doc: update async enqueue API usage
PATCH 1/1 doc: update async enqueue API usage
From: Xuan Ding <[email protected]>
This patch updates the correct usage for async enqueue APIs. The rte_vhost_poll_enqueue_completed() needs to be called in time to avoid packet loss due to full dma ring.
--- doc/guides/prog_guide/vhost_lib.rst | 9 +++++++++ 1 file changed, 9 insertions(+)
doc/guides/prog_guide/vhost_lib.rst +9 -0
@@ -488,3 +488,12 @@ For PA mode, page by page mapping may exceed IOMMU's max capability, better to use 1G guest hugepage. For UIO driver, any VFIO related error message can be ignored. + +Async enqueue API usage +----------------------- + +In async enqueue path, rte_vhost_poll_enqueue_completed() needs to be +called in time. Otherwise, calling rte_vhost_submit_enqueue_burst() +all the time will cause the DMA ring to be full, which will result +in packet loss eventually. +
Signature · 1 line
2.17.1
Hi Xuan,
Quoted 28 lines
-----Original Message-----
From: Ding, Xuan <[email protected]>
Sent: Friday, June 17, 2022 5:34 PM
To: [email protected]; Xia, Chenbo <[email protected]>
Cc: [email protected]; Hu, Jiayu <[email protected]>; Ding, Xuan <[email protected]>
Subject: [PATCH] doc: update async enqueue API usageFrom: Xuan Ding <[email protected]>
This patch updates the correct usage for async enqueue APIs. The
rte_vhost_poll_enqueue_completed()needs to be called in time to avoid packet loss due to full dma ring.Signed-off-by Xuan Ding <[email protected]>--- doc/guides/prog_guide/vhost_lib.rst | 9 +++++++++ 1 file changed, 9 insertions(+)doc/guides/prog_guide/vhost_lib.rst +0
-0@@ -488,3 +488,12 @@ For PA mode, page by page mapping may exceedIOMMU's max capability, better to use 1G guest hugepage.
For UIO driver, any VFIO related error message can be ignored.
doc/guides/prog_guide/vhost_lib.rst For PA mode, page by page mapping may exceed IOMMU's max capability, · line 493
Quoted 3 lines
doc/guides/prog_guide/vhost_lib.rst +3
-0+ +Async enqueue API usage +-----------------------
I think it's better to make it as a sub-section of "Vhost asynchronous data path".
doc/guides/prog_guide/vhost_lib.rst
Quoted 5 lines
doc/guides/prog_guide/vhost_lib.rst +5
-0+ +In async enqueue path, rte_vhost_poll_enqueue_completed() needs to be +called in time. Otherwise, calling rte_vhost_submit_enqueue_burst() all +the time will cause the DMA ring to be full, which will result in +packet loss eventually.
There are two main reasons of calling rte_vhost_poll_enqueue_completed() timely, IMO:
1. notify the guest of DMA copy completed packets;
2. in case of DMA ring size is smaller than vring size, it can avoid packet drop caused by DMA ring full.
Thanks,
Jiayu
Quoted 3 lines
+
--
2.17.1
Hi Jiayu,
Quoted 37 lines
-----Original Message-----
From: Hu, Jiayu <[email protected]>
Sent: Monday, June 20, 2022 9:43 AM
To: Ding, Xuan <[email protected]>; [email protected]; Xia, Chenbo <[email protected]>
Cc: [email protected]
Subject: RE: [PATCH] doc: update async enqueue API usageHi Xuan,
Quoted 27 lines
-----Original Message-----
From: Ding, Xuan <[email protected]>
Sent: Friday, June 17, 2022 5:34 PM
To: [email protected]; Xia, Chenbo <[email protected]>
Cc: [email protected]; Hu, Jiayu <[email protected]>; Ding, Xuan <[email protected]>
Subject: [PATCH] doc: update async enqueue API usageFrom: Xuan Ding <[email protected]>
This patch updates the correct usage for async enqueue APIs. The
rte_vhost_poll_enqueue_completed()needs to be called in time to avoid packet loss due to full dma ring.Signed-off-by Xuan Ding <[email protected]>--- doc/guides/prog_guide/vhost_lib.rst | 9 +++++++++ 1 file changed, 9 insertions(+)doc/guides/prog_guide/vhost_lib.rst +0
-0@@ -488,3 +488,12 @@ For PA mode, page by page mapping may exceedIOMMU's max capability, better to use 1G guest hugepage.
Quoted 7 lines
doc/guides/prog_guide/vhost_lib.rst For PA mode, page by page mapping may exceed IOMMU's max capability, · line 493
Quoted 4 lines
For UIO driver, any VFIO related error message can be ignored.
doc/guides/prog_guide/vhost_lib.rst +3
-0+ +Async enqueue API usage +-----------------------I think it's better to make it as a sub-section of "Vhost asynchronous data path".
That's a good idea, I will add this section as a sub-section accordingly.
Quoted 12 lines
doc/guides/prog_guide/vhost_lib.rst For PA mode, page by page mapping may exceed IOMMU's max capability,
Quoted 6 lines
doc/guides/prog_guide/vhost_lib.rst +5
-0+ +In async enqueue path, rte_vhost_poll_enqueue_completed() needs to be +called in time. Otherwise, calling rte_vhost_submit_enqueue_burst() +all the time will cause the DMA ring to be full, which will result in +packet loss eventually.There are two main reasons of calling
rte_vhost_poll_enqueue_completed()timely, IMO:
1. notify the guest of DMA copy completed packets; 2. in case of DMA ring size is smaller than vring size, it can avoid packet drop caused by DMA ring full.
Please see doc enriched in next version.
Thanks,
Xuan
Quoted 9 lines
Thanks,
JiayuQuoted 3 lines
+
--
2.17.1
Hi Xuan,
Quoted 45 lines
-----Original Message-----
From: Hu, Jiayu
Sent: Monday, June 20, 2022 9:43 AM
To: Ding, Xuan <[email protected]>; [email protected]; Xia, Chenbo <[email protected]>
Cc: [email protected]
Subject: RE: [PATCH] doc: update async enqueue API usageHi Xuan,
Quoted 28 lines
-----Original Message-----
From: Ding, Xuan <[email protected]>
Sent: Friday, June 17, 2022 5:34 PM
To: [email protected]; Xia, Chenbo <[email protected]>
Cc: [email protected]; Hu, Jiayu <[email protected]>; Ding, Xuan <[email protected]>
Subject: [PATCH] doc: update async enqueue API usageFrom: Xuan Ding <[email protected]>
This patch updates the correct usage for async enqueue APIs. The
rte_vhost_poll_enqueue_completed()needs to be called in time to avoid packet loss due to full dma ring.Signed-off-by Xuan Ding <[email protected]>--- doc/guides/prog_guide/vhost_lib.rst | 9 +++++++++ 1 file changed, 9 insertions(+)doc/guides/prog_guide/vhost_lib.rst +0
-0@@ -488,3 +488,12 @@ For PA mode, page by page mapping may exceedIOMMU's max capability, better to use 1G guest hugepage.
For UIO driver, any VFIO related error message can be ignored.
doc/guides/prog_guide/vhost_lib.rst For PA mode, page by page mapping may exceed IOMMU's max capability, · line 493
Quoted 3 lines
doc/guides/prog_guide/vhost_lib.rst +3
-0+ +Async enqueue API usage +-----------------------I think it's better to make it as a sub-section of "Vhost asynchronous data path".
Quoted 11 lines
doc/guides/prog_guide/vhost_lib.rst For PA mode, page by page mapping may exceed IOMMU's max capability,
Quoted 5 lines
doc/guides/prog_guide/vhost_lib.rst +5
-0+ +In async enqueue path, rte_vhost_poll_enqueue_completed() needs to be +called in time. Otherwise, calling rte_vhost_submit_enqueue_burst() +all the time will cause the DMA ring to be full, which will result in +packet loss eventually.There are two main reasons of calling
rte_vhost_poll_enqueue_completed()timely, IMO:
1. notify the guest of DMA copy completed packets; 2. in case of DMA ring size is smaller than vring size, it can avoid packet drop caused by DMA ring full.
