Add XEN pvSCSI support
Subsystems (3)
[email protected]
maintainerv3 cover Add XEN pvSCSI support
This series adds XEN pvSCSI support. With pvSCSI it is possible to use physical SCSI devices from a XEN domain.
The support consists of a backend in the privileged Domain-0 doing the real I/O and a frontend in the unprivileged domU passing I/O-requests to the backend.
The code is taken (and adapted) from the original pvSCSI implementation done for Linux 2.6 in 2008 by Fujitsu.
[PATCH V2 1/4] Add XEN pvSCSI protocol description [PATCH V2 2/4] Introduce xen-scsifront module [PATCH V2 3/4] Introduce XEN scsiback module [PATCH V2 4/4] add xen pvscsi maintainer
Changes in V3:
- added some comments to the protocol header file
- removed the CDB emulation from xen-scsiback, handled by core target infrastructure
- several changes in xen-scsifront after comments from Christoph Hellwig
Changes in V2:
- use core target infrastructure by backend instead of pure SCSI passthrough
- add support for larger SG lists by putting them in grant page(s)
- add command abort capability
[email protected]
maintainerPATCH v3 1/4 Add XEN pvSCSI protocol description
From: Juergen Gross <[email protected]>
Add the definition of pvSCSI protocol used between the pvSCSI frontend in a XEN domU and the pvSCSI backend in a XEN driver domain (usually Dom0).
This header was originally provided by Fujitsu for XEN based on Linux 2.6.18. Changes are:
- added comment
- adapt to Linux style guide
- add support for larger SG-lists by putting them in an own granted page
- remove stale definitions
--- include/xen/interface/io/vscsiif.h | 208 +++++++++++++++++++++++++++++++++++++ 1 file changed, 208 insertions(+) create mode 100644 include/xen/interface/io/vscsiif.h
include/xen/interface/io/vscsiif.h new file +208 -0
@@ -0,0 +1,208 @@ +/****************************************************************************** +* vscsiif.h +* +* Based on the blkif.h code. +* +* This interface is to be regarded as a stable API between XEN domains +* running potentially different Linux kernel versions. +* +* Permission is hereby granted, free of charge, to any person obtaining a copy +* of this software and associated documentation files (the "Software"), to +* deal in the Software without restriction, including without limitation the +* rights to use, copy, modify, merge, publish, distribute, sublicense, and/or +* sell copies of the Software, and to permit persons to whom the Software is +* furnished to do so, subject to the following conditions: +* +* The above copyright notice and this permission notice shall be included in +* all copies or substantial portions of the Software. +* +* THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR +* IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, +* FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE +* AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER +* LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING +* FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER +* DEALINGS IN THE SOFTWARE. +* +* Copyright(c) FUJITSU Limited 2008. +*/ + +#ifndef __XEN__PUBLIC_IO_SCSI_H__ +#define __XEN__PUBLIC_IO_SCSI_H__ + +#include "ring.h" +#include "../grant_table.h" + +/* +* Front->back notifications: When enqueuing a new request, sending a +* notification can be made conditional on req_event (i.e., the generic +* hold-off mechanism provided by the ring macros). Backends must set +* req_event appropriately (e.g., using RING_FINAL_CHECK_FOR_REQUESTS()). +* +* Back->front notifications: When enqueuing a new response, sending a +* notification can be made conditional on rsp_event (i.e., the generic +* hold-off mechanism provided by the ring macros). Frontends must set +* rsp_event appropriately (e.g., using RING_FINAL_CHECK_FOR_RESPONSES()). +*/ + +/* +* Feature and Parameter Negotiation +* ================================= +* The two halves of a Xen pvSCSI driver utilize nodes within the XenStore to +* communicate capabilities and to negotiate operating parameters. This +* section enumerates these nodes which reside in the respective front and +* backend portions of the XenStore, following the XenBus convention. +* +* All data in the XenStore is stored as strings. Nodes specifying numeric +* values are encoded in decimal. Integer value ranges listed below are +* expressed as fixed sized integer types capable of storing the conversion +* of a properly formated node string, without loss of information. +* +* Any specified default value is in effect if the corresponding XenBus node +* is not present in the XenStore. +* +* XenStore nodes in sections marked "PRIVATE" are solely for use by the +* driver side whose XenBus tree contains