Message-ID: <9f3bb60b-38da-1ac1-59bd-6f147309183c@amd.com>
Date:   Mon, 26 Sep 2022 08:29:58 +0200
From:   Christian König <christian.koenig@....com>
To:     Daniel Gomez <daniel@...c.com>
Cc:     dagmcr@...il.com, Harry Wentland <harry.wentland@....com>,
        Leo Li <sunpeng.li@....com>,
        Rodrigo Siqueira <Rodrigo.Siqueira@....com>,
        Alex Deucher <alexander.deucher@....com>,
        "Pan, Xinhui" <Xinhui.Pan@....com>,
        David Airlie <airlied@...il.com>,
        Daniel Vetter <daniel@...ll.ch>,
        Anthony Koo <Anthony.Koo@....com>,
        Alex Hung <alex.hung@....com>,
        Yi-Ling Chen <Yi-Ling.Chen2@....com>,
        Roman Li <Roman.Li@....com>, Melissa Wen <mwen@...lia.com>,
        David Zhang <dingchen.zhang@....com>,
        Agustin Gutierrez <agustin.gutierrez@....com>,
        Duncan Ma <duncan.ma@....com>,
        "Leo (Hanghong) Ma" <hanghong.ma@....com>,
        Isabella Basso <isabbasso@...eup.net>,
        Sung Joon Kim <Sungjoon.Kim@....com>,
        Alan Liu <HaoPing.Liu@....com>, Becle Lee <becle.lee@....com>,
        Bernard Zhao <bernard@...o.com>,
        Bas Nieuwenhuizen <bas@...nieuwenhuizen.nl>,
        Zhan Liu <Zhan.Liu@....com>,
        Ahmad Othman <Ahmad.Othman@....com>,
        amd-gfx@...ts.freedesktop.org, dri-devel@...ts.freedesktop.org,
        linux-kernel@...r.kernel.org
Subject: Re: [PATCH] drm/amd/display: Fix mutex lock in dcn10

Am 25.09.22 um 23:53 schrieb Daniel Gomez:
> Removal of DC_FP_* wrappers from dml (9696679bf7ac) provokes a mutex
> lock [2] on the amdgpu driver. Re-arrange the dcn10 code to avoid
> locking the mutex by placing the DC_FP_* wrappers around the proper
> functions.

Of hand that looks correct to me now, but our DC team needs to take a 
closer look.

Regards,
Christian.

>
> This fixes the following WARN/stacktrace:
>
> BUG: sleeping function called from invalid context at kernel/locking/mutex.c:283
> in_atomic(): 1, irqs_disabled(): 0, non_block: 0, pid: 227, name: systemd-udevd
> preempt_count: 1, expected: 0
> CPU: 4 PID: 227 Comm: systemd-udevd Not tainted 6.0.0-rc6-qtec-standard #2
> Hardware name: Qtechnology A/S QT5222/QT5221, BIOS v1.0.1 06/07/2021
> Call Trace:
>   <TASK>
>   dump_stack_lvl+0x33/0x42
>   __might_resched.cold.172+0xa5/0xb3
>   mutex_lock+0x1a/0x40
>   amdgpu_dpm_get_clock_by_type_with_voltage+0x38/0x70 [amdgpu]
>   dm_pp_get_clock_levels_by_type_with_voltage+0x64/0xa0 [amdgpu]
>   dcn_bw_update_from_pplib+0x70/0x340 [amdgpu]
>   dcn10_create_resource_pool+0x8c8/0xd20 [amdgpu]
>   ? __kmalloc+0x1c7/0x4a0
>   dc_create_resource_pool+0xe7/0x190 [amdgpu]
>   dc_create+0x212/0x5d0 [amdgpu]
>   amdgpu_dm_init+0x246/0x370 [amdgpu]
>   ? schedule_hrtimeout_range_clock+0x93/0x120
>   ? phm_wait_for_register_unequal.part.1+0x4a/0x80 [amdgpu]
