org>,
<linux-arm-kernel@lists.infradead.org>,
<iommu@lists.linux-foundation.org>,
<linux-kernel@vger.kernel.org>, <linuxarm@huawei.com>,
<maz@kernel.org>, "John Garry" <john.garry@huawei.com>
Subject: [PATCH RFC v2 2/4] iommu/arm-smmu-v3: Calculate bits for prod and owner
Date: Tue, 16 Jun 2020 21:59:51 +0800 [thread overview]
Message-ID: <1592315993-164290-3-git-send-email-john.garry@huawei.com> (raw)
In-Reply-To: <1592315993-164290-1-git-send-email-john.garry@huawei.com>
Since the arm_smmu_ll_queue.prod will be for counting the "owner" value
and also HW prod pointer, calculate how many bits are available for and
used by each.
This is based on the number of possible CPUs in the system. And we require
that each CPU can issue a minimum of 2 commands per batch - 1 x CMD_SYNC
and at least 1 x other.
Ignoring limits of HW max_n_shift and HW cmdq memory allocation, approx 16K
is the max supported CPUs. For this, max_n_shift would be 14.
Signed-off-by: John Garry <john.garry@huawei.com>
---
drivers/iommu/arm-smmu-v3.c | 41 ++++++++++++++++++++++++++++++++++++-
1 file changed, 40 insertions(+), 1 deletion(-)
diff --git a/drivers/iommu/arm-smmu-v3.c b/drivers/iommu/arm-smmu-v3.c
index a8e814c652fe..c3562dc35d45 100644
--- a/drivers/iommu/arm-smmu-v3.c
+++ b/drivers/iommu/arm-smmu-v3.c
@@ -532,6 +532,8 @@ struct arm_smmu_ll_queue {
u8 __pad[SMP_CACHE_BYTES];
} ____cacheline_aligned_in_smp;
u32 max_n_shift;
+ u32 max_cmd_per_batch;
+ u32 owner_count_shift;
};
struct arm_smmu_queue {
@@ -1515,7 +1517,10 @@ static void arm_smmu_cmdq_batch_add(struct arm_smmu_device *smmu,
struct arm_smmu_cmdq_batch *cmds,
struct arm_smmu_cmdq_ent *cmd)
{
- if (cmds->num == CMDQ_BATCH_ENTRIES) {
+ struct arm_smmu_cmdq *q = &smmu->cmdq;
+ struct arm_smmu_ll_queue *llq = &q->q.llq;
+
+ if (cmds->num == llq->max_cmd_per_batch) {
arm_smmu_cmdq_issue_cmdlist(smmu, cmds->cmds, cmds->num, false);
cmds->num = 0;
}
@@ -3141,8 +3146,26 @@ static int arm_smmu_init_one_queue(struct arm_smmu_device *smmu,
unsigned long cons_off,
size_t dwords, const char *name)
{
+ int cpus = num_possible_cpus();
size_t qsz;
+ /*
+ * We can get the number of bits required for owner counting by
+ * log2(nr possible cpus) + 1
+ */
+ int bits_for_cmdq_owner = ilog2(cpus) + 1;
+ /*
+ * 1-bit for overflow, 1-bit for wrap, 1-bit extra to ensure prod+wrap
+ * does not overflow into CPU count.
+ */
+ int bits_available_for_prod = 32 - 3 - bits_for_cmdq_owner;
+ int entries_for_prod;
+
+ if (bits_available_for_prod < 1) /* How many CPUs??? */
+ return -ENOMEM;
+
+ q->llq.max_n_shift = min_t(int, q->llq.max_n_shift,
+ bits_available_for_prod);
do {
qsz = ((1 << q->llq.max_n_shift) * dwords) << 3;
q->base = dmam_alloc_coherent(smmu->dev, qsz, &q->base_dma,
@@ -3152,6 +3175,22 @@ static int arm_smmu_init_one_queue(struct arm_smmu_device *smmu,
q->llq.max_n_shift--;
} while (1);
+ entries_for_prod = 1 << q->llq.max_n_shift;
+
+ /*
+ * We need at least 2 commands in a batch (1 x CMD_SYNC and 1 x
+ * whatever else).
+ */
+ if (entries_for_prod < 2 * cpus)
+ return -ENOMEM;
+
+ /*
+ * When finding max_cmd_per_batch, deduct 1 entry per batch to take
+ * account CMD_SYNC
+ */
+ q->llq.max_cmd_per_batch = min_t(u32, (entries_for_prod - cpus) / cpus,
+ CMDQ_BATCH_ENTRIES);
+ q->llq.owner_count_shift = q->llq.max_n_shift + 2;
if (!q->base) {
dev_err(smmu->dev,
--
2.26.2
next prev parent reply other threads:[~2020-06-16 14:04 UTC|newest]
Thread overview: 5+ messages / expand[flat|nested] mbox.gz Atom feed top
2020-06-16 13:59 [PATCH RFC v2 0/4] iommu/arm-smmu-v3: Improve cmdq lock efficiency John Garry
2020-06-16 13:59 ` [PATCH v2 1/4] iommu/arm-smmu-v3: Fix trivial typo John Garry
2020-06-16 13:59 ` John Garry [this message]
2020-06-16 13:59 ` [PATCH RFC v2 3/4] iommu/arm-smmu-v3: Always issue a CMD_SYNC per batch John Garry
2020-06-16 13:59 ` [PATCH RFC v2 4/4] iommu/arm-smmu-v3: Remove cmpxchg() in arm_smmu_cmdq_issue_cmdlist() John Garry
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=1592315993-164290-3-git-send-email-john.garry@huawei.com \
--to=john.garry@huawei.com \
--cc=iommu@lists.linux-foundation.org \
--cc=joro@8bytes.org \
--cc=linux-arm-kernel@lists.infradead.org \
--cc=linux-kernel@vger.kernel.org \
--cc=linuxarm@huawei.com \
--cc=maz@kernel.org \
--cc=robin.murphy@arm.com \
--cc=trivial@kernel.org \
--cc=will@kernel.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox
all inboxes | Powered by JetHome®