DPDK patches and discussions
 help / color / mirror / Atom feed
From: "Sachin Saxena (OSS)" <sachin.saxena@oss.nxp.com>
To: dev@dpdk.org, ferruh.yigit@intel.com
Cc: Sachin Saxena <sachin.saxena@oss.nxp.com>
Subject: [dpdk-dev] [PATCH v1 2/8] bus/fslmc: run secondary debug app without blocklist devices
Date: Fri, 11 Sep 2020 13:29:05 +0530
Message-ID: <20200911075911.3214-3-sachin.saxena@oss.nxp.com> (raw)
In-Reply-To: <20200911075911.3214-1-sachin.saxena@oss.nxp.com>

From: Rohit Raj <rohit.raj@nxp.com>

dpaa2 hw impose limits on some HW access devices like DPMCP(Mnagement
control Port) and DPIO (HW portal). This causes issue in their usages in
shared uses in case of multi-process applications. It can overcome by using
allowlist/blocklist in primary and secondary applications.
However it imposes restrictions on standard debugging apps like
dpdk-procinfo, which can be used to debug any existing application.

This patch introduces reserving extra DPMCP and DPIO to be used by
secondary process if devices are not blocked previously in primary
application.
This leaves the last DPMCP and DPIO for the secondary process usages.

Signed-off-by: Rohit Raj <rohit.raj@nxp.com>
Reviewed-by: Sachin Saxena <sachin.saxena@oss.nxp.com>
---
 drivers/bus/fslmc/fslmc_vfio.c | 66 +++++++++++++++++++++++++++++++---
 1 file changed, 61 insertions(+), 5 deletions(-)

diff --git a/drivers/bus/fslmc/fslmc_vfio.c b/drivers/bus/fslmc/fslmc_vfio.c
index 9134ec552..6dd797632 100644
--- a/drivers/bus/fslmc/fslmc_vfio.c
+++ b/drivers/bus/fslmc/fslmc_vfio.c
@@ -802,10 +802,29 @@ fslmc_vfio_process_group(void)
 	int ret;
 	int found_mportal = 0;
 	struct rte_dpaa2_device *dev, *dev_temp;
+	bool is_dpmcp_in_blocklist = false, is_dpio_in_blocklist = false;
+	int dpmcp_count = 0, dpio_count = 0, current_device;
+
+	TAILQ_FOREACH_SAFE(dev, &rte_fslmc_bus.device_list, next, dev_temp) {
+		if (dev->dev_type == DPAA2_MPORTAL) {
+			dpmcp_count++;
+			if (dev->device.devargs &&
+			    dev->device.devargs->policy == RTE_DEV_BLACKLISTED)
+				is_dpmcp_in_blocklist = true;
+		}
+		if (dev->dev_type == DPAA2_IO) {
+			dpio_count++;
+			if (dev->device.devargs &&
+			    dev->device.devargs->policy == RTE_DEV_BLACKLISTED)
+				is_dpmcp_in_blocklist = true;
+		}
+	}
 
 	/* Search the MCP as that should be initialized first. */
+	current_device = 0;
 	TAILQ_FOREACH_SAFE(dev, &rte_fslmc_bus.device_list, next, dev_temp) {
 		if (dev->dev_type == DPAA2_MPORTAL) {
+			current_device++;
 			if (dev->device.devargs &&
 			    dev->device.devargs->policy == RTE_DEV_BLACKLISTED) {
 				DPAA2_BUS_LOG(DEBUG, "%s Blacklisted, skipping",
@@ -815,13 +834,24 @@ fslmc_vfio_process_group(void)
 				continue;
 			}
 
-			ret = fslmc_process_mcp(dev);
-			if (ret) {
-				DPAA2_BUS_ERR("Unable to map MC Portal");
-				return -1;
+			if (rte_eal_process_type() == RTE_PROC_SECONDARY &&
+			    !is_dpmcp_in_blocklist) {
+				if (dpmcp_count == 1 ||
+				    current_device != dpmcp_count) {
+					TAILQ_REMOVE(&rte_fslmc_bus.device_list,
+						     dev, next);
+					continue;
+				}
 			}
-			if (!found_mportal)
+
+			if (!found_mportal) {
+				ret = fslmc_process_mcp(dev);
+				if (ret) {
+					DPAA2_BUS_ERR("Unable to map MC Portal");
+					return -1;
+				}
 				found_mportal = 1;
+			}
 
