Linuxarm May 2024

linuxarm@openeuler.org

1 participants
4 discussions

Re: [PATCH v6 2/5] hisi_acc_vfio_pci: modify the register location of the XQC address
by Shameerali Kolothum Thodi 13 May '24

13 May '24

> -----Original Message----- > From: liulongfang <liulongfang(a)huawei.com> > Sent: Monday, May 13, 2024 9:16 AM > To: Alex Williamson <alex.williamson(a)redhat.com>; Shameerali Kolothum > Thodi <shameerali.kolothum.thodi(a)huawei.com> > Cc: jgg(a)nvidia.com; Jonathan Cameron <jonathan.cameron(a)huawei.com>; > kvm(a)vger.kernel.org; linux-kernel(a)vger.kernel.org; > linuxarm(a)openeuler.org > Subject: Re: [PATCH v6 2/5] hisi_acc_vfio_pci: modify the register location of > the XQC address > > On 2024/5/9 22:29, Alex Williamson wrote: > > On Thu, 9 May 2024 09:37:51 +0000 > > Shameerali Kolothum Thodi <shameerali.kolothum.thodi(a)huawei.com> > wrote: > > > >>> -----Original Message----- > >>> From: Alex Williamson <alex.williamson(a)redhat.com> > >>> Sent: Wednesday, May 8, 2024 7:00 PM > >>> To: liulongfang <liulongfang(a)huawei.com> > >>> Cc: jgg(a)nvidia.com; Shameerali Kolothum Thodi > >>> <shameerali.kolothum.thodi(a)huawei.com>; Jonathan Cameron > >>> <jonathan.cameron(a)huawei.com>; kvm(a)vger.kernel.org; linux- > >>> kernel(a)vger.kernel.org; linuxarm(a)openeuler.org > >>> Subject: Re: [PATCH v6 2/5] hisi_acc_vfio_pci: modify the register > location of > >>> the XQC address > >> > >> [...] > >> > >>>> HiSilicon accelerator equipment can perform general services after > >>> completing live migration. > >>>> This kind of business is executed through the user mode driver and only > >>> needs to use SQE and CQE. > >>>> > >>>> At the same time, this device can also perform kernel-mode services in > the > >>> VM through the crypto > >>>> subsystem. This kind of service requires the use of EQE. > >>>> > >>>> Finally, if the device is abnormal, the driver needs to perform a device > >>> reset, and AEQE needs to > >>>> be used in this case. > >>>> > >>>> Therefore, a complete device live migration function needs to ensure > that > >>> device functions are > >>>> normal in all these scenarios. > >>>> Therefore, this data still needs to be migrated. > >>> > >>> Ok, I had jumped to an in-kernel host driver in reference to "kernel > >>> mode" rather than a guest kernel. Migrating with bad data only affects > >>> the current configuration of the device, reloading a guest driver to > >>> update these registers or a reset of the device would allow proper > >>> operation of the device, correct? > >> > >> Yes, after talking to Longfang, the device RAS will trigger a reset and > >> would function after reset. > >> > >>> > >>> But I think this still isn't really a complete solution, we know > >>> there's a bug in the migration data stream, so not only would we fix > >>> the data stream, but I think we should also take measures to prevent > >>> loading a known bad data stream. AIUI migration of this device while > >>> running in kernel mode (ie. a kernel driver within a guest VM) is > >>> broken. Therefore, the least we can do in a new kernel, knowing that > >>> there was previously a bug in the migration data stream, is to fail to > >>> load that migration data because it risks this scenario where the > >>> device is broken after migration. Shouldn't we then also increment a > >>> migration version field in the data stream to block migrations that > >>> risk this breakage, or barring that, change the magic data field to > >>> prevent the migration? Thanks, > >> > >> Ok. We could add a new ACC_DEV_MAGIC_V2 and prevent the migration > >> in vf_qm_check_match(). The only concern here is that, it will completely > >> block old kernel to new kernel migration and since we can recover the > >> device after the reset whether it is too restrictive or not. > > > > What's the impact to the running driver, kernel or userspace, if the > > device is reset? Migration is intended to be effectively transparent > > If the device is reset, the user's task needs to be restarted. > If an exception has been detected, the best way is not to migrate. > > > to the driver. If the driver stalls and needs to reset the device, > > what has the migration driver accomplished versus an offline migration? > > > > If there's a way to detect from the migration data if the device is > > running in kernel mode or user mode then you could potentially accept > > and send v1 magic conditional that the device is in user mode and > > require v2 magic for any migration where the device is in kernel mode. > > This all adds complication though and seems like it has corner cases > > where we might allow migration to an old kernel that might trap the > > device there if the use case changes. > > > > The driver does not support checking whether the device is running in > kernel mode or user mode. > Moreover, the device supports user-mode services and kernel-mode services > to run at the same time. > > > Essentially it comes down to what should the migration experience be > > and while restricting old->new and new->old migration is undesirable, > > it seems old->old migration is effectively already broken anyway. As > > you consider a v2 magic, perhaps consider how the migration data > > structure might be improved overall to better handle new features and > > bugs. Thanks, > > > > We discussed a plan: > Update ACC_DEV_MAGIC to ACC_DEV_MAGIC_VERSION and configure its > last byte > as version information: > > /* QM match information, last byte is version number */ > #define ACC_DEV_MAGIC_VERSION 0XACCDEVFEEDCAFE01 Oops..cant have V there. But the idea is replace magic with last byte as version info which can be used in future for handling bugs/features etc. Thanks, Shameer

