Message ID | 20180607123849.14439-1-qi.z.zhang@intel.com (mailing list archive) |
---|---|
Headers |
Return-Path: <dev-bounces@dpdk.org> X-Original-To: patchwork@dpdk.org Delivered-To: patchwork@dpdk.org Received: from [92.243.14.124] (localhost [127.0.0.1]) by dpdk.org (Postfix) with ESMTP id D08091B1A4; Thu, 7 Jun 2018 14:38:25 +0200 (CEST) Received: from mga04.intel.com (mga04.intel.com [192.55.52.120]) by dpdk.org (Postfix) with ESMTP id 40ADD1B19B for <dev@dpdk.org>; Thu, 7 Jun 2018 14:38:23 +0200 (CEST) X-Amp-Result: SKIPPED(no attachment in message) X-Amp-File-Uploaded: False Received: from orsmga001.jf.intel.com ([10.7.209.18]) by fmsmga104.fm.intel.com with ESMTP/TLS/DHE-RSA-AES256-GCM-SHA384; 07 Jun 2018 05:38:23 -0700 X-ExtLoop1: 1 X-IronPort-AV: E=Sophos;i="5.49,486,1520924400"; d="scan'208";a="62605485" Received: from dpdk51.sh.intel.com ([10.67.110.184]) by orsmga001.jf.intel.com with ESMTP; 07 Jun 2018 05:38:20 -0700 From: Qi Zhang <qi.z.zhang@intel.com> To: thomas@monjalon.net, anatoly.burakov@intel.com Cc: konstantin.ananyev@intel.com, dev@dpdk.org, bruce.richardson@intel.com, ferruh.yigit@intel.com, benjamin.h.shelton@intel.com, narender.vangati@intel.com, Qi Zhang <qi.z.zhang@intel.com> Date: Thu, 7 Jun 2018 20:38:27 +0800 Message-Id: <20180607123849.14439-1-qi.z.zhang@intel.com> X-Mailer: git-send-email 2.13.6 Subject: [dpdk-dev] [PATCH 00/22] enable hotplug on multi-process X-BeenThere: dev@dpdk.org X-Mailman-Version: 2.1.15 Precedence: list List-Id: DPDK patches and discussions <dev.dpdk.org> List-Unsubscribe: <https://dpdk.org/ml/options/dev>, <mailto:dev-request@dpdk.org?subject=unsubscribe> List-Archive: <http://dpdk.org/ml/archives/dev/> List-Post: <mailto:dev@dpdk.org> List-Help: <mailto:dev-request@dpdk.org?subject=help> List-Subscribe: <https://dpdk.org/ml/listinfo/dev>, <mailto:dev-request@dpdk.org?subject=subscribe> Errors-To: dev-bounces@dpdk.org Sender: "dev" <dev-bounces@dpdk.org> |
Series |
enable hotplug on multi-process
|
|
Message
Qi Zhang
June 7, 2018, 12:38 p.m. UTC
Background: =========== Currently secondary process will only sync ethdev from primary process at init stage, but it will not be aware if device is attached/detached on primary process at runtime. While there is the requirement from application that take primary-secondary process model. The primary process work as a resource management process, it will create/destroy virtual device at runtime, while the secondary process deal with the network stuff with these devices. Solution: ========= So the orignial intention is to fix this gap, but beyond that the patch set provide a more comprehesive solution to handle different hotplug cases in multi-process situation, it cover below scenario: 1. Attach a share device from primary 2. Detach a share device from primary 3. Attach a share device from secondary 4. Detach a share device from secondary 5. Attach a private device from secondary 6. Detach a private device from secondary 7. Detach a share device from secondary privately 8. Attach a share device from secondary privately In primary-secondary process model, we assume ethernet devices are shared by default. that means attach or detach a device on any process will broadcast to all other processes through mp channel then device information will be synchronized on all processes. Any failure during attaching process will cause inconsistent status between processes, so proper rollback action should be considered. Also, it is not safe to detach a share device when other process still use it, so a handshake mechanism is introduced. Scenario for Case 1, 2: attach device from primary a) primary attach the new device if failed goto h). b) primary send attach sync request to all secondary. c) secondary receive request and attach device and send reply. d) primary check the reply if all success go to i). e) primary send attach rollback sync request to all secondary. f) secondary receive the request and detach device and send reply. g) primary receive the reply and detach device as rollback action. h) attach fail i) attach success detach device from primary a) primary perform pre-detach check, if device is locked, goto i). b) primary send pre-detach sync request to all secondary. c) secondary perform pre-detach check and send reply. d) primary check the reply if any fail goto i). e) primary send detach sync request to all secondary f) secondary detach the device and send reply (assume no fail) g) primary detach the device. h) detach success i) detach failed Scenario for case 3, 4: attach device from secondary: a) seconary send asycn request to primary and wait on a condition which will be released by matched response from primary. b) primary receive the request and attach the new device if failed goto i). c) primary forward attach request to all secondary as async request (because this in mp thread context, use sync request will deadlock, same reason for all following async request.) d) secondary receive request and attach device and send reply. e) primary check the reply if all success go to j). f) primary send attach rollback async request to all secondary. g) secondary receive the request and detach device and send reply. h) primary receive the reply and detach device as rollback action. i) send fail response to secondary, goto k). j) send success response to secondary. k) secondary process receive response and return. detach device from secondary: a) secondary send async request to primary and wait on a condition which will be released by matched response from primary. b) primary receive the request and perform pre-detach check, if device is locked, goto j). c) primary send pre-detach async request to all secondary. d) secondary perform pre-detach check and send reply. e) primary check the reply if any fail goto j). f) primary send detach async request to all secondary g) secondary detach the device and send reply h) primary detach the device. i) send success response to secondary, goto k). j) send fail response to secondary. k) secondary process receive response and return. Case 5, 6: Secondary process can attach private device which only visible to itself, in this case no IPC is involved, primary process is not allowed to have private device so far. Case 7, 8: Secondary process can also temporally to detach a share device "privately" then attach it back later, this action also not impact other processes. APIs chenages: ============== rte_eth_dev_attach and rte_eth_dev_attach are extended to support share device attach/detach in primary-secondary process model, it will be called in case 1,2,3,4. New API rte_eth_dev_attach_private and rte_eth_dev_detach_private are introduced to cover case 5,6,7,8, this API can only be invoked in secondary process. New API rte_eth_dev_lock and rte_eth_dev_unlock are introduced to let application lock or unlock on specific ethdev, a locked device can't be detached. This help applicaiton to prevent unexpected device detaching, especially in multi-process envrionment. Aslo the new API let application to register a callback function which will be invoked before a device is going to be detached, the return value of the function will decide if device will continue be detached or not, this support application to do condition check at runtime. PMD Impact: =========== Currently device removing is not handled well in secondary process on most pmd drivers, rte_eth_dev_relase_port will be invoked and will mess up primary process since it reset all shared data. So we introduced new API rte_eth_dev_release_port_local which only reset ethdev's state to unsued but not touch shared data so other process will not be impacted. Since not all device driver is target to support primary-secondary process model, so the patch set only fix this on all Intel devices and vdev, it can be refereneced by other driver when equevalent fix is required Limitation: =========== The solution does not cover the case that primary process exit while secondary processes still be active. Though this is not a typial use case, but if this happens: 1. secondary process can't attach / detach any shared device since no primary exist. 2. secondary process still can attach / detach private device. 3. secondary process still can detach a share device privately but may not attach it back, that ethdev slot will become zombie slot. Example: ======== The patchset also contains a example to demonstrate device management in multi-process model, below are detail instructions. /* start sample code as primary then secondary */ ./devmgm_mp --proc-type=auto Command Line Example: >help >list /* attach a af_packet vdev */ >attach net_af_packet,iface=eth0 /* detach port 0 */ >detach 0 /* attach a private af_packet vdev (secondary process only)*/ >attachp net_af_packet,iface=eth0 /* detach a private device (secondary process only) */ >detachp 0 /* lock port 0 */ >lock 0 /* unlock port 0 */ >unlock 0 Qi Zhang (22): eal: introduce one device scan bus/vdev: enable one device scan ethdev: add function to release port in local process ethdev: enable hotplug on multi-process ethdev: introduce device lock ethdev: support attach or detach share device from secondary net/i40e: enable port detach on secondary process net/ixgbe: enable port detach on secondary process net/e1000: enable port detach on secondary process net/igb: enable port detach on secondary process net/fm10k: enable port detach on secondary process net/af_packet: enable port detach on secondary process net/bonding: enable port detach on secondary process net/failsafe: enable port detach on secondary process net/kni: enable port detach on secondary process net/null: enable port detach on secondary process net/octeontx: enable port detach on secondary process net/pcap: enable port detach on secondary process net/softnic: enable port detach on secondary process net/tap: enable port detach on secondary process net/vhost: enable port detach on secondary process examples/devmgm_mp: add simple device management sample drivers/bus/vdev/vdev.c | 30 ++ drivers/net/af_packet/rte_eth_af_packet.c | 11 + drivers/net/bonding/rte_eth_bond_pmd.c | 11 + drivers/net/e1000/em_ethdev.c | 9 + drivers/net/e1000/igb_ethdev.c | 9 + drivers/net/failsafe/failsafe.c | 16 + drivers/net/fm10k/fm10k_ethdev.c | 9 + drivers/net/i40e/i40e_ethdev.c | 2 + drivers/net/i40e/i40e_ethdev_vf.c | 9 + drivers/net/ixgbe/ixgbe_ethdev.c | 12 + drivers/net/kni/rte_eth_kni.c | 11 + drivers/net/null/rte_eth_null.c | 16 +- drivers/net/octeontx/octeontx_ethdev.c | 13 + drivers/net/pcap/rte_eth_pcap.c | 15 +- drivers/net/softnic/rte_eth_softnic.c | 19 +- drivers/net/tap/rte_eth_tap.c | 17 +- drivers/net/vhost/rte_eth_vhost.c | 11 + examples/devmgm_mp/Makefile | 64 +++ examples/devmgm_mp/commands.c | 383 ++++++++++++++++++ examples/devmgm_mp/commands.h | 10 + examples/devmgm_mp/main.c | 41 ++ examples/devmgm_mp/meson.build | 11 + lib/librte_eal/common/eal_common_dev.c | 17 +- lib/librte_eal/common/eal_private.h | 8 + lib/librte_eal/common/include/rte_bus.h | 4 + lib/librte_eal/linuxapp/eal/eal.c | 6 + lib/librte_ethdev/Makefile | 2 + lib/librte_ethdev/rte_ethdev.c | 246 ++++++++++- lib/librte_ethdev/rte_ethdev.h | 101 +++++ lib/librte_ethdev/rte_ethdev_core.h | 5 + lib/librte_ethdev/rte_ethdev_driver.h | 40 ++ lib/librte_ethdev/rte_ethdev_lock.c | 102 +++++ lib/librte_ethdev/rte_ethdev_lock.h | 23 ++ lib/librte_ethdev/rte_ethdev_mp.c | 653 ++++++++++++++++++++++++++++++ lib/librte_ethdev/rte_ethdev_mp.h | 45 ++ 35 files changed, 1950 insertions(+), 31 deletions(-) create mode 100644 examples/devmgm_mp/Makefile create mode 100644 examples/devmgm_mp/commands.c create mode 100644 examples/devmgm_mp/commands.h create mode 100644 examples/devmgm_mp/main.c create mode 100644 examples/devmgm_mp/meson.build create mode 100644 lib/librte_ethdev/rte_ethdev_lock.c create mode 100644 lib/librte_ethdev/rte_ethdev_lock.h create mode 100644 lib/librte_ethdev/rte_ethdev_mp.c create mode 100644 lib/librte_ethdev/rte_ethdev_mp.h
Comments
Hi Qi, I haven't read the code yet, and i'll be the first to admit that i'm not too well versed on how shared/private device data works, so my apologies in advance if all of below comments are addressed by implementation details or are way off base! On 07-Jun-18 1:38 PM, Qi Zhang wrote: > > 1. Attach a share device from primary > 2. Detach a share device from primary > 3. Attach a share device from secondary > 4. Detach a share device from secondary > 5. Attach a private device from secondary > 6. Detach a private device from secondary > 7. Detach a share device from secondary privately > 8. Attach a share device from secondary privately > <...> > Case 7, 8: > Secondary process can also temporally to detach a share device > "privately" then attach it back later, this action also not impact other > processes. > Do we really need to implement these cases? It seems to me that this "reattach it later" introduces unnecessary complexity. If secondary has detached the device, i think it is safer if we cannot reattach it, period, because it was a shared device. What if we try to attach it when a handshake has already completed and all other processes expect to detach it? (in fact, do we differentiate between non-existent device and shared device that has been "privately detached"? I would expect that we keep the device as detached as opposed to forgetting about it, so that, come handshake, we can safely reply "yeah, we can detach the device", but maybe it's OK to not treat request to detach a non-existent device as an error... thoughts? am i getting something wrong?) > APIs chenages: > ============== > > rte_eth_dev_attach and rte_eth_dev_attach are extended to support > share device attach/detach in primary-secondary process model, it will > be called in case 1,2,3,4. > > New API rte_eth_dev_attach_private and rte_eth_dev_detach_private are > introduced to cover case 5,6,7,8, this API can only be invoked in > secondary process. > > New API rte_eth_dev_lock and rte_eth_dev_unlock are introduced to let > application lock or unlock on specific ethdev, a locked device > can't be detached. This help applicaiton to prevent unexpected > device detaching, especially in multi-process envrionment. > Aslo the new API let application to register a callback function > which will be invoked before a device is going to be detached, > the return value of the function will decide if device will continue > be detached or not, this support application to do condition check > at runtime. I assume that you've added device locking to avoid having to do handshake before detach, correct? Is this a shared lock of some kind, or is it a private lock? If it's shared lock, what happens if the process holding that lock crashes? > > PMD Impact: > =========== > > Currently device removing is not handled well in secondary process on most > pmd drivers, rte_eth_dev_relase_port will be invoked and will mess up > primary process since it reset all shared data. So we introduced new API > rte_eth_dev_release_port_local which only reset ethdev's state to unsued but > not touch shared data so other process will not be impacted. > Since not all device driver is target to support primary-secondary > process model, so the patch set only fix this on all Intel devices and > vdev, it can be refereneced by other driver when equevalent fix is required Nitpick - why the naming mismatch between *_private() and *_local()? > > Limitation: > =========== > > The solution does not cover the case that primary process exit while > secondary processes still be active. Though this is not a typial use > case, but if this happens: > 1. secondary process can't attach / detach any shared device since no > primary exist. > 2. secondary process still can attach / detach private device. > 3. secondary process still can detach a share device privately but may > not attach it back, that ethdev slot will become zombie slot. I think this should be explicit and by design. Shared devices can only be communicated to all secondaries through a primary process. No primary - no shared devices. I don't think we can do anything about it unless we implement some kind of peer-to-peer IPC (which isn't happening as far as i'm aware). Thanks for your work on this patchset!
Hi Anatoly: Thanks for the review, see my reply in inline. > -----Original Message----- > From: Burakov, Anatoly > Sent: Friday, June 15, 2018 11:16 PM > To: Zhang, Qi Z <qi.z.zhang@intel.com>; thomas@monjalon.net > Cc: Ananyev, Konstantin <konstantin.ananyev@intel.com>; dev@dpdk.org; > Richardson, Bruce <bruce.richardson@intel.com>; Yigit, Ferruh > <ferruh.yigit@intel.com>; Shelton, Benjamin H > <benjamin.h.shelton@intel.com>; Vangati, Narender > <narender.vangati@intel.com> > Subject: Re: [PATCH 00/22] enable hotplug on multi-process > > Hi Qi, > > I haven't read the code yet, and i'll be the first to admit that i'm not too well > versed on how shared/private device data works, so my apologies in advance > if all of below comments are addressed by implementation details or are way > off base! > > > > On 07-Jun-18 1:38 PM, Qi Zhang wrote: > > > > 1. Attach a share device from primary > > 2. Detach a share device from primary > > 3. Attach a share device from secondary 4. Detach a share device from > > secondary 5. Attach a private device from secondary 6. Detach a > > private device from secondary 7. Detach a share device from secondary > > privately 8. Attach a share device from secondary privately > > > > <...> > > > Case 7, 8: > > Secondary process can also temporally to detach a share device > > "privately" then attach it back later, this action also not impact other > > processes. > > > > Do we really need to implement these cases? It seems to me that this > "reattach it later" introduces unnecessary complexity. I agree it's not necessary, but this looks like a free feature based on current implementation :) >If secondary has > detached the device, I think it is safer if we cannot reattach it, > period, because it was a shared device. What if we try to attach it when > a handshake has already completed and all other processes expect to > detach it? in the case: attach back a shared device already be detached will fail as expected. For PCI devices, it will fail at driver probe. For vdev, it will failed at rte_eth_dev_attach_secondary. > > (in fact, do we differentiate between non-existent device and shared > device that has been "privately detached"? I would expect that we keep > the device as detached as opposed to forgetting about it, so that, come > handshake, we can safely reply "yeah, we can detach the device", but > maybe it's OK to not treat request to detach a non-existent device as an > error... thoughts? am i getting something wrong?) > > > APIs chenages: > > ============== > > > > rte_eth_dev_attach and rte_eth_dev_attach are extended to support > > share device attach/detach in primary-secondary process model, it will > > be called in case 1,2,3,4. > > > > New API rte_eth_dev_attach_private and rte_eth_dev_detach_private are > > introduced to cover case 5,6,7,8, this API can only be invoked in > > secondary process. > > > > New API rte_eth_dev_lock and rte_eth_dev_unlock are introduced to let > > application lock or unlock on specific ethdev, a locked device > > can't be detached. This help applicaiton to prevent unexpected > > device detaching, especially in multi-process envrionment. > > Aslo the new API let application to register a callback function > > which will be invoked before a device is going to be detached, > > the return value of the function will decide if device will continue > > be detached or not, this support application to do condition check > > at runtime. > > I assume that you've added device locking to avoid having to do > handshake before detach, correct? Yes. >Is this a shared lock of some kind, or > is it a private lock? If it's shared lock, what happens if the process > holding that lock crashes? It’s a kind of process's private lock, but a shared device be locked any process will prevent it be detached. > > > > > PMD Impact: > > =========== > > > > Currently device removing is not handled well in secondary process on > most > > pmd drivers, rte_eth_dev_relase_port will be invoked and will mess up > > primary process since it reset all shared data. So we introduced new API > > rte_eth_dev_release_port_local which only reset ethdev's state to unsued > but > > not touch shared data so other process will not be impacted. > > Since not all device driver is target to support primary-secondary > > process model, so the patch set only fix this on all Intel devices and > > vdev, it can be refereneced by other driver when equevalent fix is required > > Nitpick - why the naming mismatch between *_private() and *_local()? Agree. "private" make the API more identical. > > > > > Limitation: > > =========== > > > > The solution does not cover the case that primary process exit while > > secondary processes still be active. Though this is not a typial use > > case, but if this happens: > > 1. secondary process can't attach / detach any shared device since no > > primary exist. > > 2. secondary process still can attach / detach private device. > > 3. secondary process still can detach a share device privately but may > > not attach it back, that ethdev slot will become zombie slot. > > I think this should be explicit and by design. Shared devices can only > be communicated to all secondaries through a primary process. No primary > - no shared devices. I don't think we can do anything about it unless we > implement some kind of peer-to-peer IPC (which isn't happening as far as > i'm aware). Agree. > > > > Thanks for your work on this patchset! Thanks for the design review and all the helpful inputs. Regards Qi > > -- > Thanks, > Anatoly
v6: - remove bus->scan_one, since ABI break is not necessary. - remove patch for failsafe PMD since it will not support secondary. - fix wrong implemenation on ixgbe. - add rte_eth_dev_release_port_private into rte_eth_dev_pci_generic_remove for secondary process, so we don't need to patch on PMD if PMD use the default remove function. - add release notes update. v5: - since we will keep mp thread separate from interrupt thread, it is not necessary to use temporary thread, we use rte_eal_alarm_set. - remove the change in rte_eth_dev_release_port, since there is a better way to prevent rte_eth_dev_release_port be called after rte_eth_dev_release_port_private. - fix the issue that lock does not take effect on secondary due to previous re-work - fix the issue when the first attached device is a private device from secondary. (patch 8/24) - work around for reply a sync request in separate thread, this is still an open and in discussion as below. https://mails.dpdk.org/archives/dev/2018-June/105359.html v4: - since mp thread will be merged to interrupt thread, the fix on v3 for sync IPC deadlock will not work. the new version enable the machanism to invoke a mp action callback in a temporary thread to avoid the IPC deadlock, with this, secondary to primary request impelemtation also be simplified, since we can use sync request directly in a separate thread. v3: - enable mp init callback register to help non-eal module to initialize mp channel during rte_eal_init - fix when attach share device from secondary. 1) dead lock due to sync IPC be invoked in rte_malloc in primary process when handle secondary request to attach device, the solution is primary process to issue share device attach/detach in interrupt thread. 2) return port_id not correct. - check nb_sent and nb_received in sync IPC. - fix memory leak duirng error handling at attach_on_secondary. - improve clean_lock_callback to only lock/unlock spinlock once - improve error code return in check-reply during async IPC. - remove rte_ prefix of internal function in ethdev_mp.c - sample code improvement. 1) rename sample to "hotplug_mp", and move to example/multi-process. 2) cleanup header include. 3) call rte_eal_cleanup before exit. v2: - rename rte_ethdev_mp.* to ethdev_mp.* - rename rte_ethdev_lock.* to ethdev_lock.