From c641e5b207ed7dfaa692820aeb5b6dde3de3e9b0 Mon Sep 17 00:00:00 2001 From: Johan Hovold Date: Wed, 12 Jul 2017 17:55:29 +0200 Subject: ASoC: fix pcm-creation regression This reverts commit 99b04f4c4051 ("ASoC: add Component level pcm_new/pcm_free"), which started calling the pcm_new callback for every component in a *card* when creating a new pcm, something which does not seem to make any sense. This specifically led to memory leaks in systems with more than one platform component and where DMA memory is allocated in the platform-driver callback. For example, when both mcasp devices are being used on an am335x board, DMA memory would be allocated twice for every DAI link during probe. When CONFIG_SND_VERBOSE_PROCFS was set this fortunately also led to warnings such as: WARNING: CPU: 0 PID: 565 at ../fs/proc/generic.c:346 proc_register+0x110/0x154 proc_dir_entry 'sub0/prealloc' already registered Since there seems to be no users of the new component callbacks, and the current implementation introduced a regression, let's revert the offending commit for now. Fixes: 99b04f4c4051 ("ASoC: add Component level pcm_new/pcm_free") Signed-off-by: Johan Hovold Reviewed-by: Linus Walleij Tested-by: Linus Walleij Signed-off-by: Mark Brown Cc: stable # 4.10 --- include/sound/soc.h | 6 ------ 1 file changed, 6 deletions(-) (limited to 'include') diff --git a/include/sound/soc.h b/include/sound/soc.h index 9c94b97c17f8..c4a8b1947566 100644 --- a/include/sound/soc.h +++ b/include/sound/soc.h @@ -795,10 +795,6 @@ struct snd_soc_component_driver { int (*suspend)(struct snd_soc_component *); int (*resume)(struct snd_soc_component *); - /* pcm creation and destruction */ - int (*pcm_new)(struct snd_soc_pcm_runtime *); - void (*pcm_free)(struct snd_pcm *); - /* DT */ int (*of_xlate_dai_name)(struct snd_soc_component *component, struct of_phandle_args *args, @@ -874,8 +870,6 @@ struct snd_soc_component { void (*remove)(struct snd_soc_component *); int (*suspend)(struct snd_soc_component *); int (*resume)(struct snd_soc_component *); - int (*pcm_new)(struct snd_soc_pcm_runtime *); - void (*pcm_free)(struct snd_pcm *); /* machine specific init */ int (*init)(struct snd_soc_component *component); -- cgit v1.2.3 From d9f89b4e9290e46cd9b273e9ad0bff0f93e86fae Mon Sep 17 00:00:00 2001 From: Andrew Jones Date: Sat, 1 Jul 2017 18:26:54 +0200 Subject: KVM: arm/arm64: PMU: Fix overflow interrupt injection kvm_pmu_overflow_set() is called from perf's interrupt handler, making the call of kvm_vgic_inject_irq() from it introduced with "KVM: arm/arm64: PMU: remove request-less vcpu kick" a really bad idea, as it's quite easy to try and retake a lock that the interrupted context is already holding. The fix is to use a vcpu kick, leaving the interrupt injection to kvm_pmu_sync_hwstate(), like it was doing before the refactoring. We don't just revert, though, because before the kick was request-less, leaving the vcpu exposed to the request-less vcpu kick race, and also because the kick was used unnecessarily from register access handlers. Reviewed-by: Christoffer Dall Signed-off-by: Andrew Jones Signed-off-by: Marc Zyngier --- arch/arm64/kvm/sys_regs.c | 2 +- include/kvm/arm_pmu.h | 2 -- virt/kvm/arm/pmu.c | 43 +++++++++++++++---------------------------- 3 files changed, 16 insertions(+), 31 deletions(-) (limited to 'include') diff --git a/arch/arm64/kvm/sys_regs.c b/arch/arm64/kvm/sys_regs.c index 77862881ae86..2e070d3baf9f 100644 --- a/arch/arm64/kvm/sys_regs.c +++ b/arch/arm64/kvm/sys_regs.c @@ -764,7 +764,7 @@ static bool access_pmovs(struct kvm_vcpu *vcpu, struct sys_reg_params *p, if (p->is_write) { if (r->CRm & 0x2) /* accessing PMOVSSET_EL0 */ - kvm_pmu_overflow_set(vcpu, p->regval & mask); + vcpu_sys_reg(vcpu, PMOVSSET_EL0) |= (p->regval & mask); else /* accessing PMOVSCLR_EL0 */ vcpu_sys_reg(vcpu, PMOVSSET_EL0) &= ~(p->regval & mask); diff --git a/include/kvm/arm_pmu.h b/include/kvm/arm_pmu.h index f6e030617467..f87fe20fcb05 100644 --- a/include/kvm/arm_pmu.h +++ b/include/kvm/arm_pmu.h @@ -48,7 +48,6 @@ void kvm_pmu_vcpu_reset(struct kvm_vcpu *vcpu); void kvm_pmu_vcpu_destroy(struct kvm_vcpu *vcpu); void kvm_pmu_disable_counter(struct kvm_vcpu *vcpu, u64 val); void kvm_pmu_enable_counter(struct kvm_vcpu *vcpu, u64 val); -void kvm_pmu_overflow_set(struct