From 1958f4e16864f78ab121de08ba4d7a984ed46891 Mon Sep 17 00:00:00 2001 From: Zhang Rui Date: Thu, 14 Nov 2024 15:59:42 +0800 Subject: [PATCH 01/16] tools/power turbostat: Enhance platform divergence description In various generations, platforms often share a majority of features, diverging only in a few specific aspects. The current approach of using hardcoded values in 'platform_features' structure fails to effectively represent these divergences. To improve the description of platform divergence: 1. Each newly introduced 'platform_features' structure must have a base, typically derived from the previous generation. 2. Platform feature values should be inherited from the base structure rather than being hardcoded. This approach ensures a more accurate and maintainable representation of platform-specific features across different generations. Converts `adl_features` and `lnl_features` to follow this new scheme. No functional change. Signed-off-by: Zhang Rui Signed-off-by: Len Brown --- tools/power/x86/turbostat/turbostat.c | 58 ++++++++++++++------------- 1 file changed, 30 insertions(+), 28 deletions(-) diff --git a/tools/power/x86/turbostat/turbostat.c b/tools/power/x86/turbostat/turbostat.c index 90f3119dbd14..ae841baeca85 100644 --- a/tools/power/x86/turbostat/turbostat.c +++ b/tools/power/x86/turbostat/turbostat.c @@ -735,38 +735,40 @@ static const struct platform_features cnl_features = { .enable_tsc_tweak = 1, }; +/* Copied from cnl_features, with PC7/PC9 removed */ static const struct platform_features adl_features = { - .has_msr_misc_feature_control = 1, - .has_msr_misc_pwr_mgmt = 1, - .has_nhm_msrs = 1, - .has_config_tdp = 1, - .bclk_freq = BCLK_100MHZ, - .supported_cstates = CC1 | CC6 | CC7 | PC2 | PC3 | PC6 | PC8 | PC10, - .cst_limit = CST_LIMIT_HSW, - .has_irtl_msrs = 1, - .has_msr_core_c1_res = 1, - .has_ext_cst_msrs = 1, - .trl_msrs = TRL_BASE, - .tcc_offset_bits = 6, - .rapl_msrs = RAPL_PKG_ALL | RAPL_CORE_ALL | RAPL_DRAM | RAPL_DRAM_PERF_STATUS | RAPL_GFX, - .enable_tsc_tweak = 1, + .has_msr_misc_feature_control = cnl_features.has_msr_misc_feature_control, + .has_msr_misc_pwr_mgmt = cnl_features.has_msr_misc_pwr_mgmt, + .has_nhm_msrs = cnl_features.has_nhm_msrs, + .has_config_tdp = cnl_features.has_config_tdp, + .bclk_freq = cnl_features.bclk_freq, + .supported_cstates = CC1 | CC6 | CC7 | PC2 | PC3 | PC6 | PC8 | PC10, + .cst_limit = cnl_features.cst_limit, + .has_irtl_msrs = cnl_features.has_irtl_msrs, + .has_msr_core_c1_res = cnl_features.has_msr_core_c1_res, + .has_ext_cst_msrs = cnl_features.has_ext_cst_msrs, + .trl_msrs = cnl_features.trl_msrs, + .tcc_offset_bits = cnl_features.tcc_offset_bits, + .rapl_msrs = cnl_features.rapl_msrs, + .enable_tsc_tweak = cnl_features.enable_tsc_tweak, }; +/* Copied from adl_features, with PC3/PC8 removed */ static const struct platform_features lnl_features = { - .has_msr_misc_feature_control = 1, - .has_msr_misc_pwr_mgmt = 1, - .has_nhm_msrs = 1, - .has_config_tdp = 1, - .bclk_freq = BCLK_100MHZ, - .supported_cstates = CC1 | CC6 | CC7 | PC2 | PC6 | PC10, - .cst_limit = CST_LIMIT_HSW, - .has_irtl_msrs = 1, - .has_msr_core_c1_res = 1, - .has_ext_cst_msrs = 1, - .trl_msrs = TRL_BASE, - .tcc_offset_bits = 6, - .rapl_msrs = RAPL_PKG_ALL | RAPL_CORE_ALL | RAPL_DRAM | RAPL_DRAM_PERF_STATUS | RAPL_GFX, - .enable_tsc_tweak = 1, + .has_msr_misc_feature_control = adl_features.has_msr_misc_feature_control, + .has_msr_misc_pwr_mgmt = adl_features.has_msr_misc_pwr_mgmt, + .has_nhm_msrs = adl_features.has_nhm_msrs, + .has_config_tdp = adl_features.has_config_tdp, + .bclk_freq = adl_features.bclk_freq, + .supported_cstates = CC1 | CC6 | CC7 | PC2 | PC6 | PC10, + .cst_limit = adl_features.cst_limit, + .has_irtl_msrs = adl_features.has_irtl_msrs, + .has_msr_core_c1_res = adl_features.has_msr_core_c1_res, + .has_ext_cst_msrs = adl_features.has_ext_cst_msrs, + .trl_msrs = adl_features.trl_msrs, + .tcc_offset_bits = adl_features.tcc_offset_bits, + .rapl_msrs = adl_features.rapl_msrs, + .enable_tsc_tweak = adl_features.enable_tsc_tweak, }; static const struct platform_features skx_features = { -- 2.50.1 From ba99a4fc8c24dc7d35f18edb6e3b0a65345fbfa3 Mon Sep 17 00:00:00 2001 From: Zhang Rui Date: Thu, 14 Nov 2024 15:59:43 +0800 Subject: [PATCH 02/16] tools/power turbostat: Remove unnecessary fflush() call The graphics sysfs knobs are read-only, making the use of fflush() before reading them redundant. Remove the unnecessary fflush() call. Signed-off-by: Zhang Rui Signed-off-by: Len Brown --- tools/power/x86/turbostat/turbostat.c | 7 +++---- 1 file changed, 3 insertions(+), 4 deletions(-) diff --git a/tools/power/x86/turbostat/turbostat.c b/tools/power/x86/turbostat/turbostat.c index ae841baeca85..c0596ccf92cd 100644 --- a/tools/power/x86/turbostat/turbostat.c +++ b/tools/power/x86/turbostat/turbostat.c @@ -5780,12 +5780,11 @@ int snapshot_graphics(int idx) case GFX_ACTMHz: case SAM_MHz: case SAM_ACTMHz: - if (gfx_info[idx].fp == NULL) { + if (gfx_info[idx].fp == NULL) gfx_info[idx].fp = fopen_or_die(gfx_info[idx].path, "r"); - } else { + else rewind(gfx_info[idx].fp); - fflush(gfx_info[idx].fp); - } + retval = fscanf(gfx_info[idx].fp, "%d", &gfx_info[idx].val); if (retval != 1) err(1, "MHz"); -- 2.50.1 From d071004e623b7433573019d67cba79e345d83006 Mon Sep 17 00:00:00 2001 From: Zhang Rui Date: Thu, 14 Nov 2024 15:59:44 +0800 Subject: [PATCH 03/16] tools/power turbostat: Consolidate graphics sysfs access Currently, there is an inconsistency in how graphics sysfs knobs are accessed: graphics residency sysfs knobs are opened and closed for each read, while graphics frequency sysfs knobs are opened once and remain open until turbostat exits. This inconsistency is confusing and adds unnecessary code complexity. Consolidate the access method by opening the sysfs files once and reusing the file pointers for subsequent accesses. This approach simplifies the code and ensures a consistent method for accessing graphics sysfs knobs. Signed-off-by: Zhang Rui Signed-off-by: Len Brown --- tools/power/x86/turbostat/turbostat.c | 15 ++++++--------- 1 file changed, 6 insertions(+), 9 deletions(-) diff --git a/tools/power/x86/turbostat/turbostat.c b/tools/power/x86/turbostat/turbostat.c index c0596ccf92cd..e5b100b8db24 100644 --- a/tools/power/x86/turbostat/turbostat.c +++ b/tools/power/x86/turbostat/turbostat.c @@ -5764,27 +5764,24 @@ int snapshot_proc_interrupts(void) */ int snapshot_graphics(int idx) { - FILE *fp; int retval; + if (gfx_info[idx].fp == NULL) + gfx_info[idx].fp = fopen_or_die(gfx_info[idx].path, "r"); + else + rewind(gfx_info[idx].fp); + switch (idx) { case GFX_rc6: case SAM_mc6: - fp = fopen_or_die(gfx_info[idx].path, "r"); - retval = fscanf(fp, "%lld", &gfx_info[idx].val_ull); + retval = fscanf(gfx_info[idx].fp, "%lld", &gfx_info[idx].val_ull); if (retval != 1) err(1, "rc6"); - fclose(fp); return 0; case GFX_MHz: case GFX_ACTMHz: case SAM_MHz: case SAM_ACTMHz: - if (gfx_info[idx].fp == NULL) - gfx_info[idx].fp = fopen_or_die(gfx_info[idx].path, "r"); - else - rewind(gfx_info[idx].fp); - retval = fscanf(gfx_info[idx].fp, "%d", &gfx_info[idx].val); if (retval != 1) err(1, "MHz"); -- 2.50.1 From c7538f33853b11d0ff2a81efb78bde125d1fc49f Mon Sep 17 00:00:00 2001 From: Zhang Rui Date: Thu, 14 Nov 2024 15:59:45 +0800 Subject: [PATCH 04/16] tools/power turbostat: Cache graphics sysfs file descriptors during probe Snapshots of the graphics sysfs knobs are taken based on file descriptors. To optimize this process, open the files and cache the file descriptors during the graphics probe phase. As a result, the previously cached pathnames become redundant and are removed. This change aims to streamline the code without altering its functionality. No functional change intended. Signed-off-by: Zhang Rui Signed-off-by: Len Brown --- tools/power/x86/turbostat/turbostat.c | 82 +++++++++++---------------- 1 file changed, 32 insertions(+), 50 deletions(-) diff --git a/tools/power/x86/turbostat/turbostat.c b/tools/power/x86/turbostat/turbostat.c index e5b100b8db24..28513172ffce 100644 --- a/tools/power/x86/turbostat/turbostat.c +++ b/tools/power/x86/turbostat/turbostat.c @@ -376,7 +376,6 @@ enum gfx_sysfs_idx { }; struct gfx_sysfs_info { - const char *path; FILE *fp; unsigned int val; unsigned long long val_ull; @@ -5766,10 +5765,7 @@ int snapshot_graphics(int idx) { int retval; - if (gfx_info[idx].fp == NULL) - gfx_info[idx].fp = fopen_or_die(gfx_info[idx].path, "r"); - else - rewind(gfx_info[idx].fp); + rewind(gfx_info[idx].fp); switch (idx) { case GFX_rc6: @@ -6474,6 +6470,12 @@ static void probe_intel_uncore_frequency(void) probe_intel_uncore_frequency_legacy(); } +static void set_graphics_fp(char *path, int idx) +{ + if (!access(path, R_OK)) + gfx_info[idx].fp = fopen_or_die(path, "r"); +} + static void probe_graphics(void) { /* Xe graphics sysfs knobs */ @@ -6481,7 +6483,6 @@ static void probe_graphics(void) FILE *fp; char buf[8]; bool gt0_is_gt; - int idx; fp = fopen("/sys/class/drm/card0/device/tile0/gt0/gtidle/name", "r"); if (!fp) @@ -6500,28 +6501,17 @@ static void probe_graphics(void) else goto next; - idx = gt0_is_gt ? GFX_rc6 : SAM_mc6; - gfx_info[idx].path = "/sys/class/drm/card0/device/tile0/gt0/gtidle/idle_residency_ms"; + set_graphics_fp("/sys/class/drm/card0/device/tile0/gt0/gtidle/idle_residency_ms", gt0_is_gt ? GFX_rc6 : SAM_mc6); - idx = gt0_is_gt ? GFX_MHz : SAM_MHz; - if (!access("/sys/class/drm/card0/device/tile0/gt0/freq0/cur_freq", R_OK)) - gfx_info[idx].path = "/sys/class/drm/card0/device/tile0/gt0/freq0/cur_freq"; + set_graphics_fp("/sys/class/drm/card0/device/tile0/gt0/freq0/cur_freq", gt0_is_gt ? GFX_MHz : SAM_MHz); - idx = gt0_is_gt ? GFX_ACTMHz : SAM_ACTMHz; - if (!access("/sys/class/drm/card0/device/tile0/gt0/freq0/act_freq", R_OK)) - gfx_info[idx].path = "/sys/class/drm/card0/device/tile0/gt0/freq0/act_freq"; + set_graphics_fp("/sys/class/drm/card0/device/tile0/gt0/freq0/act_freq", gt0_is_gt ? GFX_ACTMHz : SAM_ACTMHz); - idx = gt0_is_gt ? SAM_mc6 : GFX_rc6; - if (!access("/sys/class/drm/card0/device/tile0/gt1/gtidle/idle_residency_ms", R_OK)) - gfx_info[idx].path = "/sys/class/drm/card0/device/tile0/gt1/gtidle/idle_residency_ms"; + set_graphics_fp("/sys/class/drm/card0/device/tile0/gt1/gtidle/idle_residency_ms", gt0_is_gt ? SAM_mc6 : GFX_rc6); - idx = gt0_is_gt ? SAM_MHz : GFX_MHz; - if (!access("/sys/class/drm/card0/device/tile0/gt1/freq0/cur_freq", R_OK)) - gfx_info[idx].path = "/sys/class/drm/card0/device/tile0/gt1/freq0/cur_freq"; + set_graphics_fp("/sys/class/drm/card0/device/tile0/gt1/freq0/cur_freq", gt0_is_gt ? SAM_MHz : GFX_MHz); - idx = gt0_is_gt ? SAM_ACTMHz : GFX_ACTMHz; - if (!access("/sys/class/drm/card0/device/tile0/gt1/freq0/act_freq", R_OK)) - gfx_info[idx].path = "/sys/class/drm/card0/device/tile0/gt1/freq0/act_freq"; + set_graphics_fp("/sys/class/drm/card0/device/tile0/gt1/freq0/act_freq", gt0_is_gt ? SAM_ACTMHz : GFX_ACTMHz); goto end; } @@ -6529,52 +6519,44 @@ static void probe_graphics(void) next: /* New i915 graphics sysfs knobs */ if (!access("/sys/class/drm/card0/gt/gt0/rc6_residency_ms", R_OK)) { - gfx_info[GFX_rc6].path = "/sys/class/drm/card0/gt/gt0/rc6_residency_ms"; + set_graphics_fp("/sys/class/drm/card0/gt/gt0/rc6_residency_ms", GFX_rc6); - if (!access("/sys/class/drm/card0/gt/gt0/rps_cur_freq_mhz", R_OK)) - gfx_info[GFX_MHz].path = "/sys/class/drm/card0/gt/gt0/rps_cur_freq_mhz"; + set_graphics_fp("/sys/class/drm/card0/gt/gt0/rps_cur_freq_mhz", GFX_MHz); - if (!access("/sys/class/drm/card0/gt/gt0/rps_act_freq_mhz", R_OK)) - gfx_info[GFX_ACTMHz].path = "/sys/class/drm/card0/gt/gt0/rps_act_freq_mhz"; + set_graphics_fp("/sys/class/drm/card0/gt/gt0/rps_act_freq_mhz", GFX_ACTMHz); - if (!access("/sys/class/drm/card0/gt/gt1/rc6_residency_ms", R_OK)) - gfx_info[SAM_mc6].path = "/sys/class/drm/card0/gt/gt1/rc6_residency_ms"; + set_graphics_fp("/sys/class/drm/card0/gt/gt1/rc6_residency_ms", SAM_mc6); - if (!access("/sys/class/drm/card0/gt/gt1/rps_cur_freq_mhz", R_OK)) - gfx_info[SAM_MHz].path = "/sys/class/drm/card0/gt/gt1/rps_cur_freq_mhz"; + set_graphics_fp("/sys/class/drm/card0/gt/gt1/rps_cur_freq_mhz", SAM_MHz); - if (!access("/sys/class/drm/card0/gt/gt1/rps_act_freq_mhz", R_OK)) - gfx_info[SAM_ACTMHz].path = "/sys/class/drm/card0/gt/gt1/rps_act_freq_mhz"; + set_graphics_fp("/sys/class/drm/card0/gt/gt1/rps_act_freq_mhz", SAM_ACTMHz); goto end; } /* Fall back to traditional i915 graphics sysfs knobs */ - if (!access("/sys/class/drm/card0/power/rc6_residency_ms", R_OK)) - gfx_info[GFX_rc6].path = "/sys/class/drm/card0/power/rc6_residency_ms"; + set_graphics_fp("/sys/class/drm/card0/power/rc6_residency_ms", GFX_rc6); - if (!access("/sys/class/drm/card0/gt_cur_freq_mhz", R_OK)) - gfx_info[GFX_MHz].path = "/sys/class/drm/card0/gt_cur_freq_mhz"; - else if (!access("/sys/class/graphics/fb0/device/drm/card0/gt_cur_freq_mhz", R_OK)) - gfx_info[GFX_MHz].path = "/sys/class/graphics/fb0/device/drm/card0/gt_cur_freq_mhz"; + set_graphics_fp("/sys/class/drm/card0/gt_cur_freq_mhz", GFX_MHz); + if (!gfx_info[GFX_MHz].fp) + set_graphics_fp("/sys/class/graphics/fb0/device/drm/card0/gt_cur_freq_mhz", GFX_MHz); - if (!access("/sys/class/drm/card0/gt_act_freq_mhz", R_OK)) - gfx_info[GFX_ACTMHz].path = "/sys/class/drm/card0/gt_act_freq_mhz"; - else if (!access("/sys/class/graphics/fb0/device/drm/card0/gt_act_freq_mhz", R_OK)) - gfx_info[GFX_ACTMHz].path = "/sys/class/graphics/fb0/device/drm/card0/gt_act_freq_mhz"; + set_graphics_fp("/sys/class/drm/card0/gt_act_freq_mhz", GFX_ACTMHz); + if (!gfx_info[GFX_ACTMHz].fp) + set_graphics_fp("/sys/class/graphics/fb0/device/drm/card0/gt_act_freq_mhz", GFX_ACTMHz); end: - if (gfx_info[GFX_rc6].path) + if (gfx_info[GFX_rc6].fp) BIC_PRESENT(BIC_GFX_rc6); - if (gfx_info[GFX_MHz].path) + if (gfx_info[GFX_MHz].fp) BIC_PRESENT(BIC_GFXMHz); - if (gfx_info[GFX_ACTMHz].path) + if (gfx_info[GFX_ACTMHz].fp) BIC_PRESENT(BIC_GFXACTMHz); - if (gfx_info[SAM_mc6].path) + if (gfx_info[SAM_mc6].fp) BIC_PRESENT(BIC_SAM_mc6); - if (gfx_info[SAM_MHz].path) + if (gfx_info[SAM_MHz].fp) BIC_PRESENT(BIC_SAMMHz); - if (gfx_info[SAM_ACTMHz].path) + if (gfx_info[SAM_ACTMHz].fp) BIC_PRESENT(BIC_SAMACTMHz); } -- 2.50.1 From 03109e2f0d18dcb84218bd91c4fbf864193ca934 Mon Sep 17 00:00:00 2001 From: Zhang Rui Date: Thu, 14 Nov 2024 15:59:46 +0800 Subject: [PATCH 05/16] tools/power turbostat: Add support for /sys/class/drm/card1 On some machines, the graphics device is enumerated as /sys/class/drm/card1 instead of /sys/class/drm/card0. The current implementation does not handle this scenario, resulting in the loss of graphics C6 residency and frequency information. Add support for /sys/class/drm/card1, ensuring that turbostat can retrieve and display the graphics columns for these platforms. Signed-off-by: Zhang Rui Signed-off-by: Len Brown --- tools/power/x86/turbostat/turbostat.c | 38 ++++++++++++++++++++------- 1 file changed, 29 insertions(+), 9 deletions(-) diff --git a/tools/power/x86/turbostat/turbostat.c b/tools/power/x86/turbostat/turbostat.c index 28513172ffce..b250676c174e 100644 --- a/tools/power/x86/turbostat/turbostat.c +++ b/tools/power/x86/turbostat/turbostat.c @@ -6476,8 +6476,14 @@ static void set_graphics_fp(char *path, int idx) gfx_info[idx].fp = fopen_or_die(path, "r"); } +/* Enlarge this if there are /sys/class/drm/card2 ... */ +#define GFX_MAX_CARDS 2 + static void probe_graphics(void) { + char path[PATH_MAX]; + int i; + /* Xe graphics sysfs knobs */ if (!access("/sys/class/drm/card0/device/tile0/gt0/gtidle/idle_residency_ms", R_OK)) { FILE *fp; @@ -6518,22 +6524,36 @@ static void probe_graphics(void) next: /* New i915 graphics sysfs knobs */ - if (!access("/sys/class/drm/card0/gt/gt0/rc6_residency_ms", R_OK)) { - set_graphics_fp("/sys/class/drm/card0/gt/gt0/rc6_residency_ms", GFX_rc6); + for (i = 0; i < GFX_MAX_CARDS; i++) { + snprintf(path, PATH_MAX, "/sys/class/drm/card%d/gt/gt0/rc6_residency_ms", i); + if (!access(path, R_OK)) + break; + } - set_graphics_fp("/sys/class/drm/card0/gt/gt0/rps_cur_freq_mhz", GFX_MHz); + if (i == GFX_MAX_CARDS) + goto legacy_i915; - set_graphics_fp("/sys/class/drm/card0/gt/gt0/rps_act_freq_mhz", GFX_ACTMHz); + snprintf(path, PATH_MAX, "/sys/class/drm/card%d/gt/gt0/rc6_residency_ms", i); + set_graphics_fp(path, GFX_rc6); - set_graphics_fp("/sys/class/drm/card0/gt/gt1/rc6_residency_ms", SAM_mc6); + snprintf(path, PATH_MAX, "/sys/class/drm/card%d/gt/gt0/rps_cur_freq_mhz", i); + set_graphics_fp(path, GFX_MHz); - set_graphics_fp("/sys/class/drm/card0/gt/gt1/rps_cur_freq_mhz", SAM_MHz); + snprintf(path, PATH_MAX, "/sys/class/drm/card%d/gt/gt0/rps_act_freq_mhz", i); + set_graphics_fp(path, GFX_ACTMHz); - set_graphics_fp("/sys/class/drm/card0/gt/gt1/rps_act_freq_mhz", SAM_ACTMHz); + snprintf(path, PATH_MAX, "/sys/class/drm/card%d/gt/gt1/rc6_residency_ms", i); + set_graphics_fp(path, SAM_mc6); - goto end; - } + snprintf(path, PATH_MAX, "/sys/class/drm/card%d/gt/gt1/rps_cur_freq_mhz", i); + set_graphics_fp(path, SAM_MHz); + + snprintf(path, PATH_MAX, "/sys/class/drm/card%d/gt/gt1/rps_act_freq_mhz", i); + set_graphics_fp(path, SAM_ACTMHz); + + goto end; +legacy_i915: /* Fall back to traditional i915 graphics sysfs knobs */ set_graphics_fp("/sys/class/drm/card0/power/rc6_residency_ms", GFX_rc6); -- 2.50.1 From bcfab87108b33f20d847fd71a2a93114dd2ce83e Mon Sep 17 00:00:00 2001 From: Patryk Wlazlyn Date: Thu, 24 Oct 2024 15:17:45 +0200 Subject: [PATCH 06/16] tools/power turbostat: Force --no-perf in --dump mode Force the --no-perf early to prevent using it as a source. User asks for raw values, but perf returns them relative to the opening of the file descriptor. Signed-off-by: Patryk Wlazlyn Signed-off-by: Len Brown --- tools/power/x86/turbostat/turbostat.c | 6 ++++++ 1 file changed, 6 insertions(+) diff --git a/tools/power/x86/turbostat/turbostat.c b/tools/power/x86/turbostat/turbostat.c index b250676c174e..1fed799a5537 100644 --- a/tools/power/x86/turbostat/turbostat.c +++ b/tools/power/x86/turbostat/turbostat.c @@ -9897,6 +9897,12 @@ void cmdline(int argc, char **argv) break; case 'D': dump_only++; + /* + * Force the no_perf early to prevent using it as a source. + * User asks for raw values, but perf returns them relative + * to the opening of the file descriptor. + */ + no_perf = 1; break; case 'e': /* --enable specified counter */ -- 2.50.1 From 1da0daf746342dfdc114e4dc8fbf3ece28666d4f Mon Sep 17 00:00:00 2001 From: Patryk Wlazlyn Date: Wed, 13 Nov 2024 15:48:22 +0100 Subject: [PATCH 07/16] tools/power turbostat: Fix child's argument forwarding Add '+' to optstring when early scanning for --no-msr and --no-perf. It causes option processing to stop as soon as a nonoption argument is encountered, effectively skipping child's arguments. Fixes: 3e4048466c39 ("tools/power turbostat: Add --no-msr option") Signed-off-by: Patryk Wlazlyn Signed-off-by: Len Brown --- tools/power/x86/turbostat/turbostat.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/tools/power/x86/turbostat/turbostat.c b/tools/power/x86/turbostat/turbostat.c index 1fed799a5537..9025c2945737 100644 --- a/tools/power/x86/turbostat/turbostat.c +++ b/tools/power/x86/turbostat/turbostat.c @@ -9873,7 +9873,7 @@ void cmdline(int argc, char **argv) * Parse some options early, because they may make other options invalid, * like adding the MSR counter with --add and at the same time using --no-msr. */ - while ((opt = getopt_long_only(argc, argv, "MPn:", long_options, &option_index)) != -1) { + while ((opt = getopt_long_only(argc, argv, "+MPn:", long_options, &option_index)) != -1) { switch (opt) { case 'M': no_msr = 1; -- 2.50.1 From e5f687b89bc2a892256f48e14a568970b59a4812 Mon Sep 17 00:00:00 2001 From: Patryk Wlazlyn Date: Wed, 2 Oct 2024 15:05:15 +0200 Subject: [PATCH 08/16] tools/power turbostat: Add RAPL psys as a built-in counter Introduce the counter as a part of global, platform counters structure. We open the counter for only one cpu, but otherwise treat it as an ordinary RAPL counter, allowing for grouped perf read. The counter is disabled by default, because it's interpretation may require additional, platform specific information, making it unsuitable for general use. Signed-off-by: Patryk Wlazlyn Signed-off-by: Len Brown --- tools/power/x86/turbostat/turbostat.8 | 2 + tools/power/x86/turbostat/turbostat.c | 93 ++++++++++++++++++++++++--- 2 files