them. +* +***************************************************************************** +* Backend XenBus Nodes +***************************************************************************** +* +*------------------ Backend Device Identification (PRIVATE) ------------------ +* +* p-devname +* Values: string +* +* A free string used to identify the physical device (e.g. a disk name). +* +* p-dev +* Values: string +* +* A string specifying the backend device: either a 4-tuple "h:c:t:l" +* (host, controller, target, lun, all integers), or a WWN (e.g. +* "naa.60014054ac780582"). +* +* v-dev +* Values: string +* +* A string specifying the frontend device in form of a 4-tuple "h:c:t:l" +* (host, controller, target, lun, all integers). +* +*--------------------------------- Features --------------------------------- +* +* feature-sg-grant +* Values: <uint16_t> +* Default Value: 0 +* +* Specifies the maximum number of scatter/gather elements in grant pages +* supported. If not set, the backend supports up to VSCSIIF_SG_TABLESIZE +* SG elements specified directly in the request. +* +***************************************************************************** +* Frontend XenBus Nodes +***************************************************************************** +* +*----------------------- Request Transport Parameters ----------------------- +* +* event-channel +* Values: <uint32_t> +* +* The identifier of the Xen event channel used to signal activity +* in the ring buffer. +* +* ring-ref +* Values: <uint32_t> +* +* protocol +* Values: string (XEN_IO_PROTO_ABI_*) +* Default Value: XEN_IO_PROTO_ABI_NATIVE +* +* The machine ABI rules governing the format of all ring request and +* response structures. +*/ + +/* Requests from the frontend to the backend */ + +/* +* Request a SCSI operation specified via a CDB in vscsiif_request.cmnd. +* The target is specified via channel, id and lun. +*/ +#define VSCSIIF_ACT_SCSI_CDB 1 + +/* +* Request abort of a running operation for the specified target given by +* channel, id, lun and the operation's rqid in ref_rqid. +*/ +#define VSCSIIF_ACT_SCSI_ABORT 2 + +/* +* Request a device reset of the specified target (channel and id). +*/ +#define VSCSIIF_ACT_SCSI_RESET 3 + +/* +* Maximum scatter/gather segments per request. +* +* Considering balance between allocating at least 16 "vscsiif_request" +* structures on one page (4096 bytes) and the number of scatter/gather +* elements needed, we decided to use 26 as a magic number. +* +* If "feature-sg-grant" is set, more scatter/gather elements can be specified +* by placing them in one or more (up to VSCSIIF_SG_TABLESIZE) granted pages. +* In this case the vscsiif_request seg elements don't contain references to +* the user data, but to the SG elements referencing the user data. +*/ +#define VSCSIIF_SG_TABLESIZE 26 + +/* +* based on Linux kernel 2.6.18, still valid +* Changing these values requires support of multiple protocols via the rings +* as "old clients" will blindly use these values and the resulting structure +* sizes. +*/ +#define VSCSIIF_MAX_COMMAND_SIZE 16 +#define VSCSIIF_SENSE_BUFFERSIZE 96 + +struct scsiif_request_segment { +grant_ref_t gref; +uint16_t offset; +uint16_t length; +}; + +/* Size of one request is 252 bytes */ +struct vscsiif_request { +uint16_t rqid;/* private guest value, echoed in resp */ +uint8_t act;/* command between backend and frontend */ +uint8_t cmd_len;/* valid CDB bytes */ + +uint8_t cmnd[VSCSIIF_MAX_COMMAND_SIZE ];/* the CDB */ +uint16_t timeout_per_command; +uint16_t channel, id, lun;/* (virtual) device specification */ +uint16_t ref_rqid;/* command abort reference */ +uint8_t sc_data_direction;/* for DMA_TO_DEVICE(1) +DMA_FROM_DEVICE(2) +DMA_NONE(3) requests */ +uint8_t nr_segments;/* Number of pieces of scatter-gather */ +#define VSCSIIF_SG_GRANT 0x80/* flag: SG elements via grant page */ +/* nr_segments counts grant pages with +SG elements +usable if "feature-sg-grant" set */ + +struct scsiif_request_segment seg[VSCSIIF_SG_TABLESIZE ]; +uint32_t reserved[3 ]; +}; + +struct vscsiif_response { +uint16_t rqid;/* identifies request */ +uint8_t padding; +uint8_t sense_len; +uint8_t sense_buffer[VSCSIIF_SENSE_BUFFERSIZE ]; +int32_t rslt; +uint32_t residual_len;/* request bufflen - +return the value from physical device */ +uint32_t reserved[36 ]; +}; + +DEFINE_RING_TYPES(vscsiif, struct vscsiif_request, struct vscsiif_response); + +#endif /*__XEN__PUBLIC_IO_SCSI_H__*/
Signature · 1 line
1.8.4.5
[email protected]
maintainerPATCH v3 1/4 Add XEN pvSCSI protocol description
In reply to [email protected]From: Juergen Gross <[email protected]>
Add the definition of pvSCSI protocol used between the pvSCSI frontend in a XEN domU and the pvSCSI backend in a XEN driver domain (usually Dom0).
This header was originally provided by Fujitsu for XEN based on Linux 2.6.18. Changes are:
- added comment
- adapt to Linux style guide
- add support for larger SG-lists by putting them in an own granted page
- remove stale definitions
--- include/xen/interface/io/vscsiif.h | 208 +++++++++++++++++++++++++++++++++++++ 1 file changed, 208 insertions(+) create mode 100644 include/xen/interface/io/vscsiif.h
include/xen/interface/io/vscsiif.h new file +208 -0
@@ -0,0 +1,208 @@ +/****************************************************************************** +* vscsiif.h +* +* Based on the blkif.h code. +* +* This interface is to be regarded as a stable API between XEN domains +* running potentially different Linux kernel versions. +* +* Permission is hereby granted, free of charge, to any person obtaining a copy +* of this software and associated documentation files (the "Software"), to +* deal in the Software without restriction, including without limitation the +* rights to use, copy, modify, merge, publish, distribute, sublicense, and/or +* sell copies of the Software, and to permit persons to whom the Software is +* furnished to do so, subject to the following conditions: +* +* The above copyright notice and this permission notice shall be included in +* all copies or substantial portions of the Software. +* +* THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR +* IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, +* FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE +* AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER +* LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING +* FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER +* DEALINGS IN THE SOFTWARE. +* +* Copyright(c) FUJITSU Limited 2008. +*/ + +#ifndef __XEN__PUBLIC_IO_SCSI_H__ +#define __XEN__PUBLIC_IO_SCSI_H__ + +#include "ring.h" +#include "../grant_table.h" + +/* +* Front->back notifications: When enqueuing a new request, sending a +* notification can be made conditional on req_event (i.e., the generic +* hold-off mechanism provided by the ring macros). Backends must set +* req_event appropriately (e.g., using RING_FINAL_CHECK_FOR_REQUESTS()). +* +* Back->front notifications: When enqueuing a new response, sending a +* notification can be made conditional on rsp_event (i.e., the generic +* hold-off mechanism provided by the ring macros). Frontends must set +* rsp_event appropriately (e.g., using RING_FINAL_CHECK_FOR_RESPONSES()). +*/ + +/* +* Feature and Parameter Negotiation +* ================================= +* The two halves of a Xen pvSCSI driver utilize nodes within the XenStore to +* communicate capabilities and to negotiate operating parameters. This +* section enumerates these nodes which reside in the respective front and +* backend portions of the XenStore, following the XenBus convention. +* +* All data in the XenStore is stored as strings. Nodes specifying numeric +* values are encoded in decimal. Integer value ranges listed below are +* expressed as fixed sized integer types capable of storing the conversion +* of a properly formated node string, without loss of information. +* +* Any specified default value is in effect if the corresponding XenBus node +* is not present in the XenStore. +* +* XenStore nodes in sections marked "PRIVATE" are solely for use by the +* driver side whose XenBus tree contains them. +* +***************************************************************************** +* Backend XenBus Nodes +***************************************************************************** +* +*------------------ Backend Device Identification (PRIVATE) ------------------ +* +* p-devname +* Values: string +* +* A free string used to identify the physical device (e.g. a disk name). +* +* p-dev +* Values: string +* +* A string specifying the backend device: either a 4-tuple "h:c:t:l" +* (host, controller, target, lun, all integers), or a WWN (e.g. +* "naa.60014054ac780582"). +* +* v-dev +* Values: string +* +* A string specifying the frontend device in form of a 4-tuple "h:c:t:l" +* (host, controller, target, lun, all integers). +* +*--------------------------------- Features --------------------------------- +* +* feature-sg-grant +* Values: <uint16_t> +* Default Value: 0 +* +* Specifies the maximum number of scatter/gather elements in grant pages +* supported. If not set, the backend supports up to VSCSIIF_SG_TABLESIZE +* SG elements specified directly in the request. +* +***************************************************************************** +* Frontend XenBus Nodes +***************************************************************************** +* +*----------------------- Request Transport Parameters ----------------------- +* +* event-channel +* Values: <uint32_t> +* +* The identifier of the Xen event channel used to signal activity +* in the ring buffer. +* +* ring-ref +* Values: <uint32_t> +* +* protocol +* Values: string (XEN_IO_PROTO_ABI_*) +* Default Value: XEN_IO_PROTO_ABI_NATIVE +* +* The machine ABI rules governing the format of all ring request and +* response structures. +*/ + +/* Requests from the frontend to the backend */ + +/* +* Request a SCSI operation specified via a CDB in vscsiif_request.cmnd. +* The target is specified via channel, id and lun. +*/ +#define VSCSIIF_ACT_SCSI_CDB 1 + +/* +* Request abort of a running operation for the specified target given by +* channel, id, lun and the operation's rqid in ref_rqid. +*/ +#define VSCSIIF_ACT_SCSI_ABORT 2 + +/* +* Request a device reset of the specified target (channel and id). +*/ +#define VSCSIIF_ACT_SCSI_RESET 3 + +/* +* Maximum scatter/gather segments per request. +* +* Considering balance between allocating at least 16 "vscsiif_request" +* structures on one page (4096 bytes) and the number of scatter/gather +* elements needed, we decided to use 26 as a magic number. +* +* If "feature-sg-grant" is set, more scatter/gather elements can be specified +* by placing them in one or more (up to VSCSIIF_SG_TABLESIZE) granted pages. +* In this case the vscsiif_request seg elements don't contain references to +* the user data, but to the SG elements referencing the user data. +*/ +#define VSCSIIF_SG_TABLESIZE 26 + +/* +* based on Linux kernel 2.6.18, still valid +* Changing these values requires support of multiple protocols via the rings +* as "old clients" will blindly use these values and the resulting structure +* sizes. +*/ +#define VSCSIIF_MAX_COMMAND_SIZE 16 +#define VSCSIIF_SENSE_BUFFERSIZE 96 + +struct scsiif_request_segment { +grant_ref_t gref; +uint16_t offset; +uint16_t length; +}; + +/* Size of one request is 252 bytes */ +struct vscsiif_request { +uint16_t rqid;/* private guest value, echoed in resp */ +uint8_t act;/* command between backend and frontend */ +uint8_t cmd_len;/* valid CDB bytes */ + +uint8_t cmnd[VSCSIIF_MAX_COMMAND_SIZE ];/* the CDB */ +uint16_t timeout_per_command; +uint16_t channel, id, lun;/* (virtual) device specification */ +uint16_t ref_rqid;/* command abort reference */ +uint8_t sc_data_direction;/* for DMA_TO_DEVICE(1) +DMA_FROM_DEVICE(2) +DMA_NONE(3) requests */ +uint8_t nr_segments;/* Number of pieces of scatter-gather */ +#define VSCSIIF_SG_GRANT 0x80/* flag: SG elements via grant page */ +/* nr_segments counts grant pages with +SG elements +usable if "feature-sg-grant" set */ + +struct scsiif_request_segment seg[VSCSIIF_SG_TABLESIZE ]; +uint32_t reserved[3 ]; +}; + +struct vscsiif_response { +uint16_t rqid;/* identifies request */ +uint8_t padding; +uint8_t sense_len; +uint8_t sense_buffer[VSCSIIF_SENSE_BUFFERSIZE ]; +int32_t rslt; +uint32_t residual_len;/* request bufflen - +return the value from physical device */ +uint32_t reserved[36 ]; +}; + +DEFINE_RING_TYPES(vscsiif, struct vscsiif_request, struct vscsiif_response); + +#endif /*__XEN__PUBLIC_IO_SCSI_H__*/
Signature · 1 line
1.8.4.5
Re: [PATCH V3 1/4] Add XEN pvSCSI protocol description
Quoted 9 lines
Quoted 1 line
Quoted 1 line · On 04.08.14 at 10:27, <[email protected]> wrote:
include/xen/interface/io/vscsiif.h new file +8
-0+/* Requests from the frontend to the backend */ + +/* +* Request a SCSI operation specified via a CDB in vscsiif_request.cmnd. +* The target is specified via channel, id and lun. +*/ +#define VSCSIIF_ACT_SCSI_CDB 1 +
include/xen/interface/io/vscsiif.h line 141
Quoted 10 lines
include/xen/interface/io/vscsiif.h new file +10
-0+/* +* Request abort of a running operation for the specified target given by +* channel, id, lun and the operation's rqid in ref_rqid. +*/ +#define VSCSIIF_ACT_SCSI_ABORT 2 + +/* +* Request a device reset of the specified target (channel and id). +*/ +#define VSCSIIF_ACT_SCSI_RESET 3
While I realize you don't want to support it, leaving out VSCSIIF_ACT_SCSI_SG_PRESET altogether from this header may lead people to the false impression that its number is available for other use. I would generally recommend the Linux headers to not deviate from the master ones more than absolutely necessary, keeping what isn't needed/wanted by Linux at least in commented out form.
Jan
Re: [Xen-devel] [PATCH V3 1/4] Add XEN pvSCSI protocol description
In reply to [email protected]Quoted 9 lines
Quoted 1 line
Quoted 1 line · On 04.08.14 at 10:27, <[email protected]> wrote:
include/xen/interface/io/vscsiif.h new file +8
-0+/* Requests from the frontend to the backend */ + +/* +* Request a SCSI operation specified via a CDB in vscsiif_request.cmnd. +* The target is specified via channel, id and lun. +*/ +#define VSCSIIF_ACT_SCSI_CDB 1 +
include/xen/interface/io/vscsiif.h line 141
Quoted 10 lines
include/xen/interface/io/vscsiif.h new file +10
-0+/* +* Request abort of a running operation for the specified target given by +* channel, id, lun and the operation's rqid in ref_rqid. +*/ +#define VSCSIIF_ACT_SCSI_ABORT 2 + +/* +* Request a device reset of the specified target (channel and id). +*/ +#define VSCSIIF_ACT_SCSI_RESET 3
While I realize you don't want to support it, leaving out VSCSIIF_ACT_SCSI_SG_PRESET altogether from this header may lead people to the false impression that its number is available for other use. I would generally recommend the Linux headers to not deviate from the master ones more than absolutely necessary, keeping what isn't needed/wanted by Linux at least in commented out form.
