- Linux-stable-mirror - lists.linaro.org

[PATCH 6/9] drm/amd/display: Populate Power Profile In Case of Early Return

by Hamza Mahfooz

From: Austin Zheng <Austin.Zheng(a)amd.com> Early return possible if context has no clk_mgr. This will lead to an invalid power profile being returned which looks identical to a profile with the lowest power level. Add back logic that populated the power profile and overwrite the value if needed. Cc: stable(a)vger.kernel.org Fixes: fc8c959496fa ("drm/amd/display: Update Interface to Check UCLK DPM") Reviewed-by: Dillon Varone <dillon.varone(a)amd.com> Signed-off-by: Austin Zheng <Austin.Zheng(a)amd.com> Signed-off-by: Hamza Mahfooz <hamza.mahfooz(a)amd.com> --- drivers/gpu/drm/amd/display/dc/core/dc.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/drivers/gpu/drm/amd/display/dc/core/dc.c b/drivers/gpu/drm/amd/display/dc/core/dc.c index 0c1875d35a95..1dd26d5df6b9 100644 --- a/drivers/gpu/drm/amd/display/dc/core/dc.c +++ b/drivers/gpu/drm/amd/display/dc/core/dc.c @@ -6100,11 +6100,11 @@ struct dc_power_profile dc_get_power_profile_for_dc_state(const struct dc_state { struct dc_power_profile profile = { 0 }; - if (!context || !context->clk_mgr || !context->clk_mgr->ctx || !context->clk_mgr->ctx->dc) + profile.power_level = !context->bw_ctx.bw.dcn.clk.p_state_change_support; + if (!context->clk_mgr || !context->clk_mgr->ctx || !context->clk_mgr->ctx->dc) return profile; struct dc *dc = context->clk_mgr->ctx->dc; - if (dc->res_pool->funcs->get_power_profile) profile.power_level = dc->res_pool->funcs->get_power_profile(context); return profile; -- 2.46.1

3 months, 1 week

1
0
0 0

[PATCH 4/9] drm/amd/display: Enable Request rate limiter during C-State on dcn401

