September 2024 - Linux-stable-mirror

[PATCH] ASoC: codecs: Fix atomicity violation in snd_soc_component_get_drvdata()

by Qiu-ji Chen

An atomicity violation occurs when the validity of the variables da7219->clk_src and da7219->mclk_rate is being assessed. Since the entire assessment is not protected by a lock, the da7219 variable might still be in flux during the assessment, rendering this check invalid. To fix this issue, we recommend adding a lock before the block if ((da7219->clk_src == clk_id) && (da7219->mclk_rate == freq)) so that the legitimacy check for da7219->clk_src and da7219->mclk_rate is protected by the lock, ensuring the validity of the check. This possible bug is found by an experimental static analysis tool developed by our team. This tool analyzes the locking APIs to extract function pairs that can be concurrently executed, and then analyzes the instructions in the paired functions to identify possible concurrency bugs including data races and atomicity violations. Fixes: 6d817c0e9fd7 ("ASoC: codecs: Add da7219 codec driver") Cc: stable(a)vger.kernel.org Signed-off-by: Qiu-ji Chen <chenqiuji666(a)gmail.com> --- sound/soc/codecs/da7219.c | 9 ++++++--- 1 file changed, 6 insertions(+), 3 deletions(-) diff --git a/sound/soc/codecs/da7219.c b/sound/soc/codecs/da7219.c index 311ea7918b31..e2da3e317b5a 100644 --- a/sound/soc/codecs/da7219.c +++ b/sound/soc/codecs/da7219.c @@ -1167,17 +1167,20 @@ static int da7219_set_dai_sysclk(struct snd_soc_dai *codec_dai, struct da7219_priv *da7219 = snd_soc_component_get_drvdata(component); int ret = 0; - if ((da7219->clk_src == clk_id) && (da7219->mclk_rate == freq)) + mutex_lock(&da7219->pll_lock); + + if ((da7219->clk_src == clk_id) && (da7219->mclk_rate == freq)) { + mutex_unlock(&da7219->pll_lock); return 0; + } if ((freq < 2000000) || (freq > 54000000)) { + mutex_unlock(&da7219->pll_lock); dev_err(codec_dai->dev, "Unsupported MCLK value %d\n", freq); return -EINVAL; } - mutex_lock(&da7219->pll_lock); - switch (clk_id) { case DA7219_CLKSRC_MCLK_SQR: snd_soc_component_update_bits(component, DA7219_PLL_CTRL, -- 2.34.1

1 month, 2 weeks

2
1
0 0

[PATCH] Drivers: hv: kvp/vss: Avoid accessing a ringbuffer not initialized yet

by Dexuan Cui

If the KVP (or VSS) daemon starts before the VMBus channel's ringbuffer is fully initialized, we can hit the panic below: hv_utils: Registering HyperV Utility Driver hv_vmbus: registering driver hv_utils ... BUG: kernel NULL pointer dereference, address: 0000000000000000 CPU: 44 UID: 0 PID: 2552 Comm: hv_kvp_daemon Tainted: G E 6.11.0-rc3+ #1 RIP: 0010:hv_pkt_iter_first+0x12/0xd0 Call Trace: ... vmbus_recvpacket hv_kvp_onchannelcallback vmbus_on_event tasklet_action_common tasklet_action handle_softirqs irq_exit_rcu sysvec_hyperv_stimer0 </IRQ> <TASK> asm_sysvec_hyperv_stimer0 ... kvp_register_done hvt_op_read vfs_read ksys_read __x64_sys_read This can happen because the KVP/VSS channel callback can be invoked even before the channel is fully opened: 1) as soon as hv_kvp_init() -> hvutil_transport_init() creates /dev/vmbus/hv_kvp, the kvp daemon can open the device file immediately and register itself to the driver by writing a message KVP_OP_REGISTER1 to the file (which is handled by kvp_on_msg() ->kvp_handle_handshake()) and reading the file for the driver's response, which is handled by hvt_op_read(), which calls hvt->on_read(), i.e. kvp_register_done(). 2) the problem with kvp_register_done() is that it can cause the channel callback to be called even before the channel is fully opened, and when the channel callback is starting to run, util_probe()-> vmbus_open() may have not initialized the ringbuffer yet, so the callback can hit the panic of NULL pointer dereference. To reproduce the panic consistently, we can add a "ssleep(10)" for KVP in __vmbus_open(), just before the first hv_ringbuffer_init(), and then we unload and reload the driver hv_utils, and run the daemon manually within the 10 seconds. Fix the panic by checking the channel state in the channel callback. To avoid the race condition with __vmbus_open(), we disable and enable the channel callback temporarily in __vmbus_open(). The channel callbacks of the other VMBus devices don't need to check the channel state since they can't run before the channels are fully initialized. Note: we would also need to fix the fcopy driver code, but that has been removed in commit ec314f61e4fc ("Drivers: hv: Remove fcopy driver") in the mainline kernel since v6.10. For old 6.x LTS kernels, and the 5.x and 4.x LTS kernels, the fcopy driver needs to be fixed when the fix is backported to the stable kernel branches. Fixes: e0fa3e5e7df6 ("Drivers: hv: utils: fix a race on userspace daemons registration") Cc: stable(a)vger.kernel.org Signed-off-by: Dexuan Cui <decui(a)microsoft.com> --- drivers/hv/channel.c | 11 +++++++++++ drivers/hv/hv_kvp.c | 3 +++ drivers/hv/hv_snapshot.c | 3 +++ 3 files changed, 17 insertions(+) diff --git a/drivers/hv/channel.c b/drivers/hv/channel.c index fb8cd8469328..685e407a3fdf 100644 --- a/drivers/hv/channel.c +++ b/drivers/hv/channel.c @@ -657,6 +657,14 @@ static int __vmbus_open(struct vmbus_channel *newchannel, return -ENOMEM; } + /* + * The channel callbacks of KVP/VSS may run before __vmbus_open() + * finishes (see kvp_register_done() -> ... -> kvp_poll_wrapper()), so + * they check newchannel->state to tell the ringbuffer has been fully + * initialized or not. Disable and enable the tasklet to avoid the race. + */ + tasklet_disable(&newchannel->callback_event); + newchannel->state = CHANNEL_OPENING_STATE; newchannel->onchannel_callback = onchannelcallback; newchannel->channel_callback_context = context; @@ -750,6 +758,8 @@ static int __vmbus_open(struct vmbus_channel *newchannel, } newchannel->state = CHANNEL_OPENED_STATE; + tasklet_enable(&newchannel->callback_event); + kfree(open_info); return 0; @@ -766,6 +776,7 @@ static int __vmbus_open(struct vmbus_channel *newchannel, hv_ringbuffer_cleanup(&newchannel->inbound); vmbus_free_requestor(&newchannel->requestor); newchannel->state = CHANNEL_OPEN_STATE; + tasklet_enable(&newchannel->callback_event); return err; } diff --git a/drivers/hv/hv_kvp.c b/drivers/hv/hv_kvp.c index d35b60c06114..ec098067e579 100644 --- a/drivers/hv/hv_kvp.c +++ b/drivers/hv/hv_kvp.c @@ -662,6 +662,9 @@ void hv_kvp_onchannelcallback(void *context) if (kvp_transaction.state > HVUTIL_READY) return; + if (channel->state != CHANNEL_OPENED_STATE) + return; + if (vmbus_recvpacket(channel, recv_buffer, HV_HYP_PAGE_SIZE * 4, &recvlen, &requestid)) { pr_err_ratelimited("KVP request received. Could not read into recv buf\n"); return; diff --git a/drivers/hv/hv_snapshot.c b/drivers/hv/hv_snapshot.c index 0d2184be1691..f7924c2fc62e 100644 --- a/drivers/hv/hv_snapshot.c +++ b/drivers/hv/hv_snapshot.c @@ -301,6 +301,9 @@ void hv_vss_onchannelcallback(void *context) if (vss_transaction.state > HVUTIL_READY) return; + if (channel->state != CHANNEL_OPENED_STATE) + return; + if (vmbus_recvpacket(channel, recv_buffer, VSS_MAX_PKT_SIZE, &recvlen, &requestid)) { pr_err_ratelimited("VSS request received. Could not read into recv buf\n"); return; -- 2.25.1

