hw/nvme: add monitor commands for inspecting state
PATCH 2/2 hw/nvme: add 'info nvme-queues' HMP command
In reply to Mateusz NowickiAdd an 'info nvme-queues' HMP command that lists, for every emulated NVMe controller, the admin SQ/CQ and every active I/O SQ/CQ.
For each queue the command prints:
- the ring size, head and tail indices;
- the associated CQ id (for SQ) or interrupt vector (for CQ);
- the queue's PRP1 (DMA base address of the ring);
- the BAR0-relative doorbell offset (SQyTDBL / CQyHDBL);
- the current CQ phase tag.
The doorbell offsets are computed from CAP.DSTRD so they remain correct if the emulation ever exposes a non-zero stride.
Useful when debugging the Linux NVMe driver against the QEMU emulation - queue setup, doorbell rings, AERs held in the admin SQ etc. - without attaching gdb to the QEMU process.
Example, with the long PRP1 ring address and BAR0-relative doorbell offset replaced by '...'; phase tag column omitted for brevity:
(qemu) info nvme-queues
/machine/peripheral-anon/device[0]
SQ 0 size=32 head=9 tail=9 cqid=0 prp1=... SQTDBL=... CQ 0 size=32 head=8 tail=8 iv=0 prp1=... CQHDBL=... SQ 1 size=1024 head=2 tail=2 cqid=1 prp1=... SQTDBL=... CQ 1 size=1024 head=2 tail=2 iv=1 prp1=... CQHDBL=...
--- hmp-commands-info.hx | 15 +++++++++++ hw/nvme/monitor.c | 63 ++++++++++++++++++++++++++++++++++++++++++++ qapi/machine.json | 17 ++++++++++++ 3 files changed, 95 insertions(+)
hmp-commands-info.hx +15 -0
@@ -353,6 +353,21 @@ SRST Show emulated NVMe controllers. ERST + { + .name = "nvme-queues", + .args_type = "", + .params = "", + .help = "show active NVMe queues and their doorbells", + .cmd_info_hrt = qmp_x_query_nvme_queues, + }, + +SRST + ``info nvme-queues`` + Show all active NVMe submission and completion queues, including + head/tail, DMA address of the ring and BAR0-relative doorbell + offset. +ERST + { .name = "usbhost", .args_type = "",
hw/nvme/monitor.c +63 -0
66 lines not on this page · raw message
qapi/machine.json +17 -0
@@ -1789,6 +1789,23 @@ 'returns': 'HumanReadableText', 'features': [ 'unstable' ] } +## +# @x-query-nvme-queues: +# +# Query state of all active SQ/CQ for emulated NVMe controllers. +# +# Features: +# +# @unstable: This command is meant for debugging. +# +# Returns: per-queue state as human-readable text +# +# Since: 11.1 +## +{ 'command': 'x-query-nvme-queues', + 'returns': 'HumanReadableText', + 'features': [ 'unstable' ] } + ## # @SmbiosEntryPointType: #
Signature · 1 line
2.53.0
cover hw/nvme: add monitor commands for inspecting state
Add two HMP commands for inspecting emulated NVMe controllers from the QEMU monitor without attaching gdb to the QEMU process:
- 'info nvme' - per-controller summary (PCI, identify
fields, CC/CSTS/AQA, queue counts)
- 'info nvme-queues' - per-queue listing of admin and I/O SQ/CQ (size, head/tail, PRP1, doorbell offset, phase tag)
Useful for verifying queue setup, doorbell rings, AERs held in the admin SQ and similar driver/controller interaction details from a running QEMU monitor.
Mateusz Nowicki (2):
hw/nvme: add 'info nvme' HMP command
hw/nvme: add 'info nvme-queues' HMP command
hmp-commands-info.hx | 28 +++++++++ hw/nvme/meson.build | 2 +- hw/nvme/monitor.c | 139 +++++++++++++++++++++++++++++++++++++++++++ qapi/machine.json | 34 +++++++++++ 4 files changed, 202 insertions(+), 1 deletion(-) create mode 100644 hw/nvme/monitor.c
Signature · 1 line
2.53.0
PATCH 1/2 hw/nvme: add 'info nvme' HMP command
Add an 'info nvme' HMP command for inspecting emulated NVMe controllers from the QEMU monitor.
