[v2,1/5] raw/zxdh: introduce zxdh raw device driver

Message ID 20240812073209.1924286-1-zhang.yong25@zte.com.cn (mailing list archive)
State Superseded
Delegated to: Thomas Monjalon
Headers
Series [v2,1/5] raw/zxdh: introduce zxdh raw device driver |

Checks

Context Check Description
ci/checkpatch warning coding style issues

Commit Message

Yong Zhang Aug. 12, 2024, 7:31 a.m. UTC
Introduce rawdev driver support for ZXDH which
can help to connect two separate hosts with each other.

Signed-off-by: Yong Zhang <zhang.yong25@zte.com.cn>
---
 MAINTAINERS                    |   5 +
 doc/guides/rawdevs/index.rst   |   1 +
 doc/guides/rawdevs/zxdh.rst    |  30 +++++
 drivers/raw/meson.build        |   1 +
 drivers/raw/zxdh/meson.build   |   5 +
 drivers/raw/zxdh/zxdh_rawdev.c | 220 +++++++++++++++++++++++++++++++++
 drivers/raw/zxdh/zxdh_rawdev.h | 118 ++++++++++++++++++
 7 files changed, 380 insertions(+)
 create mode 100644 doc/guides/rawdevs/zxdh.rst
 create mode 100644 drivers/raw/zxdh/meson.build
 create mode 100644 drivers/raw/zxdh/zxdh_rawdev.c
 create mode 100644 drivers/raw/zxdh/zxdh_rawdev.h

-- 
2.43.0
  

Comments

Yong Zhang Sept. 2, 2024, 8:55 a.m. UTC | #1
Hi David, Stephen,
   
     Could you please provide any feedback on the patch I submitted on August 12th?

Thanks
  
Yong Zhang Sept. 6, 2024, 6:20 a.m. UTC | #2
Hi David, Stephen,
   
I recently submitted a patch based on your suggestions. Could you please review it and let me know if any changes are needed?
Looking forward to your feedback!

Thanks
  
David Marchand Sept. 9, 2024, 8:31 a.m. UTC | #3
Hello,

On Fri, Sep 6, 2024 at 8:20 AM Yong Zhang <zhang.yong25@zte.com.cn> wrote:
>
> I recently submitted a patch based on your suggestions. Could you please review it and let me know if any changes are needed?
> Looking forward to your feedback!

Sorry, I don't have enough bandwidth to do an in-depth review.
  
Yong Zhang Sept. 9, 2024, 9:26 a.m. UTC | #4
Hi thomas,
   
Could you please provide feedback on the patch I submitted on August 12?
Any suggestions for improvement would be appreciated.


Thanks
  
Yong Zhang Sept. 20, 2024, 2:07 a.m. UTC | #5
Hi thomas,
   
Could you please provide feedback on the patch I submitted on August 12?
Any suggestions for improvement would be appreciated.


Thanks
  
Yong Zhang Sept. 23, 2024, 1:49 a.m. UTC | #6
Hi thomas,
   
Could you please provide feedback on the patch I submitted on August 12?
Any suggestions for improvement would be appreciated.


Thanks
  
Yong Zhang Sept. 26, 2024, 9:44 a.m. UTC | #7
Hi thomas,
   
I hope this message finds you well.
I initially submitted the zxdh rawdev driver on June 12th, and the final version was submitted on August 12th. Since then, I have not received any feedback from the community.
I would appreciate it if you could provide any suggestions for modifications. If there is no problem with the submitted driver, please merge it as soon as possible.
Thank you for your time and assistance.

Thanks
  
Yong Zhang Sept. 29, 2024, 1:26 a.m. UTC | #8
Hi thomas,
   
I hope this message finds you well.
I initially submitted the zxdh rawdev driver on June 12th, and the final version was submitted on August 12th. Since then, I have not received any feedback from the community.
I would appreciate it if you could provide any suggestions for modifications. If there is no problem with the submitted driver, please merge it as soon as possible.
Thank you for your time and assistance.

Thanks
  
Yong Zhang Oct. 8, 2024, 1:18 a.m. UTC | #9
Hi thomas,
   
I hope this message finds you well.
I initially submitted the zxdh rawdev driver on June 12th, and the final version was submitted on August 12th. Since then, I have not received any feedback from the community.
I would appreciate it if you could provide any suggestions for modifications. If there is no problem with the submitted driver, please merge it as soon as possible.
Thank you for your time and assistance.

