From: Matthew Auld <matthew.auld@intel.com>
To: Xin Wang <x.wang@intel.com>, igt-dev@lists.freedesktop.org
Cc: shuicheng.lin@intel.com, alex.zuo@intel.com
Subject: Re: [PATCH v2 2/2] tests/intel/xe_pat: sync with Xe PAT debugfs parser
Date: Mon, 8 Dec 2025 16:48:13 +0000 [thread overview]
Message-ID: <3a75364c-528c-4298-aed6-03965e29647e@intel.com> (raw)
In-Reply-To: <20251206011241.67825-2-x.wang@intel.com>
On 06/12/2025 01:12, Xin Wang wrote:
> Pull in the Xe PAT table parsed by lib/intel_pat so Gen20+ tests
Nit: s/Gen20/graphics vesion 20+/ or just xe2+. We stopped using the
"gen" terminology on modern hw, and moved to the ip version stuff.
> use the driver-provided entries instead of hard-coded reservations.
>
> CC: Matthew Auld <matthew.auld@intel.com>
> Signed-off-by: Xin Wang <x.wang@intel.com>
> ---
> tests/intel/xe_pat.c | 38 ++++++++++++++++++++++++++++++++++++--
> 1 file changed, 36 insertions(+), 2 deletions(-)
>
> diff --git a/tests/intel/xe_pat.c b/tests/intel/xe_pat.c
> index 59dfb6b11..b911e9cbf 100644
> --- a/tests/intel/xe_pat.c
> +++ b/tests/intel/xe_pat.c
> @@ -77,6 +77,18 @@ static void userptr_coh_none(int fd)
> xe_vm_destroy(fd, vm);
> }
>
> +#define BITS_TO(n) (n >= sizeof(long) * 8 ? ~0 : (1UL << (n)) - 1)
> +#define BITMASK(high, low) (uint32_t)(BITS_TO(high+1) & ~BITS_TO(low))
> +#define FIELD_GET(val, high, low) (((val) & (BITMASK(high, low))) >> (low))
> +#define FIELD_GET_BIT(val, n) FIELD_GET(val, n, n)
> +
> +#define XE_NO_PROMOTE(val) FIELD_GET_BIT(val, 10)
> +#define XE_COMP_EN(val) FIELD_GET_BIT(val, 9)
> +#define XE_L3_CLOS(val) FIELD_GET(val, 7, 6)
> +#define XE_L3_POLICY(val) FIELD_GET(val, 5, 4)
> +#define XE_L4_POLICY(val) FIELD_GET(val, 3, 2)
> +#define XE_COH_MODE(val) FIELD_GET(val, 1, 0)
> +
> /**
> * SUBTEST: pat-index-all
> * Test category: functionality test
> @@ -86,6 +98,7 @@ static void pat_index_all(int fd)
> {
> uint16_t dev_id = intel_get_drm_devid(fd);
> size_t size = xe_get_default_alignment(fd);
> + struct xe_pat_entry xe_pat_table[XE_PAT_MAX_ENTRIES] = {0};
> uint32_t vm, bo;
> uint8_t pat_index;
>
> @@ -114,10 +127,31 @@ static void pat_index_all(int fd)
>
> igt_assert(intel_get_max_pat_index(fd));
>
> + if (intel_gen(dev_id) >= 20) {
s/intel_gen/intel_graphics_ver/
> + /* Get the Xe PAT entries from debugfs */
> + uint8_t pat_entries = xe_get_pat_entries(fd, xe_pat_table);
> +
> + for (int i = 0; i < pat_entries; i++) {
> + uint32_t pat = xe_pat_table[i].pat;
> + igt_debug("PAT[%2d] = [ %u, %u, %u, %u, %u, %u] (%#8x)%s\n", i,
> + XE_NO_PROMOTE(pat),
> + XE_COMP_EN(pat),
> + XE_L3_CLOS(pat),
> + XE_L3_POLICY(pat),
> + XE_L4_POLICY(pat),
> + XE_COH_MODE(pat),
> + pat,
> + xe_pat_table[i].rsvd ? " *" : "");
> + }
> + }
> +
> for (pat_index = 0; pat_index <= intel_get_max_pat_index(fd);
> pat_index++) {
> - if (intel_get_device_info(dev_id)->graphics_ver >= 20 &&
> - pat_index >= 16 && pat_index <= 19) { /* hw reserved */
> +
> + bool hw_reserved = intel_gen(dev_id) >= 20 ?