For each NVMe controller in the QOM tree the command prints:
- PCI BDF, vendor/device ID and BAR0 base address;
- Identify Controller fields visible to the host driver: SN, MN, FR
and CNTLID;
- the CC, CSTS and AQA registers (raw 32-bit values);
- the number of admin and active I/O submission/completion queues.
Useful when debugging the Linux NVMe driver against the QEMU emulation without attaching gdb to the QEMU process.
Example:
(qemu) info nvme
/machine/peripheral-anon/device[0]
PCI: BDF 00:04.0 VID=8086 DID=5845 BAR0=0x00000000feb50000
ID: SN=NVME0001 MN=QEMU NVMe Ctrl FR=... CNTLID=0x0000
CC: 0x00460001
CSTS: 0x00000001
AQA: 0x001f001fQueues: 1 admin + 4 IO SQ / 4 IO CQ
--- hmp-commands-info.hx | 13 ++++++++ hw/nvme/meson.build | 2 +- hw/nvme/monitor.c | 76 ++++++++++++++++++++++++++++++++++++++++++++ qapi/machine.json | 17 ++++++++++ 4 files changed, 107 insertions(+), 1 deletion(-) create mode 100644 hw/nvme/monitor.c
hmp-commands-info.hx +13 -0
@@ -340,6 +340,19 @@ SRST Show guest USB devices. ERST + { + .name = "nvme", + .args_type = "", + .params = "", + .help = "show emulated NVMe controllers", + .cmd_info_hrt = qmp_x_query_nvme, + }, + +SRST + ``info nvme`` + Show emulated NVMe controllers. +ERST + { .name = "usbhost", .args_type = "",
hw/nvme/meson.build +1 -1
@@ -1 +1 @@ -system_ss.add(when: 'CONFIG_NVME_PCI', if_true: files('ctrl.c', 'dif.c', 'ns.c', 'subsys.c', 'nguid.c')) +system_ss.add(when: 'CONFIG_NVME_PCI', if_true: files('ctrl.c', 'dif.c', 'ns.c', 'subsys.c', 'nguid.c', 'monitor.c'))
hw/nvme/monitor.c new file +76 -0
76 lines not on this page · raw message
qapi/machine.json +17 -0
@@ -1772,6 +1772,23 @@ 'returns': 'HumanReadableText', 'features': [ 'unstable' ] } +## +# @x-query-nvme: +# +# Query state of emulated NVMe controllers. +# +# Features: +# +# @unstable: This command is meant for debugging. +# +# Returns: NVMe controller state as human-readable text +# +# Since: 11.1 +## +{ 'command': 'x-query-nvme', + 'returns': 'HumanReadableText', + 'features': [ 'unstable' ] } + ## # @SmbiosEntryPointType: #
Signature · 1 line
2.53.0
Re: [PATCH 1/2] hw/nvme: add 'info nvme' HMP command
Quoted 16 lines · * Mateusz Nowicki ([email protected]) wrote:
Add an 'info nvme' HMP command for inspecting emulated NVMe controllers from the QEMU monitor.
For each NVMe controller in the QOM tree the command prints:
- PCI BDF, vendor/device ID and BAR0 base address;
- Identify Controller fields visible to the host driver: SN, MN, FRand CNTLID;- the CC, CSTS and AQA registers (raw 32-bit values);
- the number of admin and active I/O submission/completion queues.Useful when debugging the Linux NVMe driver against the QEMU emulation without attaching gdb to the QEMU process.
