From: "Teres Alexis, Alan Previn" <alan.previn.teres.alexis@intel.com>
To: "Dong, Zhanjun" <zhanjun.dong@intel.com>,
"intel-xe@lists.freedesktop.org" <intel-xe@lists.freedesktop.org>
Cc: "dri-devel@lists.freedesktop.org"
<dri-devel@lists.freedesktop.org>,
"Harrison, John C" <john.c.harrison@intel.com>,
"Brost, Matthew" <matthew.brost@intel.com>,
"Ceraolo Spurio, Daniele" <daniele.ceraolospurio@intel.com>,
"Vivi, Rodrigo" <rodrigo.vivi@intel.com>
Subject: Re: [PATCH v7 3/6] drm/xe/guc: Split engine state print between xe_hw_engine vs xe_guc_capture
Date: Wed, 12 Feb 2025 18:50:10 +0000 [thread overview]
Message-ID: <225b51b753b0dc44dfdbdd522ca9c13c90643c1c.camel@intel.com> (raw)
In-Reply-To: <0004aa7f-d6a6-4122-93b7-bae80f60009e@intel.com>
On Tue, 2025-02-11 at 17:48 -0500, Dong, Zhanjun wrote:
>
>
> On 2025-02-10 6:32 p.m., Alan Previn wrote:
> > Relocate the xe_engine_snapshot_print function from xe_guc_capture.c
> > into xe_hw_engine.c but split out the GuC-Err-Capture register printing
> > portion out into a separate helper inside xe_guc_capture.c so that
> > we can have a clear separation between printing the general engine info
> > vs GuC-Err-Capture node's register list.
> >
> > v7: - Fix function name to respect "xe_hw_engine" name space. (Rodrigo)
> > - Remove additional newline in engine dump (Jose Souza) +
> > ensure changes didn't break mesa's aubinator tool (Rodrigo)
> >
> > Signed-off-by: Alan Previn <alan.previn.teres.alexis@intel.com>
> > ---
> > drivers/gpu/drm/xe/xe_devcoredump.c | 2 +-
> > drivers/gpu/drm/xe/xe_guc_capture.c | 79 +++++++++++++----------------
> > drivers/gpu/drm/xe/xe_guc_capture.h | 4 +-
> > drivers/gpu/drm/xe/xe_hw_engine.c | 29 ++++++++++-
> > drivers/gpu/drm/xe/xe_hw_engine.h | 1 +
> > 5 files changed, 67 insertions(+), 48 deletions(-)
> >
> > diff --git a/drivers/gpu/drm/xe/xe_devcoredump.c b/drivers/gpu/drm/xe/xe_devcoredump.c
> > index 006041997550..7a4610d2ea4f 100644
> > --- a/drivers/gpu/drm/xe/xe_devcoredump.c
> > +++ b/drivers/gpu/drm/xe/xe_devcoredump.c
> > @@ -128,7 +128,7 @@ static ssize_t __xe_devcoredump_read(char *buffer, size_t count,
> > drm_puts(&p, "\n**** HW Engines ****\n");
> > for (i = 0; i < XE_NUM_HW_ENGINES; i++)
> > if (ss->hwe[i])
> > - xe_engine_snapshot_print(ss->hwe[i], &p);
> > + xe_hw_engine_snapshot_print(ss->hwe[i], &p);
> >
> > drm_puts(&p, "\n**** VM state ****\n");
> > xe_vm_snapshot_print(ss->vm, &p);
> > diff --git a/drivers/gpu/drm/xe/xe_guc_capture.c b/drivers/gpu/drm/xe/xe_guc_capture.c
> > index f118e8dd0ecb..76c20ff97864 100644
> > --- a/drivers/gpu/drm/xe/xe_guc_capture.c
> > +++ b/drivers/gpu/drm/xe/xe_guc_capture.c
> > @@ -917,9 +917,10 @@ guc_capture_init_node(struct xe_guc *guc, struct xe_guc_capture_snapshot *node)
> > * --------------------
> > * --> xe_devcoredump_read->
> > * L--> xxx_snapshot_print
> > - * L--> xe_engine_snapshot_print
> > - * Print register lists values saved at
> > - * guc->capture->outlist
> > + * L--> xe_hw_engine_print --> xe_hw_engine_snapshot_print
> > + * L--> xe_guc_capture_snapshot_print
> > + * Print register lists values saved in matching
> > + * node from guc->capture->outlist
> > *
> > */
> >
> > @@ -1655,22 +1656,16 @@ guc_capture_find_reg(struct gcap_reg_list_info *reginfo, u32 addr, u32 flags)
> > }
> >
> > static void
> > -snapshot_print_by_list_order(struct xe_hw_engine_snapshot *snapshot, struct drm_printer *p,
> > - u32 type, const struct __guc_mmio_reg_descr_group *list)
> > +print_noderegs_by_list_order(struct xe_guc *guc, struct gcap_reg_list_info *reginfo,
> > + const struct __guc_mmio_reg_descr_group *list, struct drm_printer *p)