by Hamza Mahfooz

From: Dillon Varone <dillon.varone(a)amd.com> [WHY] When C-State entry is requested, the rate limiter will be disabled which can result in high contention in the DCHUB return path. [HOW] Enable the rate limiter during C-state requests to prevent contention. Cc: stable(a)vger.kernel.org # 6.11+ Reviewed-by: Alvin Lee <alvin.lee2(a)amd.com> Signed-off-by: Dillon Varone <dillon.varone(a)amd.com> Signed-off-by: Hamza Mahfooz <hamza.mahfooz(a)amd.com> --- .../src/dml2_core/dml2_core_dcn4_calcs.c | 6 +++++ .../display/dc/hubbub/dcn10/dcn10_hubbub.h | 8 ++++++- .../display/dc/hubbub/dcn20/dcn20_hubbub.h | 1 + .../display/dc/hubbub/dcn401/dcn401_hubbub.c | 24 +++++++++++++++++-- .../display/dc/hubbub/dcn401/dcn401_hubbub.h | 7 +++++- .../amd/display/dc/hwss/dcn401/dcn401_hwseq.c | 13 ++++++---- .../gpu/drm/amd/display/dc/inc/hw/dchubbub.h | 2 +- .../dc/resource/dcn401/dcn401_resource.h | 3 ++- 8 files changed, 53 insertions(+), 11 deletions(-) diff --git a/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_core/dml2_core_dcn4_calcs.c b/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_core/dml2_core_dcn4_calcs.c index 92e43a1e4dd4..601320b1be81 100644 --- a/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_core/dml2_core_dcn4_calcs.c +++ b/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_core/dml2_core_dcn4_calcs.c @@ -11,6 +11,7 @@ #define DML2_MAX_FMT_420_BUFFER_WIDTH 4096 #define DML_MAX_NUM_OF_SLICES_PER_DSC 4 +#define ALLOW_SDPIF_RATE_LIMIT_PRE_CSTATE const char *dml2_core_internal_bw_type_str(enum dml2_core_internal_bw_type bw_type) { @@ -3886,6 +3887,10 @@ static void CalculateSwathAndDETConfiguration(struct dml2_core_internal_scratch #endif *p->hw_debug5 = false; +#ifdef ALLOW_SDPIF_RATE_LIMIT_PRE_CSTATE + if (p->NumberOfActiveSurfaces > 1) + *p->hw_debug5 = true; +#else for (unsigned int k = 0; k < p->NumberOfActiveSurfaces; ++k) { if (!(p->mrq_present) && (!(*p->UnboundedRequestEnabled)) && (TotalActiveDPP == 1) && p->display_cfg->plane_descriptors[k].surface.dcc.enable @@ -3901,6 +3906,7 @@ static void CalculateSwathAndDETConfiguration(struct dml2_core_internal_scratch dml2_printf("DML::%s: k=%u hw_debug5 = %u\n", __func__, k, *p->hw_debug5); #endif } +#endif } static enum dml2_odm_mode DecideODMMode(unsigned int HActive, diff --git a/drivers/gpu/drm/amd/display/dc/hubbub/dcn10/dcn10_hubbub.h b/drivers/gpu/drm/amd/display/dc/hubbub/dcn10/dcn10_hubbub.h index 4bd1dda07719..9fbd45c7dfef 100644 --- a/drivers/gpu/drm/amd/display/dc/hubbub/dcn10/dcn10_hubbub.h +++ b/drivers/gpu/drm/amd/display/dc/hubbub/dcn10/dcn10_hubbub.h @@ -200,6 +200,7 @@ struct dcn_hubbub_registers { uint32_t DCHUBBUB_ARB_FRAC_URG_BW_MALL_B; uint32_t DCHUBBUB_TIMEOUT_DETECTION_CTRL1; uint32_t DCHUBBUB_TIMEOUT_DETECTION_CTRL2; + uint32_t DCHUBBUB_CTRL_STATUS; }; #define HUBBUB_REG_FIELD_LIST_DCN32(type) \ @@ -320,7 +321,12 @@ struct dcn_hubbub_registers { type DCHUBBUB_TIMEOUT_REQ_STALL_THRESHOLD;\ type DCHUBBUB_TIMEOUT_PSTATE_STALL_THRESHOLD;\ type DCHUBBUB_TIMEOUT_DETECTION_EN;\ - type DCHUBBUB_TIMEOUT_TIMER_RESET + type DCHUBBUB_TIMEOUT_TIMER_RESET;\ + type ROB_UNDERFLOW_STATUS;\ + type ROB_OVERFLOW_STATUS;\ + type ROB_OVERFLOW_CLEAR;\ + type DCHUBBUB_HW_DEBUG;\ + type CSTATE_SWATH_CHK_GOOD_MODE #define HUBBUB_STUTTER_REG_FIELD_LIST(type) \ type DCHUBBUB_ARB_ALLOW_SR_ENTER_WATERMARK_A;\ diff --git a/drivers/gpu/drm/amd/display/dc/hubbub/dcn20/dcn20_hubbub.h b/drivers/gpu/drm/amd/display/dc/hubbub/dcn20/dcn20_hubbub.h index 036bb3e6c957..46d8f5c70750 100644 --- a/drivers/gpu/drm/amd/display/dc/hubbub/dcn20/dcn20_hubbub.h +++ b/drivers/gpu/drm/amd/display/dc/hubbub/dcn20/dcn20_hubbub.h @@ -96,6 +96,7 @@ struct dcn20_hubbub { unsigned int det1_size; unsigned int det2_size; unsigned int det3_size; + bool allow_sdpif_rate_limit_when_cstate_req; }; void hubbub2_construct(struct dcn20_hubbub *hubbub, diff --git a/drivers/gpu/drm/amd/display/dc/hubbub/dcn401/dcn401_hubbub.c b/drivers/gpu/drm/amd/display/dc/hubbub/dcn401/dcn401_hubbub.c index 5d658e9bef64..92fab471b183 100644 --- a/drivers/gpu/drm/amd/display/dc/hubbub/dcn401/dcn401_hubbub.c +++ b/drivers/gpu/drm/amd/display/dc/hubbub/dcn401/dcn401_hubbub.c @@ -1192,15 +1192,35 @@ static void dcn401_wait_for_det_update(struct hubbub *hubbub, int hubp_inst) } } -static void dcn401_program_timeout_thresholds(struct hubbub *hubbub, struct dml2_display_arb_regs *arb_regs) +static bool dcn401_program_arbiter(struct hubbub *hubbub, struct dml2_display_arb_regs *arb_regs, bool safe_to_lower) { struct dcn20_hubbub *hubbub2 = TO_DCN20_HUBBUB(hubbub); + bool wm_pending = false; + uint32_t temp; + /* request backpressure and outstanding return threshold (unused)*/ //REG_UPDATE(DCHUBBUB_TIMEOUT_DETECTION_CTRL1, DCHUBBUB_TIMEOUT_REQ_STALL_THRESHOLD, arb_regs->req_stall_threshold); /* P-State stall threshold */ REG_UPDATE(DCHUBBUB_TIMEOUT_DETECTION_CTRL2, DCHUBBUB_TIMEOUT_PSTATE_STALL_THRESHOLD, arb_regs->pstate_stall_threshold); + + if (safe_to_lower || arb_regs->allow_sdpif_rate_limit_when_cstate_req > hubbub2->allow_sdpif_rate_limit_when_cstate_req) { + hubbub2->allow_sdpif_rate_limit_when_cstate_req = arb_regs->allow_sdpif_rate_limit_when_cstate_req; + + /* only update the required bits */ + REG_GET(DCHUBBUB_CTRL_STATUS, DCHUBBUB_HW_DEBUG, &temp); + if (hubbub2->allow_sdpif_rate_limit_when_cstate_req) { + temp |= (1 << 5); + } else { + temp &= ~(1 << 5); + } + REG_UPDATE(DCHUBBUB_CTRL_STATUS, DCHUBBUB_HW_DEBUG, temp); + } else { + wm_pending = true; + } + + return wm_pending; } static const struct hubbub_funcs hubbub4_01_funcs = { @@ -1226,7 +1246,7 @@ static const struct hubbub_funcs hubbub4_01_funcs = { .program_det_segments = dcn401_program_det_segments, .program_compbuf_segments = dcn401_program_compbuf_segments, .wait_for_det_update = dcn401_wait_for_det_update, - .program_timeout_thresholds = dcn401_program_timeout_thresholds, + .program_arbiter = dcn401_program_arbiter, }; void hubbub401_construct(struct dcn20_hubbub *hubbub2, diff --git a/drivers/gpu/drm/amd/display/dc/hubbub/dcn401/dcn401_hubbub.h b/drivers/gpu/drm/amd/display/dc/hubbub/dcn401/dcn401_hubbub.h index 5f1960722ebd..b1d9ea9d1c3d 100644 --- a/drivers/gpu/drm/amd/display/dc/hubbub/dcn401/dcn401_hubbub.h +++ b/drivers/gpu/drm/amd/display/dc/hubbub/dcn401/dcn401_hubbub.h @@ -128,7 +128,12 @@ HUBBUB_SF(DCHUBBUB_TIMEOUT_DETECTION_CTRL1, DCHUBBUB_TIMEOUT_REQ_STALL_THRESHOLD, mask_sh),\ HUBBUB_SF(DCHUBBUB_TIMEOUT_DETECTION_CTRL2, DCHUBBUB_TIMEOUT_PSTATE_STALL_THRESHOLD, mask_sh),\ HUBBUB_SF(DCHUBBUB_TIMEOUT_DETECTION_CTRL2, DCHUBBUB_TIMEOUT_DETECTION_EN, mask_sh),\ - HUBBUB_SF(DCHUBBUB_TIMEOUT_DETECTION_CTRL2, DCHUBBUB_TIMEOUT_TIMER_RESET, mask_sh) + HUBBUB_SF(DCHUBBUB_TIMEOUT_DETECTION_CTRL2, DCHUBBUB_TIMEOUT_TIMER_RESET, mask_sh),\ + HUBBUB_SF(DCHUBBUB_CTRL_STATUS, ROB_UNDERFLOW_STATUS, mask_sh),\ + HUBBUB_SF(DCHUBBUB_CTRL_STATUS, ROB_OVERFLOW_STATUS, mask_sh),\ + HUBBUB_SF(DCHUBBUB_CTRL_STATUS, ROB_OVERFLOW_CLEAR, mask_sh),\ + HUBBUB_SF(DCHUBBUB_CTRL_STATUS, DCHUBBUB_HW_DEBUG, mask_sh),\ + HUBBUB_SF(DCHUBBUB_CTRL_STATUS, CSTATE_SWATH_CHK_GOOD_MODE, mask_sh) bool hubbub401_program_urgent_watermarks( struct hubbub *hubbub, diff --git a/drivers/gpu/drm/amd/display/dc/hwss/dcn401/dcn401_hwseq.c b/drivers/gpu/drm/amd/display/dc/hwss/dcn401/dcn401_hwseq.c index e8cc1bfa73f3..5de11e2837c0 100644 --- a/drivers/gpu/drm/amd/display/dc/hwss/dcn401/dcn401_hwseq.c +++ b/drivers/gpu/drm/amd/display/dc/hwss/dcn401/dcn401_hwseq.c @@ -1488,6 +1488,10 @@ void dcn401_prepare_bandwidth(struct dc *dc, &context->bw_ctx.bw.dcn.watermarks, dc->res_pool->ref_clocks.dchub_ref_clock_inKhz / 1000, false); + /* update timeout thresholds */ + if (hubbub->funcs->program_arbiter) { + dc->wm_optimized_required |= hubbub->funcs->program_arbiter(hubbub, &context->bw_ctx.bw.dcn.arb_regs, false); + } /* decrease compbuf size */ if (hubbub->funcs->program_compbuf_segments) { @@ -1529,6 +1533,10 @@ void dcn401_optimize_bandwidth( &context->bw_ctx.bw.dcn.watermarks, dc->res_pool->ref_clocks.dchub_ref_clock_inKhz / 1000, true); + /* update timeout thresholds */ + if (hubbub->funcs->program_arbiter) { + hubbub->funcs->program_arbiter(hubbub, &context->bw_ctx.bw.dcn.arb_regs, true); + } if (dc->clk_mgr->dc_mode_softmax_enabled) if (dc->clk_mgr->clks.dramclk_khz > dc->clk_mgr->bw_params->dc_mode_softmax_memclk * 1000 && @@ -1554,11 +1562,6 @@ void dcn401_optimize_bandwidth( pipe_ctx->dlg_regs.min_dst_y_next_start); } } - - /* update timeout thresholds */ - if (hubbub->funcs->program_timeout_thresholds) { - hubbub->funcs->program_timeout_thresholds(hubbub, &context->bw_ctx.bw.dcn.arb_regs); - } } void dcn401_fams2_global_control_lock(struct dc *dc, diff --git a/drivers/gpu/drm/amd/display/dc/inc/hw/dchubbub.h b/drivers/gpu/drm/amd/display/dc/inc/hw/dchubbub.h index 6c1d41c0f099..52b745667ef7 100644 --- a/drivers/gpu/drm/amd/display/dc/inc/hw/dchubbub.h +++ b/drivers/gpu/drm/amd/display/dc/inc/hw/dchubbub.h @@ -228,7 +228,7 @@ struct hubbub_funcs { void (*program_det_segments)(struct hubbub *hubbub, int hubp_inst, unsigned det_buffer_size_seg); void (*program_compbuf_segments)(struct hubbub *hubbub, unsigned compbuf_size_seg, bool safe_to_increase); void (*wait_for_det_update)(struct hubbub *hubbub, int hubp_inst); - void (*program_timeout_thresholds)(struct hubbub *hubbub, struct dml2_display_arb_regs *arb_regs); + bool (*program_arbiter)(struct hubbub *hubbub, struct dml2_display_arb_regs *arb_regs, bool safe_to_lower); }; struct hubbub { diff --git a/drivers/gpu/drm/amd/display/dc/resource/dcn401/dcn401_resource.h b/drivers/gpu/drm/amd/display/dc/resource/dcn401/dcn401_resource.h index 7c8d61db153d..19568c359669 100644 --- a/drivers/gpu/drm/amd/display/dc/resource/dcn401/dcn401_resource.h +++ b/drivers/gpu/drm/amd/display/dc/resource/dcn401/dcn401_resource.h @@ -612,7 +612,8 @@ void dcn401_prepare_mcache_programming(struct dc *dc, struct dc_state *context); SR(DCHUBBUB_SDPIF_CFG1), \ SR(DCHUBBUB_MEM_PWR_MODE_CTRL), \ SR(DCHUBBUB_TIMEOUT_DETECTION_CTRL1), \ - SR(DCHUBBUB_TIMEOUT_DETECTION_CTRL2) + SR(DCHUBBUB_TIMEOUT_DETECTION_CTRL2), \ + SR(DCHUBBUB_CTRL_STATUS) /* DCCG */ -- 2.46.1