Thanks
  
Stephen Hemminger Oct. 8, 2024, 2:58 a.m. UTC | #10
On Tue,  8 Oct 2024 09:18:44 +0800
Yong Zhang <zhang.yong25@zte.com.cn> wrote:

> Hi thomas,
>    
> I hope this message finds you well.
> I initially submitted the zxdh rawdev driver on June 12th, and the final version was submitted on August 12th. Since then, I have not received any feedback from the community.
> I would appreciate it if you could provide any suggestions for modifications. If there is no problem with the submitted driver, please merge it as soon as possible.
> Thank you for your time and assistance.
> 
> Thanks

Just looked at the patch today, and marked it as "Changes requested" based on
the feedback from Ferruh, and an attempt at merging it into local tree.

Check in patchwork:
https://patchwork.dpdk.org/project/dpdk/patch/20240910120020.4031035-1-wang.junlong1@zte.com.cn/
  
Yong Zhang Oct. 8, 2024, 6:40 a.m. UTC | #11
>> Hi thomas,>>    >> I hope this message finds you well.>> I initially submitted the zxdh rawdev driver on June 12th, and the final version was submitted on August 12th. Since then, I have not received any feedback from the community.>> I would appreciate it if you could provide any suggestions for modifications. If there is no problem with the submitted driver, please merge it as soon as possible.>> Thank you for your time and assistance.>> >> Thanks>Just looked at the patch today, and marked it as "Changes requested" based on>the feedback from Ferruh, and an attempt at merging it into local tree.>>Check in patchwork:>https://patchwork.dpdk.org/project/dpdk/patch/20240910120020.4031035-1-wang.junlong1@zte.com.cn/

Hi Stephen,

    I wanted to check if you might have sent your previous email to me by mistake. It seems like it might have been intended for someone else.
    Additionally, I wanted to clarify that there are two types of ZXDH drivers:



net/zxdh: Managed by Junlong Wang

raw/zxdh: Managed by me


    For your reference, here is the link to my patch: https://patches.dpdk.org/project/dpdk/patch/20240812073209.1924286-9-zhang.yong25@zte.com.cn/ 
 



Original


From: StephenHemminger <stephen@networkplumber.org>
To: 张勇10313449;
Cc: thomas@monjalon.net <thomas@monjalon.net>;techboard@dpdk.org <techboard@dpdk.org>;dev@dpdk.org <dev@dpdk.org>;汪勇10032886;李敏10314441;冉明10033339;
Date: 2024年10月08日 10:58
Subject: Re: [PATCH] raw/zxdh: introduce zxdh raw device driver

On Tue,  8 Oct 2024 09:18:44 +0800
Yong Zhang <zhang.yong25@zte.com.cn> wrote:
 
> Hi thomas,
>     
> I hope this message finds you well.
> I initially submitted the zxdh rawdev driver on June 12th, and the final version was submitted on August 12th. Since then, I have not received any feedback from the community.
> I would appreciate it if you could provide any suggestions for modifications. If there is no problem with the submitted driver, please merge it as soon as possible.
> Thank you for your time and assistance.
>  
> Thanks
 
Just looked at the patch today, and marked it as "Changes requested" based on
the feedback from Ferruh, and an attempt at merging it into local tree.
 
Check in patchwork:
https://patchwork.dpdk.org/project/dpdk/patch/20240910120020.4031035-1-wang.junlong1@zte.com.cn/
  
Stephen Hemminger Oct. 8, 2024, 4:12 p.m. UTC | #12
On Tue, 8 Oct 2024 14:40:11 +0800 (CST)
<zhang.yong25@zte.com.cn> wrote:

> >> Hi thomas,>>    >> I hope this message finds you well.>> I initially submitted the zxdh rawdev driver on June 12th, and the final version was submitted on August 12th. Since then, I have not received any feedback from the community.>> I would appreciate it if you could provide any suggestions for modifications. If there is no problem with the submitted driver, please merge it as soon as possible.>> Thank you for your time and assistance.>> >> Thanks>Just looked at the patch today, and marked it as "Changes requested" based on>the feedback from Ferruh, and an attempt at merging it into local tree.>>Check in patchwork:>https://patchwork.dpdk.org/project/dpdk/patch/20240910120020.4031035-1-wang.junlong1@zte.com.cn/  
> 
> Hi Stephen,
> 
>     I wanted to check if you might have sent your previous email to me by mistake. It seems like it might have been intended for someone else.
>     Additionally, I wanted to clarify that there are two types of ZXDH drivers:
> 
> 
> 
> net/zxdh: Managed by Junlong Wang
> 
> raw/zxdh: Managed by me

Sorry, I got confused.

The raw driver looks to be ready, I set the patch state back.
  
