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

FAILED: patch "[PATCH] x86/resctrl: Fix miscount of bandwidth event when" failed to apply to 5.10-stable tree

by gregkh＠linuxfoundation.org

The patch below does not apply to the 5.10-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>. To reproduce the conflict and resubmit, you may use the following commands: git fetch https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/ linux-5.10.y git checkout FETCH_HEAD git cherry-pick -x 15292f1b4c55a3a7c940dbcb6cb8793871ed3d92 # <resolve conflicts, build, test, etc.> git commit -s git send-email --to '<stable(a)vger.kernel.org>' --in-reply-to '2025102056-qualify-unopposed-be08@gregkh' --subject-prefix 'PATCH 5.10.y' HEAD^.. Possible dependencies: thanks, greg k-h ------------------ original commit in Linus's tree ------------------ From 15292f1b4c55a3a7c940dbcb6cb8793871ed3d92 Mon Sep 17 00:00:00 2001 From: Babu Moger <babu.moger(a)amd.com> Date: Fri, 10 Oct 2025 12:08:35 -0500 Subject: [PATCH] x86/resctrl: Fix miscount of bandwidth event when reactivating previously unavailable RMID Users can create as many monitoring groups as the number of RMIDs supported by the hardware. However, on AMD systems, only a limited number of RMIDs are guaranteed to be actively tracked by the hardware. RMIDs that exceed this limit are placed in an "Unavailable" state. When a bandwidth counter is read for such an RMID, the hardware sets MSR_IA32_QM_CTR.Unavailable (bit 62). When such an RMID starts being tracked again the hardware counter is reset to zero. MSR_IA32_QM_CTR.Unavailable remains set on first read after tracking re-starts and is clear on all subsequent reads as long as the RMID is tracked. resctrl miscounts the bandwidth events after an RMID transitions from the "Unavailable" state back to being tracked. This happens because when the hardware starts counting again after resetting the counter to zero, resctrl in turn compares the new count against the counter value stored from the previous time the RMID was tracked. This results in resctrl computing an event value that is either undercounting (when new counter is more than stored counter) or a mistaken overflow (when new counter is less than stored counter). Reset the stored value (arch_mbm_state::prev_msr) of MSR_IA32_QM_CTR to zero whenever the RMID is in the "Unavailable" state to ensure accurate counting after the RMID resets to zero when it starts to be tracked again. Example scenario that results in mistaken overflow ================================================== 1. The resctrl filesystem is mounted, and a task is assigned to a monitoring group. $mount -t resctrl resctrl /sys/fs/resctrl $mkdir /sys/fs/resctrl/mon_groups/test1/ $echo 1234 > /sys/fs/resctrl/mon_groups/test1/tasks $cat /sys/fs/resctrl/mon_groups/test1/mon_data/mon_L3_*/mbm_total_bytes 21323 <- Total bytes on domain 0 "Unavailable" <- Total bytes on domain 1 Task is running on domain 0. Counter on domain 1 is "Unavailable". 