1 month, 3 weeks

3
9
0 0

[PATCH v7 bpf-next 01/10] lib/buildid: harden build ID parsing logic

by Andrii Nakryiko

Harden build ID parsing logic, adding explicit READ_ONCE() where it's important to have a consistent value read and validated just once. Also, as pointed out by Andi Kleen, we need to make sure that entire ELF note is within a page bounds, so move the overflow check up and add an extra note_size boundaries validation. Fixes tag below points to the code that moved this code into lib/buildid.c, and then subsequently was used in perf subsystem, making this code exposed to perf_event_open() users in v5.12+. Cc: stable(a)vger.kernel.org Reviewed-by: Eduard Zingerman <eddyz87(a)gmail.com> Reviewed-by: Jann Horn <jannh(a)google.com> Suggested-by: Andi Kleen <ak(a)linux.intel.com> Fixes: bd7525dacd7e ("bpf: Move stack_map_get_build_id into lib") Signed-off-by: Andrii Nakryiko <andrii(a)kernel.org> --- lib/buildid.c | 76 +++++++++++++++++++++++++++++---------------------- 1 file changed, 44 insertions(+), 32 deletions(-) diff --git a/lib/buildid.c b/lib/buildid.c index e02b5507418b..26007cc99a38 100644 --- a/lib/buildid.c +++ b/lib/buildid.c @@ -18,31 +18,37 @@ static int parse_build_id_buf(unsigned char *build_id, const void *note_start, Elf32_Word note_size) { - Elf32_Word note_offs = 0, new_offs; - - while (note_offs + sizeof(Elf32_Nhdr) < note_size) { - Elf32_Nhdr *nhdr = (Elf32_Nhdr *)(note_start + note_offs); + const char note_name[] = "GNU"; + const size_t note_name_sz = sizeof(note_name); + u64 note_off = 0, new_off, name_sz, desc_sz; + const char *data; + + while (note_off + sizeof(Elf32_Nhdr) < note_size && + note_off + sizeof(Elf32_Nhdr) > note_off /* overflow */) { + Elf32_Nhdr *nhdr = (Elf32_Nhdr *)(note_start + note_off); + + name_sz = READ_ONCE(nhdr->n_namesz); + desc_sz = READ_ONCE(nhdr->n_descsz); + + new_off = note_off + sizeof(Elf32_Nhdr); + if (check_add_overflow(new_off, ALIGN(name_sz, 4), &new_off) || + check_add_overflow(new_off, ALIGN(desc_sz, 4), &new_off) || + new_off > note_size) + break; if (nhdr->n_type == BUILD_ID && - nhdr->n_namesz == sizeof("GNU") && - !strcmp((char *)(nhdr + 1), "GNU") && - nhdr->n_descsz > 0 && - nhdr->n_descsz <= BUILD_ID_SIZE_MAX) { - memcpy(build_id, - note_start + note_offs + - ALIGN(sizeof("GNU"), 4) + sizeof(Elf32_Nhdr), - nhdr->n_descsz); - memset(build_id + nhdr->n_descsz, 0, - BUILD_ID_SIZE_MAX - nhdr->n_descsz); + name_sz == note_name_sz && + memcmp(nhdr + 1, note_name, note_name_sz) == 0 && + desc_sz > 0 && desc_sz <= BUILD_ID_SIZE_MAX) { + data = note_start + note_off + ALIGN(note_name_sz, 