Stephen Hemminger Oct. 8, 2024, 4:21 p.m. UTC | #13
On Mon, 12 Aug 2024 15:31:24 +0800
Yong Zhang <zhang.yong25@zte.com.cn> wrote:

> diff --git a/doc/guides/rawdevs/zxdh.rst b/doc/guides/rawdevs/zxdh.rst
> new file mode 100644
> index 0000000000..fa7ada1004
> --- /dev/null
> +++ b/doc/guides/rawdevs/zxdh.rst
> @@ -0,0 +1,30 @@
> +..  SPDX-License-Identifier: BSD-3-Clause
> +    Copyright 2024 ZTE Corporation
> +
> +ZXDH Rawdev Driver
> +======================
> +
> +The ``zxdh`` rawdev driver is an implementation of the rawdev API,
> +that provides communication between two separate hosts.
> +This is achieved via using the GDMA controller of Dinghai SoC,
> +which can be configured through exposed MPF devices.

This is awkward use of passive voice and could be simplified and
shortened.

> +
> +Device Setup
> +-------------
> +
> +It is recommended to bind the ZXDH MPF kernel driver for MPF devices (Not mandatory).
> +The kernel drivers can be downloaded at `ZTE Official Website
> +<https://enterprise.zte.com.cn/>`_.

What is works (and what doesn't) without the kernel driver.
Has this driver been submitted to upstream kernel.org?
DPDK does not want to be requiring or using 3rd party kernel drivers with non GPLv2 licenses.

> +
> +Initialization
> +--------------
> +
> +The ``zxdh`` rawdev driver needs to work in IOVA PA mode.
> +Consider using ``--iova-mode=pa`` in the EAL options.
> +
> +Platform Requirement
> +~~~~~~~~~~~~~~~~~~~~
> +
> +This PMD is only supported on ZTE Neo Platforms:
> +- Neo X510/X512
> +


Adding blank line at end of file causes git to complain when merging

> diff --git a/drivers/raw/meson.build b/drivers/raw/meson.build
> index 05cad143fe..237d1bdd80 100644
> --- a/drivers/raw/meson.build
> +++ b/drivers/raw/meson.build
> @@ -12,5 +12,6 @@ drivers = [
>          'ifpga',
>          'ntb',
>          'skeleton',
> +        'zxdh',
>  ]
>  std_deps = ['rawdev']
> diff --git a/drivers/raw/zxdh/meson.build b/drivers/raw/zxdh/meson.build
> new file mode 100644
> index 0000000000..266d3db6d8
> --- /dev/null
> +++ b/drivers/raw/zxdh/meson.build
> @@ -0,0 +1,5 @@
> +#SPDX-License-Identifier: BSD-3-Clause
> +#Copyright 2024 ZTE Corporation
> +
> +deps += ['rawdev', 'kvargs', 'mbuf', 'bus_pci']
> +sources = files('zxdh_rawdev.c')
> diff --git a/drivers/raw/zxdh/zxdh_rawdev.c b/drivers/raw/zxdh/zxdh_rawdev.c
> new file mode 100644
> index 0000000000..269c4f92e0
> --- /dev/null
> +++ b/drivers/raw/zxdh/zxdh_rawdev.c
> @@ -0,0 +1,220 @@
> +/* SPDX-License-Identifier: BSD-3-Clause
> + * Copyright 2024 ZTE Corporation
> + */
> +
> +#include <assert.h>
> +#include <stdio.h>
> +#include <stdbool.h>
> +#include <errno.h>
> +#include <stdint.h>
> +#include <inttypes.h>
> +#include <string.h>
> +#include <sys/types.h>
> +#include <unistd.h>
> +#include <fcntl.h>
> +
> +#include <rte_byteorder.h>
> +#include <rte_errno.h>
> +#include <rte_common.h>
> +#include <rte_debug.h>
> +#include <rte_dev.h>
> +#include <rte_eal.h>
> +#include <rte_kvargs.h>
> +#include <rte_log.h>
> +#include <rte_malloc.h>
> +#include <rte_memory.h>
> +#include <rte_memcpy.h>
> +#include <rte_lcore.h>
> +#include <rte_cycles.h>
> +#include <rte_memzone.h>
> +#include <rte_atomic.h>
> +#include <rte_rawdev.h>
> +#include <rte_rawdev_pmd.h>
> +#include <rte_pci.h>
> +#include <bus_pci_driver.h>
> +#include <rte_eal_paging.h>

This is a large set of include and looks like it was a copy paste from elsewhere.

For example: you are including rte_kvargs.h but never use kvargs in the driver.

If I run iwyu on the driver it produces.