kvm_vcpu *vcpu, u64 val); void kvm_pmu_flush_hwstate(struct kvm_vcpu *vcpu); void kvm_pmu_sync_hwstate(struct kvm_vcpu *vcpu); bool kvm_pmu_should_notify_user(struct kvm_vcpu *vcpu); @@ -86,7 +85,6 @@ static inline void kvm_pmu_vcpu_reset(struct kvm_vcpu *vcpu) {} static inline void kvm_pmu_vcpu_destroy(struct kvm_vcpu *vcpu) {} static inline void kvm_pmu_disable_counter(struct kvm_vcpu *vcpu, u64 val) {} static inline void kvm_pmu_enable_counter(struct kvm_vcpu *vcpu, u64 val) {} -static inline void kvm_pmu_overflow_set(struct kvm_vcpu *vcpu, u64 val) {} static inline void kvm_pmu_flush_hwstate(struct kvm_vcpu *vcpu) {} static inline void kvm_pmu_sync_hwstate(struct kvm_vcpu *vcpu) {} static inline bool kvm_pmu_should_notify_user(struct kvm_vcpu *vcpu) diff --git a/virt/kvm/arm/pmu.c b/virt/kvm/arm/pmu.c index fc8a723ff387..8a9c42366db7 100644 --- a/virt/kvm/arm/pmu.c +++ b/virt/kvm/arm/pmu.c @@ -203,11 +203,15 @@ static u64 kvm_pmu_overflow_status(struct kvm_vcpu *vcpu) return reg; } -static void kvm_pmu_check_overflow(struct kvm_vcpu *vcpu) +static void kvm_pmu_update_state(struct kvm_vcpu *vcpu) { struct kvm_pmu *pmu = &vcpu->arch.pmu; - bool overflow = !!kvm_pmu_overflow_status(vcpu); + bool overflow; + + if (!kvm_arm_pmu_v3_ready(vcpu)) + return; + overflow = !!kvm_pmu_overflow_status(vcpu); if (pmu->irq_level == overflow) return; @@ -215,33 +219,11 @@ static void kvm_pmu_check_overflow(struct kvm_vcpu *vcpu) if (likely(irqchip_in_kernel(vcpu->kvm))) { int ret = kvm_vgic_inject_irq(vcpu->kvm, vcpu->vcpu_id, - pmu->irq_num, overflow, - &vcpu->arch.pmu); + pmu->irq_num, overflow, pmu); WARN_ON(ret); } } -/** - * kvm_pmu_overflow_set - set PMU overflow interrupt - * @vcpu: The vcpu pointer - * @val: the value guest writes to PMOVSSET register - */ -void kvm_pmu_overflow_set(struct kvm_vcpu *vcpu, u64 val) -{ - if (val == 0) - return; - - vcpu_sys_reg(vcpu, PMOVSSET_EL0) |= val; - kvm_pmu_check_overflow(vcpu); -} - -static void kvm_pmu_update_state(struct kvm_vcpu *vcpu) -{ - if (!kvm_arm_pmu_v3_ready(vcpu)) - return; - kvm_pmu_check_overflow(vcpu); -} - bool kvm_pmu_should_notify_user(struct kvm_vcpu *vcpu) { struct kvm_pmu *pmu = &vcpu->arch.pmu; @@ -303,7 +285,7 @@ static inline struct kvm_vcpu *kvm_pmc_to_vcpu(struct kvm_pmc *pmc) } /** - * When perf event overflows, call kvm_pmu_overflow_set to set overflow status. + * When the perf event overflows, set the overflow status and inform the vcpu. */ static void kvm_pmu_perf_overflow(struct perf_event *perf_event, struct perf_sample_data *data, @@ -313,7 +295,12 @@ static void kvm_pmu_perf_overflow(struct perf_event *perf_event, struct kvm_vcpu *vcpu = kvm_pmc_to_vcpu(pmc); int idx = pmc->idx; - kvm_pmu_overflow_set(vcpu, BIT(idx)); + vcpu_sys_reg(vcpu, PMOVSSET_EL0) |= BIT(idx); + + if (kvm_pmu_overflow_status(vcpu)) { + kvm_make_request(KVM_REQ_IRQ_PENDING, vcpu); + kvm_vcpu_kick(vcpu); + } } /** @@ -341,7 +328,7 @@ void kvm_pmu_software_increment(struct kvm_vcpu *vcpu, u64 val) reg = lower_32_bits(reg); vcpu_sys_reg(vcpu, PMEVCNTR0_EL0 + i) = reg; if (!reg) - kvm_pmu_overflow_set(vcpu, BIT(i)); + vcpu_sys_reg(vcpu, PMOVSSET_EL0) |= BIT(i); } } } -- cgit v1.2.3 From fc1ff45a07abf240aa0c6586c11465c86c8cab8d Mon Sep 17 00:00:00 2001 From: Hans Verkuil Date: Sat, 15 Jul 2017 09:32:56 -0300 Subject: media: cec-notifier: small improvements Allow calling cec_notifier_set_phys_addr and cec_notifier_set_phys_addr_from_edid with a NULL notifier, in which case these functions do nothing. Add a cec_notifier_phys_addr_invalidate helper function (the notifier equivalent of cec_phys_addr_invalidate). These changes simplify drm CEC driver support. Signed-off-by: Hans Verkuil Signed-off-by: Mauro Carvalho Chehab --- drivers/media/cec/cec-notifier.c | 6 ++++++ include/media/cec-notifier.h | 15 +++++++++++++++ 2 files changed, 21 insertions(+) (limited to 'include') diff --git a/drivers/media/cec/cec-notifier.c b/drivers/media/cec/cec-notifier.c index 74dc1c32080e..08b619d0ea1e 100644 --- a/drivers/media/cec/cec-notifier.c +++ b/drivers/media/cec/cec-notifier.c @@ -87,6 +87,9 @@ EXPORT_SYMBOL_GPL(cec_notifier_put); void cec_notifier_set_phys_addr(struct cec_notifier *n, u16 pa) { + if (n == NULL) + return; + mutex_lock(&n->lock); n->phys_addr = pa; if (n->callback) @@ -100,6 +103,9 @@ void