AI/rocm-systems
2
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

libhsakmt: fix pc sampling return of functions

Browse Source 
C Error returns are positive in user space and should check against errno
instead.
Fix declaration of return to type HSAKMT_STATUS.
KFD IOCTL should handle size return when querying capabilities so return
size to caller unconditionally.
Clean up error translations per function so that it's stylistically
clear.

Signed-off-by: Jonathan Kim <jonathan.kim@amd.com>
Change-Id: Ic37390425f370c7ad88f9ed014444decf19383a3
Signed-off-by: Chris Freehill <cfreehil@amd.com>


[ROCm/ROCR-Runtime commit: 206db80a56]
This commit is contained in:
Jonathan Kim
2024-04-18 10:39:03 -04:00
committed by Chris Freehill
parent af2a46cfd6
commit 23c0cf8727
1 changed files with 70 additions and 45 deletions
Download Patch File Download Diff File Expand all files Collapse all files
projects/rocr-runtime/libhsakmt/src/pc_sampling.c
+70 -45
View File
@@ -43,7 +43,6 @@ HSAKMT_STATUS HSAKMTAPI hsaKmtPcSamplingSupport(void)
HSAKMT_STATUS HSAKMTAPI hsaKmtPcSamplingQueryCapabilities(HSAuint32 NodeId, void *sample_info,
HSAuint32 sample_info_sz, HSAuint32 *size)
{
int ret;
struct kfd_ioctl_pc_sample_args args = {0};
uint32_t gpu_id;
@@ -53,38 +52,44 @@ HSAKMT_STATUS HSAKMTAPI hsaKmtPcSamplingQueryCapabilities(HSAuint32 NodeId, void
CHECK_KFD_OPEN();
CHECK_KFD_MINOR_VERSION(16);
ret = validate_nodeid(NodeId, &gpu_id);
HSAKMT_STATUS ret = validate_nodeid(NodeId, &gpu_id);
if (ret != HSAKMT_STATUS_SUCCESS) {
pr_err("[%s] invalid node ID: %d\n", __func__, NodeId);
return ret;
}
assert(sizeof(HsaPcSamplingInfo) == sizeof(struct kfd_pc_sample_info));
ret = HSAKMT_STATUS_SUCCESS;
args.op = KFD_IOCTL_PCS_OP_QUERY_CAPABILITIES;
args.gpu_id = gpu_id;
args.sample_info_ptr = (uint64_t)sample_info;
args.num_sample_info = sample_info_sz;
args.flags = 0;
ret = kmtIoctl(kfd_fd, AMDKFD_IOC_PC_SAMPLE, &args);
int err = kmtIoctl(kfd_fd, AMDKFD_IOC_PC_SAMPLE, &args);
if (ret) {
return (ret == -ENOSPC) ? HSAKMT_STATUS_BUFFER_TOO_SMALL :
(ret == -EINVAL) ? HSAKMT_STATUS_INVALID_PARAMETER :
(ret == -EOPNOTSUPP) ? HSAKMT_STATUS_NOT_SUPPORTED :
(ret == -EBUSY) ? HSAKMT_STATUS_UNAVAILABLE :
HSAKMT_STATUS_ERROR;
}
*size = args.num_sample_info;
if (err) {
switch (errno) {
case ENOSPC:
return HSAKMT_STATUS_BUFFER_TOO_SMALL;
case EINVAL:
return HSAKMT_STATUS_INVALID_PARAMETER;
case EOPNOTSUPP:
return HSAKMT_STATUS_NOT_SUPPORTED;
case EBUSY:
return HSAKMT_STATUS_UNAVAILABLE;
default:
return HSAKMT_STATUS_ERROR;
}
}
return HSAKMT_STATUS_SUCCESS;
}
HSAKMT_STATUS HSAKMTAPI hsaKmtPcSamplingCreate(HSAuint32 NodeId, HsaPcSamplingInfo *sample_info,
HsaPcSamplingTraceId *traceId)
{
int ret;
struct kfd_ioctl_pc_sample_args args = {0};
uint32_t gpu_id;
@@ -94,25 +99,30 @@ HSAKMT_STATUS HSAKMTAPI hsaKmtPcSamplingCreate(HSAuint32 NodeId, HsaPcSamplingIn
CHECK_KFD_OPEN();
*traceId = INVALID_TRACE_ID;
ret = validate_nodeid(NodeId, &gpu_id);
HSAKMT_STATUS ret = validate_nodeid(NodeId, &gpu_id);
if (ret != HSAKMT_STATUS_SUCCESS) {
pr_err("[%s] invalid node ID: %d\n", __func__, NodeId);
return ret;
}
ret = HSAKMT_STATUS_SUCCESS;
args.op = KFD_IOCTL_PCS_OP_CREATE;
args.gpu_id = gpu_id;
args.sample_info_ptr = (uint64_t)sample_info;
args.num_sample_info = 1;
args.trace_id = INVALID_TRACE_ID;
ret = kmtIoctl(kfd_fd, AMDKFD_IOC_PC_SAMPLE, &args);
if (ret) {
return (ret == -EINVAL) ? HSAKMT_STATUS_INVALID_PARAMETER :
(ret == -ENOMEM) ? HSAKMT_STATUS_NO_MEMORY :
(ret == -EBUSY) ? HSAKMT_STATUS_UNAVAILABLE :
HSAKMT_STATUS_ERROR;
int err = kmtIoctl(kfd_fd, AMDKFD_IOC_PC_SAMPLE, &args);
if (err) {
switch (errno) {
case EINVAL:
return HSAKMT_STATUS_INVALID_PARAMETER;
case ENOMEM:
return HSAKMT_STATUS_NO_MEMORY;
case EBUSY:
return HSAKMT_STATUS_UNAVAILABLE;
default:
return HSAKMT_STATUS_ERROR;
}
}
*traceId = args.trace_id;
@@ -121,7 +131,6 @@ HSAKMT_STATUS HSAKMTAPI hsaKmtPcSamplingCreate(HSAuint32 NodeId, HsaPcSamplingIn
HSAKMT_STATUS HSAKMTAPI hsaKmtPcSamplingDestroy(HSAuint32 NodeId, HsaPcSamplingTraceId traceId)
{
int ret;
struct kfd_ioctl_pc_sample_args args = {0};
uint32_t gpu_id;
@@ -130,7 +139,7 @@ HSAKMT_STATUS HSAKMTAPI hsaKmtPcSamplingDestroy(HSAuint32 NodeId, HsaPcSamplingT
CHECK_KFD_OPEN();
ret = validate_nodeid(NodeId, &gpu_id);
HSAKMT_STATUS ret = validate_nodeid(NodeId, &gpu_id);
if (ret != HSAKMT_STATUS_SUCCESS) {
pr_err("[%s] invalid node ID: %d\n", __func__, NodeId);
return ret;
@@ -138,20 +147,22 @@ HSAKMT_STATUS HSAKMTAPI hsaKmtPcSamplingDestroy(HSAuint32 NodeId, HsaPcSamplingT
hsaKmtPcSamplingStop(NodeId, traceId);
ret = HSAKMT_STATUS_SUCCESS;
args.op = KFD_IOCTL_PCS_OP_DESTROY;
args.gpu_id = gpu_id;
args.trace_id = traceId;
ret = kmtIoctl(kfd_fd, AMDKFD_IOC_PC_SAMPLE, &args);
return (ret == -EINVAL) ? HSAKMT_STATUS_INVALID_PARAMETER :
ret ? HSAKMT_STATUS_ERROR :
HSAKMT_STATUS_SUCCESS;
int err = kmtIoctl(kfd_fd, AMDKFD_IOC_PC_SAMPLE, &args);
if (err) {
if (errno == EINVAL)