I'd like to clarify #2 a little: not only for the case that DMA ring size < vring size. Even if DMA ring size > vring size, it's possible dropping packets as a result of DMA ring full.
Thanks,
Jiayu
Quoted 8 lines
Thanks,
JiayuQuoted 3 lines
+
--
2.17.1
PATCH v2 1/1 doc: update async enqueue API usage
In reply to [email protected]From: Xuan Ding <[email protected]>
This patch updates the correct usage for async enqueue APIs. The rte_vhost_poll_enqueue_completed() needs to be called in time to notify the guest of completed packets and avoid packet loss.
---
v2:
* refine doc and commit log
--- doc/guides/prog_guide/vhost_lib.rst | 8 ++++++++ 1 file changed, 8 insertions(+)
doc/guides/prog_guide/vhost_lib.rst +8 -0
@@ -480,6 +480,14 @@ the same vring with their own DMA virtual channels. Besides, the number of DMA devices is limited. For the purpose of scaling, it's necessary to support sharing DMA channels among vrings. +* Async enqueue API usage + + In async enqueue path, rte_vhost_poll_enqueue_completed() needs to be + called in time to notify the guest of DMA copy completed packets. + Moreover, calling rte_vhost_submit_enqueue_burst() all the time but + not poll completed will cause the DMA ring to be full, which will + result in packet loss eventually. + Recommended IOVA mode in async datapath ---------------------------------------
Signature · 1 line
2.17.1
Hi Xuan,
Quoted 33 lines
-----Original Message-----
From: Ding, Xuan <[email protected]>
Sent: Tuesday, June 21, 2022 3:22 PM
To: [email protected]; Xia, Chenbo <[email protected]>
Cc: [email protected]; Hu, Jiayu <[email protected]>; Ding, Xuan <[email protected]>
Subject: [PATCH v2] doc: update async enqueue API usageFrom: Xuan Ding <[email protected]>
This patch updates the correct usage for async enqueue APIs. The
rte_vhost_poll_enqueue_completed()needs to be called in time to notify the guest of completed packets and avoid packet loss.Signed-off-by Xuan Ding <[email protected]>---
v2:
* refine doc and commit log--- doc/guides/prog_guide/vhost_lib.rst | 8 ++++++++ 1 file changed, 8 insertions(+)doc/guides/prog_guide/vhost_lib.rst +0
-0@@ -480,6 +480,14 @@ the same vring with their own DMA virtual channels.Besides, the number
of DMA devices is limited. For the purpose of scaling, it's necessary to support sharing DMA channels among vrings.
doc/guides/prog_guide/vhost_lib.rst the same vring with their own DMA virtual channels. Besides, the number · line 490
Quoted 8 lines
doc/guides/prog_guide/vhost_lib.rst +8
-0+* Async enqueue API usage + + In async enqueue path, rte_vhost_poll_enqueue_completed() needs to be + called in time to notify the guest of DMA copy completed packets. + Moreover, calling rte_vhost_submit_enqueue_burst() all the time but + not poll completed will cause the DMA ring to be full, which will + result in packet loss eventually. +
This subsection seems cleaner, could you also help to make 'Recommended IOVA mode in async datapath' as a sub-section?
Another patch or not, either works for me.
Thanks,
Chenbo
Quoted 5 lines
Recommended IOVA mode in async datapath
-----------------------------------------
2.17.1
Hi Chenbo,
Quoted 54 lines
-----Original Message-----
From: Xia, Chenbo <[email protected]>
Sent: Tuesday, June 21, 2022 9:23 PM
To: Ding, Xuan <[email protected]>; [email protected]
Cc: [email protected]; Hu, Jiayu <[email protected]>
Subject: RE: [PATCH v2] doc: update async enqueue API usageHi Xuan,
doc/guides/prog_guide/vhost_lib.rst the same vring with their own DMA virtual
Quoted 28 lines
-----Original Message-----
From: Ding, Xuan <[email protected]>
Sent: Tuesday, June 21, 2022 3:22 PM
To: [email protected]; Xia, Chenbo <[email protected]>
Cc: [email protected]; Hu, Jiayu <[email protected]>; Ding, Xuan <[email protected]>
Subject: [PATCH v2] doc: update async enqueue API usageFrom: Xuan Ding <[email protected]>
This patch updates the correct usage for async enqueue APIs. The
rte_vhost_poll_enqueue_completed()needs to be called in time to notify the guest of completed packets and avoid packet loss.Signed-off-by Xuan Ding <[email protected]>---
v2:
* refine doc and commit log--- doc/guides/prog_guide/vhost_lib.rst | 8 ++++++++ 1 file changed, 8 insertions(+)doc/guides/prog_guide/vhost_lib.rst +0
-0@@ -480,6 +480,14 @@ the same vring with their own DMA virtualchannels.
