From: "Summers, Stuart" <stuart.summers@intel.com>
To: "Sousa, Gustavo" <gustavo.sousa@intel.com>
Cc: "intel-xe@lists.freedesktop.org" <intel-xe@lists.freedesktop.org>,
"Lin, Shuicheng" <shuicheng.lin@intel.com>,
"Vivi, Rodrigo" <rodrigo.vivi@intel.com>,
"Roper, Matthew D" <matthew.d.roper@intel.com>,
"Brost, Matthew" <matthew.brost@intel.com>,
"Wajdeczko, Michal" <michal.wajdeczko@intel.com>,
"Nerlige Ramappa, Umesh" <umesh.nerlige.ramappa@intel.com>,
"Ceraolo Spurio, Daniele" <daniele.ceraolospurio@intel.com>
Subject: Re: [PATCH 2/9] drm/xe: Sort xe_config_device fields and defaults alphabetically
Date: Mon, 4 May 2026 14:38:54 +0000 [thread overview]
Message-ID: <9869801345a0e54643dd22836f1069c71cec6aac.camel@intel.com> (raw)
In-Reply-To: <878q9ze0nz.fsf@intel.com>
On Mon, 2026-05-04 at 10:58 -0300, Gustavo Sousa wrote:
> Stuart Summers <stuart.summers@intel.com> writes:
>
> > Sort the fields in struct xe_config_device and the corresponding
> > device_defaults initializer in alphabetical order. Non-SRIOV fields
> > come first (alphabetically), followed by the SRIOV sub-struct with
> > its own fields in alphabetical order. Also sort the PRI_CUSTOM_ATTR
> > calls in dump_custom_dev_config() to match and the documentation
> > for
> > each of the fields.
>
> I think the sorting alphabetically seems fine, but I'll defer it to
> others to weigh in here on this decision though.
>
> I guess we will not have many instances of that struct, so I guess
> potential padding resulting from the sorting requirement should be
> fine.
>
> We probably want to add a note somewhere in the code asking to keep
> that
> stuff ordered alphabetically.
>
> By the way, for other people looking at this, I find it easier to
> review
> this with: git show --color-moved --histogram.
>
> Code-wise, looks good to me, so if this is the direction we take,
>
> Reviewed-by: Gustavo Sousa <gustavo.sousa@intel.com>
I had accidentally responded to the wrong patch with this... Also
thanks for the review!
Yeah this is something I wanted to discuss here too. I kind of just
blindly alphabetized things here, but we might want to sort by data
type or something else. also, we only have a few parameters here, but
I'd like to set this up to support a much larger set of parameters on
the debug side - we need much finer grained support for things like
power-on for instance. And I'd like a way that easily scales here.
Also I agree on the documentation.. I'll make sure to add that in the
next patch once we conclude on the approach here.
Thanks,
Stuart
>
> >
> > Signed-off-by: Stuart Summers <stuart.summers@intel.com>
> > Assisted-by: Copilot:claude-opus-4.7
> > ---
> > drivers/gpu/drm/xe/xe_configfs.c | 192 +++++++++++++++------------
> > ----
> > 1 file changed, 96 insertions(+), 96 deletions(-)
> >
> > diff --git a/drivers/gpu/drm/xe/xe_configfs.c
> > b/drivers/gpu/drm/xe/xe_configfs.c
> > index 69abc69ec0f3..1e134057fae8 100644
> > --- a/drivers/gpu/drm/xe/xe_configfs.c
> > +++ b/drivers/gpu/drm/xe/xe_configfs.c
> > @@ -73,53 +73,79 @@
> > * Configure Attributes
> > * ====================
> > *
> > - * Survivability mode:
> > - * -------------------
> > + * Context restore BB
> > + * ------------------
> > *
> > - * Enable survivability mode on supported cards. This setting only
> > takes
> > - * effect when probing the device. Example to enable it::
> > + * Allow to execute a batch buffer during any context switches.
> > When the
> > + * GPU is restoring the context, it executes additional commands.
> > It's useful
> > + * for testing additional workarounds and validating certain HW
> > behaviors: it's
> > + * not intended for normal execution and will taint the kernel
> > with TAINT_TEST
> > + * when used.
> > *
> > - * # echo 1 >
> > /sys/kernel/config/xe/0000:03:00.0/enable_survivability_mode
> > + * The syntax allows to pass straight instructions to be executed
> > by the engine
> > + * in a batch buffer or set specific registers.