Jan
Juergen Gross
maintainerRe: [Xen-devel] [PATCH V3 1/4] Add XEN pvSCSI protocol description
Quoted 20 lines · Jan Beulich
Quoted 9 lines
Quoted 1 line
Quoted 1 line · On 04.08.14 at 10:27, <[email protected]> wrote:
include/xen/interface/io/vscsiif.h new file +8
-0+/* Requests from the frontend to the backend */ + +/* +* Request a SCSI operation specified via a CDB in vscsiif_request.cmnd. +* The target is specified via channel, id and lun. +*/ +#define VSCSIIF_ACT_SCSI_CDB 1 +include/xen/interface/io/vscsiif.h line 141
Quoted 10 lines
include/xen/interface/io/vscsiif.h new file +10
-0+/* +* Request abort of a running operation for the specified target given by +* channel, id, lun and the operation's rqid in ref_rqid. +*/ +#define VSCSIIF_ACT_SCSI_ABORT 2 + +/* +* Request a device reset of the specified target (channel and id). +*/ +#define VSCSIIF_ACT_SCSI_RESET 3
Quoted 7 lines
While I realize you don't want to support it, leaving out
VSCSIIF_ACT_SCSI_SG_PRESETaltogether from this header may lead people to the false impression that its number is available for other use. I would generally recommend the Linux headers to not deviate from the master ones more than absolutely necessary, keeping what isn't needed/wanted by Linux at least in commented out form.
Okay. I'll re-add the #define
Juergen
Juergen Gross
maintainerRe: [PATCH V3 1/4] Add XEN pvSCSI protocol description
In reply to Jan BeulichQuoted 20 lines · Jan Beulich
Quoted 9 lines
Quoted 1 line
Quoted 1 line · On 04.08.14 at 10:27, <[email protected]> wrote:
include/xen/interface/io/vscsiif.h new file +8
-0+/* Requests from the frontend to the backend */ + +/* +* Request a SCSI operation specified via a CDB in vscsiif_request.cmnd. +* The target is specified via channel, id and lun. +*/ +#define VSCSIIF_ACT_SCSI_CDB 1 +include/xen/interface/io/vscsiif.h line 141
Quoted 10 lines
include/xen/interface/io/vscsiif.h new file +10
-0+/* +* Request abort of a running operation for the specified target given by +* channel, id, lun and the operation's rqid in ref_rqid. +*/ +#define VSCSIIF_ACT_SCSI_ABORT 2 + +/* +* Request a device reset of the specified target (channel and id). +*/ +#define VSCSIIF_ACT_SCSI_RESET 3
Quoted 7 lines
While I realize you don't want to support it, leaving out
VSCSIIF_ACT_SCSI_SG_PRESETaltogether from this header may lead people to the false impression that its number is available for other use. I would generally recommend the Linux headers to not deviate from the master ones more than absolutely necessary, keeping what isn't needed/wanted by Linux at least in commented out form.
Okay. I'll re-add the #define
Juergen
[email protected]
maintainerPATCH v3 2/4 Introduce xen-scsifront module
In reply to [email protected]From: Juergen Gross <[email protected]>
Introduces the XEN pvSCSI frontend. With pvSCSI it is possible for a XEN domU to issue SCSI commands to a SCSI LUN assigned to that domU. The SCSI commands are passed to the pvSCSI backend in a driver domain (usually Dom0) which is owner of the physical device. This allows e.g. to use SCSI tape drives in a XEN domU.
The code is taken from the pvSCSI implementation in XEN done by Fujitsu based on Linux kernel 2.6.18.
Changes from the original version are:
- port to upstream kernel
- put all code in just one source file
- move module to appropriate location in kernel tree
- adapt to Linux style guide
- some minor code simplifications
- replace constants with defines
- remove not used defines
- add support for larger SG lists by putting them in a granted page
--- drivers/scsi/Kconfig | 9 + drivers/scsi/Makefile | 1 + drivers/scsi/xen-scsifront.c | 1011 ++++++++++++++++++++++++++++++++++++++++++ 3 files changed, 1021 insertions(+) create mode 100644 drivers/scsi/xen-scsifront.c
drivers/scsi/Kconfig +9 -0
@@ -611,6 +611,15 @@ config VMWARE_PVSCSI To compile this driver as a module, choose M here: the module will be called vmw_pvscsi. +config XEN_SCSI_FRONTEND + tristate "XEN SCSI frontend driver" + depends on SCSI && XEN + help + The XEN SCSI frontend driver allows the kernel to access SCSI Devices + within another guest OS (usually Dom0). + Only needed if the kernel is running in a XEN guest and generic + SCSI access to a device is needed. + config HYPERV_STORAGE tristate "Microsoft Hyper-V virtual storage driver" depends on SCSI && HYPERV