1 0

Re: [PATCH v6 2/5] hisi_acc_vfio_pci: modify the register location of the XQC address
by Shameerali Kolothum Thodi 09 May '24

09 May '24

> -----Original Message----- > From: Alex Williamson <alex.williamson(a)redhat.com> > Sent: Wednesday, May 8, 2024 7:00 PM > To: liulongfang <liulongfang(a)huawei.com> > Cc: jgg(a)nvidia.com; Shameerali Kolothum Thodi > <shameerali.kolothum.thodi(a)huawei.com>; Jonathan Cameron > <jonathan.cameron(a)huawei.com>; kvm(a)vger.kernel.org; linux- > kernel(a)vger.kernel.org; linuxarm(a)openeuler.org > Subject: Re: [PATCH v6 2/5] hisi_acc_vfio_pci: modify the register location of > the XQC address [...] > > HiSilicon accelerator equipment can perform general services after > completing live migration. > > This kind of business is executed through the user mode driver and only > needs to use SQE and CQE. > > > > At the same time, this device can also perform kernel-mode services in the > VM through the crypto > > subsystem. This kind of service requires the use of EQE. > > > > Finally, if the device is abnormal, the driver needs to perform a device > reset, and AEQE needs to > > be used in this case. > > > > Therefore, a complete device live migration function needs to ensure that > device functions are > > normal in all these scenarios. > > Therefore, this data still needs to be migrated. > > Ok, I had jumped to an in-kernel host driver in reference to "kernel > mode" rather than a guest kernel. Migrating with bad data only affects > the current configuration of the device, reloading a guest driver to > update these registers or a reset of the device would allow proper > operation of the device, correct? Yes, after talking to Longfang, the device RAS will trigger a reset and would function after reset. > > But I think this still isn't really a complete solution, we know > there's a bug in the migration data stream, so not only would we fix > the data stream, but I think we should also take measures to prevent > loading a known bad data stream. AIUI migration of this device while > running in kernel mode (ie. a kernel driver within a guest VM) is > broken. Therefore, the least we can do in a new kernel, knowing that > there was previously a bug in the migration data stream, is to fail to > load that migration data because it risks this scenario where the > device is broken after migration. Shouldn't we then also increment a > migration version field in the data stream to block migrations that > risk this breakage, or barring that, change the magic data field to > prevent the migration? Thanks, Ok. We could add a new ACC_DEV_MAGIC_V2 and prevent the migration in vf_qm_check_match(). The only concern here is that, it will completely block old kernel to new kernel migration and since we can recover the device after the reset whether it is too restrictive or not. Thanks, Shameer

1 0

Re: [PATCH v6 2/5] hisi_acc_vfio_pci: modify the register location of the XQC address
by Shameerali Kolothum Thodi 07 May '24