Example:
Quoted 8 lines
(qemu) info nvme
/machine/peripheral-anon/device[0]PCI: BDF 00:04.0 VID=8086 DID=5845 BAR0=0x00000000feb50000 ID: SN=NVME0001 MN=QEMU NVMe Ctrl FR=... CNTLID=0x0000 CC: 0x00460001 CSTS: 0x00000001 AQA: 0x001f001fQueues: 1 admin + 4 IO SQ / 4 IO CQ
I won't claim to know what those NVMe magic numbers mean, but I'll trust you they're useful.
Quoted 6 lines
Signed-off-by Mateusz Nowicki <[email protected]>--- hmp-commands-info.hx | 13 ++++++++ hw/nvme/meson.build | 2 +- hw/nvme/monitor.c | 76 ++++++++++++++++++++++++++++++++++++++++++++ qapi/machine.json | 17 ++++++++++
So that lot looks OK from an HMP point of view, I wonder if qapi/machine.json is the right place for it. Markus?
Quoted 149 lines
4 files changed, 107 insertions(+), 1 deletion(-) create mode 100644 hw/nvme/monitor.chmp-commands-info.hx +13
-0@@ -340,6 +340,19 @@ SRST Show guest USB devices. ERST + { + .name = "nvme", + .args_type = "", + .params = "", + .help = "show emulated NVMe controllers", + .cmd_info_hrt = qmp_x_query_nvme, + }, + +SRST + ``info nvme`` + Show emulated NVMe controllers. +ERST + { .name = "usbhost", .args_type = "",hw/nvme/meson.build +1
-1@@ -1 +1 @@ -system_ss.add(when: 'CONFIG_NVME_PCI', if_true: files('ctrl.c', 'dif.c', 'ns.c', 'subsys.c', 'nguid.c')) +system_ss.add(when: 'CONFIG_NVME_PCI', if_true: files('ctrl.c', 'dif.c', 'ns.c', 'subsys.c', 'nguid.c', 'monitor.c'))hw/nvme/monitor.c new file +76
-0@@ -0,0 +1,76 @@ +/* +* QEMU NVMe Controller monitor commands +* +* SPDX-License-Identifier: GPL-2.0-or-later +*/ + +#include "qemu/osdep.h" +#include "qapi/error.h" +#include "qapi/qapi-commands-machine.h" +#include "qapi/type-helpers.h" +#include "hw/pci/pci.h" + +#include "nvme.h" + +static int collect_one (Object *obj,void *opaque) +{ +GString *buf = opaque; +NvmeCtrl *n; +PCIDevice *pci; +pcibus_t bar0; +unsigned io_sq =0 , io_cq =0 ; + +if (!object_dynamic_cast (obj,TYPE_NVME )) { +return 0 ; + } + n =NVME (obj); + pci =PCI_DEVICE (n); + bar0 =pci_get_bar_addr (pci,0 ); + +for (unsigned i =1 ; i <= n->params.max_ioqpairs; i++) { +if (n->sq && n->sq[i]) { + io_sq++; + } +if (n->cq && n->cq[i]) { + io_cq++; + } + } + +g_string_append_printf (buf,"%s\n" ,object_get_canonical_path (obj)); +g_string_append_printf (buf, +" PCI: BDF %02x:%02x.%x VID=%04x DID=%04x " , +pci_dev_bus_num (pci),PCI_SLOT (pci->devfn),PCI_FUNC (pci->devfn), +pci_get_word (pci->config +PCI_VENDOR_ID ), +pci_get_word (pci->config +PCI_DEVICE_ID )); +if (bar0 ==PCI_BAR_UNMAPPED ) { +g_string_append (buf,"BAR0=unmapped\n" ); + }else { +g_string_append_printf (buf,"BAR0=0x%016" PRIx64"\n" , + (uint64_t )bar0); + } +g_string_append_printf (buf, +" ID: SN=%.20s MN=%.40s FR=%.8s CNTLID=0x%04x\n" , + n->id_ctrl.sn, n->id_ctrl.mn, n->id_ctrl.fr, n->cntlid); +g_string_append_printf (buf," CC: 0x%08x\n" , +ldl_le_p (&n->bar.cc)); +g_string_append_printf (buf," CSTS: 0x%08x\n" , +ldl_le_p (&n->bar.csts)); +g_string_append_printf (buf," AQA: 0x%08x\n" , +ldl_le_p (&n->bar.aqa)); +g_string_append_printf (buf, +" Queues: 1 admin + %u IO SQ / %u IO CQ\n" , io_sq, io_cq); +return 0 ; +} + +HumanReadableText *qmp_x_query_nvme (Error **errp) +{ +g_autoptr (GString) buf =g_string_new ("" ); + +object_child_foreach_recursive (object_get_root (), collect_one, buf); + +if (buf->len ==0 ) { +g_string_append (buf,"no NVMe controllers\n" ); + } + +return human_readable_text_from_str (buf); +}qapi/machine.json +17
-0@@ -1772,6 +1772,23 @@ 'returns': 'HumanReadableText', 'features': [ 'unstable' ] } +## +# @x-query-nvme: +# +# Query state of emulated NVMe controllers. +# +# Features: +# +# @unstable: This command is meant for debugging. +# +# Returns: NVMe controller state as human-readable text +# +# Since: 11.1 +## +{ 'command': 'x-query-nvme', + 'returns': 'HumanReadableText', + 'features': [ 'unstable' ] } + ## # @SmbiosEntryPointType: #Signature · 1 line
2.53.0
Signature · 4 lines
-----Open up your eyes, open up your mind, open up your code ------- / Dr. David Alan Gilbert | Running GNU/Linux | Happy \ \ dave @ treblig.org | | In Hex / \ _________________________|_____ http://www.treblig.org |_______/
Re: [PATCH 2/2] hw/nvme: add 'info nvme-queues' HMP command
In reply to Mateusz NowickiQuoted 21 lines · * Mateusz Nowicki ([email protected]) wrote:
Add an 'info nvme-queues' HMP command that lists, for every emulated NVMe controller, the admin SQ/CQ and every active I/O SQ/CQ.
For each queue the command prints:
- the ring size, head and tail indices;
- the associated CQ id (for SQ) or interrupt vector (for CQ);
- the queue's PRP1 (DMA base address of the ring);
- the BAR0-relative doorbell offset (SQyTDBL / CQyHDBL);
- the current CQ phase tag.The doorbell offsets are computed from CAP.DSTRD so they remain correct if the emulation ever exposes a non-zero stride.
Useful when debugging the Linux NVMe driver against the QEMU emulation - queue setup, doorbell rings, AERs held in the admin SQ etc. - without attaching gdb to the QEMU process.
Example, with the long PRP1 ring address and BAR0-relative doorbell offset replaced by '...'; phase tag column omitted for brevity:
Quoted 8 lines
(qemu) info nvme-queues
/machine/peripheral-anon/device[0]SQ 0 size=32 head=9 tail=9 cqid=0 prp1=... SQTDBL=... CQ 0 size=32 head=8 tail=8 iv=0 prp1=... CQHDBL=... SQ 1 size=1024 head=2 tail=2 cqid=1 prp1=... SQTDBL=... CQ 1 size=1024 head=2 tail=2 iv=1 prp1=... CQHDBL=...Signed-off-by Mateusz Nowicki <[email protected]>
as the previous patch; looks OK from HMP but check where the json should live.