cec_notifier_set_phys_addr_from_edid(struct cec_notifier *n, { u16 pa = CEC_PHYS_ADDR_INVALID; + if (n == NULL) + return; + if (edid && edid->extensions) pa = cec_get_edid_phys_addr((const u8 *)edid, EDID_LENGTH * (edid->extensions + 1), NULL); diff --git a/include/media/cec-notifier.h b/include/media/cec-notifier.h index 298f996969df..a4f7429c4ae5 100644 --- a/include/media/cec-notifier.h +++ b/include/media/cec-notifier.h @@ -57,6 +57,7 @@ void cec_notifier_put(struct cec_notifier *n); * @pa: the CEC physical address * * Set a new CEC physical address. + * Does nothing if @n == NULL. */ void cec_notifier_set_phys_addr(struct cec_notifier *n, u16 pa); @@ -66,6 +67,7 @@ void cec_notifier_set_phys_addr(struct cec_notifier *n, u16 pa); * @edid: the struct edid pointer * * Parses the EDID to obtain the new CEC physical address and set it. + * Does nothing if @n == NULL. */ void cec_notifier_set_phys_addr_from_edid(struct cec_notifier *n, const struct edid *edid); @@ -118,4 +120,17 @@ static inline void cec_notifier_unregister(struct cec_notifier *n) #endif +/** + * cec_notifier_phys_addr_invalidate() - set the physical address to INVALID + * + * @n: the CEC notifier + * + * This is a simple helper function to invalidate the physical + * address. Does nothing if @n == NULL. + */ +static inline void cec_notifier_phys_addr_invalidate(struct cec_notifier *n) +{ + cec_notifier_set_phys_addr(n, CEC_PHYS_ADDR_INVALID); +} + #endif -- cgit v1.2.3 From b25db383928cecba356835583b16fa7008f97b3a Mon Sep 17 00:00:00 2001 From: Prabhakar Lad Date: Thu, 20 Jul 2017 04:56:31 -0400 Subject: media: platform: davinci: drop VPFE_CMD_S_CCDC_RAW_PARAMS drop VPFE_CMD_S_CCDC_RAW_PARAMS ioctl from dm355/dm644x following reasons: - This ioctl was never in public api and was only defined in kernel header. - The function set_params constantly mixes up pointers and phys_addr_t numbers. - This is part of a 'VPFE_CMD_S_CCDC_RAW_PARAMS' ioctl command that is described as an 'experimental ioctl that will change in future kernels'. - The code to allocate the table never gets called after we copy_from_user the user input over the kernel settings, and then compare them for inequality. - We then go on to use an address provided by user space as both the __user pointer for input and pass it through phys_to_virt to come up with a kernel pointer to copy the data to. This looks like a trivially exploitable root hole. Signed-off-by: Lad, Prabhakar Signed-off-by: Hans Verkuil Signed-off-by: Mauro Carvalho Chehab --- drivers/media/platform/davinci/ccdc_hw_device.h | 10 -- drivers/media/platform/davinci/dm355_ccdc.c | 92 +-------------- drivers/media/platform/davinci/dm644x_ccdc.c | 151 +----------------------- drivers/media/platform/davinci/vpfe_capture.c | 75 ------------ include/media/davinci/dm644x_ccdc.h | 12 -- include/media/davinci/vpfe_capture.h | 10 -- 6 files changed, 4 insertions(+), 346 deletions(-) (limited to 'include') diff --git a/drivers/media/platform/davinci/ccdc_hw_device.h b/drivers/media/platform/davinci/ccdc_hw_device.h index 8f6688a7a111..f1b521045d64 100644 --- a/drivers/media/platform/davinci/ccdc_hw_device.h +++ b/drivers/media/platform/davinci/ccdc_hw_device.h @@ -42,16 +42,6 @@ struct ccdc_hw_ops { int (*set_hw_if_params) (struct vpfe_hw_if_param *param); /* get interface parameters */ int (*get_hw_if_params) (struct vpfe_hw_if_param *param); - /* - * Pointer to function to set parameters. Used - * for implementing VPFE_S_CCDC_PARAMS - */ - int (*set_params) (void *params); - /* - * Pointer to function to get parameter. Used - * for implementing VPFE_G_CCDC_PARAMS - */ - int (*get_params) (void *params); /* Pointer to function to configure ccdc */ int (*configure) (void); diff --git a/drivers/media/platform/davinci/dm355_ccdc.c b/drivers/media/platform/davinci/dm355_ccdc.c index 73db166dc338..6d492dc4c3a9 100644 --- a/drivers/media/platform/davinci/dm355_ccdc.c +++ b/drivers/media/platform/davinci/dm355_ccdc.c @@ -17,12 +17,7 @@ * This module is for configuring DM355 CCD controller of VPFE to capture * Raw yuv or Bayer RGB data from a decoder. CCDC has several modules * such as Defect Pixel Correction, Color Space Conversion etc to - * pre-process the Bayer RGB data, before writing it to SDRAM. This - * module also allows application to configure