07 May '24

> -----Original Message----- > From: liulongfang <liulongfang(a)huawei.com> > Sent: Tuesday, May 7, 2024 9:29 AM > To: Alex Williamson <alex.williamson(a)redhat.com> > Cc: jgg(a)nvidia.com; Shameerali Kolothum Thodi > <shameerali.kolothum.thodi(a)huawei.com>; Jonathan Cameron > <jonathan.cameron(a)huawei.com>; kvm(a)vger.kernel.org; linux- > kernel(a)vger.kernel.org; linuxarm(a)openeuler.org > Subject: Re: [PATCH v6 2/5] hisi_acc_vfio_pci: modify the register location of > the XQC address > > On 2024/5/4 0:11, Alex Williamson wrote: > > On Thu, 25 Apr 2024 21:23:19 +0800 > > Longfang Liu <liulongfang(a)huawei.com> wrote: > > > >> According to the latest hardware register specification. The DMA > >> addresses of EQE and AEQE are not at the front of their respective > >> register groups, but start from the second. > >> So, previously fetching the value starting from the first register > >> would result in an incorrect address. > >> > >> Therefore, the register location from which the address is obtained > >> needs to be modified. > > > > How does this affect migration? Has it ever worked? Does this make > > The general HiSilicon accelerator task will only use SQE and CQE. > EQE is only used when user running kernel mode task and uses interrupt mode. > AEQE is only used when user running task exceptions occur and software reset > is required. > > The DMA addresses of these four queues are written to the device by the device > driver through the mailbox command during driver initialization. > The DMA addresses of EQE and AEQE are migrated through the device register. > > EQE and AEQE are not used in general task, after the live migration is > completed, > this DMA address error will not be found. until we added a new kernel-mode test > case > that we discovered that this address was abnormal. > > > the migration data incompatible? > > > > This address only affects the kernel mode interrupt mode task function and > device > exception recovery function. > They do not affect live migration functionality > > > Fixes: ??? > > OK! Hi, Could you please add the Fixes tag and resend this separately if there are no outstanding comments on this. This is not related to the debugfs support anyway. Thanks, Shameer > Thanks. > Longfang. > >> Signed-off-by: Longfang Liu <liulongfang(a)huawei.com> > >> --- > >> drivers/vfio/pci/hisilicon/hisi_acc_vfio_pci.c | 8 ++++---- > >> drivers/vfio/pci/hisilicon/hisi_acc_vfio_pci.h | 3 +++ > >> 2 files changed, 7 insertions(+), 4 deletions(-) > >> > >> diff --git a/drivers/vfio/pci/hisilicon/hisi_acc_vfio_pci.c > b/drivers/vfio/pci/hisilicon/hisi_acc_vfio_pci.c > >> index 45351be8e270..0c7e31076ff4 100644 > >> --- a/drivers/vfio/pci/hisilicon/hisi_acc_vfio_pci.c > >> +++ b/drivers/vfio/pci/hisilicon/hisi_acc_vfio_pci.c > >> @@ -516,12 +516,12 @@ static int vf_qm_state_save(struct > hisi_acc_vf_core_device *hisi_acc_vdev, > >> return -EINVAL; > >> > >> /* Every reg is 32 bit, the dma address is 64 bit. */ > >> - vf_data->eqe_dma = vf_data->qm_eqc_dw[1]; > >> + vf_data->eqe_dma = vf_data->qm_eqc_dw[QM_XQC_ADDR_HIGH]; > >> vf_data->eqe_dma <<= QM_XQC_ADDR_OFFSET; > >> - vf_data->eqe_dma |= vf_data->qm_eqc_dw[0]; > >> - vf_data->aeqe_dma = vf_data->qm_aeqc_dw[1]; > >> + vf_data->eqe_dma |= vf_data->qm_eqc_dw[QM_XQC_ADDR_LOW]; > >> + vf_data->aeqe_dma = vf_data->qm_aeqc_dw[QM_XQC_ADDR_HIGH]; > >> vf_data->aeqe_dma <<= QM_XQC_ADDR_OFFSET; > >> - vf_data->aeqe_dma |= vf_data->qm_aeqc_dw[0]; > >> + vf_data->aeqe_dma |= vf_data->qm_aeqc_dw[QM_XQC_ADDR_LOW]; > >> > >> /* Through SQC_BT/CQC_BT to get sqc and cqc address */ > >> ret = qm_get_sqc(vf_qm, &vf_data->sqc_dma); > >> diff --git a/drivers/vfio/pci/hisilicon/hisi_acc_vfio_pci.h > b/drivers/vfio/pci/hisilicon/hisi_acc_vfio_pci.h > >> index 5bab46602fad..f887ab98581c 100644 > >> --- a/drivers/vfio/pci/hisilicon/hisi_acc_vfio_pci.h > >> +++ b/drivers/vfio/pci/hisilicon/hisi_acc_vfio_pci.h > >> @@ -38,6 +38,9 @@ > >> #define QM_REG_ADDR_OFFSET 0x0004 > >> > >> #define QM_XQC_ADDR_OFFSET 32U > >> +#define QM_XQC_ADDR_LOW 0x1 > >> +#define QM_XQC_ADDR_HIGH 0x2 > >> + > >> #define QM_VF_AEQ_INT_MASK 0x0004 > >> #define QM_VF_EQ_INT_MASK 0x000c > >> #define QM_IFC_INT_SOURCE_V 0x0020 > > > > . > >