$ iwyu -I lib/eal/include -I drivers/bus/pci drivers/raw/zxdh/zxdh_rawdev.c

drivers/raw/zxdh/zxdh_rawdev.h should add these lines:
#include <generic/rte_spinlock.h>  // for rte_spinlock_t
#include <rte_log.h>               // for RTE_LOG_COMMA, RTE_LOG_LINE_PREFIX
#include <stdint.h>                // for uint32_t, uint16_t, uint8_t, uint64_t
#include "rte_common.h"            // for rte_iova_t

drivers/raw/zxdh/zxdh_rawdev.h should remove these lines:
- #include <rte_spinlock.h>  // lines 13-13

The full include-list for drivers/raw/zxdh/zxdh_rawdev.h:
#include <generic/rte_spinlock.h>  // for rte_spinlock_t
#include <rte_log.h>               // for RTE_LOG_COMMA, RTE_LOG_LINE_PREFIX
#include <rte_rawdev.h>            // for rte_rawdev
#include <stdint.h>                // for uint32_t, uint16_t, uint8_t, uint64_t
#include "rte_common.h"            // for rte_iova_t
---

drivers/raw/zxdh/zxdh_rawdev.c should add these lines:
#include <limits.h>                 // for PATH_MAX
#include <rte_spinlock.h>           // for rte_spinlock_lock, rte_spinlock_u...
#include "rte_branch_prediction.h"  // for unlikely

drivers/raw/zxdh/zxdh_rawdev.c should remove these lines:
- #include <assert.h>  // lines 5-5
- #include <rte_byteorder.h>  // lines 16-16
- #include <rte_cycles.h>  // lines 28-28
- #include <rte_debug.h>  // lines 19-19
- #include <rte_eal.h>  // lines 21-21
- #include <rte_kvargs.h>  // lines 22-22
- #include <rte_memcpy.h>  // lines 26-26
- #include <rte_memory.h>  // lines 25-25
- #include <stdbool.h>  // lines 7-7
- #include <stdint.h>  // lines 9-9
- #include <sys/types.h>  // lines 12-12

The full include-list for drivers/raw/zxdh/zxdh_rawdev.c:
#include "zxdh_rawdev.h"
#include <bus_pci_driver.h>         // for rte_pci_device, rte_pci_get_sysfs...
#include <errno.h>                  // for EINVAL, ENOMEM, EEXIST, errno
#include <fcntl.h>                  // for open, O_RDWR
#include <inttypes.h>               // for uint16_t, uint32_t, uint8_t, uint...
#include <limits.h>                 // for PATH_MAX
#include <rte_atomic.h>             // for rte_wmb
#include <rte_common.h>             // for __rte_unused, RTE_PRIORITY_LAST
#include <rte_dev.h>                // for rte_mem_resource, RTE_PMD_REGISTE...
#include <rte_eal_paging.h>         // for rte_mem_map, rte_mem_prot, rte_me...
#include <rte_errno.h>              // for rte_strerror, per_lcore__rte_errno
#include <rte_lcore.h>              // for rte_socket_id
#include <rte_log.h>                // for RTE_LOG_ERR, RTE_LOG_DEBUG, RTE_L...
#include <rte_malloc.h>             // for rte_free, rte_zmalloc
#include <rte_memzone.h>            // for rte_memzone, rte_memzone::(anonym...
#include <rte_pci.h>                // for rte_pci_addr, PCI_PRI_FMT, rte_pc...
#include <rte_rawdev.h>             // for rte_rawdev_obj_t, rte_rawdev, RTE...
#include <rte_rawdev_pmd.h>         // for rte_rawdev_pmd_allocate, rte_rawd...
#include <rte_spinlock.h>           // for rte_spinlock_lock, rte_spinlock_u...
#include <stdio.h>                  // for NULL, size_t, snprintf
#include <string.h>                 // for memset, strerror
#include <unistd.h>                 // for close
#include "rte_branch_prediction.h"  // for unlikely
  
Yong Zhang Oct. 14, 2024, 8:22 a.m. UTC | #14
Hi ALL,
   
To avoid confusion with net/zxdh, raw/zxdh is renamed to raw/gdtc.

Thanks
  
Morten Brørup Oct. 14, 2024, 8:33 a.m. UTC | #15
Dear Yong Zhang,

 

Please refrain from sending patches to the DPDK Technical Board (techboard@dpdk.org <mailto:techboard@dpdk.org> ).

 

If there is a conflict or issue with patch, which cannot be resolved on the DPDK Developer mailing list (dev@dpdk.org <mailto:dev@dpdk.org> ), the topic can be escalated to discussion by the DPDK Technical Board.