>   dm_hw_init+0xe/0x20 [amdgpu]
>   amdgpu_device_init.cold.56+0x1324/0x1653 [amdgpu]
>   ? pci_bus_read_config_word+0x43/0x80
>   amdgpu_driver_load_kms+0x15/0x120 [amdgpu]
>   amdgpu_pci_probe+0x116/0x320 [amdgpu]
>   pci_device_probe+0x97/0x110
>   really_probe+0xdd/0x340
>   __driver_probe_device+0x80/0x170
>   driver_probe_device+0x1f/0x90
>   __driver_attach+0xdc/0x180
>   ? __device_attach_driver+0x100/0x100
>   ? __device_attach_driver+0x100/0x100
>   bus_for_each_dev+0x74/0xc0
>   bus_add_driver+0x19e/0x210
>   ? kset_find_obj+0x30/0xa0
>   ? 0xffffffffa0a5b000
>   driver_register+0x6b/0xc0
>   ? 0xffffffffa0a5b000
>   do_one_initcall+0x4a/0x1f0
>   ? __vunmap+0x28e/0x2f0
>   ? __cond_resched+0x15/0x30
>   ? kmem_cache_alloc_trace+0x3d/0x440
>   do_init_module+0x4a/0x1e0
>   load_module+0x1cba/0x1e10
>   ? __do_sys_finit_module+0xb7/0x120
>   __do_sys_finit_module+0xb7/0x120
>   do_syscall_64+0x3c/0x80
>   entry_SYSCALL_64_after_hwframe+0x63/0xcd
> RIP: 0033:0x7ff2b5f5422d
> Code: 5d c3 66 2e 0f 1f 84 00 00 00 00 00 90 f3 0f 1e fa 48 89 f8 48
> 89 f7 48 89 d6 48 89 ca 4d 89 c2 4d 89 c8 4c 8b 4c 24 08 0f 05 <48>
> 3d 01 f0 ff ff 73 01 c3 48 8b 0d c3 ab 0e 00 f7 d8 64 89 01 48
> RSP: 002b:00007ffc44ab28e8 EFLAGS: 00000246 ORIG_RAX: 0000000000000139
> RAX: ffffffffffffffda RBX: 0000555c566a9240 RCX: 00007ff2b5f5422d
> RDX: 0000000000000000 RSI: 00007ff2b60bb353 RDI: 0000000000000019
> RBP: 00007ff2b60bb353 R08: 0000000000000000 R09: 0000555c566a9240
> R10: 0000000000000019 R11: 0000000000000246 R12: 0000000000000000
> R13: 0000000000020000 R14: 0000000000000000 R15: 0000000000000000
> </TASK>
>
> Fixes: 9696679bf7ac ("drm/amd/display: remove DC_FP_* wrapper from
> dml folder")
> Signed-off-by: Daniel Gomez <daniel@...c.com>
> ---
>   .../amd/display/dc/dcn10/dcn10_hw_sequencer.c |  12 +-
>   .../drm/amd/display/dc/dcn10/dcn10_resource.c |  66 +++++++++-
>   .../drm/amd/display/dc/dml/calcs/dcn_calcs.c  | 118 ++++++++----------
>   .../gpu/drm/amd/display/dc/inc/dcn_calcs.h    |  19 ++-
>   4 files changed, 138 insertions(+), 77 deletions(-)
>
> diff --git a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hw_sequencer.c b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hw_sequencer.c
> index 5b5d952b2b8c..cb1e06d62841 100644
> --- a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hw_sequencer.c
> +++ b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hw_sequencer.c
> @@ -2994,6 +2994,7 @@ void dcn10_prepare_bandwidth(
>   {
>   	struct dce_hwseq *hws = dc->hwseq;
>   	struct hubbub *hubbub = dc->res_pool->hubbub;
> +	int min_fclk_khz, min_dcfclk_khz, socclk_khz;
>   
>   	if (dc->debug.sanity_checks)
>   		hws->funcs.verify_allow_pstate_change_high(dc);