> > *
> > - * This attribute can only be set before binding to the device.
> > + * #. Generic instruction::
> > *
> > - * Allowed GT types:
> > - * -----------------
> > + * <engine-class> cmd <instr> [[dword0] [dword1] [...]]
> > *
> > - * Allow only specific types of GTs to be detected and initialized
> > by the
> > - * driver. Any combination of GT types can be enabled/disabled,
> > although
> > - * some settings will cause the device to fail to probe.
> > + * #. Simple register setting::
> > *
> > - * Writes support both comma- and newline-separated input format.
> > Reads
> > - * will always return one GT type per line. "primary" and "media"
> > are the
> > - * GT type names supported by this interface.
> > + * <engine-class> reg <address> <value>
> > *
> > - * This attribute can only be set before binding to the device.
> > + * Commands are saved per engine class: all instances of that
> > class will execute
> > + * those commands during context switch. The instruction, dword
> > arguments,
> > + * addresses and values are in hex format like in the examples
> > below.
> > *
> > - * Examples:
> > + * #. Execute a LRI command to write 0xDEADBEEF to register 0x4f10
> > after the
> > + * normal context restore::
> > *
> > - * Allow both primary and media GTs to be initialized and used.
> > This matches
> > - * the driver's default behavior::
> > + * # echo 'rcs cmd 11000001 4F100 DEADBEEF' \
> > + * >
> > /sys/kernel/config/xe/0000:03:00.0/ctx_restore_post_bb
> > *
> > - * # echo 'primary,media' >
> > /sys/kernel/config/xe/0000:03:00.0/gt_types_allowed
> > + * #. Execute a LRI command to write 0xDEADBEEF to register 0x4f10
> > at the
> > + * beginning of the context restore::
> > *
> > - * Allow only the primary GT of each tile to be initialized and
> > used,
> > - * effectively disabling the media GT if it exists on the
> > platform::
> > + * # echo 'rcs cmd 11000001 4F100 DEADBEEF' \
> > + * >
> > /sys/kernel/config/xe/0000:03:00.0/ctx_restore_mid_bb
> > +
> > + * #. Load certain values in a couple of registers (it can be used
> > as a simpler
> > + * alternative to the `cmd`) action::
> > *
> > - * # echo 'primary' >
> > /sys/kernel/config/xe/0000:03:00.0/gt_types_allowed
> > + * # cat >
> > /sys/kernel/config/xe/0000:03:00.0/ctx_restore_post_bb <<EOF
> > + * rcs reg 4F100 DEADBEEF
> > + * rcs reg 4F104 FFFFFFFF
> > + * EOF
> > *
> > - * Allow only the media GT of each tile to be initialized and
> > used,
> > - * effectively disabling the primary GT. **This configuration
> > will cause
> > - * device probe failure on all current platforms, but may be
> > allowed on
> > - * igpu platforms in the future**::
> > + * .. note::
> > *
> > - * # echo 'media' >
> > /sys/kernel/config/xe/0000:03:00.0/gt_types_allowed
> > + * When using multiple lines, make sure to use a command
> > that is
> > + * implemented with a single write syscall, like HEREDOC.
> > *
> > - * Disable all GTs. Only other GPU IP (such as display) is
> > potentially usable.
> > - * **This configuration will cause device probe failure on all
> > current
> > - * platforms, but may be allowed on igpu platforms in the
> > future**::
> > + * Currently this is implemented only for post and mid context
> > restore and
> > + * these attributes can only be set before binding to the device.
> > *
> > - * # echo '' >
> > /sys/kernel/config/xe/0000:03:00.0/gt_types_allowed
> > + * PSMI
> > + * ----
> > + *
> > + * Enable extra debugging capabilities to trace engine execution.
> > Only useful
> > + * during early platform enabling and requires additional hardware
> > connected.
> > + * Once it's enabled, additionals WAs are added and runtime
> > configuration is
> > + * done via debugfs. Example to enable it::
> > + *
> > + * # echo 1 > /sys/kernel/config/xe/0000:03:00.0/enable_psmi
> > + *
> > + * This attribute can only be set before binding to the device.