2. The task runs on domain 0 for a while and then moves to domain 1. The counter starts incrementing on domain 1. $cat /sys/fs/resctrl/mon_groups/test1/mon_data/mon_L3_*/mbm_total_bytes 7345357 <- Total bytes on domain 0 4545 <- Total bytes on domain 1 3. At some point, the RMID in domain 0 transitions to the "Unavailable" state because the task is no longer executing in that domain. $cat /sys/fs/resctrl/mon_groups/test1/mon_data/mon_L3_*/mbm_total_bytes "Unavailable" <- Total bytes on domain 0 434341 <- Total bytes on domain 1 4. Since the task continues to migrate between domains, it may eventually return to domain 0. $cat /sys/fs/resctrl/mon_groups/test1/mon_data/mon_L3_*/mbm_total_bytes 17592178699059 <- Overflow on domain 0 3232332 <- Total bytes on domain 1 In this case, the RMID on domain 0 transitions from "Unavailable" state to active state. The hardware sets MSR_IA32_QM_CTR.Unavailable (bit 62) when the counter is read and begins tracking the RMID counting from 0. Subsequent reads succeed but return a value smaller than the previously saved MSR value (7345357). Consequently, the resctrl's overflow logic is triggered, it compares the previous value (7345357) with the new, smaller value and incorrectly interprets this as a counter overflow, adding a large delta. In reality, this is a false positive: the counter did not overflow but was simply reset when the RMID transitioned from "Unavailable" back to active state. Here is the text from APM [1] available from [2]. "In PQOS Version 2.0 or higher, the MBM hardware will set the U bit on the first QM_CTR read when it begins tracking an RMID that it was not previously tracking. The U bit will be zero for all subsequent reads from that RMID while it is still tracked by the hardware. Therefore, a QM_CTR read with the U bit set when that RMID is in use by a processor can be considered 0 when calculating the difference with a subsequent read." [1] AMD64 Architecture Programmer's Manual Volume 2: System Programming Publication # 24593 Revision 3.41 section 19.3.3 Monitoring L3 Memory Bandwidth (MBM). [ bp: Split commit message into smaller paragraph chunks for better consumption. ] Fixes: 4d05bf71f157d ("x86/resctrl: Introduce AMD QOS feature") Signed-off-by: Babu Moger <babu.moger(a)amd.com> Signed-off-by: Borislav Petkov (AMD) <bp(a)alien8.de> Reviewed-by: Reinette Chatre <reinette.chatre(a)intel.com> Tested-by: Reinette Chatre <reinette.chatre(a)intel.com> Cc: stable(a)vger.kernel.org # needs adjustments for <= v6.17 Link: https://bugzilla.kernel.org/show_bug.cgi?id=206537 # [2] diff --git a/arch/x86/kernel/cpu/resctrl/monitor.c b/arch/x86/kernel/cpu/resctrl/monitor.c index c8945610d455..2cd25a0d4637 100644 --- a/arch/x86/kernel/cpu/resctrl/monitor.c +++ b/arch/x86/kernel/cpu/resctrl/monitor.c @@ -242,7 +242,9 @@ int resctrl_arch_rmid_read(struct rdt_resource *r, struct rdt_mon_domain *d, u32 unused, u32 rmid, enum resctrl_event_id eventid, u64 *val, void *ignored) { + struct rdt_hw_mon_domain *hw_dom = resctrl_to_arch_mon_dom(d); int cpu = cpumask_any(&d->hdr.cpu_mask); + struct arch_mbm_state *am; u64 msr_val; u32 prmid; int ret; @@ -251,12 +253,16 @@ int resctrl_arch_rmid_read(struct rdt_resource *r, struct rdt_mon_domain *d, prmid = logical_rmid_to_physical_rmid(cpu, rmid); ret = __rmid_read_phys(prmid, eventid, &msr_val); - if (ret) - return ret; - *val = get_corrected_val(r, d, rmid, eventid, msr_val); + if (!ret) { + *val = get_corrected_val(r, d, rmid, eventid, msr_val); + } else if (ret == -EINVAL) { + am = get_arch_mbm_state(hw_dom, rmid, eventid); + if (am) + am->prev_msr = 0; + } - return 0; + return ret; } static int __cntr_id_read(u32 cntr_id, u64 *val)

2 weeks, 2 days

2
1
0 0

[PATCH AUTOSEL 6.17-5.4] ACPI: property: Return present device nodes only on fwnode interface

by Sasha Levin

From: Sakari Ailus <sakari.ailus(a)linux.intel.com> [ Upstream commit d9f866b2bb3eec38b3734f1fed325ec7c55ccdfa ] fwnode_graph_get_next_subnode() may return fwnode backed by ACPI device nodes and there has been no check these devices are present in the system, unlike there has been on fwnode OF backend. In order to provide consistent behaviour towards callers, add a check for device presence by introducing a new function acpi_get_next_present_subnode(), used as the get_next_child_node() fwnode operation that also checks device node presence. Signed-off-by: Sakari Ailus <sakari.ailus(a)linux.intel.com> Reviewed-by: Laurent Pinchart <laurent.pinchart+renesas(a)ideasonboard.com> Reviewed-by: Jonathan Cameron <jonathan.cameron(a)huawei.com> Link: https://patch.msgid.link/20251001102636.1272722-2-sakari.ailus@linux.intel.… [ rjw: Kerneldoc comment and changelog edits ] Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki(a)intel.com> Signed-off-by: Sasha Levin <sashal(a)kernel.org> --- LLM Generated explanations, may be completely bogus: YES – this change fixes a real behavioural bug in the ACPI fwnode backend and should go to stable. - `drivers/acpi/property.c:1375` adds `acpi_get_next_present_subnode()` so `.get_next_child_node` skips ACPI child devices whose `_STA` says they are absent (`acpi_device_is_present()`), while still returning data subnodes unchanged. - The new helper is now wired into the ACPI fwnode ops (`drivers/acpi/property.c:1731`), making generic helpers such as `fwnode_get_next_child_node()` and macros like `fwnode_for_each_child_node` (`include/linux/property.h:167`) behave the same as the OF backend, which already filtered unavailable children via `of_get_next_available_child()` (`drivers/of/property.c:1070`). - Several core helpers assume disabled endpoints never surface: e.g. `fwnode_graph_get_endpoint_by_id()` in `drivers/base/property.c:1286` promises to hide endpoints on disabled devices, and higher layers such as `v4l2_fwnode_reference_get_int_prop()` (`drivers/media/v4l2-core/v4l2-fwnode.c:1064`) iterate child nodes without rechecking availability. On ACPI systems today they still see powered-off devices, leading to asynchronous notifiers that wait forever for hardware that can’t appear, or to bogus graph enumerations. This patch closes that gap. Risk is low: it only suppresses ACPI device nodes already known to be absent, aligns behaviour with DT userspace expectations, and leaves data nodes untouched. No extra dependencies are required, so the fix is self- contained and appropriate for stable backporting. Suggest running existing ACPI graph users (media/typec drivers) after backport to confirm no regressions. drivers/acpi/property.c | 24 +++++++++++++++++++++++- 1 file changed, 23 insertions(+), 1 deletion(-) diff --git a/drivers/acpi/property.c b/drivers/acpi/property.c index c086786fe84cb..d74678f0ba4af 100644 --- a/drivers/acpi/property.c +++ b/drivers/acpi/property.c @@ -1357,6 +1357,28 @@ struct fwnode_handle *acpi_get_next_subnode(const struct fwnode_handle *fwnode, return NULL; } +/* + * acpi_get_next_present_subnode - Return the next present child node handle + * @fwnode: Firmware node to find the next child node for. + * @child: Handle to one of the device's child nodes or a null handle. + * + * Like acpi_get_next_subnode(), but the device nodes returned by + * acpi_get_next_present_subnode() are guaranteed to be present. + * + * Returns: The fwnode handle of the next present sub-node. + */ +static struct fwnode_handle * +acpi_get_next_present_subnode(const struct fwnode_handle *fwnode, + struct fwnode_handle *child) +{ + do { + child = acpi_get_next_subnode(fwnode, child); + } while (is_acpi_device_node(child) && + !acpi_device_is_present(to_acpi_device_node(child))); + + return child; +} + /** * acpi_node_get_parent - Return parent fwnode of this fwnode * @fwnode: Firmware node whose parent to get @@ -1701,7 +1723,7 @@ static int acpi_fwnode_irq_get(const struct fwnode_handle *fwnode, .property_read_string_array = \ acpi_fwnode_property_read_string_array, \ .get_parent = acpi_node_get_parent, \ - .get_next_child_node = acpi_get_next_subnode, \ + .get_next_child_node = acpi_get_next_present_subnode, \ .get_named_child_node = acpi_fwnode_get_named_child_node, \ .get_name = acpi_fwnode_get_name, \ .get_name_prefix = acpi_fwnode_get_name_prefix, \ -- 2.51.0