4); + memcpy(build_id, data, desc_sz); + memset(build_id + desc_sz, 0, BUILD_ID_SIZE_MAX - desc_sz); if (size) - *size = nhdr->n_descsz; + *size = desc_sz; return 0; } - new_offs = note_offs + sizeof(Elf32_Nhdr) + - ALIGN(nhdr->n_namesz, 4) + ALIGN(nhdr->n_descsz, 4); - if (new_offs <= note_offs) /* overflow */ - break; - note_offs = new_offs; + + note_off = new_off; } return -EINVAL; @@ -71,7 +77,7 @@ static int get_build_id_32(const void *page_addr, unsigned char *build_id, { Elf32_Ehdr *ehdr = (Elf32_Ehdr *)page_addr; Elf32_Phdr *phdr; - int i; + __u32 i, phnum; /* * FIXME @@ -80,18 +86,19 @@ static int get_build_id_32(const void *page_addr, unsigned char *build_id, */ if (ehdr->e_phoff != sizeof(Elf32_Ehdr)) return -EINVAL; + + phnum = READ_ONCE(ehdr->e_phnum); /* only supports phdr that fits in one page */ - if (ehdr->e_phnum > - (PAGE_SIZE - sizeof(Elf32_Ehdr)) / sizeof(Elf32_Phdr)) + if (phnum > (PAGE_SIZE - sizeof(Elf32_Ehdr)) / sizeof(Elf32_Phdr)) return -EINVAL; phdr = (Elf32_Phdr *)(page_addr + sizeof(Elf32_Ehdr)); - for (i = 0; i < ehdr->e_phnum; ++i) { + for (i = 0; i < phnum; ++i) { if (phdr[i].p_type == PT_NOTE && !parse_build_id(page_addr, build_id, size, - page_addr + phdr[i].p_offset, - phdr[i].p_filesz)) + page_addr + READ_ONCE(phdr[i].p_offset), + READ_ONCE(phdr[i].p_filesz))) return 0; } return -EINVAL; @@ -103,7 +110,7 @@ static int get_build_id_64(const void *page_addr, unsigned char *build_id, { Elf64_Ehdr *ehdr = (Elf64_Ehdr *)page_addr; Elf64_Phdr *phdr; - int i; + __u32 i, phnum; /* * FIXME @@ -112,18 +119,19 @@ static int get_build_id_64(const void *page_addr, unsigned char *build_id, */ if (ehdr->e_phoff != sizeof(Elf64_Ehdr)) return -EINVAL; + + phnum = READ_ONCE(ehdr->e_phnum); /* only supports phdr that fits in one page */ - if (ehdr->e_phnum > - (PAGE_SIZE - sizeof(Elf64_Ehdr)) / sizeof(Elf64_Phdr)) + if (phnum > (PAGE_SIZE - sizeof(Elf64_Ehdr)) / sizeof(Elf64_Phdr)) return -EINVAL; phdr = (Elf64_Phdr *)(page_addr + sizeof(Elf64_Ehdr)); - for (i = 0; i < ehdr->e_phnum; ++i) { + for (i = 0; i < phnum; ++i) { if (phdr[i].p_type == PT_NOTE && !parse_build_id(page_addr, build_id, size, - page_addr + phdr[i].p_offset, - phdr[i].p_filesz)) + page_addr + READ_ONCE(phdr[i].p_offset), + READ_ONCE(phdr[i].p_filesz))) return 0; } return -EINVAL; @@ -152,6 +160,10 @@ int build_id_parse(struct vm_area_struct *vma, unsigned char *build_id, page = find_get_page(vma->vm_file->f_mapping, 0); if (!page) return -EFAULT; /* page not mapped */ + if (!PageUptodate(page)) { + put_page(page); + return -EFAULT; + } ret = -EINVAL; page_addr = kmap_local_page(page); -- 2.43.5

