hw/cxl/host: Add support for CXL Fixed Memory Windows.
The concept of these is introduced in [1] in terms of the
description the CEDT ACPI table. The principal is more general.
Unlike once traffic hits the CXL root bridges, the host system
memory address routing is implementation defined and effectively
static once observable by standard / generic system software.
Each CXL Fixed Memory Windows (CFMW) is a region of PA space
which has fixed system dependent routing configured so that
accesses can be routed to the CXL devices below a set of target
root bridges. The accesses may be interleaved across multiple
root bridges.
For QEMU we could have fully specified these regions in terms
of a base PA + size, but as the absolute address does not matter
it is simpler to let individual platforms place the memory regions.
ExampleS:
-cxl-fixed-memory-window targets.0=cxl.0,size=128G
-cxl-fixed-memory-window targets.0=cxl.1,size=128G
-cxl-fixed-memory-window targets.0=cxl0,targets.1=cxl.1,size=256G,interleave-granularity=2k
Specifies
* 2x 128G regions not interleaved across root bridges, one for each of
the root bridges with ids cxl.0 and cxl.1
* 256G region interleaved across root bridges with ids cxl.0 and cxl.1
with a 2k interleave granularity.
When system software enumerates the devices below a given root bridge
it can then decide which CFMW to use. If non interleave is desired
(or possible) it can use the appropriate CFMW for the root bridge in
question. If there are suitable devices to interleave across the
two root bridges then it may use the 3rd CFMS.
A number of other designs were considered but the following constraints
made it hard to adapt existing QEMU approaches to this particular problem.
1) The size must be known before a specific architecture / board brings
up it's PA memory map. We need to set up an appropriate region.
2) Using links to the host bridges provides a clean command line interface
but these links cannot be established until command line devices have
been added.
Hence the two step process used here of first establishing the size,
interleave-ways and granularity + caching the ids of the host bridges
and then, once available finding the actual host bridges so they can
be used later to support interleave decoding.
[1] CXL 2.0 ECN: CEDT CFMWS & QTG DSM (computeexpresslink.org / specifications)
Signed-off-by: Jonathan Cameron <jonathan.cameron@huawei.com>
Acked-by: Markus Armbruster <armbru@redhat.com> # QAPI Schema
Message-Id: <20220429144110.25167-28-Jonathan.Cameron@huawei.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
2022-04-29 17:40:52 +03:00
|
|
|
/*
|
|
|
|
* CXL host parameter parsing routines
|
|
|
|
*
|
|
|
|
* Copyright (c) 2022 Huawei
|
|
|
|
* Modeled loosely on the NUMA options handling in hw/core/numa.c
|
|
|
|
*/
|
|
|
|
|
|
|
|
#include "qemu/osdep.h"
|
|
|
|
#include "qemu/units.h"
|
|
|
|
#include "qemu/bitmap.h"
|
|
|
|
#include "qemu/error-report.h"
|
|
|
|
#include "qapi/error.h"
|
|
|
|
#include "sysemu/qtest.h"
|
|
|
|
#include "hw/boards.h"
|
|
|
|
|
|
|
|
#include "qapi/qapi-visit-machine.h"
|
|
|
|
#include "hw/cxl/cxl.h"
|
2022-06-08 17:54:33 +03:00
|
|
|
#include "hw/cxl/cxl_host.h"
|
2022-04-29 17:40:58 +03:00
|
|
|
#include "hw/pci/pci_bus.h"
|
|
|
|
#include "hw/pci/pci_bridge.h"
|
|
|
|
#include "hw/pci/pci_host.h"
|
|
|
|
#include "hw/pci/pcie_port.h"
|
2022-06-08 17:54:37 +03:00
|
|
|
#include "hw/pci-bridge/pci_expander_bridge.h"
|
hw/cxl/host: Add support for CXL Fixed Memory Windows.
The concept of these is introduced in [1] in terms of the
description the CEDT ACPI table. The principal is more general.
Unlike once traffic hits the CXL root bridges, the host system
memory address routing is implementation defined and effectively
static once observable by standard / generic system software.
Each CXL Fixed Memory Windows (CFMW) is a region of PA space
which has fixed system dependent routing configured so that
accesses can be routed to the CXL devices below a set of target
root bridges. The accesses may be interleaved across multiple
root bridges.
For QEMU we could have fully specified these regions in terms
of a base PA + size, but as the absolute address does not matter
it is simpler to let individual platforms place the memory regions.
ExampleS:
-cxl-fixed-memory-window targets.0=cxl.0,size=128G
-cxl-fixed-memory-window targets.0=cxl.1,size=128G
-cxl-fixed-memory-window targets.0=cxl0,targets.1=cxl.1,size=256G,interleave-granularity=2k
Specifies
* 2x 128G regions not interleaved across root bridges, one for each of
the root bridges with ids cxl.0 and cxl.1
* 256G region interleaved across root bridges with ids cxl.0 and cxl.1
with a 2k interleave granularity.
When system software enumerates the devices below a given root bridge
it can then decide which CFMW to use. If non interleave is desired
(or possible) it can use the appropriate CFMW for the root bridge in
question. If there are suitable devices to interleave across the
two root bridges then it may use the 3rd CFMS.
A number of other designs were considered but the following constraints
made it hard to adapt existing QEMU approaches to this particular problem.
1) The size must be known before a specific architecture / board brings
up it's PA memory map. We need to set up an appropriate region.
2) Using links to the host bridges provides a clean command line interface
but these links cannot be established until command line devices have
been added.
Hence the two step process used here of first establishing the size,
interleave-ways and granularity + caching the ids of the host bridges
and then, once available finding the actual host bridges so they can
be used later to support interleave decoding.