> > + *
> > + * Survivability mode:
> > + * -------------------
> > + *
> > + * Enable survivability mode on supported cards. This setting only
> > takes
> > + * effect when probing the device. Example to enable it::
> > + *
> > + * # echo 1 >
> > /sys/kernel/config/xe/0000:03:00.0/enable_survivability_mode
> > + *
> > + * This attribute can only be set before binding to the device.
> > *
> > * Allowed engines:
> > * ----------------
> > @@ -147,69 +173,43 @@
> > *
> > * This attribute can only be set before binding to the device.
> > *
> > - * PSMI
> > - * ----
> > + * Allowed GT types:
> > + * -----------------
> > *
> > - * Enable extra debugging capabilities to trace engine execution.
> > Only useful
> > - * during early platform enabling and requires additional hardware
> > connected.
> > - * Once it's enabled, additionals WAs are added and runtime
> > configuration is
> > - * done via debugfs. Example to enable it::
> > + * Allow only specific types of GTs to be detected and initialized
> > by the
> > + * driver. Any combination of GT types can be enabled/disabled,
> > although
> > + * some settings will cause the device to fail to probe.
> > *
> > - * # echo 1 > /sys/kernel/config/xe/0000:03:00.0/enable_psmi
> > + * Writes support both comma- and newline-separated input format.
> > Reads
> > + * will always return one GT type per line. "primary" and "media"
> > are the
> > + * GT type names supported by this interface.
> > *
> > * This attribute can only be set before binding to the device.
> > *
> > - * Context restore BB
> > - * ------------------
> > - *
> > - * Allow to execute a batch buffer during any context switches.
> > When the
> > - * GPU is restoring the context, it executes additional commands.
> > It's useful
> > - * for testing additional workarounds and validating certain HW
> > behaviors: it's
> > - * not intended for normal execution and will taint the kernel
> > with TAINT_TEST
> > - * when used.
> > - *
> > - * The syntax allows to pass straight instructions to be executed
> > by the engine
> > - * in a batch buffer or set specific registers.
> > - *
> > - * #. Generic instruction::
> > - *
> > - * <engine-class> cmd <instr> [[dword0] [dword1] [...]]
> > - *
> > - * #. Simple register setting::
> > - *
> > - * <engine-class> reg <address> <value>
> > - *
> > - * Commands are saved per engine class: all instances of that
> > class will execute
> > - * those commands during context switch. The instruction, dword
> > arguments,
> > - * addresses and values are in hex format like in the examples
> > below.
> > + * Examples:
> > *
> > - * #. Execute a LRI command to write 0xDEADBEEF to register 0x4f10
> > after the
> > - * normal context restore::
> > + * Allow both primary and media GTs to be initialized and used.
> > This matches
> > + * the driver's default behavior::
> > *
> > - * # echo 'rcs cmd 11000001 4F100 DEADBEEF' \
> > - * >
> > /sys/kernel/config/xe/0000:03:00.0/ctx_restore_post_bb
> > + * # echo 'primary,media' >
> > /sys/kernel/config/xe/0000:03:00.0/gt_types_allowed
> > *
> > - * #. Execute a LRI command to write 0xDEADBEEF to register 0x4f10
> > at the
> > - * beginning of the context restore::
> > + * Allow only the primary GT of each tile to be initialized and
> > used,
> > + * effectively disabling the media GT if it exists on the
> > platform::
> > *
> > - * # echo 'rcs cmd 11000001 4F100 DEADBEEF' \
> > - * >
> > /sys/kernel/config/xe/0000:03:00.0/ctx_restore_mid_bb
> > -
> > - * #. Load certain values in a couple of registers (it can be used
> > as a simpler
> > - * alternative to the `cmd`) action::
> > + * # echo 'primary' >
> > /sys/kernel/config/xe/0000:03:00.0/gt_types_allowed
> > *
> > - * # cat >
> > /sys/kernel/config/xe/0000:03:00.0/ctx_restore_post_bb <<EOF
> > - * rcs reg 4F100 DEADBEEF
> > - * rcs reg 4F104 FFFFFFFF
> > - * EOF
> > + * Allow only the media GT of each tile to be initialized and
> > used,
> > + * effectively disabling the primary GT. **This configuration
> > will cause
> > + * device probe failure on all current platforms, but may be
> > allowed on
> > + * igpu platforms in the future**::
> > *
> > - * .. note::
> > + * # echo 'media' >
> > /sys/kernel/config/xe/0000:03:00.0/gt_types_allowed
> > *
> > - * When using multiple lines, make sure to use a command
> > that is
> > - * implemented with a single write syscall, like HEREDOC.