drivers/scsi/Makefile +1 -0
@@ -144,6 +144,7 @@ obj-$(CONFIG_SCSI_ESAS2R) += esas2r/ obj-$(CONFIG_SCSI_PMCRAID) += pmcraid.o obj-$(CONFIG_SCSI_VIRTIO) += virtio_scsi.o obj-$(CONFIG_VMWARE_PVSCSI) += vmw_pvscsi.o +obj-$(CONFIG_XEN_SCSI_FRONTEND) += xen-scsifront.o obj-$(CONFIG_HYPERV_STORAGE) += hv_storvsc.o obj-$(CONFIG_ARM) += arm/
drivers/scsi/xen-scsifront.c new file +1011 -0
1,011 lines not on this page · raw message
Signature · 1 line
1.8.4.5
Re: [PATCH V3 2/4] Introduce xen-scsifront module
Quoted 36 lines · On 04/08/14 09:27, [email protected] wrote:
drivers/scsi/xen-scsifront.c new file +35
-0+ +static void scsifront_backend_changed (struct xenbus_device *dev, +enum xenbus_state backend_state) +{ +struct vscsifrnt_info *info =dev_get_drvdata (&dev->dev); + +DPRINTK ("%p %u %u\n" , dev, dev->state, backend_state); + +switch (backend_state) { +case XenbusStateUnknown: +case XenbusStateInitialising: +case XenbusStateInitWait: +case XenbusStateInitialised: +break ; + +case XenbusStateConnected: +scsifront_read_backend_params (dev, info); +if (xenbus_read_driver_state (dev->nodename) == + XenbusStateInitialised) { +scsifront_do_lun_hotplug (info,VSCSIFRONT_OP_ADD_LUN ); + } + +if (dev->state != XenbusStateConnected) +xenbus_switch_state (dev, XenbusStateConnected); +break ; + +case XenbusStateClosed: +if (dev->state == XenbusStateClosed) +break ; +/* Missed the backend's Closing state -- fallthrough */ +case XenbusStateClosing: +scsifront_disconnect (info); +break ; + +case XenbusStateReconfiguring:
drivers/scsi/xen-scsifront.c line 981
Quoted 10 lines
drivers/scsi/xen-scsifront.c new file +10
-0+scsifront_do_lun_hotplug (info,VSCSIFRONT_OP_DEL_LUN ); +xenbus_switch_state (dev, XenbusStateReconfiguring); +break ; + +case XenbusStateReconfigured: +scsifront_do_lun_hotplug (info,VSCSIFRONT_OP_ADD_LUN ); +xenbus_switch_state (dev, XenbusStateConnected); +break ; + } +}
I would like to see this state machine in common code but I'm not going to insist on it since this is a existing driver.
Xen related parts:
David
Re: [PATCH V3 2/4] Introduce xen-scsifront module
In reply to [email protected]Quoted 36 lines · On 04/08/14 09:27, [email protected] wrote:
drivers/scsi/xen-scsifront.c new file +35
-0+ +static void scsifront_backend_changed (struct xenbus_device *dev, +enum xenbus_state backend_state) +{ +struct vscsifrnt_info *info =dev_get_drvdata (&dev->dev); + +DPRINTK ("%p %u %u\n" , dev, dev->state, backend_state); + +switch (backend_state) { +case XenbusStateUnknown: +case XenbusStateInitialising: +case XenbusStateInitWait: +case XenbusStateInitialised: +break ; + +case XenbusStateConnected: +scsifront_read_backend_params (dev, info); +if (xenbus_read_driver_state (dev->nodename) == + XenbusStateInitialised) { +scsifront_do_lun_hotplug (info,VSCSIFRONT_OP_ADD_LUN ); + } + +if (dev->state != XenbusStateConnected) +xenbus_switch_state (dev, XenbusStateConnected); +break ; + +case XenbusStateClosed: +if (dev->state == XenbusStateClosed) +break ; +/* Missed the backend's Closing state -- fallthrough */ +case XenbusStateClosing: +scsifront_disconnect (info); +break ; + +case XenbusStateReconfiguring:
drivers/scsi/xen-scsifront.c line 981
Quoted 10 lines
drivers/scsi/xen-scsifront.c new file +10
-0+scsifront_do_lun_hotplug (info,VSCSIFRONT_OP_DEL_LUN ); +xenbus_switch_state (dev, XenbusStateReconfiguring); +break ; + +case XenbusStateReconfigured: +scsifront_do_lun_hotplug (info,VSCSIFRONT_OP_ADD_LUN ); +xenbus_switch_state (dev, XenbusStateConnected); +break ; + } +}
I would like to see this state machine in common code but I'm not going to insist on it since this is a existing driver.
Xen related parts:
David
[email protected]
maintainerPATCH v3 2/4 Introduce xen-scsifront module
In reply to [email protected]From: Juergen Gross <[email protected]>
Introduces the XEN pvSCSI frontend. With pvSCSI it is possible for a XEN domU to issue SCSI commands to a SCSI LUN assigned to that domU. The SCSI commands are passed to the pvSCSI backend in a driver domain (usually Dom0) which is owner of the physical device. This allows e.g. to use SCSI tape drives in a XEN domU.
The code is taken from the pvSCSI implementation in XEN done by Fujitsu based on Linux kernel 2.6.18.