[1] CXL 2.0 ECN: CEDT CFMWS & QTG DSM (computeexpresslink.org / specifications)
Signed-off-by: Jonathan Cameron <jonathan.cameron@huawei.com>
Acked-by: Markus Armbruster <armbru@redhat.com> # QAPI Schema
Message-Id: <20220429144110.25167-28-Jonathan.Cameron@huawei.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
2022-04-29 17:40:52 +03:00
|
|
|
|
2022-06-08 17:54:33 +03:00
|
|
|
static void cxl_fixed_memory_window_config(CXLState *cxl_state,
|
|
|
|
CXLFixedMemoryWindowOptions *object,
|
|
|
|
Error **errp)
|
hw/cxl/host: Add support for CXL Fixed Memory Windows.
The concept of these is introduced in [1] in terms of the
description the CEDT ACPI table. The principal is more general.
Unlike once traffic hits the CXL root bridges, the host system
memory address routing is implementation defined and effectively
static once observable by standard / generic system software.
Each CXL Fixed Memory Windows (CFMW) is a region of PA space
which has fixed system dependent routing configured so that
accesses can be routed to the CXL devices below a set of target
root bridges. The accesses may be interleaved across multiple
root bridges.
For QEMU we could have fully specified these regions in terms
of a base PA + size, but as the absolute address does not matter
it is simpler to let individual platforms place the memory regions.
ExampleS:
-cxl-fixed-memory-window targets.0=cxl.0,size=128G
-cxl-fixed-memory-window targets.0=cxl.1,size=128G
-cxl-fixed-memory-window targets.0=cxl0,targets.1=cxl.1,size=256G,interleave-granularity=2k
Specifies
* 2x 128G regions not interleaved across root bridges, one for each of
the root bridges with ids cxl.0 and cxl.1
* 256G region interleaved across root bridges with ids cxl.0 and cxl.1
with a 2k interleave granularity.
When system software enumerates the devices below a given root bridge
it can then decide which CFMW to use. If non interleave is desired
(or possible) it can use the appropriate CFMW for the root bridge in
question. If there are suitable devices to interleave across the
two root bridges then it may use the 3rd CFMS.
A number of other designs were considered but the following constraints
made it hard to adapt existing QEMU approaches to this particular problem.
1) The size must be known before a specific architecture / board brings
up it's PA memory map. We need to set up an appropriate region.
2) Using links to the host bridges provides a clean command line interface
but these links cannot be established until command line devices have
been added.
Hence the two step process used here of first establishing the size,
interleave-ways and granularity + caching the ids of the host bridges
and then, once available finding the actual host bridges so they can
be used later to support interleave decoding.
[1] CXL 2.0 ECN: CEDT CFMWS & QTG DSM (computeexpresslink.org / specifications)
Signed-off-by: Jonathan Cameron <jonathan.cameron@huawei.com>
Acked-by: Markus Armbruster <armbru@redhat.com> # QAPI Schema
Message-Id: <20220429144110.25167-28-Jonathan.Cameron@huawei.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
2022-04-29 17:40:52 +03:00
|
|
|
{
|
hw/cxl/cxl-host: Fix missing ERRP_GUARD() in cxl_fixed_memory_window_config()
As the comment in qapi/error, dereferencing @errp requires
ERRP_GUARD():
* = Why, when and how to use ERRP_GUARD() =
*
* Without ERRP_GUARD(), use of the @errp parameter is restricted:
* - It must not be dereferenced, because it may be null.
...
* ERRP_GUARD() lifts these restrictions.
*
* To use ERRP_GUARD(), add it right at the beginning of the function.
* @errp can then be used without worrying about the argument being
* NULL or &error_fatal.
*
* Using it when it's not needed is safe, but please avoid cluttering
* the source with useless code.
But in cxl_fixed_memory_window_config(), @errp is dereferenced in 2
places without ERRP_GUARD():
fw->enc_int_ways = cxl_interleave_ways_enc(fw->num_targets, errp);
if (*errp) {
return;
}
and
fw->enc_int_gran =
cxl_interleave_granularity_enc(object->interleave_granularity,
errp);
if (*errp) {
return;
}
For the above 2 places, we check "*errp", because neither function
returns a suitable error code. And since machine_set_cfmw() - the caller
of cxl_fixed_memory_window_config() - doesn't get the NULL @errp
parameter as the "set" method of object property,
cxl_fixed_memory_window_config() hasn't triggered the bug that
dereferencing the NULL @errp.
To follow the requirement of @errp, add missing ERRP_GUARD() in
cxl_fixed_memory_window_config().
Suggested-by: Markus Armbruster <armbru@redhat.com>
Signed-off-by: Zhao Liu <zhao1.liu@intel.com>
Reviewed-by: Markus Armbruster <armbru@redhat.com>
Message-ID: <20240223085653.1255438-2-zhao1.liu@linux.intel.com>
Signed-off-by: Thomas Huth <thuth@redhat.com>
2024-02-23 11:56:47 +03:00
|
|
|
ERRP_GUARD();
|
2022-08-08 15:20:50 +03:00
|
|
|
g_autofree CXLFixedWindow *fw = g_malloc0(sizeof(*fw));
|
hw/cxl/host: Add support for CXL Fixed Memory Windows.
The concept of these is introduced in [1] in terms of the
description the CEDT ACPI table. The principal is more general.
Unlike once traffic hits the CXL root bridges, the host system
memory address routing is implementation defined and effectively
static once observable by standard / generic system software.
Each CXL Fixed Memory Windows (CFMW) is a region of PA space
which has fixed system dependent routing configured so that
accesses can be routed to the CXL devices below a set of target
root bridges. The accesses may be interleaved across multiple
root bridges.
For QEMU we could have fully specified these regions in terms
of a base PA + size, but as the absolute address does not matter
it is simpler to let individual platforms place the memory regions.