> > + * Disable all GTs. Only other GPU IP (such as display) is
> > potentially usable.
> > + * **This configuration will cause device probe failure on all
> > current
> > + * platforms, but may be allowed on igpu platforms in the
> > future**::
> > *
> > - * Currently this is implemented only for post and mid context
> > restore and
> > - * these attributes can only be set before binding to the device.
> > + * # echo '' >
> > /sys/kernel/config/xe/0000:03:00.0/gt_types_allowed
> > *
> > * Max SR-IOV Virtual Functions
> > * ----------------------------
> > @@ -256,15 +256,15 @@ struct xe_config_group_device {
> > struct config_group sriov;
> >
> > struct xe_config_device {
> > - u64 gt_types_allowed;
> > - u64 engines_allowed;
> > - struct wa_bb
> > ctx_restore_post_bb[XE_ENGINE_CLASS_MAX];
> > struct wa_bb
> > ctx_restore_mid_bb[XE_ENGINE_CLASS_MAX];
> > - bool enable_survivability_mode;
> > + struct wa_bb
> > ctx_restore_post_bb[XE_ENGINE_CLASS_MAX];
> > bool enable_psmi;
> > + bool enable_survivability_mode;
> > + u64 engines_allowed;
> > + u64 gt_types_allowed;
> > struct {
> > - unsigned int max_vfs;
> > bool admin_only_pf;
> > + unsigned int max_vfs;
> > } sriov;
> > } config;
> >
> > @@ -277,13 +277,13 @@ struct xe_config_group_device {
> > };
> >
> > static const struct xe_config_device device_defaults = {
> > - .gt_types_allowed = U64_MAX,
> > - .engines_allowed = U64_MAX,
> > - .enable_survivability_mode = false,
> > .enable_psmi = false,
> > + .enable_survivability_mode = false,
> > + .engines_allowed = U64_MAX,
> > + .gt_types_allowed = U64_MAX,
> > .sriov = {
> > - .max_vfs = XE_DEFAULT_MAX_VFS,
> > .admin_only_pf = XE_DEFAULT_ADMIN_ONLY_PF,
> > + .max_vfs = XE_DEFAULT_MAX_VFS,
> > },
> > };
> >
> > @@ -814,17 +814,17 @@ static ssize_t
> > ctx_restore_post_bb_store(struct config_item *item,
> > CONFIGFS_ATTR(, ctx_restore_mid_bb);
> > CONFIGFS_ATTR(, ctx_restore_post_bb);
> > CONFIGFS_ATTR(, enable_psmi);
> > +CONFIGFS_ATTR(, enable_survivability_mode);
> > CONFIGFS_ATTR(, engines_allowed);
> > CONFIGFS_ATTR(, gt_types_allowed);
> > -CONFIGFS_ATTR(, enable_survivability_mode);
> >
> > static struct configfs_attribute *xe_config_device_attrs[] = {
> > &attr_ctx_restore_mid_bb,
> > &attr_ctx_restore_post_bb,
> > &attr_enable_psmi,
> > + &attr_enable_survivability_mode,
> > &attr_engines_allowed,
> > &attr_gt_types_allowed,
> > - &attr_enable_survivability_mode,
> > NULL,
> > };
> >
> > @@ -929,12 +929,12 @@ static ssize_t
> > sriov_admin_only_pf_store(struct config_item *item, const char *p
> > return len;
> > }
> >
> > -CONFIGFS_ATTR(sriov_, max_vfs);
> > CONFIGFS_ATTR(sriov_, admin_only_pf);
> > +CONFIGFS_ATTR(sriov_, max_vfs);
> >
> > static struct configfs_attribute *xe_config_sriov_attrs[] = {
> > - &sriov_attr_max_vfs,
> > &sriov_attr_admin_only_pf,
> > + &sriov_attr_max_vfs,
> > NULL,
> > };
> >
> > @@ -1096,10 +1096,10 @@ static void dump_custom_dev_config(struct
> > pci_dev *pdev,
> > dev->config.attr_); \
> > } while (0)
> >
> > - PRI_CUSTOM_ATTR("%llx", gt_types_allowed);