> > {
> > - struct xe_gt *gt = snapshot->hwe->gt;
> > - struct xe_guc *guc = >->uc.guc;
> > - struct gcap_reg_list_info *reginfo = NULL;
> > - u32 i, last_value = 0;
> > + u32 last_value, i;
> > bool is_ext, low32_ready = false;
> >
> > if (!list || !list->list || list->num_regs == 0)
> > return;
> >
> > - XE_WARN_ON(!snapshot->matched_node);
> > -
> > is_ext = list == guc->capture->extlists;
> > - reginfo = &snapshot->matched_node->reginfo[type];
> >
> > /*
> > * loop through descriptor first and find the register in the node
> > @@ -1740,8 +1735,8 @@ snapshot_print_by_list_order(struct xe_hw_engine_snapshot *snapshot, struct drm_
> >
> > group = FIELD_GET(GUC_REGSET_STEERING_GROUP, reg_desc->flags);
> > instance = FIELD_GET(GUC_REGSET_STEERING_INSTANCE, reg_desc->flags);
> > - dss = xe_gt_mcr_steering_info_to_dss_id(gt, group, instance);
> > -
> > + dss = xe_gt_mcr_steering_info_to_dss_id(guc_to_gt(guc), group,
> > + instance);
> > drm_printf(p, "\t%s[%u]: 0x%08x\n", reg_desc->regname, dss, value);
> > } else {
> > drm_printf(p, "\t%s: 0x%08x\n", reg_desc->regname, value);
> > @@ -1760,13 +1755,18 @@ snapshot_print_by_list_order(struct xe_hw_engine_snapshot *snapshot, struct drm_
> > }
> >
> > /**
> > - * xe_engine_snapshot_print - Print out a given Xe HW Engine snapshot.
> > - * @snapshot: Xe HW Engine snapshot object.
> > + * xe_guc_capture_snapshot_print - Print out a the contents of a provided Guc-Err-Capture node
> > + * @guc : Target GuC for operation.
> > + * @node: GuC Error Capture register dump node.
> > * @p: drm_printer where it will be printed out.
> > *
> > - * This function prints out a given Xe HW Engine snapshot object.
> > + * This function prints out a register dump of a GuC-Err-Capture node that was retrieved
> > + * earlier either by GuC-FW reporting or by manual capture depending on how the
> > + * caller (typically xe_hw_engine_snapshot) was invoked and used.
> > */
> > -void xe_engine_snapshot_print(struct xe_hw_engine_snapshot *snapshot, struct drm_printer *p)
> > +
> > +void xe_guc_capture_snapshot_print(struct xe_guc *guc, struct xe_guc_capture_snapshot *node,
> > + struct drm_printer *p)
> > {
> > const char *grptype[GUC_STATE_CAPTURE_GROUP_TYPE_MAX] = {
> > "full-capture",
> > @@ -1774,45 +1774,36 @@ void xe_engine_snapshot_print(struct xe_hw_engine_snapshot *snapshot, struct drm
> > };
> > int type;
> > const struct __guc_mmio_reg_descr_group *list;
> > - enum guc_capture_list_class_type capture_class;
> > struct xe_gt *gt;
> >
> > - if (!snapshot)
> > + if (!guc)
> > return;
> > -
> > - gt = snapshot->hwe->gt;
> > -
> > - if (!snapshot->matched_node)
> > + gt = guc_to_gt(guc);
> > + if (!node) {
> node was called snapshot befrore.
alan: yes i did - previously snapshot was one-level up of xe_hw_engine that included other things
alongside the matching node from guc-err-capture but with this patch i wanted to streamline this
this helper since it only needs access to the node and nothing else.
> > + xe_gt_warn(gt, "GuC Capture printing without node!\n");
> > return;
> > + }
> > + if (!p) {
> New printer pointer check, good.
> > + xe_gt_warn(gt, "GuC Capture printing without printer!\n");
> > + return;
> > + }
> >
> > - xe_gt_assert(gt, snapshot->hwe);
> > -
> > - capture_class = xe_engine_class_to_guc_capture_class(snapshot->hwe->class);
> > -
> > - drm_printf(p, "%s (physical), logical instance=%d\n",
> > - snapshot->name ? snapshot->name : "",
> > - snapshot->logical_instance);
> > drm_printf(p, "\tCapture_source: %s\n",
> > - snapshot->matched_node->source == XE_ENGINE_CAPTURE_SOURCE_GUC ?
