| lists.openwall.net | lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC | |
|
Open Source and information security mailing list archives
| ||
|
Message-Id: <20201002212327.18393-12-scott.branden@broadcom.com>
Date: Fri, 2 Oct 2020 14:23:24 -0700
From: Scott Branden <scott.branden@...adcom.com>
To: Arnd Bergmann <arnd@...db.de>,
Greg Kroah-Hartman <gregkh@...uxfoundation.org>,
Scott Branden <scott.branden@...adcom.com>,
Desmond Yan <desmond.yan@...adcom.com>
Cc: Kees Cook <keescook@...omium.org>, linux-kernel@...r.kernel.org,
bcm-kernel-feedback-list@...adcom.com,
Olof Johansson <olof@...om.net>
Subject: [PATCH v6 11/14] misc: bcm-vk: add BCM_VK_QSTATS
Add BCM_VK_QSTATS Kconfig option to allow for enabling debug VK
queue statistics.
These statistics keep track of max, abs_max, and average for the
messages queues.
Co-developed-by: Desmond Yan <desmond.yan@...adcom.com>
Signed-off-by: Desmond Yan <desmond.yan@...adcom.com>
Signed-off-by: Scott Branden <scott.branden@...adcom.com>
---
drivers/misc/bcm-vk/Kconfig | 14 +++++++++
drivers/misc/bcm-vk/bcm_vk_dev.c | 9 ++++++
drivers/misc/bcm-vk/bcm_vk_msg.c | 52 +++++++++++++++++++++++++++++++-
drivers/misc/bcm-vk/bcm_vk_msg.h | 12 ++++++++
4 files changed, 86 insertions(+), 1 deletion(-)
diff --git a/drivers/misc/bcm-vk/Kconfig b/drivers/misc/bcm-vk/Kconfig
index 2272e47655ed..a3a020b19e3b 100644
--- a/drivers/misc/bcm-vk/Kconfig
+++ b/drivers/misc/bcm-vk/Kconfig
@@ -13,3 +13,17 @@ config BCM_VK
accelerators via /dev/bcm-vk.N devices.
If unsure, say N.
+
+if BCM_VK
+
+config BCM_VK_QSTATS
+ bool "VK Queue Statistics"
+ help
+ Turn on to enable Queue Statistics.
+ These are useful for debugging purposes.
+ Some performance loss by enabling this debug config.
+ For properly operating PCIe hardware no need to enable this.
+
+ If unsure, say N.
+
+endif
diff --git a/drivers/misc/bcm-vk/bcm_vk_dev.c b/drivers/misc/bcm-vk/bcm_vk_dev.c
index 718badb53100..6c2370723e0a 100644
--- a/drivers/misc/bcm-vk/bcm_vk_dev.c
+++ b/drivers/misc/bcm-vk/bcm_vk_dev.c
@@ -1097,6 +1097,15 @@ static int bcm_vk_trigger_reset(struct bcm_vk *vk)
vkwrite32(vk, 0, BAR_0, BAR_INTF_VER);
memset(&vk->host_alert, 0, sizeof(vk->host_alert));
memset(&vk->peer_alert, 0, sizeof(vk->peer_alert));
+#if defined(CONFIG_BCM_VK_QSTATS)
+ /* clear qstats */
+ for (i = 0; i < VK_MSGQ_MAX_NR; i++) {
+ memset(&vk->to_v_msg_chan.qstats[i].qcnts, 0,
+ sizeof(vk->to_v_msg_chan.qstats[i].qcnts));
+ memset(&vk->to_h_msg_chan.qstats[i].qcnts, 0,
+ sizeof(vk->to_h_msg_chan.qstats[i].qcnts));
+ }
+#endif
/* clear 4096 bits of bitmap */
bitmap_clear(vk->bmap, 0, VK_MSG_ID_BITMAP_SIZE);
diff --git a/drivers/misc/bcm-vk/bcm_vk_msg.c b/drivers/misc/bcm-vk/bcm_vk_msg.c
index e31d41400199..6ba0a7a94dcc 100644
--- a/drivers/misc/bcm-vk/bcm_vk_msg.c
+++ b/drivers/misc/bcm-vk/bcm_vk_msg.c
@@ -91,6 +91,44 @@ u32 msgq_avail_space(const struct bcm_vk_msgq __iomem *msgq,
return (qinfo->q_size - msgq_occupied(msgq, qinfo) - 1);
}