ExampleS:
-cxl-fixed-memory-window targets.0=cxl.0,size=128G
-cxl-fixed-memory-window targets.0=cxl.1,size=128G
-cxl-fixed-memory-window targets.0=cxl0,targets.1=cxl.1,size=256G,interleave-granularity=2k
Specifies
* 2x 128G regions not interleaved across root bridges, one for each of
the root bridges with ids cxl.0 and cxl.1
* 256G region interleaved across root bridges with ids cxl.0 and cxl.1
with a 2k interleave granularity.
When system software enumerates the devices below a given root bridge
it can then decide which CFMW to use. If non interleave is desired
(or possible) it can use the appropriate CFMW for the root bridge in
question. If there are suitable devices to interleave across the
two root bridges then it may use the 3rd CFMS.
A number of other designs were considered but the following constraints
made it hard to adapt existing QEMU approaches to this particular problem.
1) The size must be known before a specific architecture / board brings
up it's PA memory map. We need to set up an appropriate region.
2) Using links to the host bridges provides a clean command line interface
but these links cannot be established until command line devices have
been added.
Hence the two step process used here of first establishing the size,
interleave-ways and granularity + caching the ids of the host bridges
and then, once available finding the actual host bridges so they can
be used later to support interleave decoding.
[1] CXL 2.0 ECN: CEDT CFMWS & QTG DSM (computeexpresslink.org / specifications)
Signed-off-by: Jonathan Cameron <jonathan.cameron@huawei.com>
Acked-by: Markus Armbruster <armbru@redhat.com> # QAPI Schema
Message-Id: <20220429144110.25167-28-Jonathan.Cameron@huawei.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
2022-04-29 17:40:52 +03:00
|
|
|
strList *target;
|
|
|
|
int i;
|
|
|
|
|
|
|
|
for (target = object->targets; target; target = target->next) {
|
|
|
|
fw->num_targets++;
|
|
|
|
}
|
|
|
|
|
|
|
|
fw->enc_int_ways = cxl_interleave_ways_enc(fw->num_targets, errp);
|
|
|
|
if (*errp) {
|
|
|
|
return;
|
|
|
|
}
|
|
|
|
|
|
|
|
if (object->size % (256 * MiB)) {
|
|
|
|
error_setg(errp,
|
2022-11-27 06:22:20 +03:00
|
|
|
"Size of a CXL fixed memory window must be a multiple of 256MiB");
|
hw/cxl/host: Add support for CXL Fixed Memory Windows.
The concept of these is introduced in [1] in terms of the
description the CEDT ACPI table. The principal is more general.
Unlike once traffic hits the CXL root bridges, the host system
memory address routing is implementation defined and effectively
static once observable by standard / generic system software.
Each CXL Fixed Memory Windows (CFMW) is a region of PA space
which has fixed system dependent routing configured so that
accesses can be routed to the CXL devices below a set of target
root bridges. The accesses may be interleaved across multiple
root bridges.
For QEMU we could have fully specified these regions in terms
of a base PA + size, but as the absolute address does not matter
it is simpler to let individual platforms place the memory regions.
ExampleS:
-cxl-fixed-memory-window targets.0=cxl.0,size=128G
-cxl-fixed-memory-window targets.0=cxl.1,size=128G
-cxl-fixed-memory-window targets.0=cxl0,targets.1=cxl.1,size=256G,interleave-granularity=2k
Specifies
* 2x 128G regions not interleaved across root bridges, one for each of
the root bridges with ids cxl.0 and cxl.1
* 256G region interleaved across root bridges with ids cxl.0 and cxl.1
with a 2k interleave granularity.
When system software enumerates the devices below a given root bridge
it can then decide which CFMW to use. If non interleave is desired
(or possible) it can use the appropriate CFMW for the root bridge in
question. If there are suitable devices to interleave across the
two root bridges then it may use the 3rd CFMS.
A number of other designs were considered but the following constraints
made it hard to adapt existing QEMU approaches to this particular problem.
1) The size must be known before a specific architecture / board brings
up it's PA memory map. We need to set up an appropriate region.
2) Using links to the host bridges provides a clean command line interface
but these links cannot be established until command line devices have
been added.
Hence the two step process used here of first establishing the size,
interleave-ways and granularity + caching the ids of the host bridges
and then, once available finding the actual host bridges so they can
be used later to support interleave decoding.
[1] CXL 2.0 ECN: CEDT CFMWS & QTG DSM (computeexpresslink.org / specifications)
Signed-off-by: Jonathan Cameron <jonathan.cameron@huawei.com>
Acked-by: Markus Armbruster <armbru@redhat.com> # QAPI Schema
Message-Id: <20220429144110.25167-28-Jonathan.Cameron@huawei.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
2022-04-29 17:40:52 +03:00
|
|
|
return;
|
|
|
|
}
|
|
|
|
fw->size = object->size;
|
|
|
|
|
|
|
|
if (object->has_interleave_granularity) {
|
|
|
|
fw->enc_int_gran =
|
|
|
|
cxl_interleave_granularity_enc(object->interleave_granularity,
|
|
|
|
errp);
|
|
|
|
if (*errp) {
|
|
|
|
return;
|
|
|
|
}
|
|
|
|
} else {
|
|
|
|
/* Default to 256 byte interleave */
|
|
|
|
fw->enc_int_gran = 0;
|
|
|
|
}
|
|
|
|
|
2023-09-04 16:28:03 +03:00
|
|
|
fw->targets = g_malloc0_n(fw->num_targets, sizeof(*fw->targets));
|
|
|
|
for (i = 0, target = object->targets; target; i++, target = target->next) {
|
|
|
|
/* This link cannot be resolved yet, so stash the name for now */
|
|
|
|
fw->targets[i] = g_strdup(target->value);
|
|
|
|
}
|
|
|
|
|
2022-08-08 15:20:50 +03:00
|
|
|
cxl_state->fixed_windows = g_list_append(cxl_state->fixed_windows,
|
|
|
|
g_steal_pointer(&fw));
|
hw/cxl/host: Add support for CXL Fixed Memory Windows.
