From: "Bhadane, Dnyaneshwar" <dnyaneshwar.bhadane@intel.com>
To: <intel-xe@lists.freedesktop.org>
Subject: Re: [PATCH 8/9] drm/xe: Don't initialize tile_count in xe_info_init_early()
Date: Fri, 19 Jun 2026 23:34:48 +0530 [thread overview]
Message-ID: <78cc5c4c-c588-4d4a-9cd4-94adb8a320a7@intel.com> (raw)
In-Reply-To: <20260609-xe-probe-info-v1-8-21e83e188e60@intel.com>
On 10-Jun-26 1:47 AM, Gustavo Sousa wrote:
> The value of xe->info.tile_count is only really valid after
> xe_info_probe_tile_count(). Any use of tile_count before that point
> is invalid and, consequently, initializing it in xe_info_init_early()
> is pointless.
>
> Move the initialization to xe_info_probe_tile_count().
>
> Signed-off-by: Gustavo Sousa <gustavo.sousa@intel.com>
LGTM,
Reviewed-by: Dnyaneshwar Bhadane <dnyaneshwar.bhadane@intel.com>
> ---
> drivers/gpu/drm/xe/tests/xe_pci.c | 5 +++--
> drivers/gpu/drm/xe/xe_pci.c | 10 ++++++----
> 2 files changed, 9 insertions(+), 6 deletions(-)
>
> diff --git a/drivers/gpu/drm/xe/tests/xe_pci.c b/drivers/gpu/drm/xe/tests/xe_pci.c
> index cd64b1d614c8..31ec41aa997d 100644
> --- a/drivers/gpu/drm/xe/tests/xe_pci.c
> +++ b/drivers/gpu/drm/xe/tests/xe_pci.c
> @@ -311,9 +311,10 @@ const void *xe_pci_id_gen_param(struct kunit *test, const void *prev, char *desc
> }
> EXPORT_SYMBOL_IF_KUNIT(xe_pci_id_gen_param);
>
> -static void fake_xe_info_probe_tile_count(struct xe_device *xe)
> +static void fake_xe_info_probe_tile_count(struct xe_device *xe,
> + const struct xe_device_desc *desc)
> {
> - /* Nothing to do, just use the statically defined value. */
> + xe->info.tile_count = 1 + desc->max_remote_tiles;
> }
>
> static int fake_probe_info(struct xe_device *xe,
> diff --git a/drivers/gpu/drm/xe/xe_pci.c b/drivers/gpu/drm/xe/xe_pci.c
> index 388771ef6a52..6b2daad87316 100644
> --- a/drivers/gpu/drm/xe/xe_pci.c
> +++ b/drivers/gpu/drm/xe/xe_pci.c
> @@ -811,7 +811,6 @@ static int xe_info_init_early(struct xe_device *xe,
> xe_assert(xe, desc->max_gt_per_tile > 0);
> xe_assert(xe, desc->max_gt_per_tile <= XE_MAX_GT_PER_TILE);
> xe->info.max_gt_per_tile = desc->max_gt_per_tile;
> - xe->info.tile_count = 1 + desc->max_remote_tiles;
>
> err = xe_tile_init_early(xe_device_get_root_tile(xe), xe, 0);
> if (err)
> @@ -823,13 +822,16 @@ static int xe_info_init_early(struct xe_device *xe,
> /*
> * Possibly override number of tile based on configuration register.