individual - * module parameters through VPFE_CMD_S_CCDC_RAW_PARAMS IOCTL. - * To do so, application include dm355_ccdc.h and vpfe_capture.h header - * files. The setparams() API is called by vpfe_capture driver - * to configure module parameters + * pre-process the Bayer RGB data, before writing it to SDRAM. * * TODO: 1) Raw bayer parameter settings and bayer capture * 2) Split module parameter structure to module specific ioctl structs @@ -260,90 +255,6 @@ static void ccdc_setwin(struct v4l2_rect *image_win, dev_dbg(ccdc_cfg.dev, "\nEnd of ccdc_setwin..."); } -static int validate_ccdc_param(struct ccdc_config_params_raw *ccdcparam) -{ - if (ccdcparam->datasft < CCDC_DATA_NO_SHIFT || - ccdcparam->datasft > CCDC_DATA_SHIFT_6BIT) { - dev_dbg(ccdc_cfg.dev, "Invalid value of data shift\n"); - return -EINVAL; - } - - if (ccdcparam->mfilt1 < CCDC_NO_MEDIAN_FILTER1 || - ccdcparam->mfilt1 > CCDC_MEDIAN_FILTER1) { - dev_dbg(ccdc_cfg.dev, "Invalid value of median filter1\n"); - return -EINVAL; - } - - if (ccdcparam->mfilt2 < CCDC_NO_MEDIAN_FILTER2 || - ccdcparam->mfilt2 > CCDC_MEDIAN_FILTER2) { - dev_dbg(ccdc_cfg.dev, "Invalid value of median filter2\n"); - return -EINVAL; - } - - if ((ccdcparam->med_filt_thres < 0) || - (ccdcparam->med_filt_thres > CCDC_MED_FILT_THRESH)) { - dev_dbg(ccdc_cfg.dev, - "Invalid value of median filter threshold\n"); - return -EINVAL; - } - - if (ccdcparam->data_sz < CCDC_DATA_16BITS || - ccdcparam->data_sz > CCDC_DATA_8BITS) { - dev_dbg(ccdc_cfg.dev, "Invalid value of data size\n"); - return -EINVAL; - } - - if (ccdcparam->alaw.enable) { - if (ccdcparam->alaw.gamma_wd < CCDC_GAMMA_BITS_13_4 || - ccdcparam->alaw.gamma_wd > CCDC_GAMMA_BITS_09_0) { - dev_dbg(ccdc_cfg.dev, "Invalid value of ALAW\n"); - return -EINVAL; - } - } - - if (ccdcparam->blk_clamp.b_clamp_enable) { - if (ccdcparam->blk_clamp.sample_pixel < CCDC_SAMPLE_1PIXELS || - ccdcparam->blk_clamp.sample_pixel > CCDC_SAMPLE_16PIXELS) { - dev_dbg(ccdc_cfg.dev, - "Invalid value of sample pixel\n"); - return -EINVAL; - } - if (ccdcparam->blk_clamp.sample_ln < CCDC_SAMPLE_1LINES || - ccdcparam->blk_clamp.sample_ln > CCDC_SAMPLE_16LINES) { - dev_dbg(ccdc_cfg.dev, - "Invalid value of sample lines\n"); - return -EINVAL; - } - } - return 0; -} - -/* Parameter operations */ -static int ccdc_set_params(void __user *params) -{ - struct ccdc_config_params_raw ccdc_raw_params; - int x; - - /* only raw module parameters can be set through the IOCTL */ - if (ccdc_cfg.if_type != VPFE_RAW_BAYER) - return -EINVAL; - - x = copy_from_user(&ccdc_raw_params, params, sizeof(ccdc_raw_params)); - if (x) { - dev_dbg(ccdc_cfg.dev, "ccdc_set_params: error in copying ccdcparams, %d\n", - x); - return -EFAULT; - } - - if (!validate_ccdc_param(&ccdc_raw_params)) { - memcpy(&ccdc_cfg.bayer.config_params, - &ccdc_raw_params, - sizeof(ccdc_raw_params)); - return 0; - } - return -EINVAL; -} - /* This function will configure CCDC for YCbCr video capture */ static void ccdc_config_ycbcr(void) { @@ -939,7 +850,6 @@ static struct ccdc_hw_device ccdc_hw_dev = { .enable = ccdc_enable, .enable_out_to_sdram = ccdc_enable_output_to_sdram, .set_hw_if_params = ccdc_set_hw_if_params, - .set_params = ccdc_set_params, .configure = ccdc_configure, .set_buftype = ccdc_set_buftype, .get_buftype = ccdc_get_buftype, diff --git a/drivers/media/platform/davinci/dm644x_ccdc.c b/drivers/media/platform/davinci/dm644x_ccdc.c index 740fbc7a8c14..3b2d8a9317b8 100644 --- a/drivers/media/platform/davinci/dm644x_ccdc.c +++ b/drivers/media/platform/davinci/dm644x_ccdc.c @@ -17,13 +17,9 @@ * This module is for configuring CCD controller of DM6446 VPFE to capture * Raw yuv or Bayer RGB data from a decoder. CCDC has several modules * such as Defect Pixel Correction, Color Space Conversion etc to - * pre-process the Raw Bayer RGB data, before writing it to SDRAM. This - * module also allows application to configure individual - * module parameters through VPFE_CMD_S_CCDC_RAW_PARAMS IOCTL. - * To do so, application includes dm644x_ccdc.h and vpfe_capture.h header - * files. The setparams() API is called by vpfe_capture driver - * to configure module parameters. This file is named DM644x so that other - * variants such DM6443 may be supported using the same module. + * pre-process the Raw Bayer RGB data, before writing it to SDRAM. + * This file is named DM644x so that other variants such DM6443 + * may be supported using the same module. * * TODO: Test Raw bayer parameter settings and bayer capture * Split module parameter structure to module specific ioctl structs @@ -216,96 +212,8 @@ static void ccdc_readregs(void) dev_notice(ccdc_cfg.dev, "\nReading 0x%x to VERT_LINES...