* - move internal funciton to ethdev_private.h - separate rte_eth_dev_[un]lock into rte_eth_dev_[un]lock and rte_eth_dev_[un]lock_with_callback - lock callbacks will be removed automatically after device is detached. - add experimental tag for all new APIs. - fix coding style issue. - fix wrong lisence header in sample code. - fix spelling - fix meson.build. - improve comments. Background: =========== Currently secondary process will only sync ethdev from primary process at init stage, but it will not be aware if device is attached/detached on primary process at runtime. While there is the requirement from application that take primary-secondary process model. The primary process work as a resource management process, it will create/destroy virtual device at runtime, while the secondary process deal with the network stuff with these devices. Solution: ========= So the orignial intention is to fix this gap, but beyond that the patch set provide a more comprehesive solution to handle different hotplug cases in multi-process situation, it cover below scenario: 1. Attach a share device from primary 2. Detach a share device from primary 3. Attach a share device from secondary 4. Detach a share device from secondary 5. Attach a private device from secondary 6. Detach a private device from secondary 7. Detach a share device from secondary privately 8. Attach a share device from secondary privately In primary-secondary process model, we assume ethernet devices are shared by default. that means attach or detach a device on any process will broadcast to all other processes through mp channel then device information will be synchronized on all processes. Any failure during attaching process will cause inconsistent status between processes, so proper rollback action should be considered. Also, it is not safe to detach a share device when other process still use it, so a handshake mechanism is introduced. Scenario for Case 1, 2: attach device from primary a) primary attach the new device if failed goto h). b) primary send attach sync request to all secondary. c) secondary receive request and attach device and send reply. d) primary check the reply if all success go to i). e) primary send attach rollback sync request to all secondary. f) secondary receive the request and detach device and send reply. g) primary receive the reply and detach device as rollback action. h) attach fail i) attach success detach device from primary a) primary perform pre-detach check, if device is locked, goto i). b) primary send pre-detach sync request to all secondary. c) secondary perform pre-detach check and send reply. d) primary check the reply if any fail goto i). e) primary send detach sync request to all secondary f) secondary detach the device and send reply (assume no fail) g) primary detach the device. h) detach success i) detach failed Scenario for case 3, 4: attach device from secondary: a) seconary send asycn request to primary and wait on a condition which will be released by matched response from primary. b) primary receive the request and attach the new device if failed goto i). c) primary forward attach request to all secondary as async request (because this in mp thread context, use sync request will deadlock, same reason for all following async request.) d) secondary receive request and attach device and send reply. e) primary check the reply if all success go to j). f) primary send attach rollback async request to all secondary. g) secondary receive the request and detach device and send reply. h) primary receive the reply and detach device as rollback action. i) send fail response to secondary, goto k). j) send success response to secondary. k) secondary process receive response and return. detach device from secondary: a) secondary send async request to primary and wait on a condition which will be released by matched response from primary. b) primary receive the request and perform pre-detach check, if device is locked, goto j). c) primary send pre-detach async request to all secondary. d) secondary perform pre-detach check and send reply. e) primary check the reply if any fail goto j). f) primary send detach async request to all secondary g) secondary detach the device and send reply h) primary detach the device. i) send success response to secondary, goto k). j) send fail response to secondary. k) secondary process receive response and return. Case 5, 6: Secondary process can attach private device which only visible to itself, in this case no IPC is involved, primary process is not allowed to have private device so far. Case 7, 8: Secondary process can also temporally to detach a share device "privately" then attach it back later, this action also not impact other processes. APIs chenages: ============== rte_eth_dev_attach and rte_eth_dev_attach are extended to support share device attach/detach in primary-secondary process model, it will be called in case 1,2,3,4. New API rte_eth_dev_attach_private and rte_eth_dev_detach_private are introduced to cover case 5,6,7,8, this API can only be invoked in secondary process. New API rte_eth_dev_lock and rte_eth_dev_unlock are introduced to let application lock or unlock on specific ethdev, a locked device can't be detached. This help applicaiton to prevent unexpected device detaching, especially in multi-process envrionment. Aslo the new API let application to register a callback function which will be invoked before a device is going to be detached, the return value of the function will decide if device will continue be detached or not, this support application to do condition check at runtime. PMD Impact: =========== Currently device removing is not handled well in secondary process on most pmd drivers, rte_eth_dev_relase_port will be invoked and will mess up primary process since it reset all shared data. So we introduced new API rte_eth_dev_release_port_local which only reset ethdev's state to unsued but not touch shared data so other process will not be impacted. Since not all device driver is target to support primary-secondary process model, so the patch set only fix this on all Intel devices and vdev, it can be refereneced by other driver when equevalent fix is required Limitation: =========== 1. The solution does not cover the case that primary process exit while secondary processes still be active. Though this is not a typial use case, but if this happens: a. secondary process can't attach / detach any shared device since no primary exist. b. secondary process still can attach / detach private device. c. secondary process still can detach a share device privately but may not attach it back, that ethdev slot will become zombie slot. 2. So for, for PCI bus, case 5,6 is not supported. PCI bus scan/probe mechanism can be improved to support attach private device on secondary process, but this is not the scope of this patchset. Example: ======== The patchset also contains a example to demonstrate device hotplug in multi-process model, below are detail instructions. /* start sample code as primary then secondary */ ./hotplug_mp --proc-type=auto Command Line Example: >help >list /* attach a af_packet vdev */ >attach net_af_packet,iface=eth0 /* detach port 0 */ >detach 0 /* attach a private af_packet vdev (secondary process only)*/ >attachp net_af_packet,iface=eth0 /* detach a private device (secondary process only) */ >detachp 0 /* lock port 0 */ >lock 0 /* unlock port 0 */ >unlock 0 Qi Zhang (19): ethdev: add function to release port in local process eal: enable multi process init callback ethdev: enable hotplug on multi-process ethdev: introduce device lock ethdev: support attach or detach share device from secondary ethdev: support attach private device as first net/i40e: enable port detach on secondary process net/ixgbe: enable port detach on secondary process net/af_packet: enable port detach on secondary process net/bonding: enable port detach on secondary process