The concept of these is introduced in [1] in terms of the
description the CEDT ACPI table. The principal is more general.
Unlike once traffic hits the CXL root bridges, the host system
memory address routing is implementation defined and effectively
static once observable by standard / generic system software.
Each CXL Fixed Memory Windows (CFMW) is a region of PA space
which has fixed system dependent routing configured so that
accesses can be routed to the CXL devices below a set of target
root bridges. The accesses may be interleaved across multiple
root bridges.
For QEMU we could have fully specified these regions in terms
of a base PA + size, but as the absolute address does not matter
it is simpler to let individual platforms place the memory regions.
ExampleS:
-cxl-fixed-memory-window targets.0=cxl.0,size=128G
-cxl-fixed-memory-window targets.0=cxl.1,size=128G
-cxl-fixed-memory-window targets.0=cxl0,targets.1=cxl.1,size=256G,interleave-granularity=2k
Specifies
* 2x 128G regions not interleaved across root bridges, one for each of
the root bridges with ids cxl.0 and cxl.1
* 256G region interleaved across root bridges with ids cxl.0 and cxl.1
with a 2k interleave granularity.
When system software enumerates the devices below a given root bridge
it can then decide which CFMW to use. If non interleave is desired
(or possible) it can use the appropriate CFMW for the root bridge in
question. If there are suitable devices to interleave across the
two root bridges then it may use the 3rd CFMS.
A number of other designs were considered but the following constraints
made it hard to adapt existing QEMU approaches to this particular problem.
1) The size must be known before a specific architecture / board brings
up it's PA memory map. We need to set up an appropriate region.
2) Using links to the host bridges provides a clean command line interface
but these links cannot be established until command line devices have
been added.
Hence the two step process used here of first establishing the size,
interleave-ways and granularity + caching the ids of the host bridges
and then, once available finding the actual host bridges so they can
be used later to support interleave decoding.
[1] CXL 2.0 ECN: CEDT CFMWS & QTG DSM (computeexpresslink.org / specifications)
Signed-off-by: Jonathan Cameron <jonathan.cameron@huawei.com>
Acked-by: Markus Armbruster <armbru@redhat.com> # QAPI Schema
Message-Id: <20220429144110.25167-28-Jonathan.Cameron@huawei.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
2022-04-29 17:40:52 +03:00
|
|
|
|
|
|
|
return;
|
|
|
|
}
|
|
|
|
|
2022-06-08 17:54:35 +03:00
|
|
|
void cxl_fmws_link_targets(CXLState *cxl_state, Error **errp)
|
hw/cxl/host: Add support for CXL Fixed Memory Windows.
The concept of these is introduced in [1] in terms of the
description the CEDT ACPI table. The principal is more general.
Unlike once traffic hits the CXL root bridges, the host system
memory address routing is implementation defined and effectively
static once observable by standard / generic system software.
Each CXL Fixed Memory Windows (CFMW) is a region of PA space
which has fixed system dependent routing configured so that
accesses can be routed to the CXL devices below a set of target
root bridges. The accesses may be interleaved across multiple
root bridges.
For QEMU we could have fully specified these regions in terms
of a base PA + size, but as the absolute address does not matter
it is simpler to let individual platforms place the memory regions.
ExampleS:
-cxl-fixed-memory-window targets.0=cxl.0,size=128G
-cxl-fixed-memory-window targets.0=cxl.1,size=128G
-cxl-fixed-memory-window targets.0=cxl0,targets.1=cxl.1,size=256G,interleave-granularity=2k
Specifies
* 2x 128G regions not interleaved across root bridges, one for each of
the root bridges with ids cxl.0 and cxl.1
* 256G region interleaved across root bridges with ids cxl.0 and cxl.1
with a 2k interleave granularity.
When system software enumerates the devices below a given root bridge
it can then decide which CFMW to use. If non interleave is desired
(or possible) it can use the appropriate CFMW for the root bridge in
question. If there are suitable devices to interleave across the
two root bridges then it may use the 3rd CFMS.
A number of other designs were considered but the following constraints
made it hard to adapt existing QEMU approaches to this particular problem.
1) The size must be known before a specific architecture / board brings
up it's PA memory map. We need to set up an appropriate region.
2) Using links to the host bridges provides a clean command line interface
but these links cannot be established until command line devices have
been added.
Hence the two step process used here of first establishing the size,
interleave-ways and granularity + caching the ids of the host bridges
and then, once available finding the actual host bridges so they can
be used later to support interleave decoding.
[1] CXL 2.0 ECN: CEDT CFMWS & QTG DSM (computeexpresslink.org / specifications)
Signed-off-by: Jonathan Cameron <jonathan.cameron@huawei.com>
Acked-by: Markus Armbruster <armbru@redhat.com> # QAPI Schema
Message-Id: <20220429144110.25167-28-Jonathan.Cameron@huawei.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
2022-04-29 17:40:52 +03:00
|
|
|
{
|
2022-06-08 17:54:35 +03:00
|
|
|
if (cxl_state && cxl_state->fixed_windows) {
|
hw/cxl/host: Add support for CXL Fixed Memory Windows.
The concept of these is introduced in [1] in terms of the
description the CEDT ACPI table. The principal is more general.
Unlike once traffic hits the CXL root bridges, the host system
memory address routing is implementation defined and effectively
static once observable by standard / generic system software.
Each CXL Fixed Memory Windows (CFMW) is a region of PA space
which has fixed system dependent routing configured so that
accesses can be routed to the CXL devices below a set of target
root bridges. The accesses may be interleaved across multiple
root bridges.