> > - PRI_CUSTOM_ATTR("%llx", engines_allowed);
> > PRI_CUSTOM_ATTR("%d", enable_psmi);
> > PRI_CUSTOM_ATTR("%d", enable_survivability_mode);
> > + PRI_CUSTOM_ATTR("%llx", engines_allowed);
> > + PRI_CUSTOM_ATTR("%llx", gt_types_allowed);
> > PRI_CUSTOM_ATTR("%u", sriov.admin_only_pf);
> >
> > #undef PRI_CUSTOM_ATTR
> > --
> > 2.43.0
next prev parent reply other threads:[~2026-05-04 14:39 UTC|newest]
Thread overview: 37+ messages / expand[flat|nested] mbox.gz Atom feed top
2026-05-04 4:43 [PATCH 0/9] Add new debug infrastructure for configfs Stuart Summers
2026-05-04 4:43 ` [PATCH 1/9] drm/xe: Rename survivability_mode to enable_survivability_mode Stuart Summers
2026-05-04 13:29 ` Gustavo Sousa
2026-05-04 14:32 ` Summers, Stuart
2026-05-04 14:38 ` Summers, Stuart
2026-05-04 14:40 ` Summers, Stuart
2026-05-04 18:31 ` Rodrigo Vivi
2026-05-04 18:38 ` Summers, Stuart
2026-05-04 4:43 ` [PATCH 2/9] drm/xe: Sort xe_config_device fields and defaults alphabetically Stuart Summers
2026-05-04 13:58 ` Gustavo Sousa
2026-05-04 14:38 ` Summers, Stuart [this message]
2026-05-04 15:47 ` Lin, Shuicheng
2026-05-04 15:54 ` Summers, Stuart
2026-05-04 4:43 ` [PATCH 3/9] drm/xe: Split out configfs data structures Stuart Summers
2026-05-04 4:52 ` Summers, Stuart
2026-05-04 8:47 ` Jani Nikula
2026-05-04 14:24 ` Summers, Stuart
2026-05-04 21:48 ` Matthew Brost
2026-05-04 21:51 ` Summers, Stuart
2026-05-04 4:43 ` [PATCH 4/9] drm/xe: Add a new debug focused configfs group Stuart Summers
2026-05-04 15:42 ` Gustavo Sousa
2026-05-04 15:50 ` Summers, Stuart
2026-05-04 17:28 ` Gustavo Sousa
2026-05-04 17:44 ` Summers, Stuart
2026-05-04 19:04 ` Gustavo Sousa
2026-05-05 21:45 ` Summers, Stuart
2026-05-04 4:43 ` [PATCH 5/9] drm/xe: Move debug configfs entries to xe_configfs_debug.c Stuart Summers
2026-05-04 4:43 ` [PATCH 6/9] drm/xe/guc: Add configfs support for guc_log_level Stuart Summers
2026-05-05 23:54 ` Daniele Ceraolo Spurio
2026-05-04 4:43 ` [PATCH 7/9] drm/xe/guc: Add support for NPK as a GuC log target Stuart Summers
2026-05-04 4:43 ` [PATCH 8/9] drm/xe: Add infrastructure for debug configfs parameters Stuart Summers
2026-05-04 4:43 ` [PATCH 9/9] drm/xe: Migrate existing debug configfs entries to params infrastructure Stuart Summers
2026-05-04 4:54 ` [PATCH 0/9] Add new debug infrastructure for configfs Summers, Stuart
2026-05-04 5:30 ` ✗ CI.checkpatch: warning for " Patchwork
2026-05-04 5:32 ` ✓ CI.KUnit: success " Patchwork
2026-05-04 6:44 ` ✓ Xe.CI.BAT: " Patchwork
2026-05-04 8:42 ` ✗ Xe.CI.FULL: failure " Patchwork
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=9869801345a0e54643dd22836f1069c71cec6aac.camel@intel.com \
--to=stuart.summers@intel.com \
--cc=daniele.ceraolospurio@intel.com \
--cc=gustavo.sousa@intel.com \
--cc=intel-xe@lists.freedesktop.org \
--cc=matthew.brost@intel.com \
--cc=matthew.d.roper@intel.com \
--cc=michal.wajdeczko@intel.com \
--cc=rodrigo.vivi@intel.com \
--cc=shuicheng.lin@intel.com \
--cc=umesh.nerlige.ramappa@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