 

 

Med venlig hilsen / kind regards

 

Morten Brørup

CTO

 

 

SmartShare Systems A/S

Tonsbakken 16-18

DK-2740 Skovlunde

Denmark

 

Office      +45 70 20 00 93

Direct      +45 89 93 50 22

Mobile     +45 25 40 82 12

 

mb@smartsharesystems.com <mailto:mb@smartsharesystems.com> 

www.smartsharesystems.com <https://www.smartsharesystems.com/> 

 

From: Yong Zhang [mailto:zhang.yong25@zte.com.cn] 
Sent: Monday, 14 October 2024 10.23
To: thomas@monjalon.net; techboard@dpdk.org
Cc: dev@dpdk.org; zhang.yong25@zte.com.cn; wang.yong19@zte.com.cn; li.min10@zte.com.cn; ran.ming@zte.com.cn
Subject: Re:[PATCH] raw/gdtc: introduce gdtc raw device driver

 

Hi ALL,
    
To avoid confusion with net/zxdh, raw/zxdh is renamed to raw/gdtc.

Thanks
  

Patch

diff --git a/MAINTAINERS b/MAINTAINERS
index c5a703b5c0..6dd4fbae6e 100644
--- a/MAINTAINERS
+++ b/MAINTAINERS
@@ -1511,6 +1511,11 @@  M: Gagandeep Singh <g.singh@nxp.com>
 F: drivers/raw/dpaa2_cmdif/
 F: doc/guides/rawdevs/dpaa2_cmdif.rst
 
+ZXDH
+M: Yong Zhang <zhang.yong25@zte.com.cn>
+F: drivers/raw/zxdh/
+F: doc/guides/rawdevs/zxdh.rst
+
 
