From: Tom Chung <chiahsuan.chung@amd.com>
To: James Lin <PingLei.Lin@amd.com>, igt-dev@lists.freedesktop.org
Cc: alex.hung@amd.com, sunpeng.li@amd.com
Subject: Re: [PATCH i-g-t] tests/amdgpu: Discover amdgpu backlight device dynamically instead of hardcoding amdgpu_bl0
Date: Fri, 26 Jun 2026 13:36:36 +0800 [thread overview]
Message-ID: <dcf429e7-ff2f-424a-aeec-4d566beb0b02@amd.com> (raw)
In-Reply-To: <20260626032211.38353-1-PingLei.Lin@amd.com>
Patch looks good to me.
Reviewed-by: Tom Chung <chiahsuan.chung@amd.com>
On 6/26/2026 11:21 AM, James Lin wrote:
> [Why]
> amd_abm builds the backlight sysfs path assuming the amdgpu backlight
> device is amdgpu_bl0. The amdgpu backlight device is named amdgpu_bl<N>
> where N is derived from the DRM primary node index (see
> amdgpu_dm_backlight.c). On systems where the amdgpu DRM node is not
> card0 (e.g. simpledrm takes card0 and amdgpu comes up as card1), the
> backlight device is amdgpu_bl1, not amdgpu_bl0. The hardcoded
> "/sys/class/backlight/amdgpu_bl0" path does not exist, so
> backlight_read_max_brightness() and backlight_write_brightness() fail
> before any test logic runs, preventing dpms_cycle and the other abm
> subtests from executing.
>
> [How]
> Add find_backlight_device() to scan /sys/class/backlight for the first
> entry beginning with "amdgpu_bl" and store its full path in a
> file-scope buffer. Call it from igt_fixture and route the brightness
> and max_brightness accessors through the discovered path instead of the
> hardcoded amdgpu_bl0 macro. Use igt_require_f() so a machine with no
> amdgpu backlight skips cleanly rather than asserting.
>
> Signed-off-by: James Lin <PingLei.Lin@amd.com>
> ---
> tests/amdgpu/amd_abm.c | 39 ++++++++++++++++++++++++++++++++++++---
> 1 file changed, 36 insertions(+), 3 deletions(-)
>
> diff --git a/tests/amdgpu/amd_abm.c b/tests/amdgpu/amd_abm.c
> index ba90344d3..7992dfd87 100644
> --- a/tests/amdgpu/amd_abm.c
> +++ b/tests/amdgpu/amd_abm.c
> @@ -31,13 +31,44 @@
> #include <string.h>
> #include <fcntl.h>
> #include <time.h>
> +#include <dirent.h>
>
> #define DEBUGFS_CURRENT_BACKLIGHT_PWM "amdgpu_current_backlight_pwm"
> #define DEBUGFS_TARGET_BACKLIGHT_PWM "amdgpu_target_backlight_pwm"
> -#define BACKLIGHT_PATH "/sys/class/backlight/amdgpu_bl0"
> +#define BACKLIGHT_DIR "/sys/class/backlight"
> #define PANEL_POWER_SAVINGS_ATTR "amdgpu/panel_power_savings"
> #define MK_COLOR(r, g, b) ((0 << 24) | (r << 16) | (g << 8) | b)
>
> +/*
> + * The amdgpu backlight device is named amdgpu_bl<N> where N is derived from
> + * the DRM primary node index (see amdgpu_dm_backlight.c). On systems where
> + * the amdgpu DRM node is not card0 (e.g. card1), the device is amdgpu_bl1,
> + * not amdgpu_bl0. Discover it at runtime instead of hardcoding the name.
> + */
> +static char backlight_path[PATH_MAX];
> +
> +static void find_backlight_device(void)
> +{
> + DIR *dir;
> + struct dirent *ent;
> +
> + dir = opendir(BACKLIGHT_DIR);
> + igt_require_f(dir, "Cannot open %s\n", BACKLIGHT_DIR);
> +
> + while ((ent = readdir(dir))) {
> + if (strncmp(ent->d_name, "amdgpu_bl", 9) == 0) {
> + snprintf(backlight_path, sizeof(backlight_path),
> + "%s/%s", BACKLIGHT_DIR, ent->d_name);
> + break;
> + }
> + }
> + closedir(dir);
> +
> + igt_require_f(backlight_path[0] != '\0',
> + "No amdgpu_bl* backlight device found in %s\n",
> + BACKLIGHT_DIR);
> +}
> +
> typedef struct data {
> igt_display_t display;
> igt_plane_t *primary;
> @@ -234,7 +265,7 @@ static int backlight_write_brightness(int value)
> char src[64];
> int len;
>
> - igt_assert(snprintf(full, PATH_MAX, "%s/%s", BACKLIGHT_PATH, "brightness") < PATH_MAX);
> + igt_assert(snprintf(full, PATH_MAX, "%s/%s", backlight_path, "brightness") < PATH_MAX);
> fd = open(full, O_WRONLY);
> if (fd == -1)
> return -errno;
> @@ -288,7 +319,7 @@ static int backlight_read_max_brightness(int *result)
> char dst[64];
> int r, e;
>
> - igt_assert(snprintf(full, PATH_MAX, "%s/%s", BACKLIGHT_PATH, "max_brightness") < PATH_MAX);
> + igt_assert(snprintf(full, PATH_MAX, "%s/%s", backlight_path, "max_brightness") < PATH_MAX);
>
> fd = open(full, O_RDONLY);
> if (fd == -1)
> @@ -528,6 +559,8 @@ int igt_main()
>
> igt_display_require(&data.display, data.drm_fd);
>
> + find_backlight_device();
> +
> test_init(&data);
> }
>
next prev parent reply other threads:[~2026-06-26 5:37 UTC|newest]
Thread overview: 5+ messages / expand[flat|nested] mbox.gz Atom feed top
2026-06-26 3:21 [PATCH i-g-t] tests/amdgpu: Discover amdgpu backlight device dynamically instead of hardcoding amdgpu_bl0 James Lin
2026-06-26 4:12 ` ✓ Xe.CI.BAT: success for " Patchwork
2026-06-26 4:28 ` ✓ i915.CI.BAT: " Patchwork
2026-06-26 5:36 ` Tom Chung [this message]
2026-06-26 5:50 ` ✗ 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=dcf429e7-ff2f-424a-aeec-4d566beb0b02@amd.com \
--to=chiahsuan.chung@amd.com \
--cc=PingLei.Lin@amd.com \
--cc=alex.hung@amd.com \
--cc=igt-dev@lists.freedesktop.org \
--cc=sunpeng.li@amd.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