AMD-GFX Archive on lore.kernel.org
 help / color / mirror / Atom feed
From: Jonathan Kim <jonathan.kim@amd.com>
To: <amd-gfx@lists.freedesktop.org>
Cc: Felix.Kuehling@amd.com, Lancelot.Six@amd.com, Laurent.Morichetti@amd.com
Subject: [PATCH 24/29] drm/amdkfd: add debug set flags operation
Date: Mon, 29 Aug 2022 10:30:21 -0400	[thread overview]
Message-ID: <20220829143026.1509027-25-jonathan.kim@amd.com> (raw)
In-Reply-To: <20220829143026.1509027-1-jonathan.kim@amd.com>

Allow the debugger to set single memory and single ALU operations.

Some exceptions are imprecise (memory violations, address watch) in the
sense that a trap occurs only when the exception interrupt occurs and
not at the non-halting faulty instruction.  Trap temporaries 0 & 1 save
the program counter address, which means that these values will not point
to the faulty instruction address but to whenever the interrupt was
raised.

Setting the Single Memory Operations flag will inject an automatic wait
on every memory operation instruction forcing imprecise memory exceptions
to become precise at the cost of performance.  This setting is not
permitted on debug devices that support only a global setting of this
option.

Likewise, Single ALU Operations will force in-order ALU operations.
Although this is available on current hardware, it's not required so it
will be treated as a NOP.

Return the previous set flags to the debugger as well.

Signed-off-by: Jonathan Kim <jonathan.kim@amd.com>
---
 drivers/gpu/drm/amd/amdkfd/kfd_chardev.c |  2 ++
 drivers/gpu/drm/amd/amdkfd/kfd_debug.c   | 35 ++++++++++++++++++++++++
 drivers/gpu/drm/amd/amdkfd/kfd_debug.h   |  1 +
 3 files changed, 38 insertions(+)

diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_chardev.c b/drivers/gpu/drm/amd/amdkfd/kfd_chardev.c
index 7448c3e50620..bec52bbba330 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_chardev.c
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_chardev.c
@@ -2912,6 +2912,8 @@ static int kfd_ioctl_set_debug_trap(struct file *filep, struct kfd_process *p, v
 				args->clear_node_address_watch.id);
 		break;
 	case KFD_IOC_DBG_TRAP_SET_FLAGS:
+		r = kfd_dbg_trap_set_flags(target, &args->set_flags.flags);
+		break;
 	case KFD_IOC_DBG_TRAP_QUERY_DEBUG_EVENT:
 	case KFD_IOC_DBG_TRAP_QUERY_EXCEPTION_INFO:
 	case KFD_IOC_DBG_TRAP_GET_QUEUE_SNAPSHOT:
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_debug.c b/drivers/gpu/drm/amd/amdkfd/kfd_debug.c
index 8195f1d9da8e..0219032b9ce1 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_debug.c
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_debug.c
@@ -23,6 +23,7 @@
 #include "kfd_debug.h"
 #include "kfd_device_queue_manager.h"
 #include <linux/file.h>
+#include <uapi/linux/kfd_ioctl.h>
 
 /*
  * The spinlock protects the per device dev->alloc_watch_ids for multi-process access.
@@ -355,6 +356,37 @@ static void kfd_dbg_clear_process_address_watch(struct kfd_process *target)
 			kfd_dbg_trap_clear_dev_address_watch(target->pdds[i], j);
 }
 
+int kfd_dbg_trap_set_flags(struct kfd_process *target, uint32_t *flags)
+{
+	uint32_t prev_flags = target->dbg_flags;
+	int i, r = 0;
+
+	for (i = 0; i < target->n_pdds; i++) {
+		if (!kfd_dbg_is_per_vmid_supported(target->pdds[i]->dev) &&
+			(*flags & KFD_DBG_TRAP_FLAG_SINGLE_MEM_OP)) {
+			*flags = prev_flags;
+			return -EPERM;
+		}
+	}
+
+	target->dbg_flags = *flags;
+	*flags = prev_flags;
+	for (i = 0; i < target->n_pdds; i++) {
+		struct kfd_process_device *pdd = target->pdds[i];
+
+		if (!kfd_dbg_is_per_vmid_supported(pdd->dev))
+			continue;
+
+		r = debug_refresh_runlist(target->pdds[i]->dev->dqm);
+		if (r) {
+			target->dbg_flags = prev_flags;
+			break;
+		}
+	}
+
+	return r;
+}
+
 
 /* kfd_dbg_trap_deactivate:
  *	target: target process
@@ -369,9 +401,12 @@ void kfd_dbg_trap_deactivate(struct kfd_process *target, bool unwind, int unwind
 	int i, count = 0;
 
 	if (!unwind) {
+		uint32_t flags = 0;
 		cancel_work_sync(&target->debug_event_workarea);
 		kfd_dbg_clear_process_address_watch(target);
 		kfd_dbg_trap_set_wave_launch_mode(target, 0);
+
+		kfd_dbg_trap_set_flags(target, &flags);
 	}
 
 	for (i = 0; i < target->n_pdds; i++) {
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_debug.h b/drivers/gpu/drm/amd/amdkfd/kfd_debug.h
index ad677e67e7eb..12b80b6c96d0 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_debug.h
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_debug.h
@@ -57,6 +57,7 @@ int kfd_dbg_trap_set_dev_address_watch(struct kfd_process_device *pdd,
 					uint32_t watch_address_mask,
 					uint32_t *watch_id,
 					uint32_t watch_mode);
+int kfd_dbg_trap_set_flags(struct kfd_process *target, uint32_t *flags);
 int kfd_dbg_send_exception_to_runtime(struct kfd_process *p,
 					unsigned int dev_id,
 					unsigned int queue_id,
-- 
2.25.1


  parent reply	other threads:[~2022-08-29 14:31 UTC|newest]