net/kni: enable port detach on secondary process net/null: enable port detach on secondary process net/octeontx: enable port detach on secondary process net/pcap: enable port detach on secondary process net/softnic: enable port detach on secondary process net/tap: enable port detach on secondary process net/vhost: enable port detach on secondary process examples/multi_process: add hotplug sample doc: update release notes for multi process hotplug doc/guides/rel_notes/release_18_08.rst | 21 ++ drivers/net/af_packet/rte_eth_af_packet.c | 11 + drivers/net/bonding/rte_eth_bond_pmd.c | 11 + drivers/net/i40e/i40e_ethdev.c | 2 + drivers/net/ixgbe/ixgbe_ethdev.c | 3 + drivers/net/kni/rte_eth_kni.c | 11 + drivers/net/null/rte_eth_null.c | 16 +- drivers/net/octeontx/octeontx_ethdev.c | 16 ++ drivers/net/pcap/rte_eth_pcap.c | 15 +- drivers/net/softnic/rte_eth_softnic.c | 19 +- drivers/net/tap/rte_eth_tap.c | 17 +- drivers/net/vhost/rte_eth_vhost.c | 11 + examples/multi_process/Makefile | 1 + examples/multi_process/hotplug_mp/Makefile | 23 ++ examples/multi_process/hotplug_mp/commands.c | 356 +++++++++++++++++++++++ examples/multi_process/hotplug_mp/commands.h | 10 + examples/multi_process/hotplug_mp/main.c | 41 +++ lib/librte_eal/common/eal_common_proc.c | 51 +++- lib/librte_eal/common/eal_private.h | 5 + lib/librte_eal/common/include/rte_eal.h | 34 +++ lib/librte_eal/linuxapp/eal/eal.c | 2 + lib/librte_ethdev/Makefile | 2 + lib/librte_ethdev/ethdev_lock.c | 140 +++++++++ lib/librte_ethdev/ethdev_lock.h | 31 ++ lib/librte_ethdev/ethdev_mp.c | 415 +++++++++++++++++++++++++++ lib/librte_ethdev/ethdev_mp.h | 42 +++ lib/librte_ethdev/ethdev_private.h | 42 +++ lib/librte_ethdev/meson.build | 2 + lib/librte_ethdev/rte_ethdev.c | 312 ++++++++++++++++++-- lib/librte_ethdev/rte_ethdev.h | 169 +++++++++++ lib/librte_ethdev/rte_ethdev_core.h | 5 + lib/librte_ethdev/rte_ethdev_driver.h | 13 + lib/librte_ethdev/rte_ethdev_pci.h | 3 + 33 files changed, 1815 insertions(+), 37 deletions(-) create mode 100644 examples/multi_process/hotplug_mp/Makefile create mode 100644 examples/multi_process/hotplug_mp/commands.c create mode 100644 examples/multi_process/hotplug_mp/commands.h create mode 100644 examples/multi_process/hotplug_mp/main.c create mode 100644 lib/librte_ethdev/ethdev_lock.c create mode 100644 lib/librte_ethdev/ethdev_lock.h create mode 100644 lib/librte_ethdev/ethdev_mp.c create mode 100644 lib/librte_ethdev/ethdev_mp.h create mode 100644 lib/librte_ethdev/ethdev_private.h
v13: - Since rte_eth_dev_attach/rte_eth_dev_detach will be deprecated, so, modify the sample code to use rte_eal_hotplug_add and rte_eal_hotplug_remove to attach/detach device. v12: - fix return value in eal_dev_hotplug_request_to_primary. - add more error log in rte_eal_hotplug_add. - fix return value in rte_eal_hotplug_add and rte_eal_hotplug_remove any failure due to IPC error will return -ENOMSG, but not -1. - remove unnecessary changes from previous rework. v11: - move out common code from pci_vfio_unmap_secondary and pci_vfio_unmap_primary. - move RTE_BUS_NAME_MAX_LEN and RTE_DEV_ARGS_MAX_LEN into hotplug_mp.h - fix reply check in eal_dev_hotplug_request_to_primary. - move skeleton code for attaching device from secondary from patch 6/19 to patch 5/19 to improve code readability. v10: - Since hotplug add/remove a vdev on a secondary process will sync on all processes now, it is not necessary to support private vdev for a secondary process which is identified by a not-NULL devargs in "--vdev". So re-work on all vdev driver changes to simpified device probe scenario on a secondary process, devargs will be ignored on secondary process now. - fix lisence header in example/multi-process/hotplug_mp/Makefile. v9: - Move hotplug IPC from rte_eth_dev_attach/rte_eth_dev_detach to eal_dev_hotplug_add and eal_dev_hotplug_remove, now all kinds of devices will be synced in multi-process. - Fix couple issue when a device is bound to vfio. 1) The device can't be detached clearly in a secondary process, which also cause it can't be attached again, due to the error that /dev/vfio/<group_fd> is still busy.(see Patch 3/19 and 4/19) 2) repeat detach/attach device will cause "cannot find TAILQ entry for PCI device" due to incorrect PCI address compare. (see patch 2/19). - Removed device lock. - Removed private device support. - Fix commit log grammar issue v8: - update rte_eal_version.map due to new API added. - minor reword on release note. - minor fix on commit log and code style. NOTE: Some issues which is not related with this patchset is expected when play with hotplug_mp sample as belows. - Attach a PCI device twice may cause device can't be detached below fix is required: https://patches.dpdk.org/patch/42030/ - ixgbe device can't detached, below fix is required https://patches.dpdk.org/patch/42031/ v7: - update rte_ethdev_version.map for new APIs. - improve code readability in __handle_secondary_request by use goto. - add comments to explain why need to call rte_eal_alarm_set. - add error log when process_mp_init_callbacks failed. - reword release notes base on Anatoly's suggestion. - add back previous "Acked-by" and "Reviewed-by" in commit log. NOTE: current patchset depends on below IPC fix, or it may not be able to attach a shared vdev. https://patches.dpdk.org/patch/41647/ v6: - remove bus->scan_one, since ABI break is not necessary. - remove patch for failsafe PMD since it will not support secondary. - fix wrong implemenation on ixgbe. - add rte_eth_dev_release_port_private into rte_eth_dev_pci_generic_remove for secondary process, so we don't need to patch on PMD if PMD use the default remove function. - add release notes update. - agreed to use strdup(peer) as workaround for repling a sync request in seperate thread. v5: - since we will keep mp thread separate from interrupt thread, it is not necessary to use temporary thread, we use rte_eal_alarm_set. - remove the change in rte_eth_dev_release_port, since there is a better way to prevent rte_eth_dev_release_port be called after rte_eth_dev_release_port_private. - fix the issue that lock does not take effect on secondary due to previous re-work - fix the issue when the first attached device is a private device from secondary. (patch 8/24) - work around for reply a sync request in separate thread, this is still an open and in discussion as below. https://mails.dpdk.org/archives/dev/2018-June/105359.html v4: - since mp thread will be merged to interrupt thread, the fix on v3 for sync IPC deadlock will not work. the new version enable the machanism to invoke a mp action callback in a temporary thread to avoid the IPC deadlock, with this, secondary to primary request impelemtation also be simplified, since we can use sync request directly in a separate thread. v3: - enable mp init callback register to help non-eal module to initialize mp channel during rte_eal_init - fix when attach share device from secondary. 1) dead lock due to sync IPC be invoked in rte_malloc in primary process when handle secondary request to attach device, the solution is primary process to issue share device attach/detach in interrupt thread. 2) return port_id not correct. - check nb_sent and nb_received in sync IPC. - fix memory leak duirng error handling at attach_on_secondary. - improve clean_lock_callback to only lock/unlock spinlock once - improve error code return in check-reply during async IPC. - remove rte_ prefix of internal function in ethdev_mp.c - sample code improvement. 1) rename sample to "hotplug_mp", and move to example/multi-process. 2) cleanup header include. 3) call rte_eal_cleanup before exit. v2: - rename rte_ethdev_mp.* to ethdev_mp.* - rename rte_ethdev_lock.* to ethdev_lock.* - move internal funciton to ethdev_private.h - separate rte_eth_dev_[un]lock into rte_eth_dev_[un]lock and rte_eth_dev_[un]lock_with_callback - lock callbacks will be removed automatically after device is detached. - add experimental tag for all new APIs. - fix coding style issue. - fix wrong lisence header in sample code. - fix spelling - fix meson.build. - improve comments. Background: =========== Currently secondary process will only sync ethdev from primary process at init stage, but it will not be aware if device is attached/detached on primary process at runtime. While there is the requirement from application that take primary-secondary process model. The primary process work as a resource management process, it will create/destroy virtual device at runtime, while the secondary process deal with the network stuff with these devices. Solution: ========= So the orignial intention is to fix this gap, but beyond that the patch set provide a more comprehesive solution to handle different hotplug cases in multi-process situation, it cover below scenario: 1. Attach a device from the primary 2. Detach a device from the primary 3. Attach a device from a secondary 4. Detach a device from a secondary In primary-secondary process model, we assume ethernet devices are shared by default. that means attach or detach a device on any process will broadcast to all other processes through mp channel then device information will be synchronized on all processes. Any failure during attaching or detaching process will cause inconsistent status between processes, so proper rollback action should be considered. Scenario for Case 1, 2: attach device from primary a) primary attach the new device if failed goto h). b) primary send attach sync request to all secondary. c) secondary receive request and attach device and send reply. d) primary check the reply if all success go to i). e) primary send attach rollback sync request to all secondary. f) secondary receive the request and detach device and send reply. g) primary receive the reply and detach device as rollback action. h) attach fail i) attach success detach device from primary a) primary perform pre-detach check, if device is locked, goto i). b) primary send pre-detach sync request to all secondary. c) secondary perform pre-detach check and send reply. d) primary check the reply if any fail goto i). e) primary send detach sync request to all secondary f) secondary detach the device and send reply (assume no fail) g) primary detach the device. h) detach success i) detach failed Scenario for case 3, 4: attach device from secondary: a) seconary send asycn request to primary and wait on a condition which will be released by matched response from primary. b) primary receive the request and attach the new device if failed goto i). c) primary forward attach request to all secondary as async request (because this in mp thread context, use sync request will deadlock, same reason for all following async request.) d) secondary receive request and attach device and send reply. e) primary check the reply if all success go to j). f) primary send attach rollback async request to all secondary. g) secondary receive the request and detach device and send reply. h) primary receive the reply and detach device as rollback action. i) send fail response to secondary, goto k). j) send success response to secondary. k) secondary process receive response and return. detach device from secondary: a) secondary send async request to primary and wait on a condition which will be released by matched response from primary. b) primary receive the request and perform pre-detach check, if device is locked, goto j). c) primary send pre-detach async request to all secondary. d) secondary perform pre-detach check and send reply. e) primary check the reply if any fail goto j). f) primary send detach async request to all secondary g) secondary detach the device and send reply h) primary detach the device. i) send success response to secondary, goto k). j) send fail response to secondary. k) secondary process receive response and return. APIs chenages: ============== scope of rte_eal_hotplug_add and rte_eal_hotplug_remove is extended. In primary-secondary process model, rte_eal_hotplug_add will guarantee that device be attached on all processes, while rte_eal_hotplug_remove will guarantee device be detached on all processes. PMD Impact: =========== Currently device removing is not handled well in secondary process on most pmd drivers, rte_eth_dev_relase_port will be invoked and will mess up primary process since it reset all shared data. So we introduced new API rte_eth_dev_release_port_local which only reset ethdev's state to unsued but not touch shared data so other process will not be impacted. Since not all device driver is target to support primary-secondary process model, so the patch set only fix this on all Intel devices and vdev, it can be refereneced by other driver when equevalent fix is required Example: ======== The patchset also contains a example to demonstrate device hotplug in multi-process model, below are detail instructions. /* start sample code as primary then secondary */ ./hotplug_mp --proc-type=auto Command Line Example: >help >list /* attach a pci device */ > attach 0000:81:00.0 /* detach the pci device */ > detach 0000:81:00.0 /* attach a vdev af_packet device */ > attach net_af_packet,iface=eth0 /* detach the vdev af_packet device */ > detach net_af_packet Qi Zhang (19): ethdev: add function to release port in local process bus/pci: fix PCI address compare bus/pci: enable vfio unmap resource for secondary vfio: remove uneccessary IPC for group fd clear eal: enable hotplug on multi-process eal: support attach or detach share device from secondary net/i40e: enable hotplug on secondary process net/ixgbe: enable hotplug on secondary process net/af_packet: enable hotplug on secondary process net/bonding: enable hotplug on secondary process net/kni: enable hotplug on secondary process net/null: enable hotplug on secondary process net/octeontx: enable hotplug on secondary process net/pcap: enable hotplug on secondary process net/softnic: enable hotplug on secondary process net/tap: enable hotplug on secondary process net/vhost: enable hotplug on secondary process examples/multi_process: add hotplug sample doc: update release notes for multi process hotplug doc/guides/rel_notes/release_18_08.rst | 11 + drivers/bus/pci/linux/pci_vfio.c | 129 +++++++-- drivers/net/af_packet/rte_eth_af_packet.c | 7 +- drivers/net/bonding/rte_eth_bond_pmd.c | 7 +- drivers/net/i40e/i40e_ethdev.c | 2 + drivers/net/ixgbe/ixgbe_ethdev.c | 3 + drivers/net/kni/rte_eth_kni.c | 7 +- drivers/net/null/rte_eth_null.c | 12 +- drivers/net/octeontx/octeontx_ethdev.c | 9 + drivers/net/pcap/rte_eth_pcap.c | 11 +- drivers/net/softnic/rte_eth_softnic.c | 15 +- drivers/net/tap/rte_eth_tap.c | 13 +- drivers/net/vhost/rte_eth_vhost.c | 7 +- examples/multi_process/Makefile | 1 + examples/multi_process/hotplug_mp/Makefile | 23 ++ examples/multi_process/hotplug_mp/commands.c | 214 +++++++++++++++ examples/multi_process/hotplug_mp/commands.h | 10 + examples/multi_process/hotplug_mp/main.c | 41 +++ lib/librte_eal/bsdapp/eal/Makefile | 1 + lib/librte_eal/common/eal_common_dev.c | 177 ++++++++++++- lib/librte_eal/common/eal_private.h | 37 +++ lib/librte_eal/common/hotplug_mp.c | 348 +++++++++++++++++++++++++ lib/librte_eal/common/hotplug_mp.h | 48 ++++ lib/librte_eal/common/include/rte_dev.h | 6 + lib/librte_eal/common/meson.build | 1 + lib/librte_eal/linuxapp/eal/Makefile | 1 + lib/librte_eal/linuxapp/eal/eal.c | 6 + lib/librte_eal/linuxapp/eal/eal_vfio.c | 45 +--- lib/librte_eal/linuxapp/eal/eal_vfio.h | 1 - lib/librte_eal/linuxapp/eal/eal_vfio_mp_sync.c | 8 - lib/librte_ethdev/rte_ethdev.c | 12 + lib/librte_ethdev/rte_ethdev_driver.h | 16 +- lib/librte_ethdev/rte_ethdev_pci.h | 8 + 33 files changed, 1134 insertions(+), 103 deletions(-) create mode 100644 examples/multi_process/hotplug_mp/Makefile create mode 100644 examples/multi_process/hotplug_mp/commands.c create mode 100644 examples/multi_process/hotplug_mp/commands.h create mode 100644 examples/multi_process/hotplug_mp/main.c create mode 100644 lib/librte_eal/common/hotplug_mp.c create mode 100644 lib/librte_eal/common/hotplug_mp.h
v13: - Since rte_eth_dev_attach/rte_eth_dev_detach will be deprecated, so, modify the sample code to use rte_eal_hotplug_add and rte_eal_hotplug_remove to attach/detach device. v12: - fix return value in eal_dev_hotplug_request_to_primary. - add more error log in rte_eal_hotplug_add. - fix return value in rte_eal_hotplug_add and rte_eal_hotplug_remove any failure due to IPC error will return -ENOMSG, but not -1. - remove unnecessary changes from previous rework. v11: - move out common code from pci_vfio_unmap_secondary and pci_vfio_unmap_primary. - move RTE_BUS_NAME_MAX_LEN and RTE_DEV_ARGS_MAX_LEN into hotplug_mp.h - fix reply check in eal_dev_hotplug_request_to_primary. - move skeleton code for attaching device from secondary from patch 6/19 to patch 5/19 to improve code readability. v10: - Since hotplug add/remove a vdev on a secondary process will sync on all processes now, it is not necessary to support private vdev for a secondary process which is identified by a not-NULL devargs in "--vdev". So re-work on all vdev driver changes to simpified device probe scenario on a secondary process, devargs will be ignored on secondary process now. - fix lisence header in example/multi-process/hotplug_mp/Makefile. v9: - Move hotplug IPC from rte_eth_dev_attach/rte_eth_dev_detach to eal_dev_hotplug_add and eal_dev_hotplug_remove, now all kinds of devices will be synced in multi-process. - Fix couple issue when a device is bound to vfio. 1) The device can't be detached clearly in a secondary process, which also cause it can't be attached again, due to the error that /dev/vfio/<group_fd> is still busy.(see Patch 3/19 and 4/19) 2) repeat detach/attach device will cause "cannot find TAILQ entry for PCI device" due to incorrect PCI address compare. (see patch 2/19). - Removed device lock. - Removed private device support. - Fix commit log grammar issue v8: - update rte_eal_version.map due to new API added. - minor reword on release note. - minor fix on commit log and code style. NOTE: Some issues which is not related with this patchset is expected when play with hotplug_mp sample as belows. - Attach a PCI device twice may cause device can't be detached below fix is required: https://patches.dpdk.org/patch/42030/ - ixgbe device can't detached, below fix is required https://patches.dpdk.org/patch/42031/ v7: - update rte_ethdev_version.map for new APIs. - improve code readability in __handle_secondary_request by use goto. - add comments to explain why need to call rte_eal_alarm_set. - add error log when process_mp_init_callbacks failed. - reword release notes base on Anatoly's suggestion. - add back previous "Acked-by" and "Reviewed-by" in commit log. NOTE: current patchset depends on below IPC fix, or it may not be able to attach a shared vdev. https://patches.dpdk.org/patch/41647/ v6: - remove bus->scan_one, since ABI break is not necessary. - remove patch for failsafe PMD since it will not support secondary. - fix wrong implemenation on ixgbe. - add rte_eth_dev_release_port_private into rte_eth_dev_pci_generic_remove for secondary process, so we don't need to patch on PMD if PMD use the default remove function. - add release notes update. - agreed to use strdup(peer) as workaround for repling a sync request in seperate thread. v5: - since we will keep mp thread separate from interrupt thread, it is not necessary to use temporary thread, we use rte_eal_alarm_set. - remove the change in rte_eth_dev_release_port, since there is a better way to prevent rte_eth_dev_release_port be called after rte_eth_dev_release_port_private. - fix the issue that lock does not take effect on secondary due to previous re-work - fix the issue when the first attached device is a private device from secondary. (patch 8/24) - work around for reply a sync request in separate thread, this is still an open and in discussion as below. https://mails.dpdk.org/archives/dev/2018-June/105359.html v4: - since mp thread will be merged to interrupt thread, the fix on v3 for sync IPC deadlock will not work. the new version enable the machanism to invoke a mp action callback in a temporary thread to avoid the IPC deadlock, with this, secondary to primary request impelemtation also be simplified, since we can use sync request directly in a separate thread. v3: - enable mp init callback register to help non-eal module to initialize mp channel during rte_eal_init - fix when attach share device from secondary. 1) dead lock due to sync IPC be invoked in rte_malloc in primary process when handle secondary request to attach device, the solution is primary process to issue share device attach/detach in interrupt thread. 2) return port_id not correct. - check nb_sent and nb_received in sync IPC. - fix memory leak duirng error handling at attach_on_secondary. - improve clean_lock_callback to only lock/unlock spinlock once - improve error code return in check-reply during async IPC. - remove rte_ prefix of internal function in ethdev_mp.c - sample code improvement. 1) rename sample to "hotplug_mp", and move to example/multi-process. 2) cleanup header include. 3) call rte_eal_cleanup before exit. v2: - rename rte_ethdev_mp.* to ethdev_mp.* - rename rte_ethdev_lock.* to ethdev_lock.* - move internal funciton to ethdev_private.h - separate rte_eth_dev_[un]lock into rte_eth_dev_[un]lock and rte_eth_dev_[un]lock_with_callback - lock callbacks will be removed automatically after device is detached. - add experimental tag for all new APIs. - fix coding style issue. - fix wrong lisence header in sample code. - fix spelling - fix meson.build. - improve comments. Background: =========== Currently secondary process will only sync ethdev from primary process at init stage, but it will not be aware if device is attached/detached on primary process at runtime. While there is the requirement from application that take primary-secondary process model. The primary process work as a resource management process, it will create/destroy virtual device at runtime, while the secondary process deal with the network stuff with these devices. Solution: ========= So the orignial intention is to fix this gap, but beyond that the patch set provide a more comprehesive solution to handle different hotplug cases in multi-process situation, it cover below scenario: 1. Attach a device from the primary 2. Detach a device from the primary 3. Attach a device from a secondary 4. Detach a device from a secondary In primary-secondary process model, we assume ethernet devices are shared by default. that means attach or detach a device on any process will broadcast to all other processes through mp channel then device information will be synchronized on all processes. Any failure during attaching or detaching process will cause inconsistent status between processes, so proper rollback action should be considered. Scenario for Case 1, 2: attach device from primary a) primary attach the new device if failed goto h). b) primary send attach sync request to all secondary. c) secondary receive request and attach device and send reply. d) primary check the reply if all success go to i). e) primary send attach rollback sync request to all secondary. f) secondary receive the request and detach device and send reply. g) primary receive the reply and detach device as rollback action. h) attach fail i) attach success detach device from primary a) primary perform pre-detach check, if device is locked, goto i). b) primary send pre-detach sync request to all secondary. c) secondary perform pre-detach check and send reply. d) primary check the reply if any fail goto i). e) primary send detach sync request to all secondary f) secondary detach the device and send reply (assume no fail) g) primary detach the device. h) detach success i) detach failed Scenario for case 3, 4: attach device from secondary: a) seconary send asycn request to primary and wait on a condition which will be released by matched response from primary. b) primary receive the request and attach the new device if failed goto i). c) primary forward attach request to all secondary as async request (because this in mp thread context, use sync request will deadlock, same reason for all following async request.) d) secondary receive request and attach device and send reply. e) primary check the reply if all success go to j). f) primary send attach rollback async request to all secondary. g) secondary receive the request and detach device and send reply. h) primary receive the reply and detach device as rollback action. i) send fail response to secondary, goto k). j) send success response to secondary. k) secondary process receive response and return. detach device from secondary: a) secondary send async request to primary and wait on a condition which will be released by matched response from primary. b) primary receive the request and perform pre-detach check, if device is locked, goto j). c) primary send pre-detach async request to all secondary. d) secondary perform pre-detach check and send reply. e) primary check the reply if any fail goto j). f) primary send detach async request to all secondary g) secondary detach the device and send reply h) primary detach the device. i) send success response to secondary, goto k). j) send fail response to secondary. k) secondary process receive response and return. APIs chenages: ============== scope of rte_eal_hotplug_add and rte_eal_hotplug_remove is extended. In primary-secondary process model, rte_eal_hotplug_add will guarantee that device be attached on all processes, while rte_eal_hotplug_remove will guarantee device be detached on all processes. PMD Impact: =========== Currently device removing is not handled well in secondary process on most pmd drivers, rte_eth_dev_relase_port will be invoked and will mess up primary process since it reset all shared data. So we introduced new API rte_eth_dev_release_port_local which only reset ethdev's state to unsued but not touch shared data so other process will not be impacted. Since not all device driver is target to support primary-secondary process model, so the patch set only fix this on all Intel devices and vdev, it can be refereneced by other driver when equevalent fix is required Example: ======== The patchset also contains a example to demonstrate device hotplug in multi-process model, below are detail instructions. /* start sample code as primary then secondary */ ./hotplug_mp --proc-type=auto Command Line Example: >help >list /* attach a pci device */ > attach 0000:81:00.0 /* detach the pci device */ > detach 0000:81:00.0 /* attach a vdev af_packet device */ > attach net_af_packet,iface=eth0 /* detach the vdev af_packet device */ > detach net_af_packet Qi Zhang (19): ethdev: add function to release port in local process bus/pci: fix PCI address compare bus/pci: enable vfio unmap resource for secondary vfio: remove uneccessary IPC for group fd clear eal: enable hotplug on multi-process eal: support attach or detach share device from secondary net/i40e: enable hotplug on secondary process net/ixgbe: enable hotplug on secondary process net/af_packet: enable hotplug on secondary process net/bonding: enable hotplug on secondary process net/kni: enable hotplug on secondary process net/null: enable hotplug on secondary process net/octeontx: enable hotplug on secondary process net/pcap: enable hotplug on secondary process net/softnic: enable hotplug on secondary process net/tap: enable hotplug on secondary process net/vhost: enable hotplug on secondary process examples/multi_process: add hotplug sample doc: update release notes for multi process hotplug doc/guides/rel_notes/release_18_08.rst | 11 + drivers/bus/pci/linux/pci_vfio.c | 129 +++++++-- drivers/net/af_packet/rte_eth_af_packet.c | 7 +- drivers/net/bonding/rte_eth_bond_pmd.c | 7 +- drivers/net/i40e/i40e_ethdev.c | 2 + drivers/net/ixgbe/ixgbe_ethdev.c | 3 + drivers/net/kni/rte_eth_kni.c | 7 +- drivers/net/null/rte_eth_null.c | 12 +- drivers/net/octeontx/octeontx_ethdev.c | 9 + drivers/net/pcap/rte_eth_pcap.c | 11 +- drivers/net/softnic/rte_eth_softnic.c | 15 +- drivers/net/tap/rte_eth_tap.c | 13 +- drivers/net/vhost/rte_eth_vhost.c | 7 +- examples/multi_process/Makefile | 1 + examples/multi_process/hotplug_mp/Makefile | 23 ++ examples/multi_process/hotplug_mp/commands.c | 214 +++++++++++++++ examples/multi_process/hotplug_mp/commands.h | 10 + examples/multi_process/hotplug_mp/main.c | 41 +++ lib/librte_eal/bsdapp/eal/Makefile | 1 + lib/librte_eal/common/eal_common_dev.c | 177 ++++++++++++- lib/librte_eal/common/eal_private.h | 37 +++ lib/librte_eal/common/hotplug_mp.c | 348 +++++++++++++++++++++++++ lib/librte_eal/common/hotplug_mp.h | 48 ++++ lib/librte_eal/common/include/rte_dev.h | 6 + lib/librte_eal/common/meson.build | 1 + lib/librte_eal/linuxapp/eal/Makefile | 1 + lib/librte_eal/linuxapp/eal/eal.c | 6 + lib/librte_eal/linuxapp/eal/eal_vfio.c | 45 +--- lib/librte_eal/linuxapp/eal/eal_vfio.h | 1 - lib/librte_eal/linuxapp/eal/eal_vfio_mp_sync.c | 8 - lib/librte_ethdev/rte_ethdev.c | 12 + lib/librte_ethdev/rte_ethdev_driver.h | 16 +- lib/librte_ethdev/rte_ethdev_pci.h | 8 + 33 files changed, 1134 insertions(+), 103 deletions(-) create mode 100644 examples/multi_process/hotplug_mp/Makefile create mode 100644 examples/multi_process/hotplug_mp/commands.c create mode 100644 examples/multi_process/hotplug_mp/commands.h create mode 100644 examples/multi_process/hotplug_mp/main.c create mode 100644 lib/librte_eal/common/hotplug_mp.c create mode 100644 lib/librte_eal/common/hotplug_mp.h
12/07/2018 03:14, Qi Zhang: > v13: > - Since rte_eth_dev_attach/rte_eth_dev_detach will be deprecated, > so, modify the sample code to use rte_eal_hotplug_add and > rte_eal_hotplug_remove to attach/detach device. Yes, this is what I tried to explain you. I think it is now too late for 18.08. We see that this patchset deserves more reviews.
> -----Original Message----- > From: Thomas Monjalon [mailto:thomas@monjalon.net] > Sent: Thursday, July 12, 2018 4:30 PM > To: Zhang, Qi Z <qi.z.zhang@intel.com> > Cc: dev@dpdk.org; Burakov, Anatoly <anatoly.burakov@intel.com>; Ananyev, > Konstantin <konstantin.ananyev@intel.com>; Richardson, Bruce > <bruce.richardson@intel.com>; Yigit, Ferruh <ferruh.yigit@intel.com>; Shelton, > Benjamin H <benjamin.h.shelton@intel.com>; Vangati, Narender > <narender.vangati@intel.com> > Subject: Re: [dpdk-dev] [PATCH v12 00/19] enable hotplug on multi-process > > 12/07/2018 03:14, Qi Zhang: > > v13: > > - Since rte_eth_dev_attach/rte_eth_dev_detach will be deprecated, > > so, modify the sample code to use rte_eal_hotplug_add and > > rte_eal_hotplug_remove to attach/detach device. > > Yes, this is what I tried to explain you. > > I think it is now too late for 18.08. Understand, but probably patch 2,3,4 could be considered in 18.08, since they fix general issue but not just for hotplug mp. What do you think? > We see that this patchset deserves more reviews. > >
12/07/2018 11:11, Zhang, Qi Z: > From: Thomas Monjalon [mailto:thomas@monjalon.net] > > 12/07/2018 03:14, Qi Zhang: > > > v13: > > > - Since rte_eth_dev_attach/rte_eth_dev_detach will be deprecated, > > > so, modify the sample code to use rte_eal_hotplug_add and > > > rte_eal_hotplug_remove to attach/detach device. > > > > Yes, this is what I tried to explain you. > > > > I think it is now too late for 18.08. > > Understand, but probably patch 2,3,4 could be considered in 18.08, since they fix general issue but not just for hotplug mp. > What do you think? Yes, you are right. Please send a separate patchset and try to get reviews. Gaetan, Anatoly, please review patches 2 and 3.