| lists.openwall.net | lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC | |
|
Open Source and information security mailing list archives
| ||
|
Message-ID: <74909b12-80d5-653e-cd1c-3ea6bc5dbbde@oracle.com>
Date: Wed, 23 Nov 2022 14:52:52 -0800
From: Si-Wei Liu <si-wei.liu@...cle.com>
To: Jason Wang <jasowang@...hat.com>
Cc: dsahern@...nel.org, netdev@...r.kernel.org,
virtualization@...ts.linux-foundation.org, mst@...hat.com,
eperezma@...hat.com, lingshan.zhu@...el.com, elic@...dia.com
Subject: Re: [PATCH V2] vdpa: allow provisioning device features
On 11/22/2022 7:35 PM, Jason Wang wrote:
> On Wed, Nov 23, 2022 at 6:29 AM Si-Wei Liu <si-wei.liu@...cle.com> wrote:
>>
>>
>> On 11/16/2022 7:33 PM, Jason Wang wrote:
>>> This patch allows device features to be provisioned via vdpa. This
>>> will be useful for preserving migration compatibility between source
>>> and destination:
>>>
>>> # vdpa dev add name dev1 mgmtdev pci/0000:02:00.0 device_features 0x300020000
>> Miss the actual "vdpa dev config show" command below
> Right, let me fix that.
>
>>> # dev1: mac 52:54:00:12:34:56 link up link_announce false mtu 65535
>>> negotiated_features CTRL_VQ VERSION_1 ACCESS_PLATFORM
>>>
>>> Signed-off-by: Jason Wang <jasowang@...hat.com>
>>> ---
>>> Changes since v1:
>>> - Use uint64_t instead of __u64 for device_features
>>> - Fix typos and tweak the manpage
>>> - Add device_features to the help text
>>> ---
>>> man/man8/vdpa-dev.8 | 15 +++++++++++++++
>>> vdpa/include/uapi/linux/vdpa.h | 1 +
>>> vdpa/vdpa.c | 32 +++++++++++++++++++++++++++++---
>>> 3 files changed, 45 insertions(+), 3 deletions(-)
>>>
>>> diff --git a/man/man8/vdpa-dev.8 b/man/man8/vdpa-dev.8
>>> index 9faf3838..43e5bf48 100644
>>> --- a/man/man8/vdpa-dev.8
>>> +++ b/man/man8/vdpa-dev.8
>>> @@ -31,6 +31,7 @@ vdpa-dev \- vdpa device configuration
>>> .I NAME
>>> .B mgmtdev
>>> .I MGMTDEV
>>> +.RI "[ device_features " DEVICE_FEATURES " ]"
>>> .RI "[ mac " MACADDR " ]"
>>> .RI "[ mtu " MTU " ]"
>>> .RI "[ max_vqp " MAX_VQ_PAIRS " ]"
>>> @@ -74,6 +75,15 @@ Name of the new vdpa device to add.
>>> Name of the management device to use for device addition.
>>>
>>> .PP
>>> +.BI device_features " DEVICE_FEATURES"
>>> +Specifies the virtio device features bit-mask that is provisioned for the new vdpa device.
>>> +
>>> +The bits can be found under include/uapi/linux/virtio*h.
>>> +
>>> +see macros such as VIRTIO_F_ and VIRTIO_XXX(e.g NET)_F_ for specific bit values.
>>> +
>>> +This is optional.
>> Document the behavior when this attribute is missing? For e.g. inherit
>> device features from parent device.
> This is the current behaviour but unless we've found a way to mandate
> it, I'd like to not mention it. Maybe add a description to say the
> user needs to check the features after the add if features are not
> specified.
Well, I think at least for live migration the mgmt software should get
to some consistent result between all vdpa parent drivers regarding
feature inheritance. This inheritance predates the exposure of device
features, until which user can check into specific features after
creation. Imagine the case mgmt software of live migration needs to work
with older vdpa tool stack with no device_features exposure, how does it
know what device features are provisioned - it can only tell it from
dev_features shown at the parent mgmtdev level.
IMHO it's not about whether vdpa core can or should mandate it in a
common place or not, it's that (the man page of) the CLI tool should set
user's expectation upfront for consumers (for e.g. mgmt software). I.e.
in case the parent driver doesn't follow the man page doc, it should be
considered as an implementation bug in the individual driver rather than
flexibility of its own.
>> And what is the expected behavior when feature bit mask is off but the
>> corresponding config attr (for e.g. mac, mtu, and max_vqp) is set?
> It depends totally on the parent. And this "issue" is not introduced
> by this feature. Parents can decide to provision MQ by itself even if
> max_vqp is not specified.
Sorry, maybe I wasn't clear enough. The case I referred to was that the
parent is capable of certain feature (for e.g. _F_MQ), the associated
config attr (for e.g. max_vqp) is already present in the CLI, but the
device_features bit mask doesn't have the corresponding bit set (e.g.
the _F_MQ bit). Are you saying that the failure of this apparently
invalid/ambiguous/conflicting command can't be predicated and the
resulting behavior is totally ruled by the parent driver?
Thanks,
-Siwei
>> I think the previous behavior without device_features is that any config
>> attr implies the presence of the specific corresponding feature (_F_MAC,
>> _F_MTU, and _F_MQ). Should device_features override the other config
>> attribute, or such combination is considered invalid thus should fail?
> It follows the current policy, e.g if the parent doesn't support
> _F_MQ, we can neither provision _F_MQ nor max_vqp.
>
> Thanks
>
>> Thanks,
>> -Siwei
>>
>>> +
>>> .BI mac " MACADDR"
>>> - specifies the mac address for the new vdpa device.
>>> This is applicable only for the network type of vdpa device. This is optional.
>>> @@ -127,6 +137,11 @@ vdpa dev add name foo mgmtdev vdpa_sim_net
>>> Add the vdpa device named foo on the management device vdpa_sim_net.
>>> .RE
>>> .PP
>>> +vdpa dev add name foo mgmtdev vdpa_sim_net device_features 0x300020000
>>> +.RS 4
>>> +Add the vdpa device named foo on the management device vdpa_sim_net with device_features of 0x300020000
>>> +.RE
>>> +.PP
>>> vdpa dev add name foo mgmtdev vdpa_sim_net mac 00:11:22:33:44:55
>>> .RS 4
>>> Add the vdpa device named foo on the management device vdpa_sim_net with mac address of 00:11:22:33:44:55.
>>> diff --git a/vdpa/include/uapi/linux/vdpa.h b/vdpa/include/uapi/linux/vdpa.h
>>> index 94e4dad1..7c961991 100644
>>> --- a/vdpa/include/uapi/linux/vdpa.h
>>> +++ b/vdpa/include/uapi/linux/vdpa.h
>>> @@ -51,6 +51,7 @@ enum vdpa_attr {
>>> VDPA_ATTR_DEV_QUEUE_INDEX, /* u32 */
>>> VDPA_ATTR_DEV_VENDOR_ATTR_NAME, /* string */
>>> VDPA_ATTR_DEV_VENDOR_ATTR_VALUE, /* u64 */
>>> + VDPA_ATTR_DEV_FEATURES, /* u64 */
>>>
>>> /* new attributes must be added above here */
>>> VDPA_ATTR_MAX,
>>> diff --git a/vdpa/vdpa.c b/vdpa/vdpa.c
>>> index b73e40b4..d0ce5e22 100644
>>> --- a/vdpa/vdpa.c
>>> +++ b/vdpa/vdpa.c
>>> @@ -27,6 +27,7 @@