2 weeks, 2 days

2
46
0 0

[PATCH] x86/fpu: Ensure XFD state on signal delivery

by Chang S. Bae

Sean reported [1] the following splat when running KVM tests: WARNING: CPU: 232 PID: 15391 at xfd_validate_state+0x65/0x70 Call Trace: <TASK> fpu__clear_user_states+0x9c/0x100 arch_do_signal_or_restart+0x142/0x210 exit_to_user_mode_loop+0x55/0x100 do_syscall_64+0x205/0x2c0 entry_SYSCALL_64_after_hwframe+0x4b/0x53 Chao further identified [2] a reproducible scenarios involving signal delivery: a non-AMX task is preempted by an AMX-enabled task which modifies the XFD MSR. When the non-AMX task resumes and reloads XSTATE with init values, a warning is triggered due to a mismatch between fpstate::xfd and the CPU's current XFD state. fpu__clear_user_states() does not currently re-synchronize the XFD state after such preemption. Invoke xfd_update_state() which detects and corrects the mismatch if the dynamic feature is enabled. This also benefits the sigreturn path, as fpu__restore_sig() may call fpu__clear_user_states() when the sigframe is inaccessible. Fixes: 672365477ae8a ("x86/fpu: Update XFD state where required") Reported-by: Sean Christopherson <seanjc(a)google.com> Closes: https://lore.kernel.org/lkml/aDCo_SczQOUaB2rS@google.com [1] Tested-by: Chao Gao <chao.gao(a)intel.com> Signed-off-by: Chang S. Bae <chang.seok.bae(a)intel.com> Cc: stable(a)vger.kernel.org Link: https://lore.kernel.org/all/aDWbctO%2FRfTGiCg3@intel.com [2] --- arch/x86/kernel/fpu/core.c | 3 +++ 1 file changed, 3 insertions(+) diff --git a/arch/x86/kernel/fpu/core.c b/arch/x86/kernel/fpu/core.c index ea138583dd92..5fa782a2ae7c 100644 --- a/arch/x86/kernel/fpu/core.c +++ b/arch/x86/kernel/fpu/core.c @@ -800,6 +800,9 @@ void fpu__clear_user_states(struct fpu *fpu) !fpregs_state_valid(fpu, smp_processor_id())) os_xrstor_supervisor(fpu->fpstate); + /* Ensure XFD state is in sync before reloading XSTATE */ + xfd_update_state(fpu->fpstate); + /* Reset user states in registers. */ restore_fpregs_from_init_fpstate(XFEATURE_MASK_USER_RESTORE); -- 2.48.1

2 weeks, 2 days

3
2
0 0

[PATCH] regmap: irq: Correct documentation of wake_invert flag

by Shawn Guo

From: Shawn Guo <shawnguo(a)kernel.org> Per commit 9442490a0286 ("regmap: irq: Support wake IRQ mask inversion") the wake_invert flag is to support enable register, so cleared bits are wake disabled. Fixes: 68622bdfefb9 ("regmap: irq: document mask/wake_invert flags") Cc: stable(a)vger.kernel.org Signed-off-by: Shawn Guo <shawnguo(a)kernel.org> --- include/linux/regmap.h | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/include/linux/regmap.h b/include/linux/regmap.h index 4e1ac1fbcec4..55343795644b 100644 --- a/include/linux/regmap.h +++ b/include/linux/regmap.h @@ -1643,7 +1643,7 @@ struct regmap_irq_chip_data; * @status_invert: Inverted status register: cleared bits are active interrupts. * @status_is_level: Status register is actuall signal level: Xor status * register with previous value to get active interrupts. - * @wake_invert: Inverted wake register: cleared bits are wake enabled. + * @wake_invert: Inverted wake register: cleared bits are wake disabled. * @type_in_mask: Use the mask registers for controlling irq type. Use this if * the hardware provides separate bits for rising/falling edge * or low/high level interrupts and they should be combined into -- 2.43.0