> @@ -3016,8 +3017,11 @@ void dcn10_prepare_bandwidth(
>   
>   	if (dc->debug.pplib_wm_report_mode == WM_REPORT_OVERRIDE) {
>   		DC_FP_START();
> -		dcn_bw_notify_pplib_of_wm_ranges(dc);
> +		dcn_get_soc_clks(
> +			dc, &min_fclk_khz, &min_dcfclk_khz, &socclk_khz);
>   		DC_FP_END();
> +		dcn_bw_notify_pplib_of_wm_ranges(
> +			dc, min_fclk_khz, min_dcfclk_khz, socclk_khz);
>   	}
>   
>   	if (dc->debug.sanity_checks)
> @@ -3030,6 +3034,7 @@ void dcn10_optimize_bandwidth(
>   {
>   	struct dce_hwseq *hws = dc->hwseq;
>   	struct hubbub *hubbub = dc->res_pool->hubbub;
> +	int min_fclk_khz, min_dcfclk_khz, socclk_khz;
>   
>   	if (dc->debug.sanity_checks)
>   		hws->funcs.verify_allow_pstate_change_high(dc);
> @@ -3053,8 +3058,11 @@ void dcn10_optimize_bandwidth(
>   
>   	if (dc->debug.pplib_wm_report_mode == WM_REPORT_OVERRIDE) {
>   		DC_FP_START();
> -		dcn_bw_notify_pplib_of_wm_ranges(dc);
> +		dcn_get_soc_clks(
> +			dc, &min_fclk_khz, &min_dcfclk_khz, &socclk_khz);
>   		DC_FP_END();
> +		dcn_bw_notify_pplib_of_wm_ranges(
> +			dc, min_fclk_khz, min_dcfclk_khz, socclk_khz);
>   	}
>   
>   	if (dc->debug.sanity_checks)
> diff --git a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_resource.c b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_resource.c
> index 174eebbe8b4f..a18a5b56ca7d 100644
> --- a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_resource.c
> +++ b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_resource.c
> @@ -1336,6 +1336,21 @@ static noinline void dcn10_resource_construct_fp(
>   	}
>   }
>   
> +static bool verify_clock_values(struct dm_pp_clock_levels_with_voltage *clks)
> +{
> +	int i;
> +
> +	if (clks->num_levels == 0)
> +		return false;
> +
> +	for (i = 0; i < clks->num_levels; i++)
> +		/* Ensure that the result is sane */
> +		if (clks->data[i].clocks_in_khz == 0)
> +			return false;
> +
> +	return true;
> +}
> +
>   static bool dcn10_resource_construct(
>   	uint8_t num_virtual_links,
>   	struct dc *dc,
> @@ -1345,6 +1360,9 @@ static bool dcn10_resource_construct(
>   	int j;
>   	struct dc_context *ctx = dc->ctx;
>   	uint32_t pipe_fuses = read_pipe_fuses(ctx);
> +	struct dm_pp_clock_levels_with_voltage fclks = {0}, dcfclks = {0};
> +	int min_fclk_khz, min_dcfclk_khz, socclk_khz;
> +	bool res;
>   
>   	ctx->dc_bios->regs = &bios_regs;
>   
> @@ -1505,15 +1523,53 @@ static bool dcn10_resource_construct(
>   			&& pool->base.pp_smu->rv_funcs.set_pme_wa_enable != NULL)
>   		dc->debug.az_endpoint_mute_only = false;
>   
> -	DC_FP_START();
> -	if (!dc->debug.disable_pplib_clock_request)
> -		dcn_bw_update_from_pplib(dc);
> +
> +	if (!dc->debug.disable_pplib_clock_request) {
> +		/*
> +		 * TODO: This is not the proper way to obtain
> +		 * fabric_and_dram_bandwidth, should be min(fclk, memclk).