> > + node->source == XE_ENGINE_CAPTURE_SOURCE_GUC ?
> > "GuC" : "Manual");
> > - drm_printf(p, "\tCoverage: %s\n", grptype[snapshot->matched_node->is_partial]);
> > - drm_printf(p, "\tForcewake: domain 0x%x, ref %d\n",
> > - snapshot->forcewake.domain, snapshot->forcewake.ref);
> > - drm_printf(p, "\tReserved: %s\n",
> > - str_yes_no(snapshot->kernel_reserved));
> > + drm_printf(p, "\tCoverage: %s\n", grptype[node->is_partial]);
> Yes, I see the printout order was changed:
> vcs0 (physical), logical instance=0
> Capture_source: GuC
> Coverage: full-capture
> Forcewake: domain 0x8, ref 1
> Reserved: no
> FORCEWAKE_GT: 0x00000000
> to:
> vcs0 (physical), logical instance=0
> Forcewake: domain 0x8, ref 1
> Reserved: no
> Capture_source: GuC
> Coverage: full-capture
> FORCEWAKE_GT: 0x00000000
> The xe_exec_capture igt test can handle this change, as long as it not
> cause other tools stop working, I'm fine.
>
> Reviewed-by: Zhanjun Dong <zhanjun.dong@intel.com>
alan: Thanks for the RB and yes - we tested it with mesa tool also.
alan:snip
next prev parent reply other threads:[~2025-02-12 18:50 UTC|newest]
Thread overview: 23+ messages / expand[flat|nested] mbox.gz Atom feed top
2025-02-10 23:32 [PATCH v7 0/6] Maintenence of devcoredump <-> GuC-Err-Capture plumbing Alan Previn
2025-02-10 23:32 ` [PATCH v7 1/6] drm/xe/guc: Rename __guc_capture_parsed_output Alan Previn
2025-02-10 23:32 ` [PATCH v7 2/6] drm/xe/guc: Don't store capture nodes in xe_devcoredump_snapshot Alan Previn
2025-02-10 23:32 ` [PATCH v7 3/6] drm/xe/guc: Split engine state print between xe_hw_engine vs xe_guc_capture Alan Previn
2025-02-11 22:48 ` Dong, Zhanjun
2025-02-12 18:50 ` Teres Alexis, Alan Previn [this message]
2025-02-10 23:32 ` [PATCH v7 4/6] drm/xe/guc: Move xe_hw_engine_snapshot creation back to xe_hw_engine.c Alan Previn
2025-02-12 17:19 ` Dong, Zhanjun
2025-02-10 23:32 ` [PATCH v7 5/6] drm/xe/xe_hw_engine: Update xe_hw_engine capture for debugfs/gt_reset Alan Previn
2025-02-11 17:34 ` Teres Alexis, Alan Previn
2025-02-12 18:24 ` Dong, Zhanjun
2025-02-10 23:32 ` [PATCH v7 6/6] drm/xe/guc: Update comments on GuC-Err-Capture flows Alan Previn
2025-02-11 23:09 ` Dong, Zhanjun
2025-02-12 18:54 ` Teres Alexis, Alan Previn
2025-02-11 0:38 ` ✓ CI.Patch_applied: success for Maintenence of devcoredump <-> GuC-Err-Capture plumbing Patchwork
2025-02-11 0:38 ` ✗ CI.checkpatch: warning " Patchwork
2025-02-11 0:39 ` ✓ CI.KUnit: success " Patchwork
2025-02-11 0:56 ` ✓ CI.Build: " Patchwork
2025-02-11 0:57 ` ✗ CI.Hooks: failure " Patchwork
2025-02-11 0:58 ` ✗ CI.checksparse: warning " Patchwork
2025-02-11 1:18 ` ✓ Xe.CI.BAT: success " Patchwork
2025-02-11 13:01 ` ✗ Xe.CI.Full: failure " Patchwork
2025-02-13 0:56 ` Teres Alexis, Alan Previn
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=225b51b753b0dc44dfdbdd522ca9c13c90643c1c.camel@intel.com \
--to=alan.previn.teres.alexis@intel.com \
--cc=daniele.ceraolospurio@intel.com \
--cc=dri-devel@lists.freedesktop.org \
--cc=intel-xe@lists.freedesktop.org \
--cc=john.c.harrison@intel.com \
--cc=matthew.brost@intel.com \
--cc=rodrigo.vivi@intel.com \
--cc=zhanjun.dong@intel.com \
/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