Quoted 134 lines
--- hmp-commands-info.hx | 15 +++++++++++ hw/nvme/monitor.c | 63 ++++++++++++++++++++++++++++++++++++++++++++ qapi/machine.json | 17 ++++++++++++ 3 files changed, 95 insertions(+)hmp-commands-info.hx +15
-0@@ -353,6 +353,21 @@ SRST Show emulated NVMe controllers. ERST + { + .name = "nvme-queues", + .args_type = "", + .params = "", + .help = "show active NVMe queues and their doorbells", + .cmd_info_hrt = qmp_x_query_nvme_queues, + }, + +SRST + ``info nvme-queues`` + Show all active NVMe submission and completion queues, including + head/tail, DMA address of the ring and BAR0-relative doorbell + offset. +ERST + { .name = "usbhost", .args_type = "",hw/nvme/monitor.c +63
-0@@ -74,3 +74,66 @@ HumanReadableText *qmp_x_query_nvme(Error **errp)return human_readable_text_from_str (buf); } + +static void append_sq (GString *buf,NvmeSQueue *sq,unsigned stride) +{ +hwaddr db_off =0x1000 +2 * sq->sqid * stride; + +g_string_append_printf (buf, +" SQ %u size=%-5u head=%-5u tail=%-5u cqid=%-3u " +"prp1=0x%016" PRIx64" SQTDBL=BAR0+0x%03" HWADDR_PRIx"\n" , + sq->sqid, sq->size, sq->head, sq->tail, sq->cqid, + sq->dma_addr, db_off); +} + +static void append_cq (GString *buf,NvmeCQueue *cq,unsigned stride) +{ +hwaddr db_off =0x1000 + (2 * cq->cqid +1 ) * stride; + +g_string_append_printf (buf, +" CQ %u size=%-5u head=%-5u tail=%-5u iv=%-5u " +"prp1=0x%016" PRIx64" CQHDBL=BAR0+0x%03" HWADDR_PRIx +" phaseTag=%u\n" , + cq->cqid, cq->size, cq->head, cq->tail, cq->vector, + cq->dma_addr, db_off, cq->phase); +} + +static int collect_queues (Object *obj,void *opaque) +{ +GString *buf = opaque; +NvmeCtrl *n; +unsigned stride; + +if (!object_dynamic_cast (obj,TYPE_NVME )) { +return 0 ; + } + n =NVME (obj); + stride =4u <<NVME_CAP_DSTRD (ldq_le_p (&n->bar.cap)); + +g_string_append_printf (buf,"%s\n" ,object_get_canonical_path (obj)); +append_sq (buf, &n->admin_sq, stride); +append_cq (buf, &n->admin_cq, stride); + +for (unsigned i =1 ; i <= n->params.max_ioqpairs; i++) { +if (n->sq && n->sq[i]) { +append_sq (buf, n->sq[i], stride); + } +if (n->cq && n->cq[i]) { +append_cq (buf, n->cq[i], stride); + } + } +return 0 ; +} + +HumanReadableText *qmp_x_query_nvme_queues (Error **errp) +{ +g_autoptr (GString) buf =g_string_new ("" ); + +object_child_foreach_recursive (object_get_root (), collect_queues, buf); + +if (buf->len ==0 ) { +g_string_append (buf,"no NVMe controllers\n" ); + } + +return human_readable_text_from_str (buf); +}qapi/machine.json +17
-0@@ -1789,6 +1789,23 @@ 'returns': 'HumanReadableText', 'features': [ 'unstable' ] } +## +# @x-query-nvme-queues: +# +# Query state of all active SQ/CQ for emulated NVMe controllers. +# +# Features: +# +# @unstable: This command is meant for debugging. +# +# Returns: per-queue state as human-readable text +# +# Since: 11.1 +## +{ 'command': 'x-query-nvme-queues', + 'returns': 'HumanReadableText', + 'features': [ 'unstable' ] } + ## # @SmbiosEntryPointType: #Signature · 1 line
2.53.0
Signature · 4 lines
-----Open up your eyes, open up your mind, open up your code ------- / Dr. David Alan Gilbert | Running GNU/Linux | Happy \ \ dave @ treblig.org | | In Hex / \ _________________________|_____ http://www.treblig.org |_______/
Markus Armbruster
Acked-byRe: [PATCH 1/2] hw/nvme: add 'info nvme' HMP command
In reply to Dr. David Alan GilbertQuoted 36 lines · Dr. David Alan Gilbert
Quoted 16 lines · * Mateusz Nowicki ([email protected]) wrote:
Add an 'info nvme' HMP command for inspecting emulated NVMe controllers from the QEMU monitor.