2 weeks, 2 days

2
3
0 0

[PATCH] ALSA: wavefront: fix buffer overflow in longname construction

by moonafterrain＠outlook.com

From: Junrui Luo <moonafterrain(a)outlook.com> The snd_wavefront_probe() function constructs the card->longname string using unsafe sprintf() calls that can overflow the 80-byte buffer when module parameters contain large values. The vulnerability exists at wavefront.c where multiple sprintf() operations append to card->longname without length checking. Fix by replacing all sprintf() calls with scnprintf() and proper length tracking to ensure writes never exceed sizeof(card->longname). Reported-by: Yuhao Jiang <danisjiang(a)gmail.com> Reported-by: Junrui Luo <moonafterrain(a)outlook.com> Fixes: 1da177e4c3f4 ("Linux-2.6.12-rc2") Cc: stable(a)vger.kernel.org Signed-off-by: Junrui Luo <moonafterrain(a)outlook.com> --- sound/isa/wavefront/wavefront.c | 40 ++++++++++++++++++++------------- 1 file changed, 25 insertions(+), 15 deletions(-) diff --git a/sound/isa/wavefront/wavefront.c b/sound/isa/wavefront/wavefront.c index 07c68568091d..74ea3a67620c 100644 --- a/sound/isa/wavefront/wavefront.c +++ b/sound/isa/wavefront/wavefront.c @@ -343,6 +343,7 @@ snd_wavefront_probe (struct snd_card *card, int dev) struct snd_rawmidi *ics2115_external_rmidi = NULL; struct snd_hwdep *fx_processor; int hw_dev = 0, midi_dev = 0, err; + size_t len, rem; /* --------- PCM --------------- */ @@ -492,26 +493,35 @@ snd_wavefront_probe (struct snd_card *card, int dev) length restrictions */ - sprintf(card->longname, "%s PCM 0x%lx irq %d dma %d", - card->driver, - chip->port, - cs4232_pcm_irq[dev], - dma1[dev]); + len = scnprintf(card->longname, sizeof(card->longname), + "%s PCM 0x%lx irq %d dma %d", + card->driver, + chip->port, + cs4232_pcm_irq[dev], + dma1[dev]); - if (dma2[dev] >= 0 && dma2[dev] < 8) - sprintf(card->longname + strlen(card->longname), "&%d", dma2[dev]); + if (dma2[dev] >= 0 && dma2[dev] < 8 && len < sizeof(card->longname)) { + rem = sizeof(card->longname) - len; + len += scnprintf(card->longname + len, rem, "&%d", dma2[dev]); + } if (cs4232_mpu_port[dev] > 0 && cs4232_mpu_port[dev] != SNDRV_AUTO_PORT) { - sprintf (card->longname + strlen (card->longname), - " MPU-401 0x%lx irq %d", - cs4232_mpu_port[dev], - cs4232_mpu_irq[dev]); + if (len < sizeof(card->longname)) { + rem = sizeof(card->longname) - len; + len += scnprintf(card->longname + len, rem, + " MPU-401 0x%lx irq %d", + cs4232_mpu_port[dev], + cs4232_mpu_irq[dev]); + } } - sprintf (card->longname + strlen (card->longname), - " SYNTH 0x%lx irq %d", - ics2115_port[dev], - ics2115_irq[dev]); + if (len < sizeof(card->longname)) { + rem = sizeof(card->longname) - len; + scnprintf(card->longname + len, rem, + " SYNTH 0x%lx irq %d", + ics2115_port[dev], + ics2115_irq[dev]); + } return snd_card_register(card); } -- 2.51.1.dirty

2 weeks, 2 days

2
1
0 0

[PATCH] mm/mremap: Honour writable bit in mremap pte batching

by Dev Jain

Currently mremap folio pte batch ignores the writable bit during figuring out a set of similar ptes mapping the same folio. Suppose that the first pte of the batch is writable while the others are not - set_ptes will end up setting the writable bit on the other ptes, which is a violation of mremap semantics. Therefore, use FPB_RESPECT_WRITE to check the writable bit while determining the pte batch. Cc: stable(a)vger.kernel.org #6.17 Fixes: f822a9a81a31 ("mm: optimize mremap() by PTE batching") Reported-by: David Hildenbrand <david(a)redhat.com> Debugged-by: David Hildenbrand <david(a)redhat.com> Signed-off-by: Dev Jain <dev.jain(a)arm.com> --- mm-selftests pass. Based on mm-new. Need David H. to confirm whether the repro passes. mm/mremap.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/mm/mremap.c b/mm/mremap.c index a7f531c17b79..8ad06cf50783 100644 --- a/mm/mremap.c +++ b/mm/mremap.c @@ -187,7 +187,7 @@ static int mremap_folio_pte_batch(struct vm_area_struct *vma, unsigned long addr if (!folio || !folio_test_large(folio)) return 1; - return folio_pte_batch(folio, ptep, pte, max_nr); + return folio_pte_batch_flags(folio, NULL, ptep, &pte, max_nr, FPB_RESPECT_WRITE); } static int move_ptes(struct pagetable_move_control *pmc, -- 2.30.2