 			TAILQ_REMOVE(&rte_fslmc_bus.device_list, dev, next);
 			free(dev);
@@ -838,7 +868,10 @@ fslmc_vfio_process_group(void)
 		return -1;
 	}
 
+	current_device = 0;
 	TAILQ_FOREACH_SAFE(dev, &rte_fslmc_bus.device_list, next, dev_temp) {
+		if (dev->dev_type == DPAA2_IO)
+			current_device++;
 		if (dev->device.devargs &&
 		    dev->device.devargs->policy == RTE_DEV_BLACKLISTED) {
 			DPAA2_BUS_LOG(DEBUG, "%s Blacklisted, skipping",
@@ -846,6 +879,14 @@ fslmc_vfio_process_group(void)
 			TAILQ_REMOVE(&rte_fslmc_bus.device_list, dev, next);
 			continue;
 		}
+		if (rte_eal_process_type() == RTE_PROC_SECONDARY &&
+		    dev->dev_type != DPAA2_ETH &&
+		    dev->dev_type != DPAA2_CRYPTO &&
+		    dev->dev_type != DPAA2_QDMA &&
+		    dev->dev_type != DPAA2_IO) {
+			TAILQ_REMOVE(&rte_fslmc_bus.device_list, dev, next);
+			continue;
+		}
 		switch (dev->dev_type) {
 		case DPAA2_ETH:
 		case DPAA2_CRYPTO:
@@ -882,6 +923,21 @@ fslmc_vfio_process_group(void)
 
 			break;
 		case DPAA2_IO:
+			if (!is_dpio_in_blocklist && dpio_count > 1) {
+				if (rte_eal_process_type() == RTE_PROC_SECONDARY
+				    && current_device != dpio_count) {
+					TAILQ_REMOVE(&rte_fslmc_bus.device_list,
+						     dev, next);
+					break;
+				}
+				if (rte_eal_process_type() == RTE_PROC_PRIMARY
+				    && current_device == dpio_count) {
+					TAILQ_REMOVE(&rte_fslmc_bus.device_list,
+						     dev, next);
+					break;
+				}
+			}
+
 			ret = fslmc_process_iodevices(dev);
 			if (ret) {
 				DPAA2_BUS_DEBUG("Dev (%s) init failed",
-- 
2.28.0


  parent reply	other threads:[~2020-09-11  7:59 UTC|newest]

Thread overview: 9+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2020-09-11  7:59 [dpdk-dev] [PATCH v1 0/8] enhancements and fixes for nxp dpaax platform Sachin Saxena (OSS)
2020-09-11  7:59 ` [dpdk-dev] [PATCH v1 1/8] bus/fslmc: fix atomic queues on nxp lx2 platform Sachin Saxena (OSS)
2020-09-11  7:59 ` Sachin Saxena (OSS) [this message]
2020-09-11  7:59 ` [dpdk-dev] [PATCH v1 3/8] net/dpaa2: fix check for key size Sachin Saxena (OSS)
2020-09-11  7:59 ` [dpdk-dev] [PATCH v1 4/8] bus/dpaa: RX/TX error queues Sachin Saxena (OSS)
2020-09-11  7:59 ` [dpdk-dev] [PATCH v1 5/8] bus/fslmc: fix usage of dpio blocklist flag Sachin Saxena (OSS)
2020-09-11  7:59 ` [dpdk-dev] [PATCH v1 6/8] bus/dpaa: enhance link status support Sachin Saxena (OSS)
2020-09-11  7:59 ` [dpdk-dev] [PATCH v1 7/8] bus/dpaa: send error packets to application Sachin Saxena (OSS)
2020-09-11  7:59 ` [dpdk-dev] [PATCH v1 8/8] net/dpaa: " Sachin Saxena (OSS)

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=20200911075911.3214-3-sachin.saxena@oss.nxp.com \
    --to=sachin.saxena@oss.nxp.com \
    --cc=dev@dpdk.org \
    --cc=ferruh.yigit@intel.com \
    /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

DPDK patches and discussions

This inbox may be cloned and mirrored by anyone:

	git clone --mirror https://inbox.dpdk.org/dev/0 dev/git/0.git

	# If you have public-inbox 1.1+ installed, you may
	# initialize and index your mirror using the following commands:
	public-inbox-init -V2 dev dev/ https://inbox.dpdk.org/dev \
		dev@dpdk.org
	public-inbox-index dev

Example config snippet for mirrors.
Newsgroup available over NNTP:
	nntp://inbox.dpdk.org/inbox.dpdk.dev


AGPL code for this site: git clone https://public-inbox.org/public-inbox.git