For QEMU we could have fully specified these regions in terms
of a base PA + size, but as the absolute address does not matter
it is simpler to let individual platforms place the memory regions.
ExampleS:
-cxl-fixed-memory-window targets.0=cxl.0,size=128G
-cxl-fixed-memory-window targets.0=cxl.1,size=128G
-cxl-fixed-memory-window targets.0=cxl0,targets.1=cxl.1,size=256G,interleave-granularity=2k
Specifies
* 2x 128G regions not interleaved across root bridges, one for each of
the root bridges with ids cxl.0 and cxl.1
* 256G region interleaved across root bridges with ids cxl.0 and cxl.1
with a 2k interleave granularity.
When system software enumerates the devices below a given root bridge
it can then decide which CFMW to use. If non interleave is desired
(or possible) it can use the appropriate CFMW for the root bridge in
question. If there are suitable devices to interleave across the
two root bridges then it may use the 3rd CFMS.
A number of other designs were considered but the following constraints
made it hard to adapt existing QEMU approaches to this particular problem.
1) The size must be known before a specific architecture / board brings
up it's PA memory map. We need to set up an appropriate region.
2) Using links to the host bridges provides a clean command line interface
but these links cannot be established until command line devices have
been added.
Hence the two step process used here of first establishing the size,
interleave-ways and granularity + caching the ids of the host bridges
and then, once available finding the actual host bridges so they can
be used later to support interleave decoding.
[1] CXL 2.0 ECN: CEDT CFMWS & QTG DSM (computeexpresslink.org / specifications)
Signed-off-by: Jonathan Cameron <jonathan.cameron@huawei.com>
Acked-by: Markus Armbruster <armbru@redhat.com> # QAPI Schema
Message-Id: <20220429144110.25167-28-Jonathan.Cameron@huawei.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
2022-04-29 17:40:52 +03:00
|
|
|
GList *it;
|
|
|
|
|
2022-06-08 17:54:35 +03:00
|
|
|
for (it = cxl_state->fixed_windows; it; it = it->next) {
|
hw/cxl/host: Add support for CXL Fixed Memory Windows.
The concept of these is introduced in [1] in terms of the
description the CEDT ACPI table. The principal is more general.
Unlike once traffic hits the CXL root bridges, the host system
memory address routing is implementation defined and effectively
static once observable by standard / generic system software.
Each CXL Fixed Memory Windows (CFMW) is a region of PA space
which has fixed system dependent routing configured so that
accesses can be routed to the CXL devices below a set of target
root bridges. The accesses may be interleaved across multiple
root bridges.
For QEMU we could have fully specified these regions in terms
of a base PA + size, but as the absolute address does not matter
it is simpler to let individual platforms place the memory regions.
ExampleS:
-cxl-fixed-memory-window targets.0=cxl.0,size=128G
-cxl-fixed-memory-window targets.0=cxl.1,size=128G
-cxl-fixed-memory-window targets.0=cxl0,targets.1=cxl.1,size=256G,interleave-granularity=2k
Specifies
* 2x 128G regions not interleaved across root bridges, one for each of
the root bridges with ids cxl.0 and cxl.1
* 256G region interleaved across root bridges with ids cxl.0 and cxl.1
with a 2k interleave granularity.
When system software enumerates the devices below a given root bridge
it can then decide which CFMW to use. If non interleave is desired
(or possible) it can use the appropriate CFMW for the root bridge in
question. If there are suitable devices to interleave across the
two root bridges then it may use the 3rd CFMS.
A number of other designs were considered but the following constraints
made it hard to adapt existing QEMU approaches to this particular problem.
1) The size must be known before a specific architecture / board brings
up it's PA memory map. We need to set up an appropriate region.
2) Using links to the host bridges provides a clean command line interface
but these links cannot be established until command line devices have
been added.
Hence the two step process used here of first establishing the size,
interleave-ways and granularity + caching the ids of the host bridges
and then, once available finding the actual host bridges so they can
be used later to support interleave decoding.
[1] CXL 2.0 ECN: CEDT CFMWS & QTG DSM (computeexpresslink.org / specifications)
Signed-off-by: Jonathan Cameron <jonathan.cameron@huawei.com>
Acked-by: Markus Armbruster <armbru@redhat.com> # QAPI Schema
Message-Id: <20220429144110.25167-28-Jonathan.Cameron@huawei.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
2022-04-29 17:40:52 +03:00
|
|
|
CXLFixedWindow *fw = it->data;
|
|
|
|
int i;
|
|
|
|
|
|
|
|
for (i = 0; i < fw->num_targets; i++) {
|
|
|
|
Object *o;
|
|
|
|
bool ambig;
|
|
|
|
|
|
|
|
o = object_resolve_path_type(fw->targets[i],
|
2023-04-20 17:27:50 +03:00
|
|
|
TYPE_PXB_CXL_DEV,
|
hw/cxl/host: Add support for CXL Fixed Memory Windows.
The concept of these is introduced in [1] in terms of the
description the CEDT ACPI table. The principal is more general.
Unlike once traffic hits the CXL root bridges, the host system
memory address routing is implementation defined and effectively
static once observable by standard / generic system software.
Each CXL Fixed Memory Windows (CFMW) is a region of PA space
which has fixed system dependent routing configured so that
accesses can be routed to the CXL devices below a set of target
root bridges. The accesses may be interleaved across multiple
root bridges.
For QEMU we could have fully specified these regions in terms
of a base PA + size, but as the absolute address does not matter
it is simpler to let individual platforms place the memory regions.