changed, 85 insertions(+), 10 deletions(-) diff --git a/tools/power/x86/turbostat/turbostat.8 b/tools/power/x86/turbostat/turbostat.8 index 56c7ff6efcda..95eb02346d3a 100644 --- a/tools/power/x86/turbostat/turbostat.8 +++ b/tools/power/x86/turbostat/turbostat.8 @@ -190,6 +190,8 @@ The system configuration dump (if --quiet is not used) is followed by statistics .PP \fBRAMWatt\fP Watts consumed by the DRAM DIMMS -- available only on server processors. .PP +\fBSysWatt\fP Watts consumed by the whole platform (RAPL PSYS). Disabled by default. May require platform specific information to interpret the data, making it not suitable for general use. +.PP \fBPKG_%\fP percent of the interval that RAPL throttling was active on the Package. Note that the system summary is the sum of the package throttling time, and thus may be higher than 100% on a multi-package system. Note that the meaning of this field is model specific. For example, some hardware increments this counter when RAPL responds to thermal limits, but does not increment this counter when RAPL responds to power limits. Comparing PkgWatt and PkgTmp to system limits is necessary. .PP \fBRAM_%\fP percent of the interval that RAPL throttling was active on DRAM. diff --git a/tools/power/x86/turbostat/turbostat.c b/tools/power/x86/turbostat/turbostat.c index 9025c2945737..88c7f896c5b2 100644 --- a/tools/power/x86/turbostat/turbostat.c +++ b/tools/power/x86/turbostat/turbostat.c @@ -200,6 +200,8 @@ struct msr_counter bic[] = { { 0x0, "SAMMHz", NULL, 0, 0, 0, NULL, 0 }, { 0x0, "SAMAMHz", NULL, 0, 0, 0, NULL, 0 }, { 0x0, "Die%c6", NULL, 0, 0, 0, NULL, 0 }, + { 0x0, "SysWatt", NULL, 0, 0, 0, NULL, 0 }, + { 0x0, "Sys_J", NULL, 0, 0, 0, NULL, 0 }, }; #define MAX_BIC (sizeof(bic) / sizeof(struct msr_counter)) @@ -262,6 +264,8 @@ struct msr_counter bic[] = { #define BIC_SAMMHz (1ULL << 56) #define BIC_SAMACTMHz (1ULL << 57) #define BIC_Diec6 (1ULL << 58) +#define BIC_SysWatt (1ULL << 59) +#define BIC_Sys_J (1ULL << 60) #define BIC_TOPOLOGY (BIC_Package | BIC_Node | BIC_CoreCnt | BIC_PkgCnt | BIC_Core | BIC_CPU | BIC_Die ) #define BIC_THERMAL_PWR ( BIC_CoreTmp | BIC_PkgTmp | BIC_PkgWatt | BIC_CorWatt | BIC_GFXWatt | BIC_RAMWatt | BIC_PKG__ | BIC_RAM__) @@ -269,7 +273,7 @@ struct msr_counter bic[] = { #define BIC_IDLE (BIC_sysfs | BIC_CPU_c1 | BIC_CPU_c3 | BIC_CPU_c6 | BIC_CPU_c7 | BIC_GFX_rc6 | BIC_Pkgpc2 | BIC_Pkgpc3 | BIC_Pkgpc6 | BIC_Pkgpc7 | BIC_Pkgpc8 | BIC_Pkgpc9 | BIC_Pkgpc10 | BIC_CPU_LPI | BIC_SYS_LPI | BIC_Mod_c6 | BIC_Totl_c0 | BIC_Any_c0 | BIC_GFX_c0 | BIC_CPUGFX | BIC_SAM_mc6 | BIC_Diec6) #define BIC_OTHER ( BIC_IRQ | BIC_SMI | BIC_ThreadC | BIC_CoreTmp | BIC_IPC) -#define BIC_DISABLED_BY_DEFAULT (BIC_USEC | BIC_TOD | BIC_APIC | BIC_X2APIC) +#define BIC_DISABLED_BY_DEFAULT (BIC_USEC | BIC_TOD | BIC_APIC | BIC_X2APIC | BIC_SysWatt | BIC_Sys_J) unsigned long long bic_enabled = (0xFFFFFFFFFFFFFFFFULL & ~BIC_DISABLED_BY_DEFAULT); unsigned long long bic_present = BIC_USEC | BIC_TOD | BIC_sysfs | BIC_APIC | BIC_X2APIC; @@ -507,12 +511,15 @@ enum rapl_msrs { RAPL_AMD_PWR_UNIT = BIT(14), /* 0xc0010299 MSR_AMD_RAPL_POWER_UNIT */ RAPL_AMD_CORE_ENERGY_STAT = BIT(15), /* 0xc001029a MSR_AMD_CORE_ENERGY_STATUS */ RAPL_AMD_PKG_ENERGY_STAT = BIT(16), /* 0xc001029b MSR_AMD_PKG_ENERGY_STATUS */ + RAPL_PLATFORM_ENERGY_LIMIT = BIT(17), /* 0x64c MSR_PLATFORM_ENERGY_LIMIT */ + RAPL_PLATFORM_ENERGY_STATUS = BIT(18), /* 0x64d MSR_PLATFORM_ENERGY_STATUS */ }; #define RAPL_PKG (RAPL_PKG_ENERGY_STATUS | RAPL_PKG_POWER_LIMIT) #define RAPL_DRAM (RAPL_DRAM_ENERGY_STATUS | RAPL_DRAM_POWER_LIMIT) #define RAPL_CORE (RAPL_CORE_ENERGY_STATUS | RAPL_CORE_POWER_LIMIT) #define RAPL_GFX (RAPL_GFX_POWER_LIMIT | RAPL_GFX_ENERGY_STATUS) +#define RAPL_PSYS (RAPL_PLATFORM_ENERGY_STATUS | RAPL_PLATFORM_ENERGY_LIMIT) #define RAPL_PKG_ALL (RAPL_PKG | RAPL_PKG_PERF_STATUS | RAPL_PKG_POWER_INFO) #define RAPL_DRAM_ALL (RAPL_DRAM | RAPL_DRAM_PERF_STATUS | RAPL_DRAM_POWER_INFO) @@ -713,7 +720,7 @@ static const struct platform_features skl_features = { .has_ext_cst_msrs = 1, .trl_msrs = TRL_BASE, .tcc_offset_bits = 6, - .rapl_msrs = RAPL_PKG_ALL | RAPL_CORE_ALL | RAPL_DRAM | RAPL_DRAM_PERF_STATUS | RAPL_GFX, + .rapl_msrs = RAPL_PKG_ALL | RAPL_CORE_ALL | RAPL_DRAM | RAPL_DRAM_PERF_STATUS | RAPL_GFX | RAPL_PSYS, .enable_tsc_tweak = 1, }; @@ -730,7 +737,7 @@ static const struct platform_features cnl_features = { .has_ext_cst_msrs = 1, .trl_msrs = TRL_BASE, .tcc_offset_bits = 6, - .rapl_msrs = RAPL_PKG_ALL | RAPL_CORE_ALL | RAPL_DRAM | RAPL_DRAM_PERF_STATUS | RAPL_GFX, + .rapl_msrs = RAPL_PKG_ALL | RAPL_CORE_ALL | RAPL_DRAM | RAPL_DRAM_PERF_STATUS | RAPL_GFX | RAPL_PSYS, .enable_tsc_tweak = 1, }; @@ -797,7 +804,7 @@ static const struct platform_features icx_features = { .has_irtl_msrs = 1, .has_cst_prewake_bit = 1, .trl_msrs = TRL_BASE | TRL_CORECOUNT, - .rapl_msrs = RAPL_PKG_ALL | RAPL_DRAM_ALL, + .rapl_msrs = RAPL_PKG_ALL | RAPL_DRAM_ALL | RAPL_PSYS, .has_fixed_rapl_unit = 1, }; @@ -813,7 +820,7 @@ static const struct platform_features spr_features = { .has_irtl_msrs = 1, .has_cst_prewake_bit = 1, .trl_msrs = TRL_BASE | TRL_CORECOUNT, - .rapl_msrs = RAPL_PKG_ALL | RAPL_DRAM_ALL, + .rapl_msrs = RAPL_PKG_ALL | RAPL_DRAM_ALL | RAPL_PSYS, }; static const struct platform_features srf_features = { @@ -829,7 +836,7 @@ static const struct platform_features srf_features = { .has_irtl_msrs = 1, .has_cst_prewake_bit = 1, .trl_msrs = TRL_BASE | TRL_CORECOUNT, - .rapl_msrs = RAPL_PKG_ALL | RAPL_DRAM_ALL, + .rapl_msrs = RAPL_PKG_ALL | RAPL_DRAM_ALL | RAPL_PSYS, }; static const struct platform_features grr_features = { @@ -845,7 +852,7 @@ static const struct platform_features grr_features = { .has_irtl_msrs = 1, .has_cst_prewake_bit = 1, .trl_msrs = TRL_BASE | TRL_CORECOUNT, - .rapl_msrs = RAPL_PKG_ALL | RAPL_DRAM_ALL, + .rapl_msrs = RAPL_PKG_ALL | RAPL_DRAM_ALL | RAPL_PSYS, }; static const struct platform_features slv_features = { @@ -1108,6 +1115,7 @@ enum rapl_rci_index { RAPL_RCI_INDEX_PKG_PERF_STATUS = 4, RAPL_RCI_INDEX_DRAM_PERF_STATUS = 5, RAPL_RCI_INDEX_CORE_ENERGY = 6, + RAPL_RCI_INDEX_ENERGY_PLATFORM = 7, NUM_RAPL_COUNTERS, }; @@ -1134,6 +1142,7 @@ struct rapl_counter_info_t { struct rapl_counter_info_t *rapl_counter_info_perdomain; unsigned int rapl_counter_info_perdomain_size; +#define RAPL_COUNTER_FLAG_PLATFORM_COUNTER (1u << 0) #define RAPL_COUNTER_FLAG_USE_MSR_SUM (1u << 1) struct rapl_counter_arch_info { @@ -1255,6 +1264,19 @@ static const struct rapl_counter_arch_info rapl_counter_arch_infos[] = { .compat_scale = 1.0, .flags = 0, }, + { + .feature_mask = RAPL_PSYS, + .perf_subsys = "power", + .perf_name = "energy-psys", + .msr = MSR_PLATFORM_ENERGY_STATUS, + .msr_mask = 0x00000000FFFFFFFF, + .msr_shift = 0, + .platform_rapl_msr_scale = &rapl_energy_units, + .rci_index = RAPL_RCI_INDEX_ENERGY_PLATFORM, + .bic = BIC_SysWatt | BIC_Sys_J, + .compat_scale = 1.0, + .flags = RAPL_COUNTER_FLAG_PLATFORM_COUNTER | RAPL_COUNTER_FLAG_USE_MSR_SUM, + }, }; struct rapl_counter { @@ -1682,6 +1704,7 @@ enum { IDX_PP1_ENERGY, IDX_PKG_PERF, IDX_DRAM_PERF, + IDX_PSYS_ENERGY, IDX_COUNT, }; @@ -1726,6 +1749,9 @@ off_t idx_to_offset(int idx) case IDX_DRAM_PERF: offset = MSR_DRAM_PERF_STATUS; break; + case IDX_PSYS_ENERGY: + offset = MSR_PLATFORM_ENERGY_STATUS; + break; default: offset = -1; } @@ -1756,6 +1782,9 @@ int offset_to_idx(off_t offset) case MSR_DRAM_PERF_STATUS: idx = IDX_DRAM_PERF; break; + case MSR_PLATFORM_ENERGY_STATUS: + idx = IDX_PSYS_ENERGY; + break; default: idx = -1; } @@ -1777,6 +1806,8 @@ int idx_valid(int idx) return platform->rapl_msrs & RAPL_PKG_PERF_STATUS; case IDX_DRAM_PERF: return platform->rapl_msrs & RAPL_DRAM_PERF_STATUS; + case IDX_PSYS_ENERGY: + return platform->rapl_msrs & RAPL_PSYS; default: return 0; } @@ -1848,6 +1879,10 @@ struct system_summary { struct pkg_data packages; } average; +struct platform_counters { + struct rapl_counter energy_psys; /* MSR_PLATFORM_ENERGY_STATUS */ +} platform_counters_odd, platform_counters_even; + struct cpu_topology { int physical_package_id; int die_id; @@ -2512,6 +2547,11 @@ void print_header(char *delim) ppmt = ppmt->next; } + if (DO_BIC(BIC_SysWatt)) + outp += sprintf(outp, "%sSysWatt", (printed++ ? delim : "")); + if (DO_BIC(BIC_Sys_J)) + outp += sprintf(outp, "%sSys_J", (printed++ ? delim : "")); + outp += sprintf(outp, "\n"); } @@ -2519,6 +2559,7 @@ int dump_counters(struct thread_data *t, struct core_data *c, struct pkg_data *p { int i; struct msr_counter *mp; + struct platform_counters *pplat_cnt = p == package_odd ? &platform_counters_odd : &platform_counters_even; outp += sprintf(outp, "t %p, c %p, p %p\n", t, c, p); @@ -2590,6 +2631,7 @@ int dump_counters(struct thread_data *t, struct core_data *c, struct pkg_data *p outp += sprintf(outp, "Joules COR: %0llX\n", p->energy_cores.raw_value); outp += sprintf(outp, "Joules GFX: %0llX\n", p->energy_gfx.raw_value); outp += sprintf(outp, "Joules RAM: %0llX\n", p->energy_dram.raw_value); + outp += sprintf(outp, "Joules PSYS: %0llX\n", pplat_cnt->energy_psys.raw_value); outp += sprintf(outp, "Throttle PKG: %0llX\n", p->rapl_pkg_perf_status.raw_value); outp += sprintf(outp, "Throttle RAM: %0llX\n", p->rapl_dram_perf_status.raw_value); outp += sprintf(outp, "PTM: %dC\n", p->pkg_temp_c); @@ -2628,6 +2670,9 @@ double rapl_counter_get_value(const struct rapl_counter *c, enum rapl_unit desir */ int format_counters(struct thread_data *t, struct core_data *c, struct pkg_data *p) { + static int count; + + struct platform_counters *pplat_cnt = NULL; double interval_float, tsc; char *fmt8; int i; @@ -2637,6 +2682,11 @@ int format_counters(struct thread_data *t, struct core_data *c, struct pkg_data char *delim = "\t"; int printed = 0; + if (t == &average.threads) { + pplat_cnt = count & 1 ? &platform_counters_odd : &platform_counters_even; + ++count; + } + /* if showing only 1st thread in core and this isn't one, bail out */ if (show_core_only && !is_cpu_first_thread_in_core(t, c, p)) return 0; @@ -3093,6 +3143,13 @@ int format_counters(struct thread_data *t, struct core_data *c, struct pkg_data } } + if (DO_BIC(BIC_SysWatt) && (t == &average.threads)) + outp += sprintf(outp, fmt8, (printed++ ? delim : ""), + rapl_counter_get_value(&pplat_cnt->energy_psys, RAPL_UNIT_WATTS, interval_float)); + if (DO_BIC(BIC_Sys_J) && (t == &average.threads)) + outp += sprintf(outp, fmt8, (printed++ ? delim : ""), + rapl_counter_get_value(&pplat_cnt->energy_psys, RAPL_UNIT_JOULES, interval_float)); + done: if (*(outp - 1) != '\n') outp += sprintf(outp, "\n"); @@ -3400,6 +3457,11 @@ int delta_cpu(struct thread_data *t, struct core_data *c, return retval; } +void delta_platform(struct platform_counters *new, struct platform_counters *old) +{ + old->energy_psys.raw_value = new->energy_psys.raw_value - old->energy_psys.raw_value; +} + void rapl_counter_clear(struct rapl_counter *c) { c->raw_value = 0; @@ -4129,6 +4191,9 @@ static size_t cstate_counter_info_count_perf(const struct cstate_counter_info_t void write_rapl_counter(struct rapl_counter *rc, struct rapl_counter_info_t *rci, unsigned int idx) { + if (rci->source[idx] == COUNTER_SOURCE_NONE) + return; + rc->raw_value = rci->data[idx]; rc->unit = rci->unit[idx]; rc->scale = rci->scale[idx]; @@ -4136,6 +4201,7 @@ void write_rapl_counter(struct rapl_counter *rc, struct rapl_counter_info_t *rci int get_rapl_counters(int cpu, unsigned int domain, struct core_data *c, struct pkg_data *p) { + struct platform_counters *pplat_cnt = p == package_odd ? &platform_counters_odd : &platform_counters_even; unsigned long long perf_data[NUM_RAPL_COUNTERS + 1]; struct rapl_counter_info_t *rci; @@ -4163,6 +4229,7 @@ int get_rapl_counters(int cpu, unsigned int domain, struct core_data *c, struct for (unsigned int i = 0, pi = 1; i < NUM_RAPL_COUNTERS; ++i) { switch (rci->source[i]) { case COUNTER_SOURCE_NONE: + rci->data[i] = 0; break; case COUNTER_SOURCE_PERF: @@ -4201,7 +4268,7 @@ int get_rapl_counters(int cpu, unsigned int domain, struct core_data *c, struct } } - BUILD_BUG_ON(NUM_RAPL_COUNTERS != 7); + BUILD_BUG_ON(NUM_RAPL_COUNTERS != 8); write_rapl_counter(&p->energy_pkg, rci, RAPL_RCI_INDEX_ENERGY_PKG); write_rapl_counter(&p->energy_cores, rci, RAPL_RCI_INDEX_ENERGY_CORES); write_rapl_counter(&p->energy_dram, rci, RAPL_RCI_INDEX_DRAM); @@ -4209,6 +4276,7 @@ int get_rapl_counters(int cpu, unsigned int domain, struct core_data *c, struct write_rapl_counter(&p->rapl_pkg_perf_status, rci, RAPL_RCI_INDEX_PKG_PERF_STATUS); write_rapl_counter(&p->rapl_dram_perf_status, rci, RAPL_RCI_INDEX_DRAM_PERF_STATUS); write_rapl_counter(&c->core_energy, rci, RAPL_RCI_INDEX_CORE_ENERGY); + write_rapl_counter(&pplat_cnt->energy_psys, rci, RAPL_RCI_INDEX_ENERGY_PLATFORM); return 0; } @@ -6144,6 +6212,7 @@ restart: re_initialize(); goto restart; } + delta_platform(&platform_counters_odd, &platform_counters_even); compute_average(EVEN_COUNTERS); format_all_counters(EVEN_COUNTERS); flush_output_stdout(); @@ -6167,6 +6236,7 @@ restart: re_initialize(); goto restart; } + delta_platform(&platform_counters_even, &platform_counters_odd); compute_average(ODD_COUNTERS); format_all_counters(ODD_COUNTERS); flush_output_stdout(); @@ -6945,8 +7015,8 @@ void rapl_probe_intel(void) unsigned long long msr; unsigned int time_unit; double tdp; - const unsigned long long bic_watt_bits = BIC_PkgWatt | BIC_CorWatt | BIC_RAMWatt | BIC_GFXWatt; - const unsigned long long bic_joules_bits = BIC_Pkg_J | BIC_Cor_J | BIC_RAM_J | BIC_GFX_J; + const unsigned long long bic_watt_bits = BIC_SysWatt | BIC_PkgWatt | BIC_CorWatt | BIC_RAMWatt | BIC_GFXWatt; + const unsigned long long bic_joules_bits = BIC_Sys_J | BIC_Pkg_J | BIC_Cor_J | BIC_RAM_J | BIC_GFX_J; if (rapl_joules) bic_enabled &= ~bic_watt_bits; @@ -7606,6 +7676,9 @@ void rapl_perf_init(void) domain_visited[next_domain] = 1; + if ((cai->flags & RAPL_COUNTER_FLAG_PLATFORM_COUNTER) && (cpu != base_cpu)) + continue; + struct rapl_counter_info_t *rci = &rapl_counter_info_perdomain[next_domain]; /* Check if the counter is enabled and accessible */ -- 2.50.1 From 86d237734091201d2ab2c1d2e1063893621c770f Mon Sep 17 00:00:00 2001 From: Len Brown Date: Sat, 30 Nov 2024 16:22:00 -0500 Subject: [PATCH 09/16] tools/power turbostat: 2024.11.30 since 2024.07.26: assorted minor bug fixes assorted platform specific tweaks initial RAPL PSYS (SysWatt) support Signed-off-by: Len Brown --- tools/power/x86/turbostat/turbostat.8 | 2 +- tools/power/x86/turbostat/turbostat.c | 2 +- 2 files changed, 2 insertions(+), 2 deletions(-) diff --git a/tools/power/x86/turbostat/turbostat.8 b/tools/power/x86/turbostat/turbostat.8 index 95eb02346d3a..a7f7ed01421c 100644 --- a/tools/power/x86/turbostat/turbostat.8 +++ b/tools/power/x86/turbostat/turbostat.8 @@ -190,7 +190,7 @@ The system configuration dump (if --quiet is not used) is followed by statistics .PP \fBRAMWatt\fP Watts consumed by the DRAM DIMMS -- available only on server processors. .PP -\fBSysWatt\fP Watts consumed by the whole platform (RAPL PSYS). Disabled by default. May require platform specific information to interpret the data, making it not suitable for general use. +\fBSysWatt\fP Watts consumed by the whole platform (RAPL PSYS). Disabled by default. Enable with --enable SysWatt. .PP \fBPKG_%\fP percent of the interval that RAPL throttling was active on the Package. Note that the system summary is the sum of the package throttling time, and thus may be higher than 100% on a multi-package system. Note that the meaning of this field is model specific. For example, some hardware increments this counter when RAPL responds to thermal limits, but does not increment this counter when RAPL responds to power limits. Comparing PkgWatt and PkgTmp to system limits is necessary. .PP diff --git a/tools/power/x86/turbostat/turbostat.c b/tools/power/x86/turbostat/turbostat.c index 88c7f896c5b2..58a487c225a7 100644 --- a/tools/power/x86/turbostat/turbostat.c +++ b/tools/power/x86/turbostat/turbostat.c @@ -9236,7 +9236,7 @@ int get_and_dump_counters(void) void print_version() { - fprintf(outf, "turbostat version 2024.07.26 - Len Brown \n"); + fprintf(outf, "turbostat version 2024.11.30 - Len Brown \n"); } #define COMMAND_LINE_SIZE 2048 -- 2.50.1 From f69e63756f7822fcdad8a34f9967e8b243e883ee Mon Sep 17 00:00:00 2001 From: "Dr. David Alan Gilbert" Date: Wed, 2 Oct 2024 18:31:47 +0100 Subject: [PATCH 10/16] printf: Remove unused 'bprintf' bprintf() is unused. Remove it. It was added in the commit 4370aa4aa753 ("vsprintf: add binary printf") but as far as I can see was never used, unlike the other two functions in that patch. Link: https://lore.kernel.org/20241002173147.210107-1-linux@treblig.org Reviewed-by: Andy Shevchenko Acked-by: Petr Mladek Signed-off-by: Dr. David Alan Gilbert Signed-off-by: Steven Rostedt (Google) --- include/linux/string.h | 1 - lib/vsprintf.c | 23 ----------------------- 2 files changed, 24 deletions(-) diff --git a/include/linux/string.h b/include/linux/string.h index 0dd27afcfaf7..493ac4862c77 100644 --- a/include/linux/string.h +++ b/include/linux/string.h @@ -335,7 +335,6 @@ int __sysfs_match_string(const char * const *array, size_t n, const char *s); #ifdef CONFIG_BINARY_PRINTF int vbin_printf(u32 *bin_buf, size_t size, const char *fmt, va_list args); int bstr_printf(char *buf, size_t size, const char *fmt, const u32 *bin_buf); -int bprintf(u32 *bin_buf, size_t size, const char *fmt, ...) __printf(3, 4); #endif extern ssize_t memory_read_from_buffer(void *to, size_t count, loff_t *ppos, diff --git a/lib/vsprintf.c b/lib/vsprintf.c index 6ac02bbb7df1..9d3dac38a3f4 100644 --- a/lib/vsprintf.c +++ b/lib/vsprintf.c @@ -3428,29 +3428,6 @@ out: } EXPORT_SYMBOL_GPL(bstr_printf); -/** - * bprintf - Parse a format string and place args' binary value in a buffer - * @bin_buf: The buffer to place args' binary value - * @size: The size of the buffer(by words(32bits), not characters) - * @fmt: The format string to use - * @...: Arguments for the format string - * - * The function returns the number of words(u32) written - * into @bin_buf. - */ -int bprintf(u32 *bin_buf, size_t size, const char *fmt, ...) -{ - va_list args; - int ret; - - va_start(args, fmt); - ret = vbin_printf(bin_buf, size, fmt, args); - va_end(args); - - return ret; -} -EXPORT_SYMBOL_GPL(bprintf); - #endif /* CONFIG_BINARY_PRINTF */ /** -- 2.50.1 From 9022ed0e7e65734d83a0648648589b9fbea8e8c9 Mon Sep 17 00:00:00 2001 From: Linus Torvalds Date: Sun, 1 Dec 2024 09:23:33 -0800 Subject: [PATCH 11/16] strscpy: write destination buffer only once The point behind strscpy() was to once and for all avoid all the problems with 'strncpy()' and later broken "fixed" versions like strlcpy() that just made things worse. So strscpy not only guarantees NUL-termination (unlike strncpy), it also doesn't do unnecessary padding at the destination. But at the same time also avoids byte-at-a-time reads and writes by _allowing_ some extra NUL writes - within the size, of course - so that the whole copy can be done with word operations. It is also stable in the face of a mutable source string: it explicitly does not read the source buffer multiple times (so an implementation using "strnlen()+memcpy()" would be wrong), and does not read the source buffer past the size (like the mis-design that is strlcpy does). Finally, the return value is designed to be simple and unambiguous: if the string cannot be copied fully, it returns an actual negative error, making error handling clearer and simpler (and the caller already knows the size of the buffer). Otherwise it returns the string length of the result. However, there was one final stability issue that can be important to callers: the stability of the destination buffer. In particular, the same way we shouldn't read the source buffer more than once, we should avoid doing multiple writes to the destination buffer: first writing a potentially non-terminated string, and then terminating it with NUL at the end does not result in a stable result buffer. Yes, it gives the right result in the end, but if the rule for the destination buffer was that it is _always_ NUL-terminated even when accessed concurrently with updates, the final byte of the buffer needs to always _stay_ as a NUL byte. [ Note that "final byte is NUL" here is literally about the final byte in the destination array, not