intel_graphics_ver
Otherwise lgtm. This should make things easier to maintain going
forward. Thanks for working on this,
Reviewed-by: Matthew Auld <matthew.auld@intel.com>
> + xe_pat_table[pat_index].rsvd : false;
> +
> + if (hw_reserved) {
> igt_assert_eq(__xe_vm_bind(fd, vm, 0, bo, 0, 0x40000,
> size, DRM_XE_VM_BIND_OP_MAP, 0, NULL, 0, 0,
> pat_index, 0),
next prev parent reply other threads:[~2025-12-08 16:48 UTC|newest]
Thread overview: 28+ messages / expand[flat|nested] mbox.gz Atom feed top
2025-11-13 7:05 [PATCH] tests/intel/xe_pat: add helper funtion to read PAT table Xin Wang
2025-11-13 7:57 ` ✗ Xe.CI.BAT: failure for " Patchwork
2025-11-13 14:33 ` ✓ Xe.CI.Full: success " Patchwork
2025-11-13 20:55 ` ✓ i915.CI.BAT: " Patchwork
2025-11-13 21:40 ` [PATCH] " Wang, X
2025-11-13 22:12 ` ✓ i915.CI.Full: success for " Patchwork
2025-12-06 1:12 ` [PATCH v2 1/2] lib/intel_pat: read Xe PAT config from debugfs Xin Wang
2025-12-06 1:12 ` [PATCH v2 2/2] tests/intel/xe_pat: sync with Xe PAT debugfs parser Xin Wang
2025-12-08 16:48 ` Matthew Auld [this message]
2025-12-10 5:24 ` [PATCH v3 " Xin Wang
2025-12-06 18:48 ` [PATCH v2 1/2] lib/intel_pat: read Xe PAT config from debugfs Lin, Shuicheng
2025-12-08 18:37 ` Wang, X
2025-12-08 23:27 ` Lin, Shuicheng
2025-12-09 9:36 ` Kamil Konieczny
2025-12-09 16:41 ` Wang, X
2025-12-08 19:25 ` Wang, X
2025-12-10 5:23 ` [PATCH v3 " Xin Wang
2025-12-11 7:27 ` [PATCH v4 0/2] tests/intel/xe_pat: add helper funtion to read PAT table Xin Wang
2025-12-11 7:27 ` [PATCH v4 1/2] lib/intel_pat: read Xe PAT config from debugfs Xin Wang
2025-12-11 7:27 ` [PATCH v4 2/2] tests/intel/xe_pat: sync with Xe PAT debugfs parser Xin Wang
2025-12-06 1:57 ` ✓ Xe.CI.BAT: success for tests/intel/xe_pat: add helper funtion to read PAT table (rev2) Patchwork
2025-12-06 2:05 ` ✓ i915.CI.BAT: " Patchwork
2025-12-06 13:36 ` ✗ Xe.CI.Full: failure " Patchwork
2025-12-07 4:24 ` ✗ i915.CI.Full: " Patchwork
2025-12-10 9:21 ` ✓ Xe.CI.BAT: success for tests/intel/xe_pat: add helper funtion to read PAT table (rev4) Patchwork
2025-12-10 11:05 ` ✓ i915.CI.BAT: " Patchwork
2025-12-10 12:32 ` ✗ Xe.CI.Full: failure " Patchwork
2025-12-10 12:55 ` ✓ i915.CI.Full: success " 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=3a75364c-528c-4298-aed6-03965e29647e@intel.com \
--to=matthew.auld@intel.com \
--cc=alex.zuo@intel.com \
--cc=igt-dev@lists.freedesktop.org \
--cc=shuicheng.lin@intel.com \
--cc=x.wang@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;
as well as URLs for NNTP newsgroup(s).