1 0

Re: [PATCH v6 4/5] hisi_acc_vfio_pci: register debugfs for hisilicon migration driver
by Shameerali Kolothum Thodi 07 May '24

07 May '24

> -----Original Message----- > From: Alex Williamson <alex.williamson(a)redhat.com> > Sent: Friday, May 3, 2024 6:21 PM > To: liulongfang <liulongfang(a)huawei.com> > Cc: jgg(a)nvidia.com; Shameerali Kolothum Thodi > <shameerali.kolothum.thodi(a)huawei.com>; Jonathan Cameron > <jonathan.cameron(a)huawei.com>; kvm(a)vger.kernel.org; linux- > kernel(a)vger.kernel.org; linuxarm(a)openeuler.org > Subject: Re: [PATCH v6 4/5] hisi_acc_vfio_pci: register debugfs for hisilicon > migration driver > > On Thu, 25 Apr 2024 21:23:21 +0800 > Longfang Liu <liulongfang(a)huawei.com> wrote: > > > On the debugfs framework of VFIO, if the CONFIG_VFIO_DEBUGFS macro is > > enabled, the debug function is registered for the live migration driver > > of the HiSilicon accelerator device. > > > > After registering the HiSilicon accelerator device on the debugfs > > framework of live migration of vfio, a directory file "hisi_acc" > > of debugfs is created, and then three debug function files are > > created in this directory: > > > > vfio > > | > > +---<dev_name1> > > | +---migration > > | +--state > > | +--hisi_acc > > | +--dev_data > > | +--migf_data > > | +--cmd_state > > | > > +---<dev_name2> > > +---migration > > +--state > > +--hisi_acc > > +--dev_data > > +--migf_data > > +--cmd_state > > > > dev_data file: read device data that needs to be migrated from the > > current device in real time > > migf_data file: read the migration data of the last live migration > > from the current driver. > > cmd_state: used to get the cmd channel state for the device. > > > > +----------------+ +--------------+ +---------------+ > > | migration dev | | src dev | | dst dev | > > +-------+--------+ +------+-------+ +-------+-------+ > > | | | > > | +------v-------+ +-------v-------+ > > | | saving_mif | | resuming_migf | > > read | | file | | file | > > | +------+-------+ +-------+-------+ > > | | copy | > > | +------------+----------+ > > | | > > +-------v---------+ +-------v--------+ > > | data buffer | | debug_migf | > > +-------+---------+ +-------+--------+ > > | | > > cat | cat | > > +-------v--------+ +-------v--------+ > > | dev_data | | migf_data | > > +----------------+ +----------------+ > > > > When accessing debugfs, user can obtain the real-time status data > > of the device through the "dev_data" file. It will directly read > > the device status data and will not affect the live migration > > function. Its data is stored in the allocated memory buffer, > > and the memory is released after data returning to user mode. > > > > To obtain the data of the last complete migration, user need to > > obtain it through the "migf_data" file. Since the live migration > > data only exists during the migration process, it is destroyed > > after the migration is completed. > > In order to save this data, a debug_migf file is created in the > > driver. At the end of the live migration process, copy the data > > to debug_migf. > > > > Signed-off-by: Longfang Liu <liulongfang(a)huawei.com> > > --- > > .../vfio/pci/hisilicon/hisi_acc_vfio_pci.c | 225 ++++++++++++++++++ > > .../vfio/pci/hisilicon/hisi_acc_vfio_pci.h | 7 + > > 2 files changed, 232 insertions(+) > > > > diff --git a/drivers/vfio/pci/hisilicon/hisi_acc_vfio_pci.c > b/drivers/vfio/pci/hisilicon/hisi_acc_vfio_pci.c > > index bf358ba94b5d..656b3d975940 100644 > > --- a/drivers/vfio/pci/hisilicon/hisi_acc_vfio_pci.c > > +++ b/drivers/vfio/pci/hisilicon/hisi_acc_vfio_pci.c > > @@ -627,15 +627,33 @@ static void hisi_acc_vf_disable_fd(struct > hisi_acc_vf_migration_file *migf) > > mutex_unlock(&migf->lock); > > } > > > > +static void hisi_acc_debug_migf_copy(struct hisi_acc_vf_core_device > *hisi_acc_vdev, > > + struct hisi_acc_vf_migration_file *src_migf) > > +{ > > + struct hisi_acc_vf_migration_file *dst_migf = hisi_acc_vdev- > >debug_migf; > > + > > + if (!dst_migf) > > + return; > > + > > + mutex_lock(&hisi_acc_vdev->enable_mutex); > > + dst_migf->disabled = src_migf->disabled; > > + dst_migf->total_length = src_migf->total_length; > > + memcpy(&dst_migf->vf_data, &src_migf->vf_data, > > + sizeof(struct acc_vf_data)); > > + mutex_unlock(&hisi_acc_vdev->enable_mutex); > > +} > > + > > static void hisi_acc_vf_disable_fds(struct hisi_acc_vf_core_device > *hisi_acc_vdev) > > { > > if (hisi_acc_vdev->resuming_migf) { > > + hisi_acc_debug_migf_copy(hisi_acc_vdev, hisi_acc_vdev- > >resuming_migf); > > hisi_acc_vf_disable_fd(hisi_acc_vdev->resuming_migf); > > fput(hisi_acc_vdev->resuming_migf->filp); > > hisi_acc_vdev->resuming_migf = NULL; > > } > > > > if (hisi_acc_vdev->saving_migf) { > > + hisi_acc_debug_migf_copy(hisi_acc_vdev, hisi_acc_vdev- > >saving_migf); > > hisi_acc_vf_disable_fd(hisi_acc_vdev->saving_migf); > > fput(hisi_acc_vdev->saving_migf->filp); > > hisi_acc_vdev->saving_migf = NULL; > > @@ -1144,6 +1162,7 @@ static int hisi_acc_vf_qm_init(struct > hisi_acc_vf_core_device *hisi_acc_vdev) > > if (!vf_qm->io_base) > > return -EIO; > > > > + mutex_init(&hisi_acc_vdev->enable_mutex); > > vf_qm->fun_type = QM_HW_VF; > > vf_qm->pdev = vf_dev; > > mutex_init(&vf_qm->mailbox_lock); > > @@ -1294,6 +1313,203 @@ static long hisi_acc_vfio_pci_ioctl(struct > vfio_device *core_vdev, unsigned int > > return vfio_pci_core_ioctl(core_vdev, cmd, arg); > > } > > > > +static int hisi_acc_vf_debug_check(struct seq_file *seq, struct vfio_device > *vdev) > > +{ > > + struct hisi_acc_vf_core_device *hisi_acc_vdev = > hisi_acc_get_vf_dev(vdev); > > + struct hisi_qm *vf_qm = &hisi_acc_vdev->vf_qm; > > + struct device *dev = vdev->dev; > > + int ret; > > + > > + if (!vdev->mig_ops) { > > + dev_err(dev, "device does not support live migration!\n"); > > Sorry, every error path should not spam dmesg with dev_err(). I'm > going to wait until your co-maintainer approves this before looking at > any further iterations of this series. Thanks, Sure. I will sync up with Longfang and also make sure we address all the existing comments on this before posting the next revision. Thanks, Shameer

1 0

2024

2023

2022

2021

2020

Linuxarm May 2024