the terminating NUL at the end of the string itself. There is no attempt to try to make concurrent reads and writes give any kind of consistent string length or contents, but we do want to guarantee that there is always at least that final terminating NUL character at the end of the destination array if it existed before ] This is relevant in the kernel for the tsk->comm[] array, for example. Even without locking (for either readers or writers), we want to know that while the buffer contents may be garbled, it is always a valid C string and always has a NUL character at 'comm[TASK_COMM_LEN-1]' (and never has any "out of thin air" data). So avoid any "copy possibly non-terminated string, and terminate later" behavior, and write the destination buffer only once. Signed-off-by: Linus Torvalds --- lib/string.c | 23 +++++++++++++++++------ 1 file changed, 17 insertions(+), 6 deletions(-) diff --git a/lib/string.c b/lib/string.c index 76327b51e36f..eb4486ed40d2 100644 --- a/lib/string.c +++ b/lib/string.c @@ -104,6 +104,12 @@ char *strncpy(char *dest, const char *src, size_t count) EXPORT_SYMBOL(strncpy); #endif +#ifdef __BIG_ENDIAN +# define ALLBUTLAST_BYTE_MASK (~255ul) +#else +# define ALLBUTLAST_BYTE_MASK (~0ul >> 8) +#endif + ssize_t sized_strscpy(char *dest, const char *src, size_t count) { const struct word_at_a_time constants = WORD_AT_A_TIME_CONSTANTS; @@ -147,13 +153,18 @@ ssize_t sized_strscpy(char *dest, const char *src, size_t count) *(unsigned long *)(dest+res) = c & zero_bytemask(data); return res + find_zero(data); } + count -= sizeof(unsigned long); + if (unlikely(!count)) { + c &= ALLBUTLAST_BYTE_MASK; + *(unsigned long *)(dest+res) = c; + return -E2BIG; + } *(unsigned long *)(dest+res) = c; res += sizeof(unsigned long); - count -= sizeof(unsigned long); max -= sizeof(unsigned long); } - while (count) { + while (count > 1) { char c; c = src[res]; @@ -164,11 +175,11 @@ ssize_t sized_strscpy(char *dest, const char *src, size_t count) count--; } - /* Hit buffer length without finding a NUL; force NUL-termination. */ - if (res) - dest[res-1] = '\0'; + /* Force NUL-termination. */ + dest[res] = '\0'; - return -E2BIG; + /* Return E2BIG if the source didn't stop */ + return src[res] ? -E2BIG : res; } EXPORT_SYMBOL(sized_strscpy); -- 2.50.1 From 40384c840ea1944d7c5a392e8975ed088ecf0b37 Mon Sep 17 00:00:00 2001 From: Linus Torvalds Date: Sun, 1 Dec 2024 14:28:56 -0800 Subject: [PATCH 12/16] Linux 6.13-rc1 --- Makefile | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/Makefile b/Makefile index e34a97473fb6..93ab62cef244 100644 --- a/Makefile +++ b/Makefile @@ -1,8 +1,8 @@ # SPDX-License-Identifier: GPL-2.0 VERSION = 6 -PATCHLEVEL = 12 +PATCHLEVEL = 13 SUBLEVEL = 0 -EXTRAVERSION = +EXTRAVERSION = -rc1 NAME = Baby Opossum Posse # *DOCUMENTATION* -- 2.50.1 From def13795928b82ecca4ae3ea16f375114ff88685 Mon Sep 17 00:00:00 2001 From: Andrew Morton Date: Thu, 14 Nov 2024 15:44:21 -0800 Subject: [PATCH 13/16] fs/proc/vmcore.c: fix warning when CONFIG_MMU=n >> fs/proc/vmcore.c:424:19: warning: 'mmap_vmcore_fault' defined but not used [-Wunused-function] 424 | static vm_fault_t mmap_vmcore_fault(struct vm_fault *vmf) Reported-by: kernel test robot Closes: https://lore.kernel.org/oe-kbuild-all/202411140156.2o0nS4fl-lkp@intel.com/ Cc: Qi Xi Signed-off-by: Andrew Morton --- fs/proc/vmcore.c | 56 ++++++++++++++++++++++++------------------------ 1 file changed, 28 insertions(+), 28 deletions(-) diff --git a/fs/proc/vmcore.c b/fs/proc/vmcore.c index b4521b096058..3d8a82cee63e 100644 --- a/fs/proc/vmcore.c +++ b/fs/proc/vmcore.c @@ -414,6 +414,34 @@ static ssize_t read_vmcore(struct kiocb *iocb, struct iov_iter *iter) return __read_vmcore(iter, &iocb->ki_pos); } +/** + * vmcore_alloc_buf - allocate buffer in vmalloc memory + * @size: size of buffer + * + * If CONFIG_MMU is defined, use vmalloc_user() to allow users to mmap + * the buffer to user-space by means of remap_vmalloc_range(). + * + * If CONFIG_MMU is not defined, use vzalloc() since mmap_vmcore() is + * disabled and there's no need to allow users to mmap the buffer. + */ +static inline char *vmcore_alloc_buf(size_t size) +{ +#ifdef CONFIG_MMU + return vmalloc_user(size); +#else + return vzalloc(size); +#endif +} + +/* + * Disable mmap_vmcore() if CONFIG_MMU is not defined. MMU is + * essential for mmap_vmcore() in order to map physically + * non-contiguous objects (ELF header, ELF note segment and memory + * regions in the 1st kernel pointed to by PT_LOAD entries) into + * virtually contiguous user-space in ELF layout. + */ +#ifdef CONFIG_MMU + /* * The vmcore fault handler uses the page cache and fills data using the * standard __read_vmcore() function. @@ -457,34 +485,6 @@ static vm_fault_t mmap_vmcore_fault(struct vm_fault *vmf) #endif } -/** - * vmcore_alloc_buf - allocate buffer in vmalloc memory - * @size: size of buffer - * - * If CONFIG_MMU is defined, use vmalloc_user() to allow users to mmap - * the buffer to user-space by means of remap_vmalloc_range(). - * - * If CONFIG_MMU is not defined, use vzalloc() since mmap_vmcore() is - * disabled and there's no need to allow users to mmap the buffer. - */ -static inline char *vmcore_alloc_buf(size_t size) -{ -#ifdef CONFIG_MMU - return vmalloc_user(size); -#else - return vzalloc(size); -#endif -} - -/* - * Disable mmap_vmcore() if CONFIG_MMU is not defined. MMU is - * essential for mmap_vmcore() in order to map physically - * non-contiguous objects (ELF header, ELF note segment and memory - * regions in the 1st kernel pointed to by PT_LOAD entries) into - * virtually contiguous user-space in ELF layout. - */ -#ifdef CONFIG_MMU - static const struct vm_operations_struct vmcore_mmap_ops = { .fault = mmap_vmcore_fault, }; -- 2.50.1 From a1268be280d8e484ab3606d7476edd0f14bb9961 Mon Sep 17 00:00:00 2001 From: John Hubbard Date: Wed, 20 Nov 2024 19:49:33 -0800 Subject: [PATCH 14/16] mm/gup: handle NULL pages in unpin_user_pages() The recent addition of "pofs" (pages or folios) handling to gup has a flaw: it assumes that unpin_user_pages() handles NULL pages in the pages** array. That's not the case, as I discovered when I ran on a new configuration on my test machine. Fix this by skipping NULL pages in unpin_user_pages(), just like unpin_folios() already does. Details: when booting on x86 with "numa=fake=2 movablecore=4G" on Linux 6.12, and running this: tools/testing/selftests/mm/gup_longterm ...I get the following crash: BUG: kernel NULL pointer dereference, address: 0000000000000008 RIP: 0010:sanity_check_pinned_pages+0x3a/0x2d0 ... Call Trace: ? __die_body+0x66/0xb0 ? page_fault_oops+0x30c/0x3b0 ? do_user_addr_fault+0x6c3/0x720 ? irqentry_enter+0x34/0x60 ? exc_page_fault+0x68/0x100 ? asm_exc_page_fault+0x22/0x30 ? sanity_check_pinned_pages+0x3a/0x2d0 