doc/guides/prog_guide/vhost_lib.rst the same vring with their own DMA virtual channels. Besides, the number · line 490
Quoted 12 lines
Besides, the number
of DMA devices is limited. For the purpose of scaling, it's necessary
to support sharing DMA channels among vrings.doc/guides/prog_guide/vhost_lib.rst +8
-0+* Async enqueue API usage + + In async enqueue path, rte_vhost_poll_enqueue_completed() needs to + be called in time to notify the guest of DMA copy completed packets. + Moreover, calling rte_vhost_submit_enqueue_burst() all the time but + not poll completed will cause the DMA ring to be full, which will + result in packet loss eventually. +This subsection seems cleaner, could you also help to make 'Recommended IOVA mode in async datapath' as a sub-section?
Quoted 1 line
Another patch or not, either works for me.
I will send a separate patch to move 'Recommended IOVA mode in async datapath' as a sub-section.
Thanks,
Xuan
Quoted 9 lines
Thanks,
ChenboQuoted 5 lines
Recommended IOVA mode in async datapath
-----------------------------------------
2.17.1
Quoted 45 lines
-----Original Message-----
From: Ding, Xuan <[email protected]>
Sent: Tuesday, June 21, 2022 3:22 PM
To: [email protected]; Xia, Chenbo <[email protected]>
Cc: [email protected]; Hu, Jiayu <[email protected]>; Ding, Xuan <[email protected]>
Subject: [PATCH v2] doc: update async enqueue API usageFrom: Xuan Ding <[email protected]>
This patch updates the correct usage for async enqueue APIs. The
rte_vhost_poll_enqueue_completed()needs to be called in time to notify the guest of completed packets and avoid packet loss.Signed-off-by Xuan Ding <[email protected]>---
v2:
* refine doc and commit log--- doc/guides/prog_guide/vhost_lib.rst | 8 ++++++++ 1 file changed, 8 insertions(+)doc/guides/prog_guide/vhost_lib.rst +0
-0@@ -480,6 +480,14 @@ the same vring with their own DMA virtual channels.Besides, the number
of DMA devices is limited. For the purpose of scaling, it's necessary to support sharing DMA channels among vrings.doc/guides/prog_guide/vhost_lib.rst +8
-1+* Async enqueue API usage + + In async enqueue path, rte_vhost_poll_enqueue_completed() needs to be + called in time to notify the guest of DMA copy completed packets. + Moreover, calling rte_vhost_submit_enqueue_burst() all the time but + not poll completed will cause the DMA ring to be full, which will + result in packet loss eventually. + Recommended IOVA mode in async datapath --------------------------------------- --
Quoted 1 line
2.17.1
Maxime Coquelin
AppliedQuoted 7 lines · On 6/21/22 09:21, [email protected] wrote:
From: Xuan Ding <[email protected]>
This patch updates the correct usage for async enqueue APIs. The
rte_vhost_poll_enqueue_completed()needs to be called in time to notify the guest of completed packets and avoid packet loss.
Quoted 8 lines
Signed-off-by Xuan Ding <[email protected]>---
v2:
* refine doc and commit log--- doc/guides/prog_guide/vhost_lib.rst | 8 ++++++++ 1 file changed, 8 insertions(+)
Applied to dpdk-next-virtio/main.
Thanks,
Maxime