2 weeks, 2 days

4
3
0 0

, conoce mejor a tu candidato antes de contratar

by Valeria Pérez

Evaluaciones Psicométricas para RR.HH. body { margin: 0; padding: 0; font-family: Arial, Helvetica, sans-serif; font-size: 14px; color: #333; background-color: #ffffff; } table { border-spacing: 0; width: 100%; max-width: 600px; margin: auto; } td { padding: 12px 20px; } a { color: #1a73e8; text-decoration: none; } .footer { font-size: 12px; color: #888888; text-align: center; } Mejora tus procesos de selección con evaluaciones psicométricas fáciles y confiables. Hola, , Sabemos que encontrar al candidato ideal va más allá del currículum. Por eso quiero contarte brevemente sobre PsicoSmart, una plataforma que ayuda a equipos de RR.HH. a evaluar talento con pruebas psicométricas rápidas, confiables y fáciles de aplicar. Con PsicoSmart puedes: Aplicar evaluaciones psicométricas 100% en línea. Elegir entre más de 31 pruebas psicométricas Generar reportes automáticos, visuales y fáciles de interpretar. Comparar resultados entre candidatos en segundos. Ahorrar horas valiosas del proceso de selección. Si estás buscando mejorar tus contrataciones, te lo recomiendo muchísimo. Si quieres conocer más puedes responder este correo o simplemente contactarme, mis datos están abajo. Saludos, ----------------------------- Atte.: Valeria Pérez Ciudad de México: (55) 5018 0565 WhatsApp: +52 33 1607 2089 Si no deseas recibir más correos, haz clic aquí para darte de baja. Para remover su dirección de esta lista haga <a href="https://s1.arrobamail.com/unsuscribe.php?id=yiwtsrewisppiseup">click aquí</a>

2 weeks, 2 days

1
0
0 0

[PATCH] scsi: aic94xx: fix use-after-free in device removal path

by moonafterrain＠outlook.com

From: Junrui Luo <moonafterrain(a)outlook.com> The asd_pci_remove() function fails to synchronize with pending tasklets before freeing the asd_ha structure, leading to a potential use-after-free vulnerability. When a device removal is triggered (via hot-unplug or module unload), race condition can occur. The fix adds tasklet_kill() before freeing the asd_ha structure, ensuring all scheduled tasklets complete before cleanup proceeds. Reported-by: Yuhao Jiang <danisjiang(a)gmail.com> Reported-by: Junrui Luo <moonafterrain(a)outlook.com> Fixes: 2908d778ab3e ("[SCSI] aic94xx: new driver") Cc: stable(a)vger.kernel.org Signed-off-by: Junrui Luo <moonafterrain(a)outlook.com> --- drivers/scsi/aic94xx/aic94xx_init.c | 3 +++ 1 file changed, 3 insertions(+) diff --git a/drivers/scsi/aic94xx/aic94xx_init.c b/drivers/scsi/aic94xx/aic94xx_init.c index adf3d9145606..95f3620059f7 100644 --- a/drivers/scsi/aic94xx/aic94xx_init.c +++ b/drivers/scsi/aic94xx/aic94xx_init.c @@ -882,6 +882,9 @@ static void asd_pci_remove(struct pci_dev *dev) asd_disable_ints(asd_ha); + /* Ensure all scheduled tasklets complete before freeing resources */ + tasklet_kill(&asd_ha->seq.dl_tasklet); + asd_remove_dev_attrs(asd_ha); /* XXX more here as needed */ -- 2.51.1.dirty

2 weeks, 2 days

1
0
0 0

[PATCH] agp/uninorth: fix out-of-bounds access with negative pg_start

by moonafterrain＠outlook.com

From: Junrui Luo <moonafterrain(a)outlook.com> The uninorth_insert_memory and uninorth_remove_memory functions lack proper validation of the pg_start parameter before using it as an array index into the GATT (Graphics Address Translation Table). The current bounds check fails to reject negative pg_start values and potentially causes out-of-bounds writes. Fix by explicitly checking that pg_start is non-negative before performing bounds checking. This makes the security requirement clear and does not rely on implicit type conversion behavior. The uninorth_remove_memory function has no bounds checking at all, so add the same validation there. Reported-by: Yuhao Jiang <danisjiang(a)gmail.com> Reported-by: Junrui Luo <moonafterrain(a)outlook.com> Fixes: 1da177e4c3f4 ("Linux-2.6.12-rc2") Cc: stable(a)vger.kernel.org Signed-off-by: Junrui Luo <moonafterrain(a)outlook.com> --- drivers/char/agp/uninorth-agp.c | 15 ++++++++++++++- 1 file changed, 14 insertions(+), 1 deletion(-) diff --git a/drivers/char/agp/uninorth-agp.c b/drivers/char/agp/uninorth-agp.c index b8d7115b8c9e..4e0b949016f7 100644 --- a/drivers/char/agp/uninorth-agp.c +++ b/drivers/char/agp/uninorth-agp.c @@ -169,7 +169,9 @@ static int uninorth_insert_memory(struct agp_memory *mem, off_t pg_start, int ty temp = agp_bridge->current_size; num_entries = A_SIZE_32(temp)->num_entries; - if ((pg_start + mem->page_count) > num_entries) + if (pg_start < 0 || + (pg_start + mem->page_count) > num_entries || + (pg_start + mem->page_count) < pg_start) return -EINVAL; gp = (u32 *) &agp_bridge->gatt_table[pg_start]; @@ -200,6 +202,9 @@ static int uninorth_insert_memory(struct agp_memory *mem, off_t pg_start, int ty static int uninorth_remove_memory(struct agp_memory *mem, off_t pg_start, int type) { size_t i; + int num_entries; + void *temp; + u32 *gp; int mask_type; @@ -215,6 +220,14 @@ static int uninorth_remove_memory(struct agp_memory *mem, off_t pg_start, int ty if (mem->page_count == 0) return 0; + temp = agp_bridge->current_size; + num_entries = A_SIZE_32(temp)->num_entries; + + if (pg_start < 0 || + (pg_start + mem->page_count) > num_entries || + (pg_start + mem->page_count) < pg_start) + return -EINVAL; + gp = (u32 *) &agp_bridge->gatt_table[pg_start]; for (i = 0; i < mem->page_count; ++i) { gp[i] = scratch_value; -- 2.51.1.dirty