unpin_user_pages+0x24/0xe0 check_and_migrate_movable_pages_or_folios+0x455/0x4b0 __gup_longterm_locked+0x3bf/0x820 ? mmap_read_lock_killable+0x12/0x50 ? __pfx_mmap_read_lock_killable+0x10/0x10 pin_user_pages+0x66/0xa0 gup_test_ioctl+0x358/0xb20 __se_sys_ioctl+0x6b/0xc0 do_syscall_64+0x7b/0x150 entry_SYSCALL_64_after_hwframe+0x76/0x7e Link: https://lkml.kernel.org/r/20241121034933.77502-1-jhubbard@nvidia.com Fixes: 94efde1d1539 ("mm/gup: avoid an unnecessary allocation call for FOLL_LONGTERM cases") Signed-off-by: John Hubbard Acked-by: David Hildenbrand Cc: Oscar Salvador Cc: Vivek Kasireddy Cc: Dave Airlie Cc: Gerd Hoffmann Cc: Matthew Wilcox Cc: Christoph Hellwig Cc: Jason Gunthorpe Cc: Peter Xu Cc: Arnd Bergmann Cc: Daniel Vetter Cc: Dongwon Kim Cc: Hugh Dickins Cc: Junxiao Chang Cc: Signed-off-by: Andrew Morton --- mm/gup.c | 11 ++++++++++- 1 file changed, 10 insertions(+), 1 deletion(-) diff --git a/mm/gup.c b/mm/gup.c index 746070a1d8bf..3b75e631f369 100644 --- a/mm/gup.c +++ b/mm/gup.c @@ -52,7 +52,12 @@ static inline void sanity_check_pinned_pages(struct page **pages, */ for (; npages; npages--, pages++) { struct page *page = *pages; - struct folio *folio = page_folio(page); + struct folio *folio; + + if (!page) + continue; + + folio = page_folio(page); if (is_zero_page(page) || !folio_test_anon(folio)) @@ -409,6 +414,10 @@ void unpin_user_pages(struct page **pages, unsigned long npages) sanity_check_pinned_pages(pages, npages); for (i = 0; i < npages; i += nr) { + if (!pages[i]) { + nr = 1; + continue; + } folio = gup_folio_next(pages, npages, i, &nr); gup_put_folio(folio, nr, FOLL_PIN); } -- 2.50.1 From 091c1dd2d4df6edd1beebe0e5863d4034ade9572 Mon Sep 17 00:00:00 2001 From: David Hildenbrand Date: Wed, 20 Nov 2024 21:11:51 +0100 Subject: [PATCH 15/16] mm/mempolicy: fix migrate_to_node() assuming there is at least one VMA in a MM We currently assume that there is at least one VMA in a MM, which isn't true. So we might end up having find_vma() return NULL, to then de-reference NULL. So properly handle find_vma() returning NULL. This fixes the report: Oops: general protection fault, probably for non-canonical address 0xdffffc0000000000: 0000 [#1] PREEMPT SMP KASAN PTI KASAN: null-ptr-deref in range [0x0000000000000000-0x0000000000000007] CPU: 1 UID: 0 PID: 6021 Comm: syz-executor284 Not tainted 6.12.0-rc7-syzkaller-00187-gf868cd251776 #0 Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS Google 10/30/2024 RIP: 0010:migrate_to_node mm/mempolicy.c:1090 [inline] RIP: 0010:do_migrate_pages+0x403/0x6f0 mm/mempolicy.c:1194 Code: ... RSP: 0018:ffffc9000375fd08 EFLAGS: 00010246 RAX: 0000000000000000 RBX: ffffc9000375fd78 RCX: 0000000000000000 RDX: ffff88807e171300 RSI: dffffc0000000000 RDI: ffff88803390c044 RBP: ffff88807e171428 R08: 0000000000000014 R09: fffffbfff2039ef1 R10: ffffffff901cf78f R11: 0000000000000000 R12: 0000000000000003 R13: ffffc9000375fe90 R14: ffffc9000375fe98 R15: ffffc9000375fdf8 FS: 00005555919e1380(0000) GS:ffff8880b8700000(0000) knlGS:0000000000000000 CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 CR2: 00005555919e1ca8 CR3: 000000007f12a000 CR4: 00000000003526f0 DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000 DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400 Call Trace: kernel_migrate_pages+0x5b2/0x750 mm/mempolicy.c:1709 __do_sys_migrate_pages mm/mempolicy.c:1727 [inline] __se_sys_migrate_pages mm/mempolicy.c:1723 [inline] __x64_sys_migrate_pages+0x96/0x100 mm/mempolicy.c:1723 do_syscall_x64 arch/x86/entry/common.c:52 [inline] do_syscall_64+0xcd/0x250 arch/x86/entry/common.c:83 entry_SYSCALL_64_after_hwframe+0x77/0x7f [akpm@linux-foundation.org: add unlikely()] Link: https://lkml.kernel.org/r/20241120201151.9518-1-david@redhat.com Fixes: 39743889aaf7 ("[PATCH] Swap Migration V5: sys_migrate_pages interface") Signed-off-by: David Hildenbrand Reported-by: syzbot+3511625422f7aa637f0d@syzkaller.appspotmail.com Closes: https://lore.kernel.org/lkml/673d2696.050a0220.3c9d61.012f.GAE@google.com/T/ Reviewed-by: Liam R. Howlett Reviewed-by: Christoph Lameter Cc: Liam R. Howlett Cc: Signed-off-by: Andrew Morton --- mm/mempolicy.c | 4 ++++ 1 file changed, 4 insertions(+) diff --git a/mm/mempolicy.c b/mm/mempolicy.c index bb37cd1a51d8..04f35659717a 100644 --- a/mm/mempolicy.c +++ b/mm/mempolicy.c @@ -1080,6 +1080,10 @@ static long migrate_to_node(struct mm_struct *mm, int source, int dest, mmap_read_lock(mm); vma = find_vma(mm, 0); + if (unlikely(!vma)) { + mmap_read_unlock(mm); + return 0; + } /* * This does not migrate the range, but isolates all pages that -- 2.50.1 From e30a0361b8515d424c73c67de1a43e45a13b8ba2 Mon Sep 17 00:00:00 2001 From: Jared Kangas Date: Tue, 19 Nov 2024 13:02:34 -0800 Subject: [PATCH 16/16] kasan: make report_lock a raw spinlock If PREEMPT_RT is enabled, report_lock is a sleeping spinlock and must not be locked when IRQs are disabled. However, KASAN reports may be triggered in such contexts. For example: char *s = kzalloc(1, GFP_KERNEL); kfree(s); local_irq_disable(); char c = *s; /* KASAN report here leads to spin_lock() */ local_irq_enable(); Make report_spinlock a raw spinlock to prevent rescheduling when PREEMPT_RT is enabled. Link: https://lkml.kernel.org/r/20241119210234.1602529-1-jkangas@redhat.com Fixes: 342a93247e08 ("locking/spinlock: Provide RT variant header: ") Signed-off-by: Jared Kangas Cc: Alexander Potapenko Cc: Andrey Konovalov Cc: Andrey Ryabinin Cc: Dmitry Vyukov Cc: Vincenzo Frascino Cc: Signed-off-by: Andrew Morton --- mm/kasan/report.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) diff --git a/mm/kasan/report.c b/mm/kasan/report.c index 50fb19ad4388..3fe77a360f1c 100644 --- a/mm/kasan/report.c +++ b/mm/kasan/report.c @@ -201,7 +201,7 @@ static inline void fail_non_kasan_kunit_test(void) { } #endif /* CONFIG_KUNIT */ -static DEFINE_SPINLOCK(report_lock); +static DEFINE_RAW_SPINLOCK(report_lock); static void start_report(unsigned long *flags, bool sync) { @@ -212,7 +212,7 @@ static void start_report(unsigned long *flags, bool sync) lockdep_off(); /* Make sure we don't end up in loop. */ report_suppress_start(); - spin_lock_irqsave(&report_lock, *flags); + raw_spin_lock_irqsave(&report_lock, *flags); pr_err("==================================================================\n"); } @@ -222,7 +222,7 @@ static void end_report(unsigned long *flags, const void *addr, bool is_write) trace_error_report_end(ERROR_DETECTOR_KASAN, (unsigned long)addr); pr_err("==================================================================\n"); - spin_unlock_irqrestore(&report_lock, *flags); + raw_spin_unlock_irqrestore(&report_lock, *flags); if (!test_bit(KASAN_BIT_MULTI_SHOT, &kasan_flags)) check_panic_on_warn("KASAN"); switch (kasan_arg_fault) { -- 2.50.1