For each NVMe controller in the QOM tree the command prints:
- PCI BDF, vendor/device ID and BAR0 base address;
- Identify Controller fields visible to the host driver: SN, MN, FRand CNTLID;- the CC, CSTS and AQA registers (raw 32-bit values);
- the number of admin and active I/O submission/completion queues.Useful when debugging the Linux NVMe driver against the QEMU emulation without attaching gdb to the QEMU process.
Example:
Quoted 8 lines
(qemu) info nvme
/machine/peripheral-anon/device[0]PCI: BDF 00:04.0 VID=8086 DID=5845 BAR0=0x00000000feb50000 ID: SN=NVME0001 MN=QEMU NVMe Ctrl FR=... CNTLID=0x0000 CC: 0x00460001 CSTS: 0x00000001 AQA: 0x001f001fQueues: 1 admin + 4 IO SQ / 4 IO CQ
I won't claim to know what those NVMe magic numbers mean, but I'll trust you they're useful.
Quoted 6 lines
Signed-off-by Mateusz Nowicki <[email protected]>--- hmp-commands-info.hx | 13 ++++++++ hw/nvme/meson.build | 2 +- hw/nvme/monitor.c | 76 ++++++++++++++++++++++++++++++++++++++++++++ qapi/machine.json | 17 ++++++++++
Quoted 2 lines
So that lot looks OK from an HMP point of view, I wonder if
qapi/machine.jsonis the right place for it. Markus?
qapi/machine.json has become a bit of a dumping ground, I'm afraid.
Precedence for putting queries for the machine's device frontends of certain kinds there: x-query-irq, x-query-interrupt-controllers, and x-query-usb.
All the other NVMe QAPI stuff is in qapi/block-core.json, but that's the block backend, not frontend. Not a good home for x-query-nvme.
Let's take the patch as is.
One nitpick below.
Quoted 30 lines
Quoted 28 lines
4 files changed, 107 insertions(+), 1 deletion(-) create mode 100644 hw/nvme/monitor.chmp-commands-info.hx +13
-0@@ -340,6 +340,19 @@ SRST Show guest USB devices. ERST + { + .name = "nvme", + .args_type = "", + .params = "", + .help = "show emulated NVMe controllers", + .cmd_info_hrt = qmp_x_query_nvme, + }, + +SRST + ``info nvme`` + Show emulated NVMe controllers. +ERST + { .name = "usbhost", .args_type = "",hw/nvme/meson.build +0
-0
Quoted 8 lines
hw/nvme/meson.build line 1
Quoted 8 lines
index
7d5caa53c2..3017f058f9100644hw/nvme/meson.build +1
-1@@ -1 +1 @@ -system_ss.add(when: 'CONFIG_NVME_PCI', if_true: files('ctrl.c', 'dif.c', 'ns.c', 'subsys.c', 'nguid.c')) +system_ss.add(when: 'CONFIG_NVME_PCI', if_true: files('ctrl.c', 'dif.c', 'ns.c', 'subsys.c', 'nguid.c', 'monitor.c'))
Please use the opportunity to add the newline.
[...]
QAPI schema
Markus Armbruster
Acked-byRe: [PATCH 2/2] hw/nvme: add 'info nvme-queues' HMP command
In reply to Dr. David Alan GilbertQuoted 29 lines · Dr. David Alan Gilbert
Quoted 28 lines · * Mateusz Nowicki ([email protected]) wrote:
Add an 'info nvme-queues' HMP command that lists, for every emulated NVMe controller, the admin SQ/CQ and every active I/O SQ/CQ.
For each queue the command prints:
- the ring size, head and tail indices;
- the associated CQ id (for SQ) or interrupt vector (for CQ);
- the queue's PRP1 (DMA base address of the ring);
- the BAR0-relative doorbell offset (SQyTDBL / CQyHDBL);
- the current CQ phase tag.The doorbell offsets are computed from CAP.DSTRD so they remain correct if the emulation ever exposes a non-zero stride.