return HSAKMT_STATUS_INVALID_PARAMETER;
return HSAKMT_STATUS_ERROR;
}
return HSAKMT_STATUS_SUCCESS;
}
HSAKMT_STATUS HSAKMTAPI hsaKmtPcSamplingStart(HSAuint32 NodeId, HsaPcSamplingTraceId traceId)
{
int ret;
struct kfd_ioctl_pc_sample_args args = {0};
uint32_t gpu_id;
@@ -160,29 +171,37 @@ HSAKMT_STATUS HSAKMTAPI hsaKmtPcSamplingStart(HSAuint32 NodeId, HsaPcSamplingTra
CHECK_KFD_OPEN();
ret = validate_nodeid(NodeId, &gpu_id);
HSAKMT_STATUS ret = validate_nodeid(NodeId, &gpu_id);
if (ret != HSAKMT_STATUS_SUCCESS) {
pr_err("[%s] invalid node ID: %d\n", __func__, NodeId);
return ret;
}
ret = HSAKMT_STATUS_SUCCESS;
args.op = KFD_IOCTL_PCS_OP_START;
args.gpu_id = gpu_id;
args.trace_id = traceId;
ret = kmtIoctl(kfd_fd, AMDKFD_IOC_PC_SAMPLE, &args);
return (ret == -EINVAL) ? HSAKMT_STATUS_INVALID_PARAMETER :
(ret == -ENOMEM) ? HSAKMT_STATUS_OUT_OF_RESOURCES :
(ret == -EBUSY) ? HSAKMT_STATUS_UNAVAILABLE :
(ret == -EALREADY) ? HSAKMT_STATUS_KERNEL_ALREADY_OPENED :
ret ? HSAKMT_STATUS_ERROR :
HSAKMT_STATUS_SUCCESS;
int err = kmtIoctl(kfd_fd, AMDKFD_IOC_PC_SAMPLE, &args);
if (err) {
switch (errno) {
case EINVAL:
return HSAKMT_STATUS_INVALID_PARAMETER;
case ENOMEM:
return HSAKMT_STATUS_OUT_OF_RESOURCES;
case EBUSY:
return HSAKMT_STATUS_UNAVAILABLE;
case EALREADY:
return HSAKMT_STATUS_KERNEL_ALREADY_OPENED;
default:
return HSAKMT_STATUS_ERROR;
}
}
return HSAKMT_STATUS_SUCCESS;
}
HSAKMT_STATUS HSAKMTAPI hsaKmtPcSamplingStop(HSAuint32 NodeId, HsaPcSamplingTraceId traceId)
{
int ret;
struct kfd_ioctl_pc_sample_args args = {0};
uint32_t gpu_id;
@@ -191,20 +210,26 @@ HSAKMT_STATUS HSAKMTAPI hsaKmtPcSamplingStop(HSAuint32 NodeId, HsaPcSamplingTrac
CHECK_KFD_OPEN();
ret = validate_nodeid(NodeId, &gpu_id);
HSAKMT_STATUS ret = validate_nodeid(NodeId, &gpu_id);
if (ret != HSAKMT_STATUS_SUCCESS) {
pr_err("[%s] invalid node ID: %d\n", __func__, NodeId);
return ret;
}
ret = HSAKMT_STATUS_SUCCESS;
args.op = KFD_IOCTL_PCS_OP_STOP;
args.gpu_id = gpu_id;
args.trace_id = traceId;
ret = kmtIoctl(kfd_fd, AMDKFD_IOC_PC_SAMPLE, &args);
return (ret == -EINVAL) ? HSAKMT_STATUS_INVALID_PARAMETER :
(ret == -EALREADY) ? HSAKMT_STATUS_KERNEL_ALREADY_OPENED :
ret ? HSAKMT_STATUS_ERROR :
HSAKMT_STATUS_SUCCESS;
int err = kmtIoctl(kfd_fd, AMDKFD_IOC_PC_SAMPLE, &args);
if (err) {
switch (errno) {
case EINVAL:
return HSAKMT_STATUS_INVALID_PARAMETER;
case EALREADY:
return HSAKMT_STATUS_KERNEL_ALREADY_OPENED;
default:
return HSAKMT_STATUS_ERROR;
}
}
return HSAKMT_STATUS_SUCCESS;
}