+#if defined(CONFIG_BCM_VK_QSTATS)
+
+/* Use default value of 20000 rd/wr per update */
+#if !defined(BCM_VK_QSTATS_ACC_CNT)
+#define BCM_VK_QSTATS_ACC_CNT 20000
+#endif
+
+static void bcm_vk_update_qstats(struct bcm_vk *vk,
+ const char *tag,
+ struct bcm_vk_qstats *qstats,
+ u32 occupancy)
+{
+ struct bcm_vk_qs_cnts *qcnts = &qstats->qcnts;
+
+ if (occupancy > qcnts->max_occ) {
+ qcnts->max_occ = occupancy;
+ if (occupancy > qcnts->max_abs)
+ qcnts->max_abs = occupancy;
+ }
+
+ qcnts->acc_sum += occupancy;
+ if (++qcnts->cnt >= BCM_VK_QSTATS_ACC_CNT) {
+ /* log average and clear counters */
+ dev_dbg(&vk->pdev->dev,
+ "%s[%d]: Max: [%3d/%3d] Acc %d num %d, Aver %d\n",
+ tag, qstats->q_num,
+ qcnts->max_occ, qcnts->max_abs,
+ qcnts->acc_sum,
+ qcnts->cnt,
+ qcnts->acc_sum / qcnts->cnt);
+
+ qcnts->cnt = 0;
+ qcnts->max_occ = 0;
+ qcnts->acc_sum = 0;
+ }
+}
+#endif
+
/* number of retries when enqueue message fails before returning EAGAIN */
#define BCM_VK_H2VK_ENQ_RETRY 10
#define BCM_VK_H2VK_ENQ_RETRY_DELAY_MS 50
@@ -495,8 +533,12 @@ static int bcm_vk_msg_chan_init(struct bcm_vk_msg_chan *chan)
mutex_init(&chan->msgq_mutex);
spin_lock_init(&chan->pendq_lock);
- for (i = 0; i < VK_MSGQ_MAX_NR; i++)
+ for (i = 0; i < VK_MSGQ_MAX_NR; i++) {
INIT_LIST_HEAD(&chan->pendq[i]);
+#if defined(CONFIG_BCM_VK_QSTATS)
+ chan->qstats[i].q_num = i;
+#endif
+ }
return 0;
}
@@ -605,6 +647,10 @@ static int bcm_to_v_msg_enqueue(struct bcm_vk *vk, struct bcm_vk_wkent *entry)
avail = msgq_avail_space(msgq, qinfo);
+#if defined(CONFIG_BCM_VK_QSTATS)
+ bcm_vk_update_qstats(vk, "to_v", &chan->qstats[q_num],
+ qinfo->q_size - avail);
+#endif
/* if not enough space, return EAGAIN and let app handles it */
retry = 0;
while ((avail < entry->to_v_blks) &&
@@ -818,6 +864,10 @@ s32 bcm_to_h_msg_dequeue(struct bcm_vk *vk)
goto idx_err;
}
+#if defined(CONFIG_BCM_VK_QSTATS)
+ bcm_vk_update_qstats(vk, "to_h", &chan->qstats[q_num],
+ msgq_occupied(msgq, qinfo));
+#endif
num_blks = src_size + 1;
data = kzalloc(num_blks * VK_MSGQ_BLK_SIZE, GFP_KERNEL);
if (data) {
diff --git a/drivers/misc/bcm-vk/bcm_vk_msg.h b/drivers/misc/bcm-vk/bcm_vk_msg.h
index 637c5d662eb7..d00d9707bd01 100644
--- a/drivers/misc/bcm-vk/bcm_vk_msg.h
+++ b/drivers/misc/bcm-vk/bcm_vk_msg.h
@@ -125,6 +125,14 @@ struct bcm_vk_qs_cnts {
u32 max_abs; /* the abs max since reset */
};
+#if defined(CONFIG_BCM_VK_QSTATS)
+/* stats structure */
+struct bcm_vk_qstats {
+ u32 q_num;
+ struct bcm_vk_qs_cnts qcnts;
+};
+#endif
+
/* control channel structure for either to_v or to_h communication */
struct bcm_vk_msg_chan {
u32 q_nr;
@@ -138,6 +146,10 @@ struct bcm_vk_msg_chan {
struct list_head pendq[VK_MSGQ_MAX_NR];
/* static queue info from the sync */
struct bcm_vk_sync_qinfo sync_qinfo[VK_MSGQ_MAX_NR];
+#if defined(CONFIG_BCM_VK_QSTATS)
+ /* qstats */
+ struct bcm_vk_qstats qstats[VK_MSGQ_MAX_NR];
+#endif
};
/* total number of supported ctx, 32 ctx each for 5 components */
--
2.17.1
Powered by blists - more mailing lists