Useful when debugging the Linux NVMe driver against the QEMU emulation - queue setup, doorbell rings, AERs held in the admin SQ etc. - without attaching gdb to the QEMU process.
Example, with the long PRP1 ring address and BAR0-relative doorbell offset replaced by '...'; phase tag column omitted for brevity:
(qemu) info nvme-queues
/machine/peripheral-anon/device[0]SQ 0 size=32 head=9 tail=9 cqid=0 prp1=... SQTDBL=... CQ 0 size=32 head=8 tail=8 iv=0 prp1=... CQHDBL=... SQ 1 size=1024 head=2 tail=2 cqid=1 prp1=... SQTDBL=... CQ 1 size=1024 head=2 tail=2 iv=1 prp1=... CQHDBL=...
Quoted 5 lines
Quoted 1 line
Signed-off-by Mateusz Nowicki <[email protected]>as the previous patch; looks OK from HMP but check where the json should live.
Same answer.
QAPI schema
Re: [PATCH 1/2] hw/nvme: add 'info nvme' HMP command
In reply to Markus ArmbrusterThanks for the reviews. v2 sent, with the meson.build trailing newline fixed and your Acked-by tags picked up:
https://lore.kernel.org/qemu-devel/[email protected]/
Mateusz
Re: [PATCH 1/2] hw/nvme: add 'info nvme' HMP command
In reply to Mateusz NowickiQuoted 93 lines · Mateusz Nowicki
Add an 'info nvme' HMP command for inspecting emulated NVMe controllers from the QEMU monitor.
For each NVMe controller in the QOM tree the command prints:
- PCI BDF, vendor/device ID and BAR0 base address;
- Identify Controller fields visible to the host driver: SN, MN, FRand CNTLID;- the CC, CSTS and AQA registers (raw 32-bit values);
- the number of admin and active I/O submission/completion queues.Useful when debugging the Linux NVMe driver against the QEMU emulation without attaching gdb to the QEMU process.
Example:
(qemu) info nvme
/machine/peripheral-anon/device[0]PCI: BDF 00:04.0 VID=8086 DID=5845 BAR0=0x00000000feb50000 ID: SN=NVME0001 MN=QEMU NVMe Ctrl FR=... CNTLID=0x0000 CC: 0x00460001 CSTS: 0x00000001 AQA: 0x001f001fQueues: 1 admin + 4 IO SQ / 4 IO CQ
Signed-off-by Mateusz Nowicki <[email protected]>--- hmp-commands-info.hx | 13 ++++++++ hw/nvme/meson.build | 2 +- hw/nvme/monitor.c | 76 ++++++++++++++++++++++++++++++++++++++++++++ qapi/machine.json | 17 ++++++++++ 4 files changed, 107 insertions(+), 1 deletion(-) create mode 100644 hw/nvme/monitor.chmp-commands-info.hx +13
-0@@ -340,6 +340,19 @@ SRST Show guest USB devices. ERST + { + .name = "nvme", + .args_type = "", + .params = "", + .help = "show emulated NVMe controllers", + .cmd_info_hrt = qmp_x_query_nvme, + }, + +SRST + ``info nvme`` + Show emulated NVMe controllers. +ERST + { .name = "usbhost", .args_type = "",hw/nvme/meson.build +1
-1@@ -1 +1 @@ -system_ss.add(when: 'CONFIG_NVME_PCI', if_true: files('ctrl.c', 'dif.c', 'ns.c', 'subsys.c', 'nguid.c')) +system_ss.add(when: 'CONFIG_NVME_PCI', if_true: files('ctrl.c', 'dif.c', 'ns.c', 'subsys.c', 'nguid.c', 'monitor.c'))hw/nvme/monitor.c new file +20
-0@@ -0,0 +1,76 @@ +/* +* QEMU NVMe Controller monitor commands +* +* SPDX-License-Identifier: GPL-2.0-or-later +*/ + +#include "qemu/osdep.h" +#include "qapi/error.h" +#include "qapi/qapi-commands-machine.h" +#include "qapi/type-helpers.h" +#include "hw/pci/pci.h" + +#include "nvme.h" + +static int collect_one (Object *obj,void *opaque) +{ +GString *buf = opaque; +NvmeCtrl *n; +PCIDevice *pci; +pcibus_t bar0;
hw/nvme/monitor.c line 30
Quoted 10 lines
hw/nvme/monitor.c new file +10
-0+unsigned io_sq =0 , io_cq =0 ; + +if (!object_dynamic_cast (obj,TYPE_NVME )) { +return 0 ; + } + n =NVME (obj); + pci =PCI_DEVICE (n); + bar0 =pci_get_bar_addr (pci,0 ); + +for (unsigned i =1 ; i <= n->params.max_ioqpairs; i++) {
I think this needs to run over n->conf_ioqpairs to also work for SR-IOV VFs? Same goes for patch 2 I think.