ExampleS:
-cxl-fixed-memory-window targets.0=cxl.0,size=128G
-cxl-fixed-memory-window targets.0=cxl.1,size=128G
-cxl-fixed-memory-window targets.0=cxl0,targets.1=cxl.1,size=256G,interleave-granularity=2k
Specifies
* 2x 128G regions not interleaved across root bridges, one for each of
the root bridges with ids cxl.0 and cxl.1
* 256G region interleaved across root bridges with ids cxl.0 and cxl.1
with a 2k interleave granularity.
When system software enumerates the devices below a given root bridge
it can then decide which CFMW to use. If non interleave is desired
(or possible) it can use the appropriate CFMW for the root bridge in
question. If there are suitable devices to interleave across the
two root bridges then it may use the 3rd CFMS.
A number of other designs were considered but the following constraints
made it hard to adapt existing QEMU approaches to this particular problem.
1) The size must be known before a specific architecture / board brings
up it's PA memory map. We need to set up an appropriate region.
2) Using links to the host bridges provides a clean command line interface
but these links cannot be established until command line devices have
been added.
Hence the two step process used here of first establishing the size,
interleave-ways and granularity + caching the ids of the host bridges
and then, once available finding the actual host bridges so they can
be used later to support interleave decoding.
[1] CXL 2.0 ECN: CEDT CFMWS & QTG DSM (computeexpresslink.org / specifications)
Signed-off-by: Jonathan Cameron <jonathan.cameron@huawei.com>
Acked-by: Markus Armbruster <armbru@redhat.com> # QAPI Schema
Message-Id: <20220429144110.25167-28-Jonathan.Cameron@huawei.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
2022-04-29 17:40:52 +03:00
|
|
|
&ambig);
|
|
|
|
if (!o) {
|
|
|
|
error_setg(errp, "Could not resolve CXLFM target %s",
|
|
|
|
fw->targets[i]);
|
|
|
|
return;
|
|
|
|
}
|
|
|
|
fw->target_hbs[i] = PXB_CXL_DEV(o);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
2022-04-29 17:40:58 +03:00
|
|
|
|
|
|
|
static bool cxl_hdm_find_target(uint32_t *cache_mem, hwaddr addr,
|
|
|
|
uint8_t *target)
|
|
|
|
{
|
2023-09-13 16:25:22 +03:00
|
|
|
int hdm_inc = R_CXL_HDM_DECODER1_BASE_LO - R_CXL_HDM_DECODER0_BASE_LO;
|
2023-09-13 16:25:23 +03:00
|
|
|
unsigned int hdm_count;
|
|
|
|
bool found = false;
|
|
|
|
int i;
|
|
|
|
uint32_t cap;
|
|
|
|
|
|
|
|
cap = ldl_le_p(cache_mem + R_CXL_HDM_DECODER_CAPABILITY);
|
|
|
|
hdm_count = cxl_decoder_count_dec(FIELD_EX32(cap,
|
|
|
|
CXL_HDM_DECODER_CAPABILITY,
|
|
|
|
DECODER_COUNT));
|
|
|
|
for (i = 0; i < hdm_count; i++) {
|
|
|
|
uint32_t ctrl, ig_enc, iw_enc, target_idx;
|
|
|
|
uint32_t low, high;
|
|
|
|
uint64_t base, size;
|
|
|
|
|
|
|
|
low = ldl_le_p(cache_mem + R_CXL_HDM_DECODER0_BASE_LO + i * hdm_inc);
|
|
|
|
high = ldl_le_p(cache_mem + R_CXL_HDM_DECODER0_BASE_HI + i * hdm_inc);
|
|
|
|
base = (low & 0xf0000000) | ((uint64_t)high << 32);
|
|
|
|
low = ldl_le_p(cache_mem + R_CXL_HDM_DECODER0_SIZE_LO + i * hdm_inc);
|
|
|
|
high = ldl_le_p(cache_mem + R_CXL_HDM_DECODER0_SIZE_HI + i * hdm_inc);
|
|
|
|
size = (low & 0xf0000000) | ((uint64_t)high << 32);
|
|
|
|
if (addr < base || addr >= base + size) {
|
|
|
|
continue;
|
|
|
|
}
|
2022-04-29 17:40:58 +03:00
|
|
|
|
2023-09-13 16:25:23 +03:00
|
|
|
ctrl = ldl_le_p(cache_mem + R_CXL_HDM_DECODER0_CTRL + i * hdm_inc);
|
|
|
|
if (!FIELD_EX32(ctrl, CXL_HDM_DECODER0_CTRL, COMMITTED)) {
|
|
|
|
return false;
|
|
|
|
}
|
|
|
|
found = true;
|
|
|
|
ig_enc = FIELD_EX32(ctrl, CXL_HDM_DECODER0_CTRL, IG);
|
|
|
|
iw_enc = FIELD_EX32(ctrl, CXL_HDM_DECODER0_CTRL, IW);
|
|
|
|
target_idx = (addr / cxl_decode_ig(ig_enc)) % (1 << iw_enc);
|
|
|
|
|
|
|
|
if (target_idx < 4) {
|
|
|
|
