[pve-devel] [PATCH/RFC cluster/common/... many] add cluster-wide hardware device mapping
Dominik Csapak
d.csapak at proxmox.com
Mon Jun 21 15:55:13 CEST 2021
this series aims to add a cluster-wide device mapping for (atm) pci and
usb devices. so that an admin can configure a device to be availble
for migration and configuring for uses that are non-root
built-in are some additional safety checks in contrast to current
passthrough, e.g. if pci addresses shift, with the mapping
we can detect that and prevent a vm to boot with the wrong device
(in most cases, there are some edge cases when one has multiple
of the same device, e.g. the same gpu, that we cannot detect)
all in all the series should be functional, but there are some parts
that are not finished/rough:
* the config format
for now i used a section config with 'hostname:id' as section
header, but this is not optimal. after some discussion with several
colleagues, i'll probably settle for a config file with pure json
in it, since there is no good way to have the necessary nesting
level (a mapping per node per id + properties per id) and have
it reasonably integrate in the section config, while maintaining
ease of use with a normal editor (ideally we would have anoter
config format like json with comments/trailing commas, or
rust-object-notation)
if there are some better ideas, please tell!
* node api
here the section config helps us, but we have the usual weirdness
with all options everytime available. also currently the api is
limited to the node level, and i have no good solution for
distributing the current state of the devices yet
(we can use the pvestatd as usual, but it does so many things
already; the rewrite really becomes a blocker for many things)
* cluster-wide gui
depends on the cluster wide api call for listing, which i omitted
for now for the reasons i gave above
* vm editing gui
not completely happy with the result, but could not come up
with something better in the time i had until now
* config file location
for now it lives in pve-common, though after looking at the
dependency chain, it would be possible to move it to qemu-server
does not really matter to me, and if we sometime want to extend
it to containers it needs to be at least in pve-guest-common
* auto-generated roles:
i did not find a way to have a 'PVEHardwareUser' role without an
'PVEHardwareAdmin', but the 'Admin' would have the same privs.
so i settled on only having an Admin. Not sure if that makes sense..
dependencies:
manager depends on qemu-server,pve-access-control,pve-common
qemu-server depends on pve-access-control,pve-common
pve-common depends on pve-cluster
additioanl info:
pve-common 1/3 and 2/3 are necssary but could be independetly applied,
no breaking change
pve-cluster:
Dominik Csapak (1):
add nodes/hardware-map.conf
data/PVE/Cluster.pm | 1 +
data/src/status.c | 1 +
2 files changed, 2 insertions(+)
pve-common:
Dominik Csapak (3):
SysFSTools: add verbose flag to pci_device_info
SysFSTools: change 'product' to 'device'
add PVE/HardwareMap and Plugins
src/Makefile | 4 ++
src/PVE/HardwareMap.pm | 54 ++++++++++++++++++++
src/PVE/HardwareMap/PCIPlugin.pm | 87 ++++++++++++++++++++++++++++++++
src/PVE/HardwareMap/Plugin.pm | 82 ++++++++++++++++++++++++++++++
src/PVE/HardwareMap/USBPlugin.pm | 69 +++++++++++++++++++++++++
src/PVE/SysFSTools.pm | 33 +++++++++---
6 files changed, 323 insertions(+), 6 deletions(-)
create mode 100644 src/PVE/HardwareMap.pm
create mode 100644 src/PVE/HardwareMap/PCIPlugin.pm
create mode 100644 src/PVE/HardwareMap/Plugin.pm
create mode 100644 src/PVE/HardwareMap/USBPlugin.pm
pve-access-control:
Dominik Csapak (2):
PVE/AccessControl: add Hardware.* privileges and /hardware/ paths
PVE/RPCEnvironment: add helper for checking hw permissions
src/PVE/API2/AccessControl.pm | 3 ++-
src/PVE/AccessControl.pm | 13 +++++++++++++
src/PVE/RPCEnvironment.pm | 8 ++++++++
3 files changed, 23 insertions(+), 1 deletion(-)
qemu-server:
Dominik Csapak (7):
PVE/QemuServer: allow mapped usb devices in config
PVE/QemuServer: allow mapped pci deviced in config
PVE/API2/Qemu: add permission checks for mapped usb devices
PVE/API2/Qemu: add permission checks for mapped pci devices
PVE/QemuServer: extend 'check_local_resources' for mapped resources
PVE/API2/Qemu: migrate preconditions: use new check_local_resources
info
PVE/QemuMigrate: check for mapped resources on migration
PVE/API2/Qemu.pm | 108 ++++++++++++++++++++++++++++++++++++++----
PVE/QemuMigrate.pm | 13 ++++-
PVE/QemuServer.pm | 38 ++++++++++++++-
PVE/QemuServer/PCI.pm | 22 ++++++++-
PVE/QemuServer/USB.pm | 22 ++++++++-
5 files changed, 188 insertions(+), 15 deletions(-)
pve-manager:
Dominik Csapak (8):
PVE/API2/Hardware: add Mapping.pm
ui: form/USBSelector: make it more flexible with nodename
ui: form: add PCIMapSelector
ui: form: add USBMapSelector
ui: node: add HardwareView and relevant edit windows
ui: qemu/PCIEdit: rework panel to add a mapped configuration
ui: qemu/USBEdit: add 'mapped' device case
ui: window/Migrate: allow mapped devices
PVE/API2/Hardware.pm | 6 +
PVE/API2/Hardware/Makefile | 1 +
PVE/API2/Hardware/Mapping.pm | 292 +++++++++++++
www/manager6/Makefile | 3 +
www/manager6/form/PCIMapSelector.js | 95 +++++
www/manager6/form/USBMapSelector.js | 73 ++++
www/manager6/form/USBSelector.js | 32 +-
www/manager6/node/Config.js | 8 +
www/manager6/node/HardwareView.js | 641 ++++++++++++++++++++++++++++
www/manager6/qemu/PCIEdit.js | 231 ++++++----
www/manager6/qemu/USBEdit.js | 34 +-
www/manager6/window/Migrate.js | 37 +-
12 files changed, 1361 insertions(+), 92 deletions(-)
create mode 100644 PVE/API2/Hardware/Mapping.pm
create mode 100644 www/manager6/form/PCIMapSelector.js
create mode 100644 www/manager6/form/USBMapSelector.js
create mode 100644 www/manager6/node/HardwareView.js
--
2.20.1
More information about the pve-devel
mailing list