Quoted 78 lines
hw/nvme/monitor.c new file +46
-0+if (n->sq && n->sq[i]) { + io_sq++; + } +if (n->cq && n->cq[i]) { + io_cq++; + } + } + +g_string_append_printf (buf,"%s\n" ,object_get_canonical_path (obj)); +g_string_append_printf (buf, +" PCI: BDF %02x:%02x.%x VID=%04x DID=%04x " , +pci_dev_bus_num (pci),PCI_SLOT (pci->devfn),PCI_FUNC (pci->devfn), +pci_get_word (pci->config +PCI_VENDOR_ID ), +pci_get_word (pci->config +PCI_DEVICE_ID )); +if (bar0 ==PCI_BAR_UNMAPPED ) { +g_string_append (buf,"BAR0=unmapped\n" ); + }else { +g_string_append_printf (buf,"BAR0=0x%016" PRIx64"\n" , + (uint64_t )bar0); + } +g_string_append_printf (buf, +" ID: SN=%.20s MN=%.40s FR=%.8s CNTLID=0x%04x\n" , + n->id_ctrl.sn, n->id_ctrl.mn, n->id_ctrl.fr, n->cntlid); +g_string_append_printf (buf," CC: 0x%08x\n" , +ldl_le_p (&n->bar.cc)); +g_string_append_printf (buf," CSTS: 0x%08x\n" , +ldl_le_p (&n->bar.csts)); +g_string_append_printf (buf," AQA: 0x%08x\n" , +ldl_le_p (&n->bar.aqa)); +g_string_append_printf (buf, +" Queues: 1 admin + %u IO SQ / %u IO CQ\n" , io_sq, io_cq); +return 0 ; +} + +HumanReadableText *qmp_x_query_nvme (Error **errp) +{ +g_autoptr (GString) buf =g_string_new ("" ); + +object_child_foreach_recursive (object_get_root (), collect_one, buf); + +if (buf->len ==0 ) { +g_string_append (buf,"no NVMe controllers\n" ); + } + +return human_readable_text_from_str (buf); +}qapi/machine.json +17
-0@@ -1772,6 +1772,23 @@ 'returns': 'HumanReadableText', 'features': [ 'unstable' ] } +## +# @x-query-nvme: +# +# Query state of emulated NVMe controllers. +# +# Features: +# +# @unstable: This command is meant for debugging. +# +# Returns: NVMe controller state as human-readable text +# +# Since: 11.1 +## +{ 'command': 'x-query-nvme', + 'returns': 'HumanReadableText', + 'features': [ 'unstable' ] } + ## # @SmbiosEntryPointType: #Signature · 1 line
2.53.0
Re: [PATCH 1/2] hw/nvme: add 'info nvme' HMP command
Quoted 3 lines
I think this needs to run over
n->conf_ioqpairsto also work for SR-IOV VFs? Same goes for patch 2 I think.
Good catch, agreed. Sending v3 with the fix applied to both patches. Thanks,
Mateusz