>>> #define VDPA_OPT_VDEV_MTU BIT(5)
>>> #define VDPA_OPT_MAX_VQP BIT(6)
>>> #define VDPA_OPT_QUEUE_INDEX BIT(7)
>>> +#define VDPA_OPT_VDEV_FEATURES BIT(8)
>>>
>>> struct vdpa_opts {
>>> uint64_t present; /* flags of present items */
>>> @@ -38,6 +39,7 @@ struct vdpa_opts {
>>> uint16_t mtu;
>>> uint16_t max_vqp;
>>> uint32_t queue_idx;
>>> + uint64_t device_features;
>>> };
>>>
>>> struct vdpa {
>>> @@ -187,6 +189,17 @@ static int vdpa_argv_u32(struct vdpa *vdpa, int argc, char **argv,
>>> return get_u32(result, *argv, 10);
>>> }
>>>
>>> +static int vdpa_argv_u64_hex(struct vdpa *vdpa, int argc, char **argv,
>>> + uint64_t *result)
>>> +{
>>> + if (argc <= 0 || !*argv) {
>>> + fprintf(stderr, "number expected\n");
>>> + return -EINVAL;
>>> + }
>>> +
>>> + return get_u64(result, *argv, 16);
>>> +}
>>> +
>>> struct vdpa_args_metadata {
>>> uint64_t o_flag;
>>> const char *err_msg;
>>> @@ -244,6 +257,10 @@ static void vdpa_opts_put(struct nlmsghdr *nlh, struct vdpa *vdpa)
>>> mnl_attr_put_u16(nlh, VDPA_ATTR_DEV_NET_CFG_MAX_VQP, opts->max_vqp);
>>> if (opts->present & VDPA_OPT_QUEUE_INDEX)
>>> mnl_attr_put_u32(nlh, VDPA_ATTR_DEV_QUEUE_INDEX, opts->queue_idx);
>>> + if (opts->present & VDPA_OPT_VDEV_FEATURES) {
>>> + mnl_attr_put_u64(nlh, VDPA_ATTR_DEV_FEATURES,
>>> + opts->device_features);
>>> + }
>>> }
>>>
>>> static int vdpa_argv_parse(struct vdpa *vdpa, int argc, char **argv,
>>> @@ -329,6 +346,14 @@ static int vdpa_argv_parse(struct vdpa *vdpa, int argc, char **argv,
>>>
>>> NEXT_ARG_FWD();
>>> o_found |= VDPA_OPT_QUEUE_INDEX;
>>> + } else if (!strcmp(*argv, "device_features") &&
>>> + (o_optional & VDPA_OPT_VDEV_FEATURES)) {
>>> + NEXT_ARG_FWD();
>>> + err = vdpa_argv_u64_hex(vdpa, argc, argv,
>>> + &opts->device_features);
>>> + if (err)
>>> + return err;
>>> + o_found |= VDPA_OPT_VDEV_FEATURES;
>>> } else {
>>> fprintf(stderr, "Unknown option \"%s\"\n", *argv);
>>> return -EINVAL;
>>> @@ -615,8 +640,9 @@ static int cmd_mgmtdev(struct vdpa *vdpa, int argc, char **argv)
>>> static void cmd_dev_help(void)
>>> {
>>> fprintf(stderr, "Usage: vdpa dev show [ DEV ]\n");
>>> - fprintf(stderr, " vdpa dev add name NAME mgmtdev MANAGEMENTDEV [ mac MACADDR ] [ mtu MTU ]\n");
>>> - fprintf(stderr, " [ max_vqp MAX_VQ_PAIRS ]\n");
>>> + fprintf(stderr, " vdpa dev add name NAME mgmtdevMANAGEMENTDEV [ device_features DEVICE_FEATURES]\n");
>>> + fprintf(stderr, " [ mac MACADDR ] [ mtu MTU ]\n");
>>> + fprintf(stderr, " [ max_vqp MAX_VQ_PAIRS ]\n");
>>> fprintf(stderr, " vdpa dev del DEV\n");
>>> fprintf(stderr, "Usage: vdpa dev config COMMAND [ OPTIONS ]\n");
>>> fprintf(stderr, "Usage: vdpa dev vstats COMMAND\n");
>>> @@ -708,7 +734,7 @@ static int cmd_dev_add(struct vdpa *vdpa, int argc, char **argv)
>>> err = vdpa_argv_parse_put(nlh, vdpa, argc, argv,
>>> VDPA_OPT_VDEV_MGMTDEV_HANDLE | VDPA_OPT_VDEV_NAME,
>>> VDPA_OPT_VDEV_MAC | VDPA_OPT_VDEV_MTU |
>>> - VDPA_OPT_MAX_VQP);
>>> + VDPA_OPT_MAX_VQP | VDPA_OPT_VDEV_FEATURES);
>>> if (err)
>>> return err;
>>>
Powered by blists - more mailing lists