> +		 */
> +		res = dm_pp_get_clock_levels_by_type_with_voltage(
> +				ctx, DM_PP_CLOCK_TYPE_FCLK, &fclks);
> +
> +		DC_FP_START();
> +
> +		if (res)
> +			res = verify_clock_values(&fclks);
> +
> +		if (res)
> +			dcn_bw_update_from_pplib_fclks(dc, &fclks);
> +		else
> +			BREAK_TO_DEBUGGER();
> +
> +		DC_FP_END();
> +
> +		res = dm_pp_get_clock_levels_by_type_with_voltage(
> +			ctx, DM_PP_CLOCK_TYPE_DCFCLK, &dcfclks);
> +
> +		DC_FP_START();
> +
> +		if (res)
> +			res = verify_clock_values(&dcfclks);
> +
> +		if (res)
> +			dcn_bw_update_from_pplib_dcfclks(dc, &dcfclks);
> +		else
> +			BREAK_TO_DEBUGGER();
> +
> +		DC_FP_END();
> +	}
> +
>   	dcn_bw_sync_calcs_and_dml(dc);
>   	if (!dc->debug.disable_pplib_wm_range) {
>   		dc->res_pool = &pool->base;
> -		dcn_bw_notify_pplib_of_wm_ranges(dc);
> +		DC_FP_START();
> +		dcn_get_soc_clks(
> +			dc, &min_fclk_khz, &min_dcfclk_khz, &socclk_khz);
> +		DC_FP_END();
> +		dcn_bw_notify_pplib_of_wm_ranges(
> +			dc, min_fclk_khz, min_dcfclk_khz, socclk_khz);
>   	}
> -	DC_FP_END();
>   
>   	{
>   		struct irq_service_init_data init_data;
> diff --git a/drivers/gpu/drm/amd/display/dc/dml/calcs/dcn_calcs.c b/drivers/gpu/drm/amd/display/dc/dml/calcs/dcn_calcs.c
> index db3b16b77034..7d3394470352 100644
> --- a/drivers/gpu/drm/amd/display/dc/dml/calcs/dcn_calcs.c
> +++ b/drivers/gpu/drm/amd/display/dc/dml/calcs/dcn_calcs.c
> @@ -1464,81 +1464,67 @@ unsigned int dcn_find_dcfclk_suits_all(
>   	return dcf_clk;
>   }
>   
> -static bool verify_clock_values(struct dm_pp_clock_levels_with_voltage *clks)
> +void dcn_bw_update_from_pplib_fclks(
> +	struct dc *dc,
> +	struct dm_pp_clock_levels_with_voltage *fclks)
>   {
> -	int i;
> -
> -	if (clks->num_levels == 0)
> -		return false;
> -
> -	for (i = 0; i < clks->num_levels; i++)
> -		/* Ensure that the result is sane */
> -		if (clks->data[i].clocks_in_khz == 0)
> -			return false;
> +	unsigned vmin0p65_idx, vmid0p72_idx, vnom0p8_idx, vmax0p9_idx;
>   
> -	return true;
> +	ASSERT(fclks->num_levels);
> +
> +	vmin0p65_idx = 0;
> +	vmid0p72_idx = fclks->num_levels -
> +		(fclks->num_levels > 2 ? 3 : (fclks->num_levels > 1 ? 2 : 1));
> +	vnom0p8_idx = fclks->num_levels - (fclks->num_levels > 1 ? 2 : 1);
> +	vmax0p9_idx = fclks->num_levels - 1;
> +
> +	dc->dcn_soc->fabric_and_dram_bandwidth_vmin0p65 =
> +		32 * (fclks->data[vmin0p65_idx].clocks_in_khz / 1000.0) / 1000.0;
> +	dc->dcn_soc->fabric_and_dram_bandwidth_vmid0p72 =
> +		dc->dcn_soc->number_of_channels *
> +		(fclks->data[vmid0p72_idx].clocks_in_khz / 1000.0)
> +		* ddr4_dram_factor_single_Channel / 1000.0;
> +	dc->dcn_soc->fabric_and_dram_bandwidth_vnom0p8 =
> +		dc->dcn_soc->number_of_channels *
> +		(fclks->data[vnom0p8_idx].clocks_in_khz / 1000.0)
> +		* ddr4_dram_factor_single_Channel / 1000.0;
> +	dc->dcn_soc->fabric_and_dram_bandwidth_vmax0p9 =
> +		dc->dcn_soc->number_of_channels *
> +		(fclks->data[vmax0p9_idx].clocks_in_khz / 1000.0)
> +		* ddr4_dram_factor_single_Channel / 1000.0;
>   }
>   
> -void dcn_bw_update_from_pplib(struct dc *dc)
> +void dcn_bw_update_from_pplib_dcfclks(
> +	struct dc *dc,
> +	struct dm_pp_clock_levels_with_voltage *dcfclks)
>   {
> -	struct dc_context *ctx = dc->ctx;
> -	struct dm_pp_clock_levels_with_voltage fclks = {0}, dcfclks = {0};
> -	bool res;
> -	unsigned vmin0p65_idx, vmid0p72_idx, vnom0p8_idx, vmax0p9_idx;
> -
> -	/* TODO: This is not the proper way to obtain fabric_and_dram_bandwidth, should be min(fclk, memclk) */
> -	res = dm_pp_get_clock_levels_by_type_with_voltage(
> -			ctx, DM_PP_CLOCK_TYPE_FCLK, &fclks);
> -
> -	if (res)
> -		res = verify_clock_values(&fclks);
> -
> -	if (res) {
> -		ASSERT(fclks.num_levels);
> -
> -		vmin0p65_idx = 0;
> -		vmid0p72_idx = fclks.num_levels -
> -			(fclks.num_levels > 2 ? 3 : (fclks.num_levels > 1 ? 2 : 1));
> -		vnom0p8_idx = fclks.num_levels - (fclks.num_levels > 1 ? 2 : 1);
> -		vmax0p9_idx = fclks.num_levels - 1;
> -
> -		dc->dcn_soc->fabric_and_dram_bandwidth_vmin0p65 =
> -			32 * (fclks.data[vmin0p65_idx].clocks_in_khz / 1000.0) / 1000.0;
> -		dc->dcn_soc->fabric_and_dram_bandwidth_vmid0p72 =
> -			dc->dcn_soc->number_of_channels *
> -			(fclks.data[vmid0p72_idx].clocks_in_khz / 1000.0)
> -			* ddr4_dram_factor_single_Channel / 1000.0;
> -		dc->dcn_soc->fabric_and_dram_bandwidth_vnom0p8 =
> -			dc->dcn_soc->number_of_channels *
> -			(fclks.data[vnom0p8_idx].clocks_in_khz / 1000.0)
> -			* ddr4_dram_factor_single_Channel / 1000.0;
> -		dc->dcn_soc->fabric_and_dram_bandwidth_vmax0p9 =
> -			dc->dcn_soc->number_of_channels *
> -			(fclks.data[vmax0p9_idx].clocks_in_khz / 1000.0)
> -			* ddr4_dram_factor_single_Channel / 1000.0;
> -	} else
> -		BREAK_TO_DEBUGGER();
> -
> -	res = dm_pp_get_clock_levels_by_type_with_voltage(
> -			ctx, DM_PP_CLOCK_TYPE_DCFCLK, &dcfclks);
> -
> -	if (res)
> -		res = verify_clock_values(&dcfclks);
> +	if (dcfclks->num_levels >= 3) {
> +		dc->dcn_soc->dcfclkv_min0p65 = dcfclks->data[0].clocks_in_khz / 1000.0;
> +		dc->dcn_soc->dcfclkv_mid0p72 = dcfclks->data[dcfclks->num_levels - 3].clocks_in_khz / 1000.0;
> +		dc->dcn_soc->dcfclkv_nom0p8 = dcfclks->data[dcfclks->num_levels - 2].clocks_in_khz / 1000.0;
> +		dc->dcn_soc->dcfclkv_max0p9 = dcfclks->data[dcfclks->num_levels - 1].clocks_in_khz / 1000.0;
> +	}
> +}
>   
> -	if (res && dcfclks.num_levels >= 3) {
> -		dc->dcn_soc->dcfclkv_min0p65 = dcfclks.data[0].clocks_in_khz / 1000.0;
> -		dc->dcn_soc->dcfclkv_mid0p72 = dcfclks.data[dcfclks.num_levels - 3].clocks_in_khz / 1000.0;