3 months, 1 week

1
0
0 0

[PATCH 3/9] drm/amd/display: Fix handling of plane refcount

by Hamza Mahfooz

From: Joshua Aberback <joshua.aberback(a)amd.com> [Why] The mechanism to backup and restore plane states doesn't maintain refcount, which can cause issues if the refcount of the plane changes in between backup and restore operations, such as memory leaks if the refcount was supposed to go down, or double frees / invalid memory accesses if the refcount was supposed to go up. [How] Cache and re-apply current refcount when restoring plane states. Cc: stable(a)vger.kernel.org Reviewed-by: Josip Pavic <josip.pavic(a)amd.com> Signed-off-by: Joshua Aberback <joshua.aberback(a)amd.com> Signed-off-by: Hamza Mahfooz <hamza.mahfooz(a)amd.com> --- drivers/gpu/drm/amd/display/dc/core/dc.c | 3 +++ 1 file changed, 3 insertions(+) diff --git a/drivers/gpu/drm/amd/display/dc/core/dc.c b/drivers/gpu/drm/amd/display/dc/core/dc.c index 7872c6cabb14..0c1875d35a95 100644 --- a/drivers/gpu/drm/amd/display/dc/core/dc.c +++ b/drivers/gpu/drm/amd/display/dc/core/dc.c @@ -3141,7 +3141,10 @@ static void restore_planes_and_stream_state( return; for (i = 0; i < status->plane_count; i++) { + /* refcount will always be valid, restore everything else */ + struct kref refcount = status->plane_states[i]->refcount; *status->plane_states[i] = scratch->plane_states[i]; + status->plane_states[i]->refcount = refcount; } *stream = scratch->stream_state; } -- 2.46.1

3 months, 1 week

1
0
0 0

[PATCH 2/9] drm/amd/display: Ignore scalar validation failure if pipe is phantom

by Hamza Mahfooz

From: Chris Park <chris.park(a)amd.com> [Why] There are some pipe scaler validation failure when the pipe is phantom and causes crash in DML validation. Since, scalar parameters are not as important in phantom pipe and we require this plane to do successful MCLK switches, the failure condition can be ignored. [How] Ignore scalar validation failure if the pipe validation is marked as phantom pipe. Cc: stable(a)vger.kernel.org # 6.11+ Reviewed-by: Dillon Varone <dillon.varone(a)amd.com> Signed-off-by: Chris Park <chris.park(a)amd.com> Signed-off-by: Hamza Mahfooz <hamza.mahfooz(a)amd.com> --- drivers/gpu/drm/amd/display/dc/core/dc_resource.c | 8 ++++++++ 1 file changed, 8 insertions(+) diff --git a/drivers/gpu/drm/amd/display/dc/core/dc_resource.c b/drivers/gpu/drm/amd/display/dc/core/dc_resource.c index 33125b95c3a1..619fad17de55 100644 --- a/drivers/gpu/drm/amd/display/dc/core/dc_resource.c +++ b/drivers/gpu/drm/amd/display/dc/core/dc_resource.c @@ -1501,6 +1501,10 @@ bool resource_build_scaling_params(struct pipe_ctx *pipe_ctx) res = spl_calculate_scaler_params(spl_in, spl_out); // Convert respective out params from SPL to scaler data translate_SPL_out_params_to_pipe_ctx(pipe_ctx, spl_out); + + /* Ignore scaler failure if pipe context plane is phantom plane */ + if (!res && plane_state->is_phantom) + res = true; } else { #endif /* depends on h_active */ @@ -1571,6 +1575,10 @@ bool resource_build_scaling_params(struct pipe_ctx *pipe_ctx) &plane_state->scaling_quality); } + /* Ignore scaler failure if pipe context plane is phantom plane */ + if (!res && plane_state->is_phantom) + res = true; + if (res && (pipe_ctx->plane_res.scl_data.taps.v_taps != temp.v_taps || pipe_ctx->plane_res.scl_data.taps.h_taps != temp.h_taps || pipe_ctx->plane_res.scl_data.taps.v_taps_c != temp.v_taps_c || -- 2.46.1

3 months, 1 week

1
0
0 0

[PATCH 1/9] drm/amd/display: update pipe selection policy to check head pipe

by Hamza Mahfooz

From: Yihan Zhu <Yihan.Zhu(a)amd.com> [Why] No check on head pipe during the dml to dc hw mapping will allow illegal pipe usage. This will result in a wrong pipe topology to cause mpcc tree totally mess up then cause a display hang. [How] Avoid to use the pipe is head in all check and avoid ODM slice during preferred pipe check. v2: Added pipe type check for DPP pipe type before executing head pipe check in the pipe selection logic in DML2 to avoid NULL pointer de-reference. Cc: stable(a)vger.kernel.org Reviewed-by: Nicholas Kazlauskas <nicholas.kazlauskas(a)amd.com> Signed-off-by: Yihan Zhu <Yihan.Zhu(a)amd.com> Signed-off-by: Hamza Mahfooz <hamza.mahfooz(a)amd.com> --- .../display/dc/dml2/dml2_dc_resource_mgmt.c | 23 ++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) diff --git a/drivers/gpu/drm/amd/display/dc/dml2/dml2_dc_resource_mgmt.c b/drivers/gpu/drm/amd/display/dc/dml2/dml2_dc_resource_mgmt.c index 6eccf0241d85..1ed21c1b86a5 100644 --- a/drivers/gpu/drm/amd/display/dc/dml2/dml2_dc_resource_mgmt.c +++ b/drivers/gpu/drm/amd/display/dc/dml2/dml2_dc_resource_mgmt.c @@ -258,12 +258,25 @@ static unsigned int find_preferred_pipe_candidates(const struct dc_state *existi * However this condition comes with a caveat. We need to ignore pipes that will * require a change in OPP but still have the same stream id. For example during * an MPC to ODM transiton. + * + * Adding check to avoid pipe select on the head pipe by utilizing dc resource + * helper function resource_get_primary_dpp_pipe and comparing the pipe index. */ if (existing_state) { for (i = 0; i < pipe_count; i++) { if (existing_state->res_ctx.pipe_ctx[i].stream && existing_state->res_ctx.pipe_ctx[i].stream->stream_id == stream_id) { + struct pipe_ctx *head_pipe = + resource_is_pipe_type(&existing_state->res_ctx.pipe_ctx[i], DPP_PIPE) ? + resource_get_primary_dpp_pipe(&existing_state->res_ctx.pipe_ctx[i]) : + NULL; + + // we should always respect the head pipe from selection + if (head_pipe && head_pipe->pipe_idx == i) + continue; if (existing_state->res_ctx.pipe_ctx[i].plane_res.hubp && - existing_state->res_ctx.pipe_ctx[i].plane_res.hubp->opp_id != i) + existing_state->res_ctx.pipe_ctx[i].plane_res.hubp->opp_id != i && + (existing_state->res_ctx.pipe_ctx[i].prev_odm_pipe || + existing_state->res_ctx.pipe_ctx[i].next_odm_pipe)) continue; preferred_pipe_candidates[num_preferred_candidates++] = i; @@ -292,6 +305,14 @@ static unsigned int find_last_resort_pipe_candidates(const struct dc_state *exis */ if (existing_state) { for (i = 0; i < pipe_count; i++) { + struct pipe_ctx *head_pipe = + resource_is_pipe_type(&existing_state->res_ctx.pipe_ctx[i], DPP_PIPE) ? + resource_get_primary_dpp_pipe(&existing_state->res_ctx.pipe_ctx[i]) : + NULL; + + // we should always respect the head pipe from selection + if (head_pipe && head_pipe->pipe_idx == i) + continue; if ((existing_state->res_ctx.pipe_ctx[i].plane_res.hubp && existing_state->res_ctx.pipe_ctx[i].plane_res.hubp->opp_id != i) || existing_state->res_ctx.pipe_ctx[i].stream_res.tg) -- 2.46.1

3 months, 1 week

1
0
0 0

[PATCH net] netfilter: ipset: add missing range check in bitmap_ip_uadt

by Jeongjun Park

In the bitmap_ip_uadt function, if ip is greater than ip_to, they are swapped. However, there is no check to see if ip is smaller than map->first, which causes an out-of-bounds vulnerability. Therefore, you need to add a missing bounds check to prevent out-of-bounds. Cc: <stable(a)vger.kernel.org> Reported-by: syzbot+58c872f7790a4d2ac951(a)syzkaller.appspotmail.com Fixes: 72205fc68bd1 ("netfilter: ipset: bitmap:ip set type support") Signed-off-by: Jeongjun Park <aha310510(a)gmail.com> --- net/netfilter/ipset/ip_set_bitmap_ip.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/net/netfilter/ipset/ip_set_bitmap_ip.c b/net/netfilter/ipset/ip_set_bitmap_ip.c index e4fa00abde6a..705c316b001a 100644 --- a/net/netfilter/ipset/ip_set_bitmap_ip.c +++ b/net/netfilter/ipset/ip_set_bitmap_ip.c @@ -178,7 +178,7 @@ bitmap_ip_uadt(struct ip_set *set, struct nlattr *tb[], ip_to = ip; } - if (ip_to > map->last_ip) + if (ip < map->first_ip || ip_to > map->last_ip) return -IPSET_ERR_BITMAP_RANGE; for (; !before(ip_to, ip); ip += map->hosts) { --

3 months, 1 week

2
1
0 0

[PATCH] Revert "mmc: dw_mmc: Fix IDMAC operation with pages bigger than 4K"

by Aurelien Jarno

The commit 8396c793ffdf ("mmc: dw_mmc: Fix IDMAC operation with pages bigger than 4K") increased the max_req_size, even for 4K pages, causing various issues: - Panic booting the kernel/rootfs from an SD card on Rockchip RK3566 - Panic booting the kernel/rootfs from an SD card on StarFive JH7100 - "swiotlb buffer is full" and data corruption on StarFive JH7110 At this stage no fix have been found, so it's probably better to just revert the change. This reverts commit 8396c793ffdf28bb8aee7cfe0891080f8cab7890. Cc: stable(a)vger.kernel.org Cc: Sam Protsenko <semen.protsenko(a)linaro.org> Fixes: 8396c793ffdf ("mmc: dw_mmc: Fix IDMAC operation with pages bigger than 4K") Closes: https://lore.kernel.org/linux-mmc/614692b4-1dbe-31b8-a34d-cb6db1909bb7@w6rz… Closes: https://lore.kernel.org/linux-mmc/CAC8uq=Ppnmv98mpa1CrWLawWoPnu5abtU69v-=G-… Signed-off-by: Aurelien Jarno <aurelien(a)aurel32.net> --- drivers/mmc/host/dw_mmc.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) I have posted a patch to fix the issue, but unfortunately it only fixes the JH7110 case: https://lore.kernel.org/linux-mmc/20241020142931.138277-1-aurelien@aurel32.… diff --git a/drivers/mmc/host/dw_mmc.c b/drivers/mmc/host/dw_mmc.c index 41e451235f637..e9f6e4e622901 100644 --- a/drivers/mmc/host/dw_mmc.c +++ b/drivers/mmc/host/dw_mmc.c @@ -2957,8 +2957,8 @@ static int dw_mci_init_slot(struct dw_mci *host) if (host->use_dma == TRANS_MODE_IDMAC) { mmc->max_segs = host->ring_size; mmc->max_blk_size = 65535; - mmc->max_req_size = DW_MCI_DESC_DATA_LENGTH * host->ring_size; - mmc->max_seg_size = mmc->max_req_size; + mmc->max_seg_size = 0x1000; + mmc->max_req_size = mmc->max_seg_size * host->ring_size; mmc->max_blk_count = mmc->max_req_size / 512; } else if (host->use_dma == TRANS_MODE_EDMAC) { mmc->max_segs = 64; -- 2.45.2

3 months, 1 week

2
1
0 0

[PATCH] mmc: sunxi-mmc: Fix A100 compatible description

by Andre Przywara

It turns out that the Allwinner A100/A133 SoC only supports 8K DMA blocks (13 bits wide), for both the SD/SDIO and eMMC instances. And while this alone would make a trivial fix, the H616 falls back to the A100 compatible string, so we have to now match the H616 compatible string explicitly against the description advertising 64K DMA blocks. As the A100 is now compatible with the D1 description, let the A100 compatible string point to that block instead, and introduce an explicit match against the H616 string, pointing to the old description. Also remove the redundant setting of clk_delays to NULL on the way. Fixes: 3536b82e5853 ("mmc: sunxi: add support for A100 mmc controller") Cc: stable(a)vger.kernel.org Signed-off-by: Andre Przywara <andre.przywara(a)arm.com> --- drivers/mmc/host/sunxi-mmc.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) diff --git a/drivers/mmc/host/sunxi-mmc.c b/drivers/mmc/host/sunxi-mmc.c index d3bd0ac99ec46..e0ab5fd635e6c 100644 --- a/drivers/mmc/host/sunxi-mmc.c +++ b/drivers/mmc/host/sunxi-mmc.c @@ -1191,10 +1191,9 @@ static const struct sunxi_mmc_cfg sun50i_a64_emmc_cfg = { .needs_new_timings = true, }; -static const struct sunxi_mmc_cfg sun50i_a100_cfg = { +static const struct sunxi_mmc_cfg sun50i_h616_cfg = { .idma_des_size_bits = 16, .idma_des_shift = 2, - .clk_delays = NULL, .can_calibrate = true, .mask_data0 = true, .needs_new_timings = true, @@ -1217,8 +1216,9 @@ static const struct of_device_id sunxi_mmc_of_match[] = { { .compatible = "allwinner,sun20i-d1-mmc", .data = &sun20i_d1_cfg }, { .compatible = "allwinner,sun50i-a64-mmc", .data = &sun50i_a64_cfg }, { .compatible = "allwinner,sun50i-a64-emmc", .data = &sun50i_a64_emmc_cfg }, - { .compatible = "allwinner,sun50i-a100-mmc", .data = &sun50i_a100_cfg }, + { .compatible = "allwinner,sun50i-a100-mmc", .data = &sun20i_d1_cfg }, { .compatible = "allwinner,sun50i-a100-emmc", .data = &sun50i_a100_emmc_cfg }, + { .compatible = "allwinner,sun50i-h616-mmc", .data = &sun50i_h616_cfg }, { /* sentinel */ } }; MODULE_DEVICE_TABLE(of, sunxi_mmc_of_match); -- 2.46.2

3 months, 1 week

4
5
0 0

[merged mm-hotfixes-stable] selftests-hugetlb_dio-fixup-check-for-initial-conditions-to-skip-in-the-start.patch removed from -mm tree

by Andrew Morton

The quilt patch titled Subject: selftests: hugetlb_dio: fixup check for initial conditions to skip in the start has been removed from the -mm tree. Its filename was selftests-hugetlb_dio-fixup-check-for-initial-conditions-to-skip-in-the-start.patch This patch was dropped because it was merged into the mm-hotfixes-stable branch of git://git.kernel.org/pub/scm/linux/kernel/git/akpm/mm ------------------------------------------------------ From: Donet Tom <donettom(a)linux.ibm.com> Subject: selftests: hugetlb_dio: fixup check for initial conditions to skip in the start Date: Sun, 10 Nov 2024 00:49:03 -0600 This test verifies that a hugepage, used as a user buffer for DIO operations, is correctly freed upon unmapping. To test this, we read the count of free hugepages before and after the mmap, DIO, and munmap operations, then check if the free hugepage count is the same. Reading free hugepages before the test was removed by commit 0268d4579901 ('selftests: hugetlb_dio: check for initial conditions to skip at the start'), causing the test to always fail. This patch adds back reading the free hugepages before starting the test. With this patch, the tests are now passing. Test results without this patch: ./tools/testing/selftests/mm/hugetlb_dio TAP version 13 1..4 # No. Free pages before allocation : 0 # No. Free pages after munmap : 100 not ok 1 : Huge pages not freed! # No. Free pages before allocation : 0 # No. Free pages after munmap : 100 not ok 2 : Huge pages not freed! # No. Free pages before allocation : 0 # No. Free pages after munmap : 100 not ok 3 : Huge pages not freed! # No. Free pages before allocation : 0 # No. Free pages after munmap : 100 not ok 4 : Huge pages not freed! # Totals: pass:0 fail:4 xfail:0 xpass:0 skip:0 error:0 Test results with this patch: /tools/testing/selftests/mm/hugetlb_dio TAP version 13 1..4 # No. Free pages before allocation : 100 # No. Free pages after munmap : 100 ok 1 : Huge pages freed successfully ! # No. Free pages before allocation : 100 # No. Free pages after munmap : 100 ok 2 : Huge pages freed successfully ! # No. Free pages before allocation : 100 # No. Free pages after munmap : 100 ok 3 : Huge pages freed successfully ! # No. Free pages before allocation : 100 # No. Free pages after munmap : 100 ok 4 : Huge pages freed successfully ! # Totals: pass:4 fail:0 xfail:0 xpass:0 skip:0 error:0 Link: https://lkml.kernel.org/r/20241110064903.23626-1-donettom@linux.ibm.com Fixes: 0268d4579901 ("selftests: hugetlb_dio: check for initial conditions to skip in the start") Signed-off-by: Donet Tom <donettom(a)linux.ibm.com> Cc: Muhammad Usama Anjum <usama.anjum(a)collabora.com> Cc: Shuah Khan <shuah(a)kernel.org> Cc: <stable(a)vger.kernel.org> Signed-off-by: Andrew Morton <akpm(a)linux-foundation.org> --- tools/testing/selftests/mm/hugetlb_dio.c | 7 +++++++ 1 file changed, 7 insertions(+) --- a/tools/testing/selftests/mm/hugetlb_dio.c~selftests-hugetlb_dio-fixup-check-for-initial-conditions-to-skip-in-the-start +++ a/tools/testing/selftests/mm/hugetlb_dio.c @@ -44,6 +44,13 @@ void run_dio_using_hugetlb(unsigned int if (fd < 0) ksft_exit_fail_perror("Error opening file\n"); + /* Get the free huge pages before allocation */ + free_hpage_b = get_free_hugepages(); + if (free_hpage_b == 0) { + close(fd); + ksft_exit_skip("No free hugepage, exiting!\n"); + } + /* Allocate a hugetlb page */ orig_buffer = mmap(NULL, h_pagesize, mmap_prot, mmap_flags, -1, 0); if (orig_buffer == MAP_FAILED) { _ Patches currently in -mm which might be from donettom(a)linux.ibm.com are

3 months, 1 week

1
0
0 0

[merged mm-hotfixes-stable] mm-gup-avoid-an-unnecessary-allocation-call-for-foll_longterm-cases.patch removed from -mm tree

by Andrew Morton

The quilt patch titled Subject: mm/gup: avoid an unnecessary allocation call for FOLL_LONGTERM cases has been removed from the -mm tree. Its filename was mm-gup-avoid-an-unnecessary-allocation-call-for-foll_longterm-cases.patch This patch was dropped because it was merged into the mm-hotfixes-stable branch of git://git.kernel.org/pub/scm/linux/kernel/git/akpm/mm ------------------------------------------------------ From: John Hubbard <jhubbard(a)nvidia.com> Subject: mm/gup: avoid an unnecessary allocation call for FOLL_LONGTERM cases Date: Mon, 4 Nov 2024 19:29:44 -0800 commit 53ba78de064b ("mm/gup: introduce check_and_migrate_movable_folios()") created a new constraint on the pin_user_pages*() API family: a potentially large internal allocation must now occur, for FOLL_LONGTERM cases. A user-visible consequence has now appeared: user space can no longer pin more than 2GB of memory anymore on x86_64. That's because, on a 4KB PAGE_SIZE system, when user space tries to (indirectly, via a device driver that calls pin_user_pages()) pin 2GB, this requires an allocation of a folio pointers array of MAX_PAGE_ORDER size, which is the limit for kmalloc(). In addition to the directly visible effect described above, there is also the problem of adding an unnecessary allocation. The **pages array argument has already been allocated, and there is no need for a redundant **folios array allocation in this case. Fix this by avoiding the new allocation entirely. This is done by referring to either the original page[i] within **pages, or to the associated folio. Thanks to David Hildenbrand for suggesting this approach and for providing the initial implementation (which I've tested and adjusted slightly) as well. [jhubbard(a)nvidia.com: whitespace tweak, per David] Link: https://lkml.kernel.org/r/131cf9c8-ebc0-4cbb-b722-22fa8527bf3c@nvidia.com [jhubbard(a)nvidia.com: bypass pofs_get_folio(), per Oscar] Link: https://lkml.kernel.org/r/c1587c7f-9155-45be-bd62-1e36c0dd6923@nvidia.com Link: https://lkml.kernel.org/r/20241105032944.141488-2-jhubbard@nvidia.com Fixes: 53ba78de064b ("mm/gup: introduce check_and_migrate_movable_folios()") Signed-off-by: John Hubbard <jhubbard(a)nvidia.com> Suggested-by: David Hildenbrand <david(a)redhat.com> Acked-by: David Hildenbrand <david(a)redhat.com> Reviewed-by: Oscar Salvador <osalvador(a)suse.de> Cc: Vivek Kasireddy <vivek.kasireddy(a)intel.com> Cc: Dave Airlie <airlied(a)redhat.com> Cc: Gerd Hoffmann <kraxel(a)redhat.com> Cc: Matthew Wilcox <willy(a)infradead.org> Cc: Christoph Hellwig <hch(a)infradead.org> Cc: Jason Gunthorpe <jgg(a)nvidia.com> Cc: Peter Xu <peterx(a)redhat.com> Cc: Arnd Bergmann <arnd(a)arndb.de> Cc: Daniel Vetter <daniel.vetter(a)ffwll.ch> Cc: Dongwon Kim <dongwon.kim(a)intel.com> Cc: Hugh Dickins <hughd(a)google.com> Cc: Junxiao Chang <junxiao.chang(a)intel.com> Cc: <stable(a)vger.kernel.org> Signed-off-by: Andrew Morton <akpm(a)linux-foundation.org> --- mm/gup.c | 116 +++++++++++++++++++++++++++++++++++------------------ 1 file changed, 77 insertions(+), 39 deletions(-) --- a/mm/gup.c~mm-gup-avoid-an-unnecessary-allocation-call-for-foll_longterm-cases +++ a/mm/gup.c @@ -2273,20 +2273,57 @@ struct page *get_dump_page(unsigned long #endif /* CONFIG_ELF_CORE */ #ifdef CONFIG_MIGRATION + +/* + * An array of either pages or folios ("pofs"). Although it may seem tempting to + * avoid this complication, by simply interpreting a list of folios as a list of + * pages, that approach won't work in the longer term, because eventually the + * layouts of struct page and struct folio will become completely different. + * Furthermore, this pof approach avoids excessive page_folio() calls. + */ +struct pages_or_folios { + union { + struct page **pages; + struct folio **folios; + void **entries; + }; + bool has_folios; + long nr_entries; +}; + +static struct folio *pofs_get_folio(struct pages_or_folios *pofs, long i) +{ + if (pofs->has_folios) + return pofs->folios[i]; + return page_folio(pofs->pages[i]); +} + +static void pofs_clear_entry(struct pages_or_folios *pofs, long i) +{ + pofs->entries[i] = NULL; +} + +static void pofs_unpin(struct pages_or_folios *pofs) +{ + if (pofs->has_folios) + unpin_folios(pofs->folios, pofs->nr_entries); + else + unpin_user_pages(pofs->pages, pofs->nr_entries); +} + /* * Returns the number of collected folios. Return value is always >= 0. */ static unsigned long collect_longterm_unpinnable_folios( - struct list_head *movable_folio_list, - unsigned long nr_folios, - struct folio **folios) + struct list_head *movable_folio_list, + struct pages_or_folios *pofs) { unsigned long i, collected = 0; struct folio *prev_folio = NULL; bool drain_allow = true; - for (i = 0; i < nr_folios; i++) { - struct folio *folio = folios[i]; + for (i = 0; i < pofs->nr_entries; i++) { + struct folio *folio = pofs_get_folio(pofs, i); if (folio == prev_folio) continue; @@ -2327,16 +2364,15 @@ static unsigned long collect_longterm_un * Returns -EAGAIN if all folios were successfully migrated or -errno for * failure (or partial success). */ -static int migrate_longterm_unpinnable_folios( - struct list_head *movable_folio_list, - unsigned long nr_folios, - struct folio **folios) +static int +migrate_longterm_unpinnable_folios(struct list_head *movable_folio_list, + struct pages_or_folios *pofs) { int ret; unsigned long i; - for (i = 0; i < nr_folios; i++) { - struct folio *folio = folios[i]; + for (i = 0; i < pofs->nr_entries; i++) { + struct folio *folio = pofs_get_folio(pofs, i); if (folio_is_device_coherent(folio)) { /* @@ -2344,7 +2380,7 @@ static int migrate_longterm_unpinnable_f * convert the pin on the source folio to a normal * reference. */ - folios[i] = NULL; + pofs_clear_entry(pofs, i); folio_get(folio); gup_put_folio(folio, 1, FOLL_PIN); @@ -2363,8 +2399,8 @@ static int migrate_longterm_unpinnable_f * calling folio_isolate_lru() which takes a reference so the * folio won't be freed if it's migrating. */ - unpin_folio(folios[i]); - folios[i] = NULL; + unpin_folio(folio); + pofs_clear_entry(pofs, i); } if (!list_empty(movable_folio_list)) { @@ -2387,12 +2423,26 @@ static int migrate_longterm_unpinnable_f return -EAGAIN; err: - unpin_folios(folios, nr_folios); + pofs_unpin(pofs); putback_movable_pages(movable_folio_list); return ret; } +static long +check_and_migrate_movable_pages_or_folios(struct pages_or_folios *pofs) +{ + LIST_HEAD(movable_folio_list); + unsigned long collected; + + collected = collect_longterm_unpinnable_folios(&movable_folio_list, + pofs); + if (!collected) + return 0; + + return migrate_longterm_unpinnable_folios(&movable_folio_list, pofs); +} + /* * Check whether all folios are *allowed* to be pinned indefinitely (long term). * Rather confusingly, all folios in the range are required to be pinned via @@ -2417,16 +2467,13 @@ err: static long check_and_migrate_movable_folios(unsigned long nr_folios, struct folio **folios) { - unsigned long collected; - LIST_HEAD(movable_folio_list); + struct pages_or_folios pofs = { + .folios = folios, + .has_folios = true, + .nr_entries = nr_folios, + }; - collected = collect_longterm_unpinnable_folios(&movable_folio_list, - nr_folios, folios); - if (!collected) - return 0; - - return migrate_longterm_unpinnable_folios(&movable_folio_list, - nr_folios, folios); + return check_and_migrate_movable_pages_or_folios(&pofs); } /* @@ -2436,22 +2483,13 @@ static long check_and_migrate_movable_fo static long check_and_migrate_movable_pages(unsigned long nr_pages, struct page **pages) { - struct folio **folios; - long i, ret; - - folios = kmalloc_array(nr_pages, sizeof(*folios), GFP_KERNEL); - if (!folios) { - unpin_user_pages(pages, nr_pages); - return -ENOMEM; - } - - for (i = 0; i < nr_pages; i++) - folios[i] = page_folio(pages[i]); + struct pages_or_folios pofs = { + .pages = pages, + .has_folios = false, + .nr_entries = nr_pages, + }; - ret = check_and_migrate_movable_folios(nr_pages, folios); - - kfree(folios); - return ret; + return check_and_migrate_movable_pages_or_folios(&pofs); } #else static long check_and_migrate_movable_pages(unsigned long nr_pages, _ Patches currently in -mm which might be from jhubbard(a)nvidia.com are

3 months, 1 week

1
0
0 0

2025

2024

2023

2022

2021

2020

2019

2018

2017

Linux-stable-mirror