uint32_t val = ldl_le_p(cache_mem +
|
|
|
|
R_CXL_HDM_DECODER0_TARGET_LIST_LO +
|
|
|
|
i * hdm_inc);
|
|
|
|
*target = extract32(val, target_idx * 8, 8);
|
|
|
|
} else {
|
|
|
|
uint32_t val = ldl_le_p(cache_mem +
|
|
|
|
R_CXL_HDM_DECODER0_TARGET_LIST_HI +
|
|
|
|
i * hdm_inc);
|
|
|
|
*target = extract32(val, (target_idx - 4) * 8, 8);
|
|
|
|
}
|
|
|
|
break;
|
2022-04-29 17:40:58 +03:00
|
|
|
}
|
|
|
|
|
2023-09-13 16:25:23 +03:00
|
|
|
return found;
|
2022-04-29 17:40:58 +03:00
|
|
|
}
|
|
|
|
|
|
|
|
static PCIDevice *cxl_cfmws_find_device(CXLFixedWindow *fw, hwaddr addr)
|
|
|
|
{
|
2022-06-16 17:51:25 +03:00
|
|
|
CXLComponentState *hb_cstate, *usp_cstate;
|
2022-04-29 17:40:58 +03:00
|
|
|
PCIHostState *hb;
|
2022-06-16 17:51:25 +03:00
|
|
|
CXLUpstreamPort *usp;
|
2022-04-29 17:40:58 +03:00
|
|
|
int rb_index;
|
|
|
|
uint32_t *cache_mem;
|
|
|
|
uint8_t target;
|
|
|
|
bool target_found;
|
|
|
|
PCIDevice *rp, *d;
|
|
|
|
|
|
|
|
/* Address is relative to memory region. Convert to HPA */
|
|
|
|
addr += fw->base;
|
|
|
|
|
|
|
|
rb_index = (addr / cxl_decode_ig(fw->enc_int_gran)) % fw->num_targets;
|
2023-04-20 17:27:50 +03:00
|
|
|
hb = PCI_HOST_BRIDGE(fw->target_hbs[rb_index]->cxl_host_bridge);
|
2022-04-29 17:40:58 +03:00
|
|
|
if (!hb || !hb->bus || !pci_bus_is_cxl(hb->bus)) {
|
|
|
|
return NULL;
|
|
|
|
}
|
|
|
|
|
2023-02-27 18:31:28 +03:00
|
|
|
if (cxl_get_hb_passthrough(hb)) {
|
|
|
|
rp = pcie_find_port_first(hb->bus);
|
|
|
|
if (!rp) {
|
|
|
|
return NULL;
|
|
|
|
}
|
|
|
|
} else {
|
|
|
|
hb_cstate = cxl_get_hb_cstate(hb);
|
|
|
|
if (!hb_cstate) {
|
|
|
|
return NULL;
|
|
|
|
}
|
2022-04-29 17:40:58 +03:00
|
|
|
|
2023-02-27 18:31:28 +03:00
|
|
|
cache_mem = hb_cstate->crb.cache_mem_registers;
|
2022-04-29 17:40:58 +03:00
|
|
|
|
2023-02-27 18:31:28 +03:00
|
|
|
target_found = cxl_hdm_find_target(cache_mem, addr, &target);
|
|
|
|
if (!target_found) {
|
|
|
|
return NULL;
|
|
|
|
}
|
2022-04-29 17:40:58 +03:00
|
|
|
|
2023-02-27 18:31:28 +03:00
|
|
|
rp = pcie_find_port_by_pn(hb->bus, target);
|
|
|
|
if (!rp) {
|
|
|
|
return NULL;
|
|
|
|
}
|
2022-04-29 17:40:58 +03:00
|
|
|
}
|
|
|
|
|
|
|
|
d = pci_bridge_get_sec_bus(PCI_BRIDGE(rp))->devices[0];
|
2022-06-16 17:51:25 +03:00
|
|
|
if (!d) {
|
|
|
|
return NULL;
|
|
|
|
}
|
|
|
|
|
|
|
|
if (object_dynamic_cast(OBJECT(d), TYPE_CXL_TYPE3)) {
|
|
|
|
return d;
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Could also be a switch. Note only one level of switching currently
|
|
|
|
* supported.
|
|
|
|
*/
|
|
|
|
if (!object_dynamic_cast(OBJECT(d), TYPE_CXL_USP)) {
|
|
|
|
return NULL;
|
|
|
|
}
|
|
|
|
usp = CXL_USP(d);
|
|
|
|
|
|
|
|
usp_cstate = cxl_usp_to_cstate(usp);
|
|
|
|
if (!usp_cstate) {
|
|
|
|
return NULL;
|
|
|
|
}
|
|
|
|
|
|
|
|
cache_mem = usp_cstate->crb.cache_mem_registers;
|
|
|
|
|
|
|
|
target_found = cxl_hdm_find_target(cache_mem, addr, &target);
|
|
|
|
if (!target_found) {
|
|
|
|
return NULL;
|
|
|
|
}
|
|
|
|
|
|
|
|
d = pcie_find_port_by_pn(&PCI_BRIDGE(d)->sec_bus, target);
|
|
|
|
if (!d) {
|
|
|
|
return NULL;
|
|
|
|
}
|
|
|
|
|
|
|
|
d = pci_bridge_get_sec_bus(PCI_BRIDGE(d))->devices[0];
|
|
|
|
if (!d) {
|
|
|
|
return NULL;
|
|
|
|
}
|
2022-04-29 17:40:58 +03:00
|
|
|
|
2022-06-16 17:51:25 +03:00
|
|
|
if (!object_dynamic_cast(OBJECT(d), TYPE_CXL_TYPE3)) {
|
2022-04-29 17:40:58 +03:00
|
|
|
return NULL;
|
|
|
|
}
|
|
|
|
|
|
|
|
return d;
|
|
|
|
}
|
|
|
|
|
|
|
|
static MemTxResult cxl_read_cfmws(void *opaque, hwaddr addr, uint64_t *data,
|
|
|
|
unsigned size, MemTxAttrs attrs)
|
|
|
|
{
|
|
|
|
CXLFixedWindow *fw = opaque;
|
|
|
|
PCIDevice *d;
|
|
|
|
|
|
|
|
d = cxl_cfmws_find_device(fw, addr);
|
|
|
|
if (d == NULL) {
|
|
|
|
*data = 0;
|
|
|
|
/* Reads to invalid address return poison */
|
|
|