Thread overview: 34+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2022-08-29 14:29 [PATCH 0/29] Introduce AMD GPU ISA Debugging for HSA Compute Jonathan Kim
2022-08-29 14:29 ` [PATCH 01/29] drm/amdkfd: add debug and runtime enable interface Jonathan Kim
2022-08-30 22:46   ` Felix Kuehling
2022-08-29 14:29 ` [PATCH 02/29] drm/amdkfd: display debug capabilities Jonathan Kim
2022-08-29 14:30 ` [PATCH 03/29] drm/amdkfd: prepare per-process debug enable and disable Jonathan Kim
2022-08-29 14:30 ` [PATCH 04/29] drm/amdgpu: add kgd hw debug mode setting interface Jonathan Kim
2022-08-29 14:30 ` [PATCH 05/29] drm/amdgpu: setup hw debug registers on driver initialization Jonathan Kim
2022-08-29 14:30 ` [PATCH 06/29] drm/amdgpu: add gfx9 hw debug mode enable and disable calls Jonathan Kim
2022-08-29 14:30 ` [PATCH 07/29] drm/amdgpu: add gfx9.4.1 " Jonathan Kim
2022-08-29 14:30 ` [PATCH 08/29] drm/amdgpu: add gfx10 " Jonathan Kim
2022-08-29 14:30 ` [PATCH 09/29] drm/amdgpu: add gfx9.4.2 " Jonathan Kim
2022-08-29 14:30 ` [PATCH 10/29] drm/amdgpu: add configurable grace period for unmap queues Jonathan Kim
2022-08-29 14:30 ` [PATCH 11/29] drm/amdkfd: prepare map process for single process debug devices Jonathan Kim
2022-08-29 14:30 ` [PATCH 12/29] drm/amdgpu: prepare map process for multi-process " Jonathan Kim
2022-08-29 14:30 ` [PATCH 13/29] drm/amdkfd: add per process hw trap enable and disable functions Jonathan Kim
2022-08-29 14:30 ` [PATCH 14/29] drm/amdkfd: add raise exception event function Jonathan Kim
2022-08-29 14:30 ` [PATCH 15/29] drm/amdkfd: add send exception operation Jonathan Kim
2022-08-29 14:30 ` [PATCH 16/29] drm/amdkfd: add runtime enable operation Jonathan Kim
2022-08-29 14:30 ` [PATCH 17/29] drm/amdkfd: Add debug trap enabled flag to TMA Jonathan Kim
2022-08-29 14:30 ` [PATCH 18/29] drm/amdkfd: update process interrupt handling for debug events Jonathan Kim
2022-08-30 21:56   ` Felix Kuehling
2022-08-29 14:30 ` [PATCH 19/29] drm/amdkfd: add debug set exceptions enabled operation Jonathan Kim
2022-08-29 14:30 ` [PATCH 20/29] drm/amdkfd: add debug wave launch override operation Jonathan Kim
2022-08-29 14:30 ` [PATCH 21/29] drm/amdkfd: add debug wave launch mode operation Jonathan Kim
2022-08-29 14:30 ` [PATCH 22/29] drm/amdkfd: add debug suspend and resume process queues operation Jonathan Kim
2022-08-29 14:30 ` [PATCH 23/29] drm/amdkfd: add debug set and clear address watch points operation Jonathan Kim
2022-08-29 14:30 ` Jonathan Kim [this message]
2022-08-29 14:30 ` [PATCH 25/29] drm/amdkfd: add debug query event operation Jonathan Kim
2022-08-29 14:30 ` [PATCH 26/29] drm/amdkfd: add debug query exception info operation Jonathan Kim
2022-08-29 14:30 ` [PATCH 27/29] drm/amdkfd: add debug queue snapshot operation Jonathan Kim
2022-08-29 14:30 ` [PATCH 28/29] drm/amdkfd: add debug device " Jonathan Kim
2022-08-29 14:30 ` [PATCH 29/29] drm/amdkfd: bump kfd ioctl minor version for debug api availability Jonathan Kim
  -- strict thread matches above, loose matches on Subject: below --
2022-10-31 16:23 [PATCH 01/29] drm/amdkfd: add debug and runtime enable interface Jonathan Kim
2022-10-31 16:23 ` [PATCH 24/29] drm/amdkfd: add debug set flags operation Jonathan Kim
2022-11-30  0:39   ` Felix Kuehling

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20220829143026.1509027-25-jonathan.kim@amd.com \
    --to=jonathan.kim@amd.com \
    --cc=Felix.Kuehling@amd.com \
    --cc=Lancelot.Six@amd.com \
    --cc=Laurent.Morichetti@amd.com \
    --cc=amd-gfx@lists.freedesktop.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox