January 2022 - Linux-stable-mirror

[PATCH v1 3/3] ASoC: ops: Reject out of bounds values in snd_soc_put_xr_sx()

by Mark Brown

We don't currently validate that the values being set are within the range we advertised to userspace as being valid, do so and reject any values that are out of range. Signed-off-by: Mark Brown <broonie(a)kernel.org> Cc: stable(a)vger.kernel.org --- sound/soc/soc-ops.c | 2 ++ 1 file changed, 2 insertions(+) diff --git a/sound/soc/soc-ops.c b/sound/soc/soc-ops.c index c31e63b27193..dc0e7c8d31f3 100644 --- a/sound/soc/soc-ops.c +++ b/sound/soc/soc-ops.c @@ -879,6 +879,8 @@ int snd_soc_put_xr_sx(struct snd_kcontrol *kcontrol, long val = ucontrol->value.integer.value[0]; unsigned int i; + if (val < mc->min || val > mc->max) + return -EINVAL; if (invert) val = max - val; val &= mask; -- 2.30.2

2 years, 11 months

1
0
0 0

[PATCH v1 2/3] ASoC: ops: Reject out of bounds values in snd_soc_put_volsw_sx()

by Mark Brown

We don't currently validate that the values being set are within the range we advertised to userspace as being valid, do so and reject any values that are out of range. Signed-off-by: Mark Brown <broonie(a)kernel.org> Cc: stable(a)vger.kernel.org --- sound/soc/soc-ops.c | 9 ++++++++- 1 file changed, 8 insertions(+), 1 deletion(-) diff --git a/sound/soc/soc-ops.c b/sound/soc/soc-ops.c index fbe5d326b0f2..c31e63b27193 100644 --- a/sound/soc/soc-ops.c +++ b/sound/soc/soc-ops.c @@ -423,8 +423,15 @@ int snd_soc_put_volsw_sx(struct snd_kcontrol *kcontrol, int err = 0; unsigned int val, val_mask; + val = ucontrol->value.integer.value[0]; + if (mc->platform_max && val > mc->platform_max) + return -EINVAL; + if (val > max - min) + return -EINVAL; + if (val < 0) + return -EINVAL; val_mask = mask << shift; - val = (ucontrol->value.integer.value[0] + min) & mask; + val = (val + min) & mask; val = val << shift; err = snd_soc_component_update_bits(component, reg, val_mask, val); -- 2.30.2

2 years, 11 months

1
0
0 0

[PATCH v1 1/3] ASoC: ops: Reject out of bounds values in snd_soc_put_volsw()

by Mark Brown

We don't currently validate that the values being set are within the range we advertised to userspace as being valid, do so and reject any values that are out of range. Signed-off-by: Mark Brown <broonie(a)kernel.org> Cc: stable(a)vger.kernel.org --- sound/soc/soc-ops.c | 18 ++++++++++++++++-- 1 file changed, 16 insertions(+), 2 deletions(-) diff --git a/sound/soc/soc-ops.c b/sound/soc/soc-ops.c index 08eaa9ddf191..fbe5d326b0f2 100644 --- a/sound/soc/soc-ops.c +++ b/sound/soc/soc-ops.c @@ -316,13 +316,27 @@ int snd_soc_put_volsw(struct snd_kcontrol *kcontrol, if (sign_bit) mask = BIT(sign_bit + 1) - 1; - val = ((ucontrol->value.integer.value[0] + min) & mask); + val = ucontrol->value.integer.value[0]; + if (mc->platform_max && val > mc->platform_max) + return -EINVAL; + if (val > max - min) + return -EINVAL; + if (val < 0) + return -EINVAL; + val = (val + min) & mask; if (invert) val = max - val; val_mask = mask << shift; val = val << shift; if (snd_soc_volsw_is_stereo(mc)) { - val2 = ((ucontrol->value.integer.value[1] + min) & mask); + val2 = ucontrol->value.integer.value[1]; + if (mc->platform_max && val2 > mc->platform_max) + return -EINVAL; + if (val2 > max - min) + return -EINVAL; + if (val2 < 0) + return -EINVAL; + val2 = (val2 + min) & mask; if (invert) val2 = max - val2; if (reg == reg2) { -- 2.30.2

2 years, 11 months

1
0
0 0

[PATCH] arm64: Mark start_backtrace() notrace and NOKPROBE_SYMBOL

by Masami Hiramatsu

Mark the start_backtrace() as notrace and NOKPROBE_SYMBOL because this function is called from ftrace and lockdep to get the caller address via return_address(). The lockdep is used in kprobes, it should also be NOKPROBE_SYMBOL. Fixes: b07f3499661c ("arm64: stacktrace: Move start_backtrace() out of the header") Cc: stable(a)vger.kernel.org Signed-off-by: Masami Hiramatsu <mhiramat(a)kernel.org> --- arch/arm64/kernel/stacktrace.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/arch/arm64/kernel/stacktrace.c b/arch/arm64/kernel/stacktrace.c index 94f83cd44e50..b0f21677764d 100644 --- a/arch/arm64/kernel/stacktrace.c +++ b/arch/arm64/kernel/stacktrace.c @@ -33,7 +33,7 @@ */ -void start_backtrace(struct stackframe *frame, unsigned long fp, +void notrace start_backtrace(struct stackframe *frame, unsigned long fp, unsigned long pc) { frame->fp = fp; @@ -55,6 +55,7 @@ void start_backtrace(struct stackframe *frame, unsigned long fp, frame->prev_fp = 0; frame->prev_type = STACK_TYPE_UNKNOWN; } +NOKPROBE_SYMBOL(start_backtrace); /* * Unwind from one frame record (A) to the next frame record (B).

2 years, 11 months

3
2
0 0

[PATCH v2 5.4.y] Revert "ia64: kprobes: Use generic kretprobe trampoline handler"

by Masami Hiramatsu

This reverts commit 77fa5e15c933a1ec812de61ad709c00aa51e96ae. Since the upstream commit e792ff804f49720ce003b3e4c618b5d996256a18 depends on the generic kretprobe trampoline handler, which was introduced by commit 66ada2ccae4e ("kprobes: Add generic kretprobe trampoline handler") but that is not ported to the stable kernel because it is not a bugfix series. So revert this commit to fix a build error. NOTE: I keep commit a7fe2378454c ("ia64: kprobes: Fix to pass correct trampoline address to the handler") on the tree, that seems just a cleanup without the original reverted commit, but it would be better to use dereference_function_descriptor() macro instead of accessing descriptor's field directly. Fixes: 77fa5e15c933 ("ia64: kprobes: Use generic kretprobe trampoline handler") Reported-by: kernel test robot <lkp(a)intel.com> Signed-off-by: Masami Hiramatsu <mhiramat(a)kernel.org> --- Changes in v2: - fix the lack of type casting for dereference_function_descriptor(). --- arch/ia64/kernel/kprobes.c | 78 ++++++++++++++++++++++++++++++++++++++++++-- 1 file changed, 75 insertions(+), 3 deletions(-) diff --git a/arch/ia64/kernel/kprobes.c b/arch/ia64/kernel/kprobes.c index 8a223d0e4918..fa10d51f6217 100644 --- a/arch/ia64/kernel/kprobes.c +++ b/arch/ia64/kernel/kprobes.c @@ -396,10 +396,83 @@ static void kretprobe_trampoline(void) { } +/* + * At this point the target function has been tricked into + * returning into our trampoline. Lookup the associated instance + * and then: + * - call the handler function + * - cleanup by marking the instance as unused + * - long jump back to the original return address + */ int __kprobes trampoline_probe_handler(struct kprobe *p, struct pt_regs *regs) { - regs->cr_iip = __kretprobe_trampoline_handler(regs, - dereference_function_descriptor(kretprobe_trampoline), NULL); + struct kretprobe_instance *ri = NULL; + struct hlist_head *head, empty_rp; + struct hlist_node *tmp; + unsigned long flags, orig_ret_address = 0; + unsigned long trampoline_address = + (unsigned long)dereference_function_descriptor(kretprobe_trampoline); + + INIT_HLIST_HEAD(&empty_rp); + kretprobe_hash_lock(current, &head, &flags); + + /* + * It is possible to have multiple instances associated with a given + * task either because an multiple functions in the call path + * have a return probe installed on them, and/or more than one return + * return probe was registered for a target function. + * + * We can handle this because: + * - instances are always inserted at the head of the list + * - when multiple return probes are registered for the same + * function, the first instance's ret_addr will point to the + * real return address, and all the rest will point to + * kretprobe_trampoline + */ + hlist_for_each_entry_safe(ri, tmp, head, hlist) { + if (ri->task != current) + /* another task is sharing our hash bucket */ + continue; + + orig_ret_address = (unsigned long)ri->ret_addr; + if (orig_ret_address != trampoline_address) + /* + * This is the real return address. Any other + * instances associated with this task are for + * other calls deeper on the call stack + */ + break; + } + + regs->cr_iip = orig_ret_address; + + hlist_for_each_entry_safe(ri, tmp, head, hlist) { + if (ri->task != current) + /* another task is sharing our hash bucket */ + continue; + + if (ri->rp && ri->rp->handler) + ri->rp->handler(ri, regs); + + orig_ret_address = (unsigned long)ri->ret_addr; + recycle_rp_inst(ri, &empty_rp); + + if (orig_ret_address != trampoline_address) + /* + * This is the real return address. Any other + * instances associated with this task are for + * other calls deeper on the call stack + */ + break; + } + kretprobe_assert(ri, orig_ret_address, trampoline_address); + + kretprobe_hash_unlock(current, &flags); + + hlist_for_each_entry_safe(ri, tmp, &empty_rp, hlist) { + hlist_del(&ri->hlist); + kfree(ri); + } /* * By returning a non-zero value, we are telling * kprobe_handler() that we don't want the post_handler @@ -412,7 +485,6 @@ void __kprobes arch_prepare_kretprobe(struct kretprobe_instance *ri, struct pt_regs *regs) { ri->ret_addr = (kprobe_opcode_t *)regs->b0; - ri->fp = NULL; /* Replace the return addr with trampoline addr */ regs->b0 = (unsigned long)dereference_function_descriptor(kretprobe_trampoline);

2 years, 11 months

2
1
0 0

[PATCH 0/7] rwsem enhancement patches for 5.10

by Jaegeuk Kim

Per discussion [1], can we merge these patches in 5.10 first? [1] https://lore.kernel.org/linux-f2fs-devel/CAEe=Sx=6FCvrp_6x2Bqp3YTzep2s=aWdC… Peter Zijlstra (3): locking/rwsem: Better collate rwsem_read_trylock() locking/rwsem: Introduce rwsem_write_trylock() locking/rwsem: Fold __down_{read,write}*() Waiman Long (4): locking/rwsem: Pass the current atomic count to rwsem_down_read_slowpath() locking/rwsem: Prevent potential lock starvation locking/rwsem: Enable reader optimistic lock stealing locking/rwsem: Remove reader optimistic spinning kernel/locking/lock_events_list.h | 6 +- kernel/locking/rwsem.c | 359 +++++++++--------------------- 2 files changed, 106 insertions(+), 259 deletions(-) -- 2.34.1.703.g22d0c6ccf7-goog

2 years, 11 months

3
10
0 0

Yours Faithful,

by Mr Sal Kavar

Dear friend. This being a wide world in which it can be difficult to make new acquaintances and because it is virtually impossible to know who is trustworthy and who can be believed, I have decided to repose confidence in you after much fasting and prayer. It is only because of this that I have decided to confide in you and to share with you this confidential business. overdue and unclaimed sum of $15.5m, (Fifteen Million Five Hundred Thousand Dollars Only) when the account holder suddenly passed on, he left no beneficiary who would be entitled to the receipt of this fund. For this reason, I have found it expedient to transfer this fund to a trustworthy individual with capacity to act as foreign business partner. Yours Faithful, Mr.Sal Kavar.

2 years, 11 months

1
0
0 0

FAILED: patch "[PATCH] ASoC: SOF: handle paused streams during system suspend" failed to apply to 5.16-stable tree

by gregkh＠linuxfoundation.org

The patch below does not apply to the 5.16-stable tree. If someone wants it applied there, or to any other stable or longterm tree, then please email the backport, including the original git commit id to <stable(a)vger.kernel.org>. thanks, greg k-h ------------------ original commit in Linus's tree ------------------ From 96da174024b9c63bd5d3358668d0bc12677be877 Mon Sep 17 00:00:00 2001 From: Ranjani Sridharan <ranjani.sridharan(a)linux.intel.com> Date: Tue, 23 Nov 2021 19:16:06 +0200 Subject: [PATCH] ASoC: SOF: handle paused streams during system suspend During system suspend, paused streams do not get suspended. Therefore, we need to explicitly free these PCMs in the DSP and free the associated DAPM widgets so that they can be set up again during resume. Fixes: 5fcdbb2d45df ("ASoC: SOF: Add support for dynamic pipelines") Signed-off-by: Ranjani Sridharan <ranjani.sridharan(a)linux.intel.com> Reviewed-by: Paul Olaru <paul.olaru(a)oss.nxp.com> Reviewed-by: Bard Liao <bard.liao(a)intel.com> Reviewed-by: Pierre-Louis Bossart <pierre-louis.bossart(a)linux.intel.com> Signed-off-by: Kai Vehmanen <kai.vehmanen(a)linux.intel.com> Link: https://lore.kernel.org/r/20211123171606.129350-3-kai.vehmanen@linux.intel.… Signed-off-by: Mark Brown <broonie(a)kernel.org> diff --git a/sound/soc/sof/pcm.c b/sound/soc/sof/pcm.c index 31dd79b794f1..0ceb1a9cbf73 100644 --- a/sound/soc/sof/pcm.c +++ b/sound/soc/sof/pcm.c @@ -100,9 +100,8 @@ void snd_sof_pcm_period_elapsed(struct snd_pcm_substream *substream) } EXPORT_SYMBOL(snd_sof_pcm_period_elapsed); -static int sof_pcm_dsp_pcm_free(struct snd_pcm_substream *substream, - struct snd_sof_dev *sdev, - struct snd_sof_pcm *spcm) +int sof_pcm_dsp_pcm_free(struct snd_pcm_substream *substream, struct snd_sof_dev *sdev, + struct snd_sof_pcm *spcm) { struct sof_ipc_stream stream; struct sof_ipc_reply reply; diff --git a/sound/soc/sof/sof-audio.c b/sound/soc/sof/sof-audio.c index f4e142ec0fbd..e00ce275052f 100644 --- a/sound/soc/sof/sof-audio.c +++ b/sound/soc/sof/sof-audio.c @@ -129,6 +129,14 @@ int sof_widget_free(struct snd_sof_dev *sdev, struct snd_sof_widget *swidget) case snd_soc_dapm_buffer: ipc_free.hdr.cmd |= SOF_IPC_TPLG_BUFFER_FREE; break; + case snd_soc_dapm_dai_in: + case snd_soc_dapm_dai_out: + { + struct snd_sof_dai *dai = swidget->private; + + dai->configured = false; + fallthrough; + } default: ipc_free.hdr.cmd |= SOF_IPC_TPLG_COMP_FREE; break; @@ -720,6 +728,55 @@ int sof_set_up_pipelines(struct snd_sof_dev *sdev, bool verify) return 0; } +/* + * Free the PCM, its associated widgets and set the prepared flag to false for all PCMs that + * did not get suspended(ex: paused streams) so the widgets can be set up again during resume. + */ +static int sof_tear_down_left_over_pipelines(struct snd_sof_dev *sdev) +{ + struct snd_sof_widget *swidget; + struct snd_sof_pcm *spcm; + int dir, ret; + + /* + * free all PCMs and their associated DAPM widgets if their connected DAPM widget + * list is not NULL. This should only be true for paused streams at this point. + * This is equivalent to the handling of FE DAI suspend trigger for running streams. + */ + list_for_each_entry(spcm, &sdev->pcm_list, list) + for_each_pcm_streams(dir) { + struct snd_pcm_substream *substream = spcm->stream[dir].substream; + + if (!substream || !substream->runtime) + continue; + + if (spcm->stream[dir].list) { + ret = sof_pcm_dsp_pcm_free(substream, sdev, spcm); + if (ret < 0) + return ret; + + ret = sof_widget_list_free(sdev, spcm, dir); + if (ret < 0) { + dev_err(sdev->dev, "failed to free widgets during suspend\n"); + return ret; + } + } + } + + /* + * free any left over DAI widgets. This is equivalent to the handling of suspend trigger + * for the BE DAI for running streams. + */ + list_for_each_entry(swidget, &sdev->widget_list, list) + if (WIDGET_IS_DAI(swidget->id) && swidget->use_count == 1) { + ret = sof_widget_free(sdev, swidget); + if (ret < 0) + return ret; + } + + return 0; +} + /* * For older firmware, this function doesn't free widgets for static pipelines during suspend. * It only resets use_count for all widgets. @@ -734,8 +791,8 @@ int sof_tear_down_pipelines(struct snd_sof_dev *sdev, bool verify) /* * This function is called during suspend and for one-time topology verification during * first boot. In both cases, there is no need to protect swidget->use_count and - * sroute->setup because during suspend all streams are suspended and during topology - * loading the sound card unavailable to open PCMs. + * sroute->setup because during suspend all running streams are suspended and during + * topology loading the sound card unavailable to open PCMs. */ list_for_each_entry(swidget, &sdev->widget_list, list) { if (swidget->dynamic_pipeline_widget) @@ -754,6 +811,19 @@ int sof_tear_down_pipelines(struct snd_sof_dev *sdev, bool verify) return ret; } + /* + * Tear down all pipelines associated with PCMs that did not get suspended + * and unset the prepare flag so that they can be set up again during resume. + * Skip this step for older firmware. + */ + if (!verify && v->abi_version >= SOF_ABI_VER(3, 19, 0)) { + ret = sof_tear_down_left_over_pipelines(sdev); + if (ret < 0) { + dev_err(sdev->dev, "failed to tear down paused pipelines\n"); + return ret; + } + } + list_for_each_entry(sroute, &sdev->route_list, list) sroute->setup = false; diff --git a/sound/soc/sof/sof-audio.h b/sound/soc/sof/sof-audio.h index 389d56ac3aba..1c4f59d34717 100644 --- a/sound/soc/sof/sof-audio.h +++ b/sound/soc/sof/sof-audio.h @@ -265,4 +265,6 @@ int sof_widget_free(struct snd_sof_dev *sdev, struct snd_sof_widget *swidget); /* PCM */ int sof_widget_list_setup(struct snd_sof_dev *sdev, struct snd_sof_pcm *spcm, int dir); int sof_widget_list_free(struct snd_sof_dev *sdev, struct snd_sof_pcm *spcm, int dir); +int sof_pcm_dsp_pcm_free(struct snd_pcm_substream *substream, struct snd_sof_dev *sdev, + struct snd_sof_pcm *spcm); #endif

2 years, 11 months

3
2
0 0

FAILED: patch "[PATCH] ASoC: SOF: sof-audio: setup sched widgets during pipeline" failed to apply to 5.16-stable tree

by gregkh＠linuxfoundation.org

The patch below does not apply to the 5.16-stable tree. If someone wants it applied there, or to any other stable or longterm tree, then please email the backport, including the original git commit id to <stable(a)vger.kernel.org>. thanks, greg k-h ------------------ original commit in Linus's tree ------------------ From 01429183f479c54c1b5d15453a8ce574ea43e525 Mon Sep 17 00:00:00 2001 From: Pierre-Louis Bossart <pierre-louis.bossart(a)linux.intel.com> Date: Tue, 23 Nov 2021 19:16:04 +0200 Subject: [PATCH] ASoC: SOF: sof-audio: setup sched widgets during pipeline complete step Older firmware prior to ABI 3.19 has a dependency where the scheduler widgets need to be setup last. Moving the call to sof_widget_setup() before the pipeline_complete() call also helps remove the need for the 'reverse' direction when walking through the widget list - this was only working because of the topology macros but the topology does not require any order. Fixes: 5fcdbb2d45df ("ASoC: SOF: Add support for dynamic pipelines") Signed-off-by: Pierre-Louis Bossart <pierre-louis.bossart(a)linux.intel.com> Reviewed-by: Ranjani Sridharan <ranjani.sridharan(a)linux.intel.com> Signed-off-by: Kai Vehmanen <kai.vehmanen(a)linux.intel.com> Link: https://lore.kernel.org/r/20211123171606.129350-1-kai.vehmanen@linux.intel.… Signed-off-by: Mark Brown <broonie(a)kernel.org> diff --git a/sound/soc/sof/sof-audio.c b/sound/soc/sof/sof-audio.c index 0f2566f7c094..f4e142ec0fbd 100644 --- a/sound/soc/sof/sof-audio.c +++ b/sound/soc/sof/sof-audio.c @@ -637,16 +637,25 @@ const struct sof_ipc_pipe_new *snd_sof_pipeline_find(struct snd_sof_dev *sdev, int sof_set_up_pipelines(struct snd_sof_dev *sdev, bool verify) { + struct sof_ipc_fw_version *v = &sdev->fw_ready.version; struct snd_sof_widget *swidget; struct snd_sof_route *sroute; int ret; /* restore pipeline components */ - list_for_each_entry_reverse(swidget, &sdev->widget_list, list) { + list_for_each_entry(swidget, &sdev->widget_list, list) { /* only set up the widgets belonging to static pipelines */ if (!verify && swidget->dynamic_pipeline_widget) continue; + /* + * For older firmware, skip scheduler widgets in this loop, + * sof_widget_setup() will be called in the 'complete pipeline' loop + */ + if (v->abi_version < SOF_ABI_VER(3, 19, 0) && + swidget->id == snd_soc_dapm_scheduler) + continue; + /* update DAI config. The IPC will be sent in sof_widget_setup() */ if (WIDGET_IS_DAI(swidget->id)) { struct snd_sof_dai *dai = swidget->private; @@ -694,6 +703,12 @@ int sof_set_up_pipelines(struct snd_sof_dev *sdev, bool verify) if (!verify && swidget->dynamic_pipeline_widget) continue; + if (v->abi_version < SOF_ABI_VER(3, 19, 0)) { + ret = sof_widget_setup(sdev, swidget); + if (ret < 0) + return ret; + } + swidget->complete = snd_sof_complete_pipeline(sdev, swidget); break; @@ -722,7 +737,7 @@ int sof_tear_down_pipelines(struct snd_sof_dev *sdev, bool verify) * sroute->setup because during suspend all streams are suspended and during topology * loading the sound card unavailable to open PCMs. */ - list_for_each_entry_reverse(swidget, &sdev->widget_list, list) { + list_for_each_entry(swidget, &sdev->widget_list, list) { if (swidget->dynamic_pipeline_widget) continue;

2 years, 11 months

3
2
0 0

FAILED: patch "[PATCH] dma/pool: create dma atomic pool only if dma zone has managed" failed to apply to 5.4-stable tree

by gregkh＠linuxfoundation.org

The patch below does not apply to the 5.4-stable tree. If someone wants it applied there, or to any other stable or longterm tree, then please email the backport, including the original git commit id to <stable(a)vger.kernel.org>. thanks, greg k-h ------------------ original commit in Linus's tree ------------------ From a674e48c5443d12a8a43c3ac42367aa39505d506 Mon Sep 17 00:00:00 2001 From: Baoquan He <bhe(a)redhat.com> Date: Fri, 14 Jan 2022 14:07:41 -0800 Subject: [PATCH] dma/pool: create dma atomic pool only if dma zone has managed pages Currently three dma atomic pools are initialized as long as the relevant kernel codes are built in. While in kdump kernel of x86_64, this is not right when trying to create atomic_pool_dma, because there's no managed pages in DMA zone. In the case, DMA zone only has low 1M memory presented and locked down by memblock allocator. So no pages are added into buddy of DMA zone. Please check commit f1d4d47c5851 ("x86/setup: Always reserve the first 1M of RAM"). Then in kdump kernel of x86_64, it always prints below failure message: DMA: preallocated 128 KiB GFP_KERNEL pool for atomic allocations swapper/0: page allocation failure: order:5, mode:0xcc1(GFP_KERNEL|GFP_DMA), nodemask=(null),cpuset=/,mems_allowed=0 CPU: 0 PID: 1 Comm: swapper/0 Not tainted 5.13.0-0.rc5.20210611git929d931f2b40.42.fc35.x86_64 #1 Hardware name: Dell Inc. PowerEdge R910/0P658H, BIOS 2.12.0 06/04/2018 Call Trace: dump_stack+0x7f/0xa1 warn_alloc.cold+0x72/0xd6 __alloc_pages_slowpath.constprop.0+0xf29/0xf50 __alloc_pages+0x24d/0x2c0 alloc_page_interleave+0x13/0xb0 atomic_pool_expand+0x118/0x210 __dma_atomic_pool_init+0x45/0x93 dma_atomic_pool_init+0xdb/0x176 do_one_initcall+0x67/0x320 kernel_init_freeable+0x290/0x2dc kernel_init+0xa/0x111 ret_from_fork+0x22/0x30 Mem-Info: ...... DMA: failed to allocate 128 KiB GFP_KERNEL|GFP_DMA pool for atomic allocation DMA: preallocated 128 KiB GFP_KERNEL|GFP_DMA32 pool for atomic allocations Here, let's check if DMA zone has managed pages, then create atomic_pool_dma if yes. Otherwise just skip it. Link: https://lkml.kernel.org/r/20211223094435.248523-3-bhe@redhat.com Fixes: 6f599d84231f ("x86/kdump: Always reserve the low 1M when the crashkernel option is specified") Signed-off-by: Baoquan He <bhe(a)redhat.com> Reviewed-by: Christoph Hellwig <hch(a)lst.de> Acked-by: John Donnelly <john.p.donnelly(a)oracle.com> Reviewed-by: David Hildenbrand <david(a)redhat.com> Cc: Marek Szyprowski <m.szyprowski(a)samsung.com> Cc: Robin Murphy <robin.murphy(a)arm.com> Cc: Borislav Petkov <bp(a)alien8.de> Cc: Christoph Lameter <cl(a)linux.com> Cc: David Laight <David.Laight(a)ACULAB.COM> Cc: David Rientjes <rientjes(a)google.com> Cc: Hyeonggon Yoo <42.hyeyoo(a)gmail.com> Cc: Joonsoo Kim <iamjoonsoo.kim(a)lge.com> Cc: Pekka Enberg <penberg(a)kernel.org> Cc: Vlastimil Babka <vbabka(a)suse.cz> Cc: <stable(a)vger.kernel.org> Signed-off-by: Andrew Morton <akpm(a)linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds(a)linux-foundation.org> diff --git a/kernel/dma/pool.c b/kernel/dma/pool.c index 5f84e6cdb78e..4d40dcce7604 100644 --- a/kernel/dma/pool.c +++ b/kernel/dma/pool.c @@ -203,7 +203,7 @@ static int __init dma_atomic_pool_init(void) GFP_KERNEL); if (!atomic_pool_kernel) ret = -ENOMEM; - if (IS_ENABLED(CONFIG_ZONE_DMA)) { + if (has_managed_dma()) { atomic_pool_dma = __dma_atomic_pool_init(atomic_pool_size, GFP_KERNEL | GFP_DMA); if (!atomic_pool_dma) @@ -226,7 +226,7 @@ static inline struct gen_pool *dma_guess_pool(struct gen_pool *prev, gfp_t gfp) if (prev == NULL) { if (IS_ENABLED(CONFIG_ZONE_DMA32) && (gfp & GFP_DMA32)) return atomic_pool_dma32; - if (IS_ENABLED(CONFIG_ZONE_DMA) && (gfp & GFP_DMA)) + if (atomic_pool_dma && (gfp & GFP_DMA)) return atomic_pool_dma; return atomic_pool_kernel; }

2 years, 11 months

3
2
0 0

2025

2024

2023

2022

2021

2020

2019

2018

2017

Linux-stable-mirror January 2022