Changes from the original version are:
- port to upstream kernel
- put all code in just one source file
- move module to appropriate location in kernel tree
- adapt to Linux style guide
- some minor code simplifications
- replace constants with defines
- remove not used defines
- add support for larger SG lists by putting them in a granted page
--- drivers/scsi/Kconfig | 9 + drivers/scsi/Makefile | 1 + drivers/scsi/xen-scsifront.c | 1011 ++++++++++++++++++++++++++++++++++++++++++ 3 files changed, 1021 insertions(+) create mode 100644 drivers/scsi/xen-scsifront.c
drivers/scsi/Kconfig +9 -0
@@ -611,6 +611,15 @@ config VMWARE_PVSCSI To compile this driver as a module, choose M here: the module will be called vmw_pvscsi. +config XEN_SCSI_FRONTEND + tristate "XEN SCSI frontend driver" + depends on SCSI && XEN + help + The XEN SCSI frontend driver allows the kernel to access SCSI Devices + within another guest OS (usually Dom0). + Only needed if the kernel is running in a XEN guest and generic + SCSI access to a device is needed. + config HYPERV_STORAGE tristate "Microsoft Hyper-V virtual storage driver" depends on SCSI && HYPERV
drivers/scsi/Makefile +1 -0
@@ -144,6 +144,7 @@ obj-$(CONFIG_SCSI_ESAS2R) += esas2r/ obj-$(CONFIG_SCSI_PMCRAID) += pmcraid.o obj-$(CONFIG_SCSI_VIRTIO) += virtio_scsi.o obj-$(CONFIG_VMWARE_PVSCSI) += vmw_pvscsi.o +obj-$(CONFIG_XEN_SCSI_FRONTEND) += xen-scsifront.o obj-$(CONFIG_HYPERV_STORAGE) += hv_storvsc.o obj-$(CONFIG_ARM) += arm/
drivers/scsi/xen-scsifront.c new file +1011 -0
1,011 lines not on this page · raw message
Signature · 1 line
1.8.4.5
[email protected]
maintainerPATCH v3 3/4 Introduce XEN scsiback module
In reply to [email protected]From: Juergen Gross <[email protected]>
Introduces the XEN pvSCSI backend. With pvSCSI it is possible for a XEN domU to issue SCSI commands to a SCSI LUN assigned to that domU. The SCSI commands are passed to the pvSCSI backend in a driver domain (usually Dom0) which is owner of the physical device. This allows e.g. to use SCSI tape drives in a XEN domU.
The code is taken from the pvSCSI implementation in XEN done by Fujitsu based on Linux kernel 2.6.18.
Changes from the original version are:
- port to upstream kernel
- put all code in just one source file
- adapt to Linux style guide
- use target core infrastructure instead doing pure pass-through
- enable module unloading
- support SG-list in grant page(s)
- support task abort
- remove redundant struct backend
- allocate resources dynamically
- correct minor error in scsiback_fast_flush_area
- free allocated resources in case of error during I/O preparation
- remove CDB emulation, now handled by target core infrastructure
--- drivers/xen/Kconfig | 9 + drivers/xen/Makefile | 1 + drivers/xen/xen-scsiback.c | 2282 ++++++++++++++++++++++++++++++++++++++++++++ 3 files changed, 2292 insertions(+) create mode 100644 drivers/xen/xen-scsiback.c
drivers/xen/Kconfig +9 -0
@@ -172,6 +172,15 @@ config XEN_PCIDEV_BACKEND If in doubt, say m. +config XEN_SCSI_BACKEND + tristate "XEN SCSI backend driver" + depends on XEN && XEN_BACKEND && TARGET_CORE + help + The SCSI backend driver allows the kernel to export its SCSI Devices + to other guests via a high-performance shared-memory interface. + Only needed for systems running as XEN driver domains (e.g. Dom0) and + if guests need generic access to SCSI devices. + config XEN_PRIVCMD tristate depends on XEN
drivers/xen/Makefile +1 -0
@@ -33,6 +33,7 @@ obj-$(CONFIG_XEN_STUB) += xen-stub.o obj-$(CONFIG_XEN_ACPI_HOTPLUG_MEMORY) += xen-acpi-memhotplug.o obj-$(CONFIG_XEN_ACPI_HOTPLUG_CPU) += xen-acpi-cpuhotplug.o obj-$(CONFIG_XEN_ACPI_PROCESSOR) += xen-acpi-processor.o +obj-$(CONFIG_XEN_SCSI_BACKEND) += xen-scsiback.o xen-evtchn-y := evtchn.o xen-gntdev-y := gntdev.o xen-gntalloc-y := gntalloc.o
drivers/xen/xen-scsiback.c new file +2282 -0
2,282 lines not on this page · raw message
Signature · 1 line
1.8.4.5
Re: [PATCH V3 3/4] Introduce XEN scsiback module
Quoted 7 lines · On 04/08/14 09:27, [email protected] wrote:
From: Juergen Gross <[email protected]>
Introduces the XEN pvSCSI backend. With pvSCSI it is possible for a XEN domU to issue SCSI commands to a SCSI LUN assigned to that domU. The SCSI commands are passed to the pvSCSI backend in a driver domain (usually Dom0) which is owner of the physical device. This allows e.g. to use SCSI tape drives in a XEN domU.