> */
> -static void xe_info_probe_tile_count(struct xe_device *xe)
> +static void xe_info_probe_tile_count(struct xe_device *xe,
> + const struct xe_device_desc *desc)
> {
> struct xe_mmio *mmio;
> u8 tile_count;
> u32 mtcfg;
>
> - KUNIT_STATIC_STUB_REDIRECT(xe_info_probe_tile_count, xe);
> + KUNIT_STATIC_STUB_REDIRECT(xe_info_probe_tile_count, xe, desc);
> +
> + xe->info.tile_count = 1 + desc->max_remote_tiles;
>
> /*
> * Probe for tile count only for platforms that support multiple
> @@ -1035,7 +1037,7 @@ static int xe_info_init(struct xe_device *xe,
> xe->info.has_soc_remapper_telem = 0;
> }
>
> - xe_info_probe_tile_count(xe);
> + xe_info_probe_tile_count(xe, desc);
>
> for_each_remote_tile(tile, xe, id) {
> int err;
>
next prev parent reply other threads:[~2026-06-19 18:04 UTC|newest]
Thread overview: 36+ messages / expand[flat|nested] mbox.gz Atom feed top
2026-06-09 20:17 [PATCH 0/9] drm/xe: Probe info outside of xe_info_init() and xe_info_init_early() Gustavo Sousa
2026-06-09 20:17 ` [PATCH 1/9] drm/xe: Add framework for info probing Gustavo Sousa
2026-06-16 18:51 ` Violet Monti
2026-06-16 21:56 ` Matt Roper
2026-06-17 12:43 ` Gustavo Sousa
2026-06-19 17:45 ` Bhadane, Dnyaneshwar
2026-06-09 20:17 ` [PATCH 2/9] drm/xe/step: Pass xe_step_info to xe_step_*_get() functions Gustavo Sousa
2026-06-16 19:31 ` Violet Monti
2026-06-19 17:47 ` Bhadane, Dnyaneshwar
2026-06-09 20:17 ` [PATCH 3/9] drm/xe: Add devid and revid to xe_probed_info Gustavo Sousa
2026-06-16 19:33 ` Violet Monti
2026-06-19 17:54 ` Bhadane, Dnyaneshwar
2026-06-09 20:17 ` [PATCH 4/9] drm/xe/step: Make xe_step_platform_get() independent from xe->info Gustavo Sousa
2026-06-16 19:37 ` Violet Monti
2026-06-09 20:17 ` [PATCH 5/9] drm/xe: Add platform-level step info to xe_probed_info Gustavo Sousa
2026-06-16 19:50 ` Violet Monti
2026-06-19 18:01 ` Bhadane, Dnyaneshwar
2026-06-09 20:17 ` [PATCH 6/9] drm/xe/tests: Set non-GMDID graphics step in xe_pci_fake_device_init() Gustavo Sousa
2026-06-16 19:58 ` Violet Monti
2026-06-19 18:02 ` Bhadane, Dnyaneshwar
2026-06-09 20:17 ` [PATCH 7/9] drm/xe: Add graphics/media IPs and their step info to xe_probed_info Gustavo Sousa
2026-06-16 20:22 ` Violet Monti
2026-06-19 18:04 ` Bhadane, Dnyaneshwar
2026-06-09 20:17 ` [PATCH 8/9] drm/xe: Don't initialize tile_count in xe_info_init_early() Gustavo Sousa
2026-06-16 21:07 ` Violet Monti
2026-06-19 18:04 ` Bhadane, Dnyaneshwar [this message]
2026-06-09 20:17 ` [PATCH 9/9] drm/xe: Add tile_count to xe_probed_info Gustavo Sousa
2026-06-16 21:16 ` Violet Monti
2026-06-19 18:05 ` Bhadane, Dnyaneshwar
2026-06-09 20:58 ` ✓ CI.KUnit: success for drm/xe: Probe info outside of xe_info_init() and xe_info_init_early() Patchwork
2026-06-09 22:06 ` ✓ Xe.CI.BAT: " Patchwork
2026-06-10 12:22 ` ✗ Xe.CI.FULL: failure " Patchwork
2026-07-01 16:00 ` ✓ CI.KUnit: success for drm/xe: Probe info outside of xe_info_init() and xe_info_init_early() (rev2) Patchwork
2026-07-01 17:03 ` ✓ Xe.CI.BAT: " Patchwork
2026-07-02 11:47 ` ✓ Xe.CI.FULL: " Patchwork
2026-07-02 22:13 ` [PATCH 0/9] drm/xe: Probe info outside of xe_info_init() and xe_info_init_early() Gustavo Sousa
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=78cc5c4c-c588-4d4a-9cd4-94adb8a320a7@intel.com \
--to=dnyaneshwar.bhadane@intel.com \
--cc=intel-xe@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