 Packet processing
 -----------------
diff --git a/doc/guides/rawdevs/index.rst b/doc/guides/rawdevs/index.rst
index f34315f051..d85a4b7148 100644
--- a/doc/guides/rawdevs/index.rst
+++ b/doc/guides/rawdevs/index.rst
@@ -16,3 +16,4 @@  application through rawdev API.
     dpaa2_cmdif
     ifpga
     ntb
+    zxdh
diff --git a/doc/guides/rawdevs/zxdh.rst b/doc/guides/rawdevs/zxdh.rst
new file mode 100644
index 0000000000..fa7ada1004
--- /dev/null
+++ b/doc/guides/rawdevs/zxdh.rst
@@ -0,0 +1,30 @@ 
+..  SPDX-License-Identifier: BSD-3-Clause
+    Copyright 2024 ZTE Corporation
+
+ZXDH Rawdev Driver
+======================
+
+The ``zxdh`` rawdev driver is an implementation of the rawdev API,
+that provides communication between two separate hosts.
+This is achieved via using the GDMA controller of Dinghai SoC,
+which can be configured through exposed MPF devices.
+
+Device Setup
+-------------
+
+It is recommended to bind the ZXDH MPF kernel driver for MPF devices (Not mandatory).
+The kernel drivers can be downloaded at `ZTE Official Website
+<https://enterprise.zte.com.cn/>`_.
+
+Initialization
+--------------
+
+The ``zxdh`` rawdev driver needs to work in IOVA PA mode.
+Consider using ``--iova-mode=pa`` in the EAL options.
+
+Platform Requirement
+~~~~~~~~~~~~~~~~~~~~
+
+This PMD is only supported on ZTE Neo Platforms:
+- Neo X510/X512
+
diff --git a/drivers/raw/meson.build b/drivers/raw/meson.build
index 05cad143fe..237d1bdd80 100644
--- a/drivers/raw/meson.build
+++ b/drivers/raw/meson.build
@@ -12,5 +12,6 @@  drivers = [
         'ifpga',
         'ntb',
         'skeleton',
+        'zxdh',
 ]
 std_deps = ['rawdev']
diff --git a/drivers/raw/zxdh/meson.build b/drivers/raw/zxdh/meson.build
new file mode 100644
index 0000000000..266d3db6d8
--- /dev/null
+++ b/drivers/raw/zxdh/meson.build
@@ -0,0 +1,5 @@ 
+#SPDX-License-Identifier: BSD-3-Clause
+#Copyright 2024 ZTE Corporation
+
+deps += ['rawdev', 'kvargs', 'mbuf', 'bus_pci']
+sources = files('zxdh_rawdev.c')
diff --git a/drivers/raw/zxdh/zxdh_rawdev.c b/drivers/raw/zxdh/zxdh_rawdev.c
new file mode 100644
index 0000000000..269c4f92e0
--- /dev/null
+++ b/drivers/raw/zxdh/zxdh_rawdev.c
@@ -0,0 +1,220 @@ 
+/* SPDX-License-Identifier: BSD-3-Clause
+ * Copyright 2024 ZTE Corporation
+ */
+
+#include <assert.h>
+#include <stdio.h>
+#include <stdbool.h>
+#include <errno.h>
+#include <stdint.h>
+#include <inttypes.h>
+#include <string.h>
+#include <sys/types.h>
+#include <unistd.h>
+#include <fcntl.h>
+
+#include <rte_byteorder.h>
+#include <rte_errno.h>
+#include <rte_common.h>
+#include <rte_debug.h>
+#include <rte_dev.h>
+#include <rte_eal.h>
+#include <rte_kvargs.h>
+#include <rte_log.h>
+#include <rte_malloc.h>
+#include <rte_memory.h>
+#include <rte_memcpy.h>
+#include <rte_lcore.h>
+#include <rte_cycles.h>
+#include <rte_memzone.h>
+#include <rte_atomic.h>
+#include <rte_rawdev.h>
+#include <rte_rawdev_pmd.h>
+#include <rte_pci.h>
+#include <bus_pci_driver.h>
+#include <rte_eal_paging.h>
+
+#include "zxdh_rawdev.h"
+
+/* Register offset */
+#define ZXDH_GDMA_BASE_OFFSET                   0x100000
+
+#define ZXDH_GDMA_CHAN_SHIFT                    0x80
+char zxdh_gdma_driver_name[] = "rawdev_zxdh_gdma";
+char dev_name[] = "zxdh_gdma";
+
+uint32_t
+zxdh_gdma_read_reg(struct rte_rawdev *dev, uint16_t queue_id, uint32_t offset)
+{
+	struct zxdh_gdma_rawdev *gdmadev = zxdh_gdma_rawdev_get_priv(dev);
+	uint32_t addr = 0;
+	uint32_t val = 0;
+
+	addr = offset + queue_id * ZXDH_GDMA_CHAN_SHIFT;
+	val = *(uint32_t *)(gdmadev->base_addr + addr);
+
+	return val;
+}
+
+void
+zxdh_gdma_write_reg(struct rte_rawdev *dev, uint16_t queue_id, uint32_t offset, uint32_t val)
+{
+	struct zxdh_gdma_rawdev *gdmadev = zxdh_gdma_rawdev_get_priv(dev);
+	uint32_t addr = 0;
+
+	addr = offset + queue_id * ZXDH_GDMA_CHAN_SHIFT;
+	*(uint32_t *)(gdmadev->base_addr + addr) = val;
+}
+
+static const struct rte_rawdev_ops zxdh_gdma_rawdev_ops = {
+};
+
+static int
+zxdh_gdma_map_resource(struct rte_pci_device *dev)
+{
+	int fd = -1;
+	char devname[PATH_MAX];
+	void *mapaddr = NULL;
+	struct rte_pci_addr *loc;
+
+	loc = &dev->addr;
+	snprintf(devname, sizeof(devname), "%s/" PCI_PRI_FMT "/resource0",
+		rte_pci_get_sysfs_path(),
+		loc->domain, loc->bus, loc->devid,
+		loc->function);
+
+		fd = open(devname, O_RDWR);
+		if (fd < 0) {
+			ZXDH_PMD_LOG(ERR, "Cannot open %s: %s", devname, strerror(errno));
+			return -1;
+		}
+
+	/* Map the PCI memory resource of device */
+	mapaddr = rte_mem_map(NULL, (size_t)dev->mem_resource[0].len,
+				RTE_PROT_READ | RTE_PROT_WRITE,
+				RTE_MAP_SHARED, fd, 0);
+	if (mapaddr == NULL) {
+		ZXDH_PMD_LOG(ERR, "cannot map resource(%d, 0x%zx): %s (%p)",
+				fd, (size_t)dev->mem_resource[0].len,
+				rte_strerror(rte_errno), mapaddr);
+		close(fd);
+		return -1;
+	}
+
+	close(fd);
+	dev->mem_resource[0].addr = mapaddr;
+
+	return 0;
+}
+
+static void
+zxdh_gdma_unmap_resource(void *requested_addr, size_t size)
+{
+	if (requested_addr == NULL)
+		return;
+
+	/* Unmap the PCI memory resource of device */
+	if (rte_mem_unmap(requested_addr, size))
+		ZXDH_PMD_LOG(ERR, "cannot mem unmap(%p, %#zx): %s",
+			requested_addr, size, rte_strerror(rte_errno));
+	else
+		ZXDH_PMD_LOG(DEBUG, "PCI memory unmapped at %p", requested_addr);
+}
+
+static int
+zxdh_gdma_rawdev_probe(struct rte_pci_driver *pci_drv __rte_unused,
+						struct rte_pci_device *pci_dev)
+{
+	struct rte_rawdev *dev = NULL;
+	struct zxdh_gdma_rawdev *gdmadev = NULL;
+	struct zxdh_gdma_queue *queue = NULL;
+	uint8_t i = 0;
+	int ret;
+
+	if (pci_dev->mem_resource[0].phys_addr == 0) {
+		ZXDH_PMD_LOG(ERR, "PCI bar0 resource is invalid");
+		return -1;
+	}
+
+	ret = zxdh_gdma_map_resource(pci_dev);
+	if (ret != 0) {
+		ZXDH_PMD_LOG(ERR, "Failed to mmap pci device(%s)", pci_dev->name);
+		return -1;
+	}
+	ZXDH_PMD_LOG(INFO, "%s bar0 0x%"PRIx64" mapped at %p",
+				pci_dev->name, pci_dev->mem_resource[0].phys_addr,
+				pci_dev->mem_resource[0].addr);
+
+	dev = rte_rawdev_pmd_allocate(dev_name, sizeof(struct zxdh_gdma_rawdev), rte_socket_id());
+	if (dev == NULL) {
+		ZXDH_PMD_LOG(ERR, "Unable to allocate gdma rawdev");
+		goto err_out;
+	}
+	ZXDH_PMD_LOG(INFO, "Init %s on NUMA node %d, dev_id is %d",
+						dev_name, rte_socket_id(), dev->dev_id);
+
+	dev->dev_ops = &zxdh_gdma_rawdev_ops;
+	dev->device = &pci_dev->device;
+	dev->driver_name = zxdh_gdma_driver_name;
+	gdmadev = zxdh_gdma_rawdev_get_priv(dev);
+	gdmadev->device_state = ZXDH_GDMA_DEV_STOPPED;
+	gdmadev->rawdev = dev;
+	gdmadev->queue_num = ZXDH_GDMA_TOTAL_CHAN_NUM;
+	gdmadev->used_num = 0;
+	gdmadev->base_addr = (uintptr_t)pci_dev->mem_resource[0].addr + ZXDH_GDMA_BASE_OFFSET;
+
+	for (i = 0; i < ZXDH_GDMA_TOTAL_CHAN_NUM; i++) {
+		queue = &(gdmadev->vqs[i]);
+		queue->enable = 0;
+		queue->queue_size = ZXDH_GDMA_QUEUE_SIZE;
+		rte_spinlock_init(&(queue->enqueue_lock));
+	}
+
+	return 0;
+
+err_out:
+	zxdh_gdma_unmap_resource(pci_dev->mem_resource[0].addr,
+		(size_t)pci_dev->mem_resource[0].len);
+	return -1;
+}
+
+static int
+zxdh_gdma_rawdev_remove(struct rte_pci_device *pci_dev)
+{
+	struct rte_rawdev *dev = NULL;
+	int ret = 0;
+
+	dev = rte_rawdev_pmd_get_named_dev(dev_name);
+	if (dev == NULL)
+		return -EINVAL;
+
+	/* rte_rawdev_close is called by pmd_release */
+	ret = rte_rawdev_pmd_release(dev);
+	if (ret != 0) {
+		ZXDH_PMD_LOG(ERR, "Device cleanup failed");
+		return -1;
+	}
+