|
return MEMTX_ERROR;
|
|
|
|
}
|
|
|
|
|
|
|
|
return cxl_type3_read(d, addr + fw->base, data, size, attrs);
|
|
|
|
}
|
|
|
|
|
|
|
|
static MemTxResult cxl_write_cfmws(void *opaque, hwaddr addr,
|
|
|
|
uint64_t data, unsigned size,
|
|
|
|
MemTxAttrs attrs)
|
|
|
|
{
|
|
|
|
CXLFixedWindow *fw = opaque;
|
|
|
|
PCIDevice *d;
|
|
|
|
|
|
|
|
d = cxl_cfmws_find_device(fw, addr);
|
|
|
|
if (d == NULL) {
|
|
|
|
/* Writes to invalid address are silent */
|
|
|
|
return MEMTX_OK;
|
|
|
|
}
|
|
|
|
|
|
|
|
return cxl_type3_write(d, addr + fw->base, data, size, attrs);
|
|
|
|
}
|
|
|
|
|
|
|
|
const MemoryRegionOps cfmws_ops = {
|
|
|
|
.read_with_attrs = cxl_read_cfmws,
|
|
|
|
.write_with_attrs = cxl_write_cfmws,
|
|
|
|
.endianness = DEVICE_LITTLE_ENDIAN,
|
|
|
|
.valid = {
|
|
|
|
.min_access_size = 1,
|
|
|
|
.max_access_size = 8,
|
|
|
|
.unaligned = true,
|
|
|
|
},
|
|
|
|
.impl = {
|
|
|
|
.min_access_size = 1,
|
|
|
|
.max_access_size = 8,
|
|
|
|
.unaligned = true,
|
|
|
|
},
|
|
|
|
};
|
2022-06-08 17:54:33 +03:00
|
|
|
|
|
|
|
static void machine_get_cxl(Object *obj, Visitor *v, const char *name,
|
|
|
|
void *opaque, Error **errp)
|
|
|
|
{
|
|
|
|
CXLState *cxl_state = opaque;
|
|
|
|
bool value = cxl_state->is_enabled;
|
|
|
|
|
|
|
|
visit_type_bool(v, name, &value, errp);
|
|
|
|
}
|
|
|
|
|
|
|
|
static void machine_set_cxl(Object *obj, Visitor *v, const char *name,
|
|
|
|
void *opaque, Error **errp)
|
|
|
|
{
|
|
|
|
CXLState *cxl_state = opaque;
|
|
|
|
bool value;
|
|
|
|
|
|
|
|
if (!visit_type_bool(v, name, &value, errp)) {
|
|
|
|
return;
|
|
|
|
}
|
|
|
|
cxl_state->is_enabled = value;
|
|
|
|
}
|
|
|
|
|
|
|
|
static void machine_get_cfmw(Object *obj, Visitor *v, const char *name,
|
|
|
|
void *opaque, Error **errp)
|
|
|
|
{
|
2024-07-05 14:39:54 +03:00
|
|
|
CXLState *state = opaque;
|
|
|
|
CXLFixedMemoryWindowOptionsList **list = &state->cfmw_list;
|
2022-06-08 17:54:33 +03:00
|
|
|
|
|
|
|
visit_type_CXLFixedMemoryWindowOptionsList(v, name, list, errp);
|
|
|
|
}
|
|
|
|
|
|
|
|
static void machine_set_cfmw(Object *obj, Visitor *v, const char *name,
|
|
|
|
void *opaque, Error **errp)
|
|
|
|
{
|
|
|
|
CXLState *state = opaque;
|
|
|
|
CXLFixedMemoryWindowOptionsList *cfmw_list = NULL;
|
|
|
|
CXLFixedMemoryWindowOptionsList *it;
|
|
|
|
|
|
|
|
visit_type_CXLFixedMemoryWindowOptionsList(v, name, &cfmw_list, errp);
|
|
|
|
if (!cfmw_list) {
|
|
|
|
return;
|
|
|
|
}
|
|
|
|
|
|
|
|
for (it = cfmw_list; it; it = it->next) {
|
|
|
|
cxl_fixed_memory_window_config(state, it->value, errp);
|
|
|
|
}
|
|
|
|
state->cfmw_list = cfmw_list;
|
|
|
|
}
|
|
|
|
|
|
|
|
void cxl_machine_init(Object *obj, CXLState *state)
|
|
|
|
{
|
|
|
|
object_property_add(obj, "cxl", "bool", machine_get_cxl,
|
|
|
|
machine_set_cxl, NULL, state);
|
|
|
|
object_property_set_description(obj, "cxl",
|
|
|
|
"Set on/off to enable/disable "
|
|
|
|
"CXL instantiation");
|
|
|
|
|
|
|
|
object_property_add(obj, "cxl-fmw", "CXLFixedMemoryWindow",
|
|
|
|
machine_get_cfmw, machine_set_cfmw,
|
|
|
|
NULL, state);
|
|
|
|
object_property_set_description(obj, "cxl-fmw",
|
|
|
|
"CXL Fixed Memory Windows (array)");
|
|
|
|
}
|
2022-06-08 17:54:37 +03:00
|
|
|
|
|
|
|
void cxl_hook_up_pxb_registers(PCIBus *bus, CXLState *state, Error **errp)
|
|
|
|
{
|
|
|
|
/* Walk the pci busses looking for pxb busses to hook up */
|
|
|
|
if (bus) {
|
|
|
|
QLIST_FOREACH(bus, &bus->child, sibling) {
|
|
|
|
if (!pci_bus_is_root(bus)) {
|
|
|
|
continue;
|
|
|
|
}
|
|
|
|
if (pci_bus_is_cxl(bus)) {
|
|
|
|
if (!state->is_enabled) {
|
|
|
|
error_setg(errp, "CXL host bridges present, but cxl=off");
|
|
|
|
return;
|
|
|
|
}
|
|
|
|
pxb_cxl_hook_up_registers(state, bus, errp);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|