1 month, 3 weeks

3
2
0 0

[PATCH] efistub/tpm: Use ACPI reclaim memory for event log to avoid corruption

by Ard Biesheuvel

From: Ard Biesheuvel <ardb(a)kernel.org> The TPM event log table is a Linux specific construct, where the data produced by the GetEventLog() boot service is cached in memory, and passed on to the OS using a EFI configuration table. The use of EFI_LOADER_DATA here results in the region being left unreserved in the E820 memory map constructed by the EFI stub, and this is the memory description that is passed on to the incoming kernel by kexec, which is therefore unaware that the region should be reserved. Even though the utility of the TPM2 event log after a kexec is questionable, any corruption might send the parsing code off into the weeds and crash the kernel. So let's use EFI_ACPI_RECLAIM_MEMORY instead, which is always treated as reserved by the E820 conversion logic. Cc: <stable(a)vger.kernel.org> Reported-by: Breno Leitao <leitao(a)debian.org> Tested-by: Usama Arif <usamaarif642(a)gmail.com> Signed-off-by: Ard Biesheuvel <ardb(a)kernel.org> --- drivers/firmware/efi/libstub/tpm.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/firmware/efi/libstub/tpm.c b/drivers/firmware/efi/libstub/tpm.c index df3182f2e63a..1fd6823248ab 100644 --- a/drivers/firmware/efi/libstub/tpm.c +++ b/drivers/firmware/efi/libstub/tpm.c @@ -96,7 +96,7 @@ static void efi_retrieve_tcg2_eventlog(int version, efi_physical_addr_t log_loca } /* Allocate space for the logs and copy them. */ - status = efi_bs_call(allocate_pool, EFI_LOADER_DATA, + status = efi_bs_call(allocate_pool, EFI_ACPI_RECLAIM_MEMORY, sizeof(*log_tbl) + log_size, (void **)&log_tbl); if (status != EFI_SUCCESS) { -- 2.46.0.662.g92d0881bb0-goog

1 month, 3 weeks

7
19
0 0

[PATCH 1/2] drm/i915: Schedule the HPD poll init work on an unbound workqueue

by Imre Deak

Disabling HPD polling from i915_hpd_poll_init_work() involves probing all display connectors explicitly to account for lost hotplug interrupts. On some platforms (mostly pre-ICL) with HDMI connectors the I2C EDID bit-banging using udelay() triggers in turn the workqueue: i915_hpd_poll_init_work [i915] hogged CPU for >10000us 4 times, consider switching to WQ_UNBOUND warning. Fix the above by scheduling i915_hpd_poll_init_work() on a WQ_UNBOUND workqueue. It's ok to use a system WQ, since i915_hpd_poll_init_work() is properly flushed in intel_hpd_cancel_work(). The connector probing from drm_mode_config::output_poll_work resulting in the same warning is fixed by the next patch. Cc: Tejun Heo <tj(a)kernel.org> Cc: Heiner Kallweit <hkallweit1(a)gmail.com> CC: stable(a)vger.kernel.org # 6.5 Suggested-by: Tejun Heo <tj(a)kernel.org> Suggested-by: Heiner Kallweit <hkallweit1(a)gmail.com> Reported-by: Heiner Kallweit <hkallweit1(a)gmail.com> Closes: https://gitlab.freedesktop.org/drm/intel/-/issues/9245 Link: https://lore.kernel.org/all/f7e21caa-e98d-e5b5-932a-fe12d27fde9b@gmail.com Signed-off-by: Imre Deak <imre.deak(a)intel.com> --- drivers/gpu/drm/i915/display/intel_hotplug.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/drivers/gpu/drm/i915/display/intel_hotplug.c b/drivers/gpu/drm/i915/display/intel_hotplug.c index e8562f6f8bb44..accc2fec562a0 100644 --- a/drivers/gpu/drm/i915/display/intel_hotplug.c +++ b/drivers/gpu/drm/i915/display/intel_hotplug.c @@ -774,7 +774,7 @@ void intel_hpd_poll_enable(struct drm_i915_private *dev_priv) * As well, there's no issue if we race here since we always reschedule * this worker anyway */ - queue_work(dev_priv->unordered_wq, + queue_work(system_unbound_wq, &dev_priv->display.hotplug.poll_init_work); } @@ -803,7 +803,7 @@ void intel_hpd_poll_disable(struct drm_i915_private *dev_priv) return; WRITE_ONCE(dev_priv->display.hotplug.poll_enabled, false); - queue_work(dev_priv->unordered_wq, + queue_work(system_unbound_wq, &dev_priv->display.hotplug.poll_init_work); } -- 2.37.2

1 month, 3 weeks

3
4
0 0

[PATCH 0/2] usb: dwc3: Disable susphy during initialization

by Thinh Nguyen

We notice some platforms set "snps,dis_u3_susphy_quirk" and "snps,dis_u2_susphy_quirk" when they should not need to. Just make sure that the GUSB3PIPECTL.SUSPENDENABLE and GUSB2PHYCFG.SUSPHY are clear during initialization. The host initialization involved xhci. So the dwc3 needs to implement the xhci_plat_priv->plat_start() for xhci to re-enable the suspend bits. Since there's a prerequisite patch to drivers/usb/host/xhci-plat.h that's not a fix patch, this series should go on Greg's usb-testing branch instead of usb-linus. Thinh Nguyen (2): usb: xhci-plat: Don't include xhci.h usb: dwc3: core: Prevent phy suspend during init drivers/usb/dwc3/core.c | 90 +++++++++++++++--------------------- drivers/usb/dwc3/core.h | 1 + drivers/usb/dwc3/gadget.c | 2 + drivers/usb/dwc3/host.c | 27 +++++++++++ drivers/usb/host/xhci-plat.h | 4 +- 5 files changed, 71 insertions(+), 53 deletions(-) base-commit: 3d122e6d27e417a9fa91181922743df26b2cd679 -- 2.28.0

1 month, 4 weeks

5
12
0 0

[PATCH] coresight: etm4x: Fix PID tracing when perf is run in an init PID namespace

by Julien Meunier

The previous implementation limited the tracing capabilities when perf was run in the init PID namespace, making it impossible to trace applications in non-init PID namespaces. This update improves the tracing process by verifying the event owner. This allows us to determine whether the user has the necessary permissions to trace the application. Cc: stable(a)vger.kernel.org Fixes: aab473867fed ("coresight: etm4x: Don't trace PID for non-root PID namespace") Signed-off-by: Julien Meunier <julien.meunier(a)nokia.com> --- drivers/hwtracing/coresight/coresight-etm4x-core.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/drivers/hwtracing/coresight/coresight-etm4x-core.c b/drivers/hwtracing/coresight/coresight-etm4x-core.c index bf01f01964cf..8365307b1aec 100644 --- a/drivers/hwtracing/coresight/coresight-etm4x-core.c +++ b/drivers/hwtracing/coresight/coresight-etm4x-core.c @@ -695,7 +695,7 @@ static int etm4_parse_event_config(struct coresight_device *csdev, /* Only trace contextID when runs in root PID namespace */ if ((attr->config & BIT(ETM_OPT_CTXTID)) && - task_is_in_init_pid_ns(current)) + task_is_in_init_pid_ns(event->owner)) /* bit[6], Context ID tracing bit */ config->cfg |= TRCCONFIGR_CID; @@ -710,7 +710,7 @@ static int etm4_parse_event_config(struct coresight_device *csdev, goto out; } /* Only trace virtual contextID when runs in root PID namespace */ - if (task_is_in_init_pid_ns(current)) + if (task_is_in_init_pid_ns(event->owner)) config->cfg |= TRCCONFIGR_VMID | TRCCONFIGR_VMIDOPT; } -- 2.34.1

2 months

4
9
0 0

[PATCH] nvmet-auth: assign dh_key to NULL after kfree_sensitive

by Vitaliy Shevtsov

ctrl->dh_key might be used across multiple calls to nvmet_setup_dhgroup() for the same controller. So it's better to nullify it after release on error path in order to avoid double free later in nvmet_destroy_auth(). Found by Linux Verification Center (linuxtesting.org) with Svace. Fixes: 7a277c37d352 ("nvmet-auth: Diffie-Hellman key exchange support") Cc: stable(a)vger.kernel.org Signed-off-by: Vitaliy Shevtsov <v.shevtsov(a)maxima.ru> --- drivers/nvme/target/auth.c | 1 + 1 file changed, 1 insertion(+) diff --git a/drivers/nvme/target/auth.c b/drivers/nvme/target/auth.c index e900525b7866..7bca64de4a2f 100644 --- a/drivers/nvme/target/auth.c +++ b/drivers/nvme/target/auth.c @@ -101,6 +101,7 @@ int nvmet_setup_dhgroup(struct nvmet_ctrl *ctrl, u8 dhgroup_id) pr_debug("%s: ctrl %d failed to generate private key, err %d\n", __func__, ctrl->cntlid, ret); kfree_sensitive(ctrl->dh_key); + ctrl->dh_key = NULL; return ret; } ctrl->dh_keysize = crypto_kpp_maxsize(ctrl->dh_tfm); -- 2.46.1

2 months

5
4
0 0

Re: [REGRESSION] kexec does firmware reboot in kernel v6.7.6

by Steve Wahl

On Tue, Mar 12, 2024 at 07:04:10AM -0400, Eric Hagberg wrote: > On Thu, Mar 7, 2024 at 11:33 AM Steve Wahl <steve.wahl(a)hpe.com> wrote: > > What Linux Distribution are you running on that machine? My guess > > would be that this is not distro related; if you are running something > > quite different from Pavin that would confirm this. > > Distro in use is Rocky 8, so it’s pretty clear not to be distro-specific. > > > I found an AMD based system to try to reproduce this on. > > yeah, it probably requires either a specific cpu or set or devices plus cpu > to trigger… found that it also affects Dell R7625 servers in addition to > the R6615s I agree that it's likely the CPU or particular set of surrounding devices that trigger the problem. I have not succeeded in reproducing the problem yet. I tried an AMD based system lent to me, but it's probably the wrong generation (AMD EPYC 7251) and I didn't see the problem. I have a line on a system that's more in line with the systems the bug was reported on that I should be able to try tomorrow. I would love to have some direction from the community at large on this. The fact that nogbpages on the command line causes the same problem without my patch suggests it's not bad code directly in my patch, but something in the way kexec reacts to the resulting identity map. One quick solution would be a kernel command line parameter to select between the previous identity map creation behavior and the new behavior. E.g. in addition to "nogbpages", we could have "somegbpages" and "allgbpages" -- or gbpages=[all, some, none] with nogbpages a synonym for backwards compatibility. But I don't want to introduce a new command line parameter if the actual problem can be understood and fixed. The question is how much time do I have to persue a direct fix before some other action needs to be taken? Thanks, --> Steve Wahl -- Steve Wahl, Hewlett Packard Enterprise

2 months

6
15
0 0

[PATCH] mm: avoid unconditional one-tick sleep when swapcache_prepare fails

by Barry Song

From: Barry Song <v-songbaohua(a)oppo.com> Commit 13ddaf26be32 ("mm/swap: fix race when skipping swapcache") introduced an unconditional one-tick sleep when `swapcache_prepare()` fails, which has led to reports of UI stuttering on latency-sensitive Android devices. To address this, we can use a waitqueue to wake up tasks that fail `swapcache_prepare()` sooner, instead of always sleeping for a full tick. While tasks may occasionally be woken by an unrelated `do_swap_page()`, this method is preferable to two scenarios: rapid re-entry into page faults, which can cause livelocks, and multiple millisecond sleeps, which visibly degrade user experience. Oven's testing shows that a single waitqueue resolves the UI stuttering issue. If a 'thundering herd' problem becomes apparent later, a waitqueue hash similar to `folio_wait_table[PAGE_WAIT_TABLE_SIZE]` for page bit locks can be introduced. Fixes: 13ddaf26be32 ("mm/swap: fix race when skipping swapcache") Cc: Kairui Song <kasong(a)tencent.com> Cc: "Huang, Ying" <ying.huang(a)intel.com> Cc: Yu Zhao <yuzhao(a)google.com> Cc: David Hildenbrand <david(a)redhat.com> Cc: Chris Li <chrisl(a)kernel.org> Cc: Hugh Dickins <hughd(a)google.com> Cc: Johannes Weiner <hannes(a)cmpxchg.org> Cc: Matthew Wilcox (Oracle) <willy(a)infradead.org> Cc: Michal Hocko <mhocko(a)suse.com> Cc: Minchan Kim <minchan(a)kernel.org> Cc: Yosry Ahmed <yosryahmed(a)google.com> Cc: SeongJae Park <sj(a)kernel.org> Cc: Kalesh Singh <kaleshsingh(a)google.com> Cc: Suren Baghdasaryan <surenb(a)google.com> Cc: <stable(a)vger.kernel.org> Reported-by: Oven Liyang <liyangouwen1(a)oppo.com> Tested-by: Oven Liyang <liyangouwen1(a)oppo.com> Signed-off-by: Barry Song <v-songbaohua(a)oppo.com> --- mm/memory.c | 13 +++++++++++-- 1 file changed, 11 insertions(+), 2 deletions(-) diff --git a/mm/memory.c b/mm/memory.c index 2366578015ad..6913174f7f41 100644 --- a/mm/memory.c +++ b/mm/memory.c @@ -4192,6 +4192,8 @@ static struct folio *alloc_swap_folio(struct vm_fault *vmf) } #endif /* CONFIG_TRANSPARENT_HUGEPAGE */ +static DECLARE_WAIT_QUEUE_HEAD(swapcache_wq); + /* * We enter with non-exclusive mmap_lock (to exclude vma changes, * but allow concurrent faults), and pte mapped but not yet locked. @@ -4204,6 +4206,7 @@ vm_fault_t do_swap_page(struct vm_fault *vmf) { struct vm_area_struct *vma = vmf->vma; struct folio *swapcache, *folio = NULL; + DECLARE_WAITQUEUE(wait, current); struct page *page; struct swap_info_struct *si = NULL; rmap_t rmap_flags = RMAP_NONE; @@ -4302,7 +4305,9 @@ vm_fault_t do_swap_page(struct vm_fault *vmf) * Relax a bit to prevent rapid * repeated page faults. */ + add_wait_queue(&swapcache_wq, &wait); schedule_timeout_uninterruptible(1); + remove_wait_queue(&swapcache_wq, &wait); goto out_page; } need_clear_cache = true; @@ -4609,8 +4614,10 @@ vm_fault_t do_swap_page(struct vm_fault *vmf) pte_unmap_unlock(vmf->pte, vmf->ptl); out: /* Clear the swap cache pin for direct swapin after PTL unlock */ - if (need_clear_cache) + if (need_clear_cache) { swapcache_clear(si, entry, nr_pages); + wake_up(&swapcache_wq); + } if (si) put_swap_device(si); return ret; @@ -4625,8 +4632,10 @@ vm_fault_t do_swap_page(struct vm_fault *vmf) folio_unlock(swapcache); folio_put(swapcache); } - if (need_clear_cache) + if (need_clear_cache) { swapcache_clear(si, entry, nr_pages); + wake_up(&swapcache_wq); + } if (si) put_swap_device(si); return ret; -- 2.34.1

2 months

4
20
0 0

2024

2023

2022

2021

2020

2019

2018

2017

Linux-stable-mirror September 2024