2 weeks, 2 days

1
0
0 0

FAILED: patch "[PATCH] x86/resctrl: Fix miscount of bandwidth event when" failed to apply to 5.15-stable tree

by gregkh＠linuxfoundation.org

The patch below does not apply to the 5.15-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>. To reproduce the conflict and resubmit, you may use the following commands: git fetch https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/ linux-5.15.y git checkout FETCH_HEAD git cherry-pick -x 15292f1b4c55a3a7c940dbcb6cb8793871ed3d92 # <resolve conflicts, build, test, etc.> git commit -s git send-email --to '<stable(a)vger.kernel.org>' --in-reply-to '2025102054-slacks-ambush-f774@gregkh' --subject-prefix 'PATCH 5.15.y' HEAD^.. Possible dependencies: thanks, greg k-h ------------------ original commit in Linus's tree ------------------ From 15292f1b4c55a3a7c940dbcb6cb8793871ed3d92 Mon Sep 17 00:00:00 2001 From: Babu Moger <babu.moger(a)amd.com> Date: Fri, 10 Oct 2025 12:08:35 -0500 Subject: [PATCH] x86/resctrl: Fix miscount of bandwidth event when reactivating previously unavailable RMID Users can create as many monitoring groups as the number of RMIDs supported by the hardware. However, on AMD systems, only a limited number of RMIDs are guaranteed to be actively tracked by the hardware. RMIDs that exceed this limit are placed in an "Unavailable" state. When a bandwidth counter is read for such an RMID, the hardware sets MSR_IA32_QM_CTR.Unavailable (bit 62). When such an RMID starts being tracked again the hardware counter is reset to zero. MSR_IA32_QM_CTR.Unavailable remains set on first read after tracking re-starts and is clear on all subsequent reads as long as the RMID is tracked. resctrl miscounts the bandwidth events after an RMID transitions from the "Unavailable" state back to being tracked. This happens because when the hardware starts counting again after resetting the counter to zero, resctrl in turn compares the new count against the counter value stored from the previous time the RMID was tracked. This results in resctrl computing an event value that is either undercounting (when new counter is more than stored counter) or a mistaken overflow (when new counter is less than stored counter). Reset the stored value (arch_mbm_state::prev_msr) of MSR_IA32_QM_CTR to zero whenever the RMID is in the "Unavailable" state to ensure accurate counting after the RMID resets to zero when it starts to be tracked again. Example scenario that results in mistaken overflow ================================================== 1. The resctrl filesystem is mounted, and a task is assigned to a monitoring group. $mount -t resctrl resctrl /sys/fs/resctrl $mkdir /sys/fs/resctrl/mon_groups/test1/ $echo 1234 > /sys/fs/resctrl/mon_groups/test1/tasks $cat /sys/fs/resctrl/mon_groups/test1/mon_data/mon_L3_*/mbm_total_bytes 21323 <- Total bytes on domain 0 "Unavailable" <- Total bytes on domain 1 Task is running on domain 0. Counter on domain 1 is "Unavailable". 