> -		dc->dcn_soc->dcfclkv_nom0p8 = dcfclks.data[dcfclks.num_levels - 2].clocks_in_khz / 1000.0;
> -		dc->dcn_soc->dcfclkv_max0p9 = dcfclks.data[dcfclks.num_levels - 1].clocks_in_khz / 1000.0;
> -	} else
> -		BREAK_TO_DEBUGGER();
> +void dcn_get_soc_clks(
> +	struct dc *dc,
> +	int *min_fclk_khz,
> +	int *min_dcfclk_khz,
> +	int *socclk_khz)
> +{
> +	*min_fclk_khz = dc->dcn_soc->fabric_and_dram_bandwidth_vmin0p65 * 1000000 / 32;
> +	*min_dcfclk_khz = dc->dcn_soc->dcfclkv_min0p65 * 1000;
> +	*socclk_khz = dc->dcn_soc->socclk * 1000;
>   }
>   
> -void dcn_bw_notify_pplib_of_wm_ranges(struct dc *dc)
> +void dcn_bw_notify_pplib_of_wm_ranges(
> +	struct dc *dc,
> +	int min_fclk_khz,
> +	int min_dcfclk_khz,
> +	int socclk_khz)
>   {
>   	struct pp_smu_funcs_rv *pp = NULL;
>   	struct pp_smu_wm_range_sets ranges = {0};
> -	int min_fclk_khz, min_dcfclk_khz, socclk_khz;
>   	const int overdrive = 5000000; /* 5 GHz to cover Overdrive */
>   
>   	if (dc->res_pool->pp_smu)
> @@ -1546,10 +1532,6 @@ void dcn_bw_notify_pplib_of_wm_ranges(struct dc *dc)
>   	if (!pp || !pp->set_wm_ranges)
>   		return;
>   
> -	min_fclk_khz = dc->dcn_soc->fabric_and_dram_bandwidth_vmin0p65 * 1000000 / 32;
> -	min_dcfclk_khz = dc->dcn_soc->dcfclkv_min0p65 * 1000;
> -	socclk_khz = dc->dcn_soc->socclk * 1000;
> -
>   	/* Now notify PPLib/SMU about which Watermarks sets they should select
>   	 * depending on DPM state they are in. And update BW MGR GFX Engine and
>   	 * Memory clock member variables for Watermarks calculations for each
> diff --git a/drivers/gpu/drm/amd/display/dc/inc/dcn_calcs.h b/drivers/gpu/drm/amd/display/dc/inc/dcn_calcs.h
> index 806f3041db14..9e4ddc985240 100644
> --- a/drivers/gpu/drm/amd/display/dc/inc/dcn_calcs.h
> +++ b/drivers/gpu/drm/amd/display/dc/inc/dcn_calcs.h
> @@ -628,8 +628,23 @@ unsigned int dcn_find_dcfclk_suits_all(
>   	const struct dc *dc,
>   	struct dc_clocks *clocks);
>   
> -void dcn_bw_update_from_pplib(struct dc *dc);
> -void dcn_bw_notify_pplib_of_wm_ranges(struct dc *dc);
> +void dcn_get_soc_clks(
> +		struct dc *dc,
> +		int *min_fclk_khz,
> +		int *min_dcfclk_khz,
> +		int *socclk_khz);
> +
> +void dcn_bw_update_from_pplib_fclks(
> +		struct dc *dc,
> +		struct dm_pp_clock_levels_with_voltage *fclks);
> +void dcn_bw_update_from_pplib_dcfclks(
> +		struct dc *dc,
> +		struct dm_pp_clock_levels_with_voltage *dcfclks);
> +void dcn_bw_notify_pplib_of_wm_ranges(
> +		struct dc *dc,
> +		int min_fclk_khz,
> +		int min_dcfclk_khz,
> +		int socclk_khz);
>   void dcn_bw_sync_calcs_and_dml(struct dc *dc);
>   
>   enum source_macro_tile_size swizzle_mode_to_macro_tile_size(enum swizzle_mode_values sw_mode);

Powered by blists - more mailing lists