+	zxdh_gdma_unmap_resource(pci_dev->mem_resource[0].addr,
+		(size_t)pci_dev->mem_resource[0].len);
+
+	ZXDH_PMD_LOG(DEBUG, "rawdev %s remove done!", dev_name);
+
+	return ret;
+}
+
+static const struct rte_pci_id zxdh_gdma_rawdev_map[] = {
+	{ RTE_PCI_DEVICE(ZXDH_GDMA_VENDORID, ZXDH_GDMA_DEVICEID) },
+	{ .vendor_id = 0, /* sentinel */ },
+};
+
+static struct rte_pci_driver zxdh_gdma_rawdev_pmd = {
+	.id_table = zxdh_gdma_rawdev_map,
+	.drv_flags = 0,
+	.probe = zxdh_gdma_rawdev_probe,
+	.remove = zxdh_gdma_rawdev_remove,
+};
+
+RTE_PMD_REGISTER_PCI(zxdh_gdma_rawdev_pci_driver, zxdh_gdma_rawdev_pmd);
+RTE_PMD_REGISTER_PCI_TABLE(zxdh_gdma_rawdev_pci_driver, zxdh_gdma_rawdev_map);
+RTE_LOG_REGISTER_DEFAULT(zxdh_gdma_rawdev_logtype, NOTICE);
diff --git a/drivers/raw/zxdh/zxdh_rawdev.h b/drivers/raw/zxdh/zxdh_rawdev.h
new file mode 100644
index 0000000000..b4d977ce54
--- /dev/null
+++ b/drivers/raw/zxdh/zxdh_rawdev.h
@@ -0,0 +1,118 @@ 
+/* SPDX-License-Identifier: BSD-3-Clause
+ * Copyright 2024 ZTE Corporation
+ */
+
+#ifndef __ZXDH_RAWDEV_H__
+#define __ZXDH_RAWDEV_H__
+
+#ifdef __cplusplus
+extern "C" {
+#endif
+
+#include <rte_rawdev.h>
+#include <rte_spinlock.h>
+
+extern int zxdh_gdma_rawdev_logtype;
+#define RTE_LOGTYPE_ZXDH_GDMA                   zxdh_gdma_rawdev_logtype
+
+#define ZXDH_PMD_LOG(level, ...) \
+	RTE_LOG_LINE_PREFIX(level, ZXDH_GDMA, \
+		"%s() line %u: ", __func__ RTE_LOG_COMMA __LINE__, __VA_ARGS__)
+
+#define ZXDH_GDMA_VENDORID                      0x1cf2
+#define ZXDH_GDMA_DEVICEID                      0x8044
+
+#define ZXDH_GDMA_TOTAL_CHAN_NUM                58
+#define ZXDH_GDMA_QUEUE_SIZE                    16384
+#define ZXDH_GDMA_RING_SIZE                     32768
+
+enum zxdh_gdma_device_state {
+	ZXDH_GDMA_DEV_RUNNING,
+	ZXDH_GDMA_DEV_STOPPED
+};
+
+struct zxdh_gdma_buff_desc {
+	uint32_t SrcAddr_L;
+	uint32_t DstAddr_L;
+	uint32_t Xpara;
+	uint32_t ZY_para;
+	uint32_t ZY_SrcStep;
+	uint32_t ZY_DstStep;
+	uint32_t ExtAddr;
+	uint32_t LLI_Addr_L;
+	uint32_t LLI_Addr_H;
+	uint32_t ChCont;
+	uint32_t LLI_User;
+	uint32_t ErrAddr;
+	uint32_t Control;
+	uint32_t SrcAddr_H;
+	uint32_t DstAddr_H;
+	uint32_t Reserved;
+};
+
+struct zxdh_gdma_job {
+	uint64_t src;
+	uint64_t dest;
+	uint32_t len;
+	uint32_t flags;
+	uint64_t cnxt;
+	uint16_t status;
+	uint16_t vq_id;
+	void *usr_elem;
+	uint8_t ep_id;
+	uint8_t pf_id;
+	uint16_t vf_id;
+};
+
+struct zxdh_gdma_queue {
+	uint8_t   enable;
+	uint8_t   is_txq;
+	uint16_t  vq_id;
+	uint16_t  queue_size;
+	/* 0:GDMA needs to be configured through the APB interface */
+	uint16_t  flag;
+	uint32_t  user;
+	uint16_t  tc_cnt;
+	rte_spinlock_t enqueue_lock;
+	struct {
+		uint16_t avail_idx;
+		uint16_t last_avail_idx;
+		rte_iova_t ring_mem;
+		const struct rte_memzone *ring_mz;
+		struct zxdh_gdma_buff_desc *desc;
+	} ring;
+	struct {
+		uint16_t  free_cnt;
+		uint16_t  deq_cnt;
+		uint16_t  pend_cnt;
+		uint16_t  enq_idx;
+		uint16_t  deq_idx;
+		uint16_t  used_idx;
+		struct zxdh_gdma_job **job;
+	} sw_ring;
+};
+
+struct zxdh_gdma_rawdev {
+	struct rte_device *device;
+	struct rte_rawdev *rawdev;
+	uintptr_t base_addr;
+	uint8_t queue_num; /* total queue num */
+	uint8_t used_num;  /* used  queue num */
+	enum zxdh_gdma_device_state device_state;
+	struct zxdh_gdma_queue vqs[ZXDH_GDMA_TOTAL_CHAN_NUM];
+};
+
+static inline struct zxdh_gdma_rawdev *
+zxdh_gdma_rawdev_get_priv(const struct rte_rawdev *rawdev)
+{
+	return rawdev->dev_private;
+}
+
+uint32_t zxdh_gdma_read_reg(struct rte_rawdev *dev, uint16_t qidx, uint32_t offset);
+void zxdh_gdma_write_reg(struct rte_rawdev *dev, uint16_t qidx, uint32_t offset, uint32_t val);
+
+#ifdef __cplusplus
+}
+#endif
+
+#endif /* __ZXDH_RAWDEV_H__ */