Xen related parts
David
Re: [PATCH V3 3/4] Introduce XEN scsiback module
In reply to [email protected]Quoted 7 lines · On 04/08/14 09:27, [email protected] wrote:
From: Juergen Gross <[email protected]>
Introduces the XEN pvSCSI backend. With pvSCSI it is possible for a XEN domU to issue SCSI commands to a SCSI LUN assigned to that domU. The SCSI commands are passed to the pvSCSI backend in a driver domain (usually Dom0) which is owner of the physical device. This allows e.g. to use SCSI tape drives in a XEN domU.
Xen related parts
David
[email protected]
maintainerPATCH v3 3/4 Introduce XEN scsiback module
In reply to [email protected]From: Juergen Gross <[email protected]>
Introduces the XEN pvSCSI backend. With pvSCSI it is possible for a XEN domU to issue SCSI commands to a SCSI LUN assigned to that domU. The SCSI commands are passed to the pvSCSI backend in a driver domain (usually Dom0) which is owner of the physical device. This allows e.g. to use SCSI tape drives in a XEN domU.
The code is taken from the pvSCSI implementation in XEN done by Fujitsu based on Linux kernel 2.6.18.
Changes from the original version are:
- port to upstream kernel
- put all code in just one source file
- adapt to Linux style guide
- use target core infrastructure instead doing pure pass-through
- enable module unloading
- support SG-list in grant page(s)
- support task abort
- remove redundant struct backend
- allocate resources dynamically
- correct minor error in scsiback_fast_flush_area
- free allocated resources in case of error during I/O preparation
- remove CDB emulation, now handled by target core infrastructure
--- drivers/xen/Kconfig | 9 + drivers/xen/Makefile | 1 + drivers/xen/xen-scsiback.c | 2282 ++++++++++++++++++++++++++++++++++++++++++++ 3 files changed, 2292 insertions(+) create mode 100644 drivers/xen/xen-scsiback.c
drivers/xen/Kconfig +9 -0
@@ -172,6 +172,15 @@ config XEN_PCIDEV_BACKEND If in doubt, say m. +config XEN_SCSI_BACKEND + tristate "XEN SCSI backend driver" + depends on XEN && XEN_BACKEND && TARGET_CORE + help + The SCSI backend driver allows the kernel to export its SCSI Devices + to other guests via a high-performance shared-memory interface. + Only needed for systems running as XEN driver domains (e.g. Dom0) and + if guests need generic access to SCSI devices. + config XEN_PRIVCMD tristate depends on XEN
drivers/xen/Makefile +1 -0
@@ -33,6 +33,7 @@ obj-$(CONFIG_XEN_STUB) += xen-stub.o obj-$(CONFIG_XEN_ACPI_HOTPLUG_MEMORY) += xen-acpi-memhotplug.o obj-$(CONFIG_XEN_ACPI_HOTPLUG_CPU) += xen-acpi-cpuhotplug.o obj-$(CONFIG_XEN_ACPI_PROCESSOR) += xen-acpi-processor.o +obj-$(CONFIG_XEN_SCSI_BACKEND) += xen-scsiback.o xen-evtchn-y := evtchn.o xen-gntdev-y := gntdev.o xen-gntalloc-y := gntalloc.o
drivers/xen/xen-scsiback.c new file +2282 -0
2,282 lines not on this page · raw message
Signature · 1 line
1.8.4.5
[email protected]
maintainerPATCH v3 4/4 add xen pvscsi maintainer
In reply to [email protected]From: Juergen Gross <[email protected]>
Add myself as maintainer for the Xen pvSCSI stuff.
--- MAINTAINERS | 8 ++++++++ 1 file changed, 8 insertions(+)
MAINTAINERS +8 -0
@@ -10046,6 +10046,14 @@ S: Supported F: arch/x86/pci/*xen* F: drivers/pci/*xen* +XEN PVSCSI DRIVERS +M: Juergen Gross <[email protected]> +L: [email protected] (moderated for non-subscribers) +S: Supported +F: drivers/scsi/xen-scsifront.c +F: drivers/xen/xen-scsiback.c +F: include/xen/interface/io/vscsiif.h + XEN SWIOTLB SUBSYSTEM M: Konrad Rzeszutek Wilk <[email protected]> L: [email protected] (moderated for non-subscribers)
Signature · 1 line
1.8.4.5
[email protected]
maintainerPATCH v3 4/4 add xen pvscsi maintainer
In reply to [email protected]From: Juergen Gross <[email protected]>
Add myself as maintainer for the Xen pvSCSI stuff.
--- MAINTAINERS | 8 ++++++++ 1 file changed, 8 insertions(+)
MAINTAINERS +8 -0
@@ -10046,6 +10046,14 @@ S: Supported F: arch/x86/pci/*xen* F: drivers/pci/*xen* +XEN PVSCSI DRIVERS +M: Juergen Gross <[email protected]> +L: [email protected] (moderated for non-subscribers) +S: Supported +F: drivers/scsi/xen-scsifront.c +F: drivers/xen/xen-scsiback.c +F: include/xen/interface/io/vscsiif.h + XEN SWIOTLB SUBSYSTEM M: Konrad Rzeszutek Wilk <[email protected]> L: [email protected] (moderated for non-subscribers)
Signature · 1 line
1.8.4.5