\n", val); } -static int validate_ccdc_param(struct ccdc_config_params_raw *ccdcparam) -{ - if (ccdcparam->alaw.enable) { - u8 max_gamma = ccdc_gamma_width_max_bit(ccdcparam->alaw.gamma_wd); - u8 max_data = ccdc_data_size_max_bit(ccdcparam->data_sz); - - if ((ccdcparam->alaw.gamma_wd > CCDC_GAMMA_BITS_09_0) || - (ccdcparam->alaw.gamma_wd < CCDC_GAMMA_BITS_15_6) || - (max_gamma > max_data)) { - dev_dbg(ccdc_cfg.dev, "\nInvalid data line select"); - return -1; - } - } - return 0; -} - -static int ccdc_update_raw_params(struct ccdc_config_params_raw *raw_params) -{ - struct ccdc_config_params_raw *config_params = - &ccdc_cfg.bayer.config_params; - unsigned int *fpc_virtaddr = NULL; - unsigned int *fpc_physaddr = NULL; - - memcpy(config_params, raw_params, sizeof(*raw_params)); - /* - * allocate memory for fault pixel table and copy the user - * values to the table - */ - if (!config_params->fault_pxl.enable) - return 0; - - fpc_physaddr = (unsigned int *)config_params->fault_pxl.fpc_table_addr; - fpc_virtaddr = (unsigned int *)phys_to_virt( - (unsigned long)fpc_physaddr); - /* - * Allocate memory for FPC table if current - * FPC table buffer is not big enough to - * accommodate FPC Number requested - */ - if (raw_params->fault_pxl.fp_num != config_params->fault_pxl.fp_num) { - if (fpc_physaddr != NULL) { - free_pages((unsigned long)fpc_virtaddr, - get_order - (config_params->fault_pxl.fp_num * - FP_NUM_BYTES)); - } - - /* Allocate memory for FPC table */ - fpc_virtaddr = - (unsigned int *)__get_free_pages(GFP_KERNEL | GFP_DMA, - get_order(raw_params-> - fault_pxl.fp_num * - FP_NUM_BYTES)); - - if (fpc_virtaddr == NULL) { - dev_dbg(ccdc_cfg.dev, - "\nUnable to allocate memory for FPC"); - return -EFAULT; - } - fpc_physaddr = - (unsigned int *)virt_to_phys((void *)fpc_virtaddr); - } - - /* Copy number of fault pixels and FPC table */ - config_params->fault_pxl.fp_num = raw_params->fault_pxl.fp_num; - if (copy_from_user(fpc_virtaddr, - (void __user *)raw_params->fault_pxl.fpc_table_addr, - config_params->fault_pxl.fp_num * FP_NUM_BYTES)) { - dev_dbg(ccdc_cfg.dev, "\n copy_from_user failed"); - return -EFAULT; - } - config_params->fault_pxl.fpc_table_addr = (unsigned long)fpc_physaddr; - return 0; -} - static int ccdc_close(struct device *dev) { - struct ccdc_config_params_raw *config_params = - &ccdc_cfg.bayer.config_params; - unsigned int *fpc_physaddr = NULL, *fpc_virtaddr = NULL; - - fpc_physaddr = (unsigned int *)config_params->fault_pxl.fpc_table_addr; - - if (fpc_physaddr != NULL) { - fpc_virtaddr = (unsigned int *) - phys_to_virt((unsigned long)fpc_physaddr); - free_pages((unsigned long)fpc_virtaddr, - get_order(config_params->fault_pxl.fp_num * - FP_NUM_BYTES)); - } return 0; } @@ -339,29 +247,6 @@ static void ccdc_sbl_reset(void) vpss_clear_wbl_overflow(VPSS_PCR_CCDC_WBL_O); } -/* Parameter operations */ -static int ccdc_set_params(void __user *params) -{ - struct ccdc_config_params_raw ccdc_raw_params; - int x; - - if (ccdc_cfg.if_type != VPFE_RAW_BAYER) - return -EINVAL; - - x = copy_from_user(&ccdc_raw_params, params, sizeof(ccdc_raw_params)); - if (x) { - dev_dbg(ccdc_cfg.dev, "ccdc_set_params: error in copyingccdc params, %d\n", - x); - return -EFAULT; - } - - if (!validate_ccdc_param(&ccdc_raw_params)) { - if (!ccdc_update_raw_params(&ccdc_raw_params)) - return 0; - } - return -EINVAL; -} - /* * ccdc_config_ycbcr() * This function will configure CCDC for YCbCr video capture @@ -489,32 +374,6 @@ static void ccdc_config_black_compense(struct ccdc_black_compensation *bcomp) regw(val, CCDC_BLKCMP); } -static void ccdc_config_fpc(struct ccdc_fault_pixel *fpc) -{ - u32 val; - - /* Initially disable FPC */ - val = CCDC_FPC_DISABLE; - regw(val, CCDC_FPC); - - if (!fpc->enable) - return; - - /* Configure Fault pixel if needed */ - regw(fpc->fpc_table_addr, CCDC_FPC_ADDR); - dev_dbg(ccdc_cfg.dev, "\nWriting 0x%lx to FPC_ADDR...