2. The task runs on domain 0 for a while and then moves to domain 1. The counter starts incrementing on domain 1. $cat /sys/fs/resctrl/mon_groups/test1/mon_data/mon_L3_*/mbm_total_bytes 7345357 <- Total bytes on domain 0 4545 <- Total bytes on domain 1 3. At some point, the RMID in domain 0 transitions to the "Unavailable" state because the task is no longer executing in that domain. $cat /sys/fs/resctrl/mon_groups/test1/mon_data/mon_L3_*/mbm_total_bytes "Unavailable" <- Total bytes on domain 0 434341 <- Total bytes on domain 1 4. Since the task continues to migrate between domains, it may eventually return to domain 0. $cat /sys/fs/resctrl/mon_groups/test1/mon_data/mon_L3_*/mbm_total_bytes 17592178699059 <- Overflow on domain 0 3232332 <- Total bytes on domain 1 In this case, the RMID on domain 0 transitions from "Unavailable" state to active state. The hardware sets MSR_IA32_QM_CTR.Unavailable (bit 62) when the counter is read and begins tracking the RMID counting from 0. Subsequent reads succeed but return a value smaller than the previously saved MSR value (7345357). Consequently, the resctrl's overflow logic is triggered, it compares the previous value (7345357) with the new, smaller value and incorrectly interprets this as a counter overflow, adding a large delta. In reality, this is a false positive: the counter did not overflow but was simply reset when the RMID transitioned from "Unavailable" back to active state. Here is the text from APM [1] available from [2]. "In PQOS Version 2.0 or higher, the MBM hardware will set the U bit on the first QM_CTR read when it begins tracking an RMID that it was not previously tracking. The U bit will be zero for all subsequent reads from that RMID while it is still tracked by the hardware. Therefore, a QM_CTR read with the U bit set when that RMID is in use by a processor can be considered 0 when calculating the difference with a subsequent read." [1] AMD64 Architecture Programmer's Manual Volume 2: System Programming Publication # 24593 Revision 3.41 section 19.3.3 Monitoring L3 Memory Bandwidth (MBM). [ bp: Split commit message into smaller paragraph chunks for better consumption. ] Fixes: 4d05bf71f157d ("x86/resctrl: Introduce AMD QOS feature") Signed-off-by: Babu Moger <babu.moger(a)amd.com> Signed-off-by: Borislav Petkov (AMD) <bp(a)alien8.de> Reviewed-by: Reinette Chatre <reinette.chatre(a)intel.com> Tested-by: Reinette Chatre <reinette.chatre(a)intel.com> Cc: stable(a)vger.kernel.org # needs adjustments for <= v6.17 Link: https://bugzilla.kernel.org/show_bug.cgi?id=206537 # [2] diff --git a/arch/x86/kernel/cpu/resctrl/monitor.c b/arch/x86/kernel/cpu/resctrl/monitor.c index c8945610d455..2cd25a0d4637 100644 --- a/arch/x86/kernel/cpu/resctrl/monitor.c +++ b/arch/x86/kernel/cpu/resctrl/monitor.c @@ -242,7 +242,9 @@ int resctrl_arch_rmid_read(struct rdt_resource *r, struct rdt_mon_domain *d, u32 unused, u32 rmid, enum resctrl_event_id eventid, u64 *val, void *ignored) { + struct rdt_hw_mon_domain *hw_dom = resctrl_to_arch_mon_dom(d); int cpu = cpumask_any(&d->hdr.cpu_mask); + struct arch_mbm_state *am; u64 msr_val; u32 prmid; int ret; @@ -251,12 +253,16 @@ int resctrl_arch_rmid_read(struct rdt_resource *r, struct rdt_mon_domain *d, prmid = logical_rmid_to_physical_rmid(cpu, rmid); ret = __rmid_read_phys(prmid, eventid, &msr_val); - if (ret) - return ret; - *val = get_corrected_val(r, d, rmid, eventid, msr_val); + if (!ret) { + *val = get_corrected_val(r, d, rmid, eventid, msr_val); + } else if (ret == -EINVAL) { + am = get_arch_mbm_state(hw_dom, rmid, eventid); + if (am) + am->prev_msr = 0; + } - return 0; + return ret; } static int __cntr_id_read(u32 cntr_id, u64 *val)

2 weeks, 2 days

2
1
0 0

2025

2024

2023

2022

2021

2020

2019

2018

2017

Linux-stable-mirror