\n", - (fpc->fpc_table_addr)); - /* Write the FPC params with FPC disable */ - val = fpc->fp_num & CCDC_FPC_FPC_NUM_MASK; - regw(val, CCDC_FPC); - - dev_dbg(ccdc_cfg.dev, "\nWriting 0x%x to FPC...\n", val); - /* read the FPC register */ - val = regr(CCDC_FPC) | CCDC_FPC_ENABLE; - regw(val, CCDC_FPC); - dev_dbg(ccdc_cfg.dev, "\nWriting 0x%x to FPC...\n", val); -} - /* * ccdc_config_raw() * This function will configure CCDC for Raw capture mode @@ -569,9 +428,6 @@ static void ccdc_config_raw(void) /* Configure Black level compensation */ ccdc_config_black_compense(&config_params->blk_comp); - /* Configure Fault Pixel Correction */ - ccdc_config_fpc(&config_params->fault_pxl); - /* If data size is 8 bit then pack the data */ if ((config_params->data_sz == CCDC_DATA_8BITS) || config_params->alaw.enable) @@ -929,7 +785,6 @@ static struct ccdc_hw_device ccdc_hw_dev = { .reset = ccdc_sbl_reset, .enable = ccdc_enable, .set_hw_if_params = ccdc_set_hw_if_params, - .set_params = ccdc_set_params, .configure = ccdc_configure, .set_buftype = ccdc_set_buftype, .get_buftype = ccdc_get_buftype, diff --git a/drivers/media/platform/davinci/vpfe_capture.c b/drivers/media/platform/davinci/vpfe_capture.c index 1831bf5ccca5..b1bf4a7e8eb7 100644 --- a/drivers/media/platform/davinci/vpfe_capture.c +++ b/drivers/media/platform/davinci/vpfe_capture.c @@ -280,45 +280,6 @@ void vpfe_unregister_ccdc_device(struct ccdc_hw_device *dev) } EXPORT_SYMBOL(vpfe_unregister_ccdc_device); -/* - * vpfe_get_ccdc_image_format - Get image parameters based on CCDC settings - */ -static int vpfe_get_ccdc_image_format(struct vpfe_device *vpfe_dev, - struct v4l2_format *f) -{ - struct v4l2_rect image_win; - enum ccdc_buftype buf_type; - enum ccdc_frmfmt frm_fmt; - - memset(f, 0, sizeof(*f)); - f->type = V4L2_BUF_TYPE_VIDEO_OUTPUT; - ccdc_dev->hw_ops.get_image_window(&image_win); - f->fmt.pix.width = image_win.width; - f->fmt.pix.height = image_win.height; - f->fmt.pix.bytesperline = ccdc_dev->hw_ops.get_line_length(); - f->fmt.pix.sizeimage = f->fmt.pix.bytesperline * - f->fmt.pix.height; - buf_type = ccdc_dev->hw_ops.get_buftype(); - f->fmt.pix.pixelformat = ccdc_dev->hw_ops.get_pixel_format(); - frm_fmt = ccdc_dev->hw_ops.get_frame_format(); - if (frm_fmt == CCDC_FRMFMT_PROGRESSIVE) - f->fmt.pix.field = V4L2_FIELD_NONE; - else if (frm_fmt == CCDC_FRMFMT_INTERLACED) { - if (buf_type == CCDC_BUFTYPE_FLD_INTERLEAVED) - f->fmt.pix.field = V4L2_FIELD_INTERLACED; - else if (buf_type == CCDC_BUFTYPE_FLD_SEPARATED) - f->fmt.pix.field = V4L2_FIELD_SEQ_TB; - else { - v4l2_err(&vpfe_dev->v4l2_dev, "Invalid buf_type\n"); - return -EINVAL; - } - } else { - v4l2_err(&vpfe_dev->v4l2_dev, "Invalid frm_fmt\n"); - return -EINVAL; - } - return 0; -} - /* * vpfe_config_ccdc_image_format() * For a pix format, configure ccdc to setup the capture @@ -1697,41 +1658,6 @@ unlock_out: return ret; } - -static long vpfe_param_handler(struct file *file, void *priv, - bool valid_prio, unsigned int cmd, void *param) -{ - struct vpfe_device *vpfe_dev = video_drvdata(file); - int ret; - - v4l2_dbg(2, debug, &vpfe_dev->v4l2_dev, "vpfe_param_handler\n"); - - if (vpfe_dev->started) { - /* only allowed if streaming is not started */ - v4l2_dbg(1, debug, &vpfe_dev->v4l2_dev, - "device already started\n"); - return -EBUSY; - } - - ret = mutex_lock_interruptible(&vpfe_dev->lock); - if (ret) - return ret; - - switch (cmd) { - case VPFE_CMD_S_CCDC_RAW_PARAMS: - ret = -EINVAL; - v4l2_warn(&vpfe_dev->v4l2_dev, - "VPFE_CMD_S_CCDC_RAW_PARAMS not supported\n"); - break; - default: - ret = -ENOTTY; - } -unlock_out: - mutex_unlock(&vpfe_dev->lock); - return ret; -} - - /* vpfe capture ioctl operations */ static const struct v4l2_ioctl_ops vpfe_ioctl_ops = { .vidioc_querycap = vpfe_querycap, @@ -1754,7 +1680,6 @@ static const struct v4l2_ioctl_ops vpfe_ioctl_ops = { .vidioc_cropcap = vpfe_cropcap, .vidioc_g_selection = vpfe_g_selection, .vidioc_s_selection = vpfe_s_selection, - .vidioc_default = vpfe_param_handler, }; static struct vpfe_device *vpfe_initialize(void) diff --git a/include/media/davinci/dm644x_ccdc.h b/include/media/davinci/dm644x_ccdc.h index 7c909da29d43..6ea2ce241851 100644 --- a/include/media/davinci/dm644x_ccdc.h +++ b/include/media/davinci/dm644x_ccdc.h @@ -103,16 +103,6 @@ struct ccdc_black_compensation { char gb; }; -/* structure for fault pixel correction */ -struct ccdc_fault_pixel { - /* Enable or Disable fault pixel correction */ - unsigned char enable; - /* Number of fault pixel */ - unsigned short fp_num; - /* Address of fault pixel table */ - unsigned long fpc_table_addr; -}; - /* Structure for CCDC configuration parameters for raw capture mode passed * by application */ @@ -125,8 +115,6 @@ struct ccdc_config_params_raw { struct ccdc_black_clamp blk_clamp; /* Structure for Black Compensation */ struct ccdc_black_compensation blk_comp; - /* Structure for Fault Pixel Module Configuration */ - struct ccdc_fault_pixel fault_pxl; }; diff --git a/include/media/davinci/vpfe_capture.h b/include/media/davinci/vpfe_capture.h index 8e1a4d88daa0..f003533602d0 100644 --- a/include/media/davinci/vpfe_capture.h +++ b/include/media/davinci/vpfe_capture.h @@ -183,14 +183,4 @@ struct vpfe_config_params { }; #endif /* End of __KERNEL__ */ -/** - * VPFE_CMD_S_CCDC_RAW_PARAMS - EXPERIMENTAL IOCTL to set raw capture params - * This can be used to configure modules such as defect pixel correction, - * color space conversion, culling etc. This is an experimental ioctl that - * will change in future kernels. So use this ioctl with care ! - * TODO: This is to be split into multiple ioctls and also explore the - * possibility of extending the v4l2 api to include this - **/ -#define VPFE_CMD_S_CCDC_RAW_PARAMS _IOW('V', BASE_VIDIOC_PRIVATE + 1, \ - void *) #endif /* _DAVINCI_VPFE_H */ -- cgit v1.2.3 From bb67b496c338e15813f075f482067da930f52e39 Mon Sep 17 00:00:00 2001 From: Murilo Opsfelder Araujo Date: Tue, 18 Jul 2017 14:22:20 -0300 Subject: include/linux/vfio.h: Guard powerpc-specific functions with CONFIG_VFIO_SPAPR_EEH When CONFIG_EEH=y and CONFIG_VFIO_SPAPR_EEH=n, build fails with the following: drivers/vfio/pci/vfio_pci.o: In function `.vfio_pci_release': vfio_pci.c:(.text+0xa98): undefined reference to `.vfio_spapr_pci_eeh_release' drivers/vfio/pci/vfio_pci.o: In function `.vfio_pci_open': vfio_pci.c:(.text+0x1420): undefined reference to `.vfio_spapr_pci_eeh_open' In this case, vfio_pci.c should use the empty definitions of vfio_spapr_pci_eeh_open and vfio_spapr_pci_eeh_release functions. This patch fixes it by guarding these function definitions with CONFIG_VFIO_SPAPR_EEH, the symbol that controls whether vfio_spapr_eeh.c is built, which is where the non-empty versions of these functions are. We need to make use of IS_ENABLED() macro because CONFIG_VFIO_SPAPR_EEH is a tristate option. This issue was found during a randconfig build. Logs are here: http://kisskb.ellerman.id.au/kisskb/buildresult/12982362/ Signed-off-by: Murilo Opsfelder Araujo Reviewed-by: Alexey Kardashevskiy Reviewed-by: David Gibson Signed-off-by: Alex Williamson --- include/linux/vfio.h | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'include') diff --git a/include/linux/vfio.h b/include/linux/vfio.h index 586809abb273..a47b985341d1 100644 --- a/include/linux/vfio.h +++ b/include/linux/vfio.h @@ -152,7 +152,7 @@ extern int vfio_set_irqs_validate_and_prepare(struct vfio_irq_set *hdr, size_t *data_size); struct pci_dev; -#ifdef CONFIG_EEH +#if IS_ENABLED(CONFIG_VFIO_SPAPR_EEH) extern void vfio_spapr_pci_eeh_open(struct pci_dev *pdev); extern void vfio_spapr_pci_eeh_release(struct pci_dev *pdev); extern long vfio_spapr_iommu_eeh_ioctl(struct iommu_group *group, @@ -173,7 +173,7 @@ static inline long vfio_spapr_iommu_eeh_ioctl(struct iommu_group *group, { return -ENOTTY; } -#endif /* CONFIG_EEH */ +#endif /* CONFIG_VFIO_SPAPR_EEH */ /* * IRQfd - generic -- cgit v1.2.3 From a627b0a7c15ee4d2c87a86d5be5c8167382e8d0d Mon Sep 17 00:00:00 2001 From: Eric Whitney Date: Sun, 30 Jul 2017 22:30:11 -0400 Subject: ext4: remove unused metadata accounting variables Two variables in ext4_inode_info, i_reserved_meta_blocks and i_allocated_meta_blocks, are unused. Removing them saves a little memory per in-memory inode and cleans up clutter in several tracepoints. Adjust tracepoint output from ext4_alloc_da_blocks() for consistency and fix a typo and whitespace near these changes. Signed-off-by: Eric Whitney Signed-off-by: Theodore Ts'o Reviewed-by: Jan Kara --- fs/ext4/ext4.h | 6 ++---- fs/ext4/super.c | 2 -- include/trace/events/ext4.h | 35 ++++++++--------------------------- 3 files changed, 10 insertions(+), 33 deletions(-) (limited to 'include') diff --git a/fs/ext4/ext4.h b/fs/ext4/ext4.h index 9ebde0cd632e..dcbcd9d444d1 100644 --- a/fs/ext4/ext4.h +++ b/fs/ext4/ext4.h @@ -961,7 +961,7 @@ struct ext4_inode_info { /* * i_block_group is the number of the block group which contains * this file's inode. Constant across the lifetime of the inode, - * it is ued for making block allocation decisions - we try to + * it is used for making block allocation decisions - we try to * place a file's data blocks near its inode block, and new inodes * near to their parent directory's inode. */ @@ -1049,10 +1049,8 @@ struct ext4_inode_info { ext4_group_t i_last_alloc_group; /* allocation reservation info for delalloc */ - /* In case of bigalloc, these refer to clusters rather than blocks */ + /* In case of bigalloc, this refer to clusters rather than blocks */ unsigned int i_reserved_data_blocks; - unsigned int i_reserved_meta_blocks; - unsigned int i_allocated_meta_blocks; ext4_lblk_t i_da_metadata_calc_last_lblock; int i_da_metadata_calc_len; diff --git a/fs/ext4/super.c b/fs/ext4/super.c index 0886fe82e9c4..d61a70e2193a 100644 --- a/fs/ext4/super.c +++ b/fs/ext4/super.c @@ -978,8 +978,6 @@ static struct inode *ext4_alloc_inode(struct super_block *sb) ei->i_es_shk_nr = 0; ei->i_es_shrink_lblk = 0; ei->i_reserved_data_blocks = 0; - ei->i_reserved_meta_blocks = 0; - ei->i_allocated_meta_blocks = 0; ei->i_da_metadata_calc_len = 0; ei->i_da_metadata_calc_last_lblock = 0; spin_lock_init(&(ei->i_block_reservation_lock)); diff --git a/include/trace/events/ext4.h b/include/trace/events/ext4.h index dfae175ddebc..9c3bc3883d2f 100644 --- a/include/trace/events/ext4.h +++ b/include/trace/events/ext4.h @@ -937,21 +937,19 @@ TRACE_EVENT(ext4_alloc_da_blocks, TP_STRUCT__entry( __field( dev_t, dev ) __field( ino_t, ino ) - __field( unsigned int, data_blocks ) - __field( unsigned int, meta_blocks ) + __field( unsigned int, data_blocks ) ), TP_fast_assign( __entry->dev = inode->i_sb->s_dev; __entry->ino = inode->i_ino; __entry->data_blocks = EXT4_I(inode)->i_reserved_data_blocks; - __entry->meta_blocks = EXT4_I(inode)->i_reserved_meta_blocks; ), - TP_printk("dev %d,%d ino %lu data_blocks %u meta_blocks %u", + TP_printk("dev %d,%d ino %lu reserved_data_blocks %u", MAJOR(__entry->dev), MINOR(__entry->dev), (unsigned long) __entry->ino, - __entry->data_blocks, __entry->meta_blocks) + __entry->data_blocks) ); TRACE_EVENT(ext4_mballoc_alloc, @@ -1153,8 +1151,6 @@ TRACE_EVENT(ext4_da_update_reserve_space, __field( __u64, i_blocks ) __field( int, used_blocks ) __field( int, reserved_data_blocks ) - __field( int, reserved_meta_blocks ) - __field( int, allocated_meta_blocks ) __field( int, quota_claim ) __field( __u16, mode ) ), @@ -1166,22 +1162,16 @@ TRACE_EVENT(ext4_da_update_reserve_space, __entry->used_blocks = used_blocks; __entry->reserved_data_blocks = EXT4_I(inode)->i_reserved_data_blocks; - __entry->reserved_meta_blocks = - EXT4_I(inode)->i_reserved_meta_blocks; - __entry->allocated_meta_blocks = - EXT4_I(inode)->i_allocated_meta_blocks; __entry->quota_claim = quota_claim; __entry->mode = inode->i_mode; ), TP_printk("dev %d,%d ino %lu mode 0%o i_blocks %llu used_blocks %d " - "reserved_data_blocks %d reserved_meta_blocks %d " - "allocated_meta_blocks %d quota_claim %d", + "reserved_data_blocks %d quota_claim %d", MAJOR(__entry->dev), MINOR(__entry->dev), (unsigned long) __entry->ino, __entry->mode, __entry->i_blocks, __entry->used_blocks, __entry->reserved_data_blocks, - __entry->reserved_meta_blocks, __entry->allocated_meta_blocks, __entry->quota_claim) ); @@ -1195,7 +1185,6 @@ TRACE_EVENT(ext4_da_reserve_space, __field( ino_t, ino ) __field( __u64, i_blocks ) __field( int, reserved_data_blocks ) - __field( int, reserved_meta_blocks ) __field( __u16, mode ) ), @@ -1204,17 +1193,15 @@ TRACE_EVENT(ext4_da_reserve_space, __entry->ino = inode->i_ino; __entry->i_blocks = inode->i_blocks; __entry->reserved_data_blocks = EXT4_I(inode)->i_reserved_data_blocks; - __entry->reserved_meta_blocks = EXT4_I(inode)->i_reserved_meta_blocks; __entry->mode = inode->i_mode; ), TP_printk("dev %d,%d ino %lu mode 0%o i_blocks %llu " - "reserved_data_blocks %d reserved_meta_blocks %d", + "reserved_data_blocks %d", MAJOR(__entry->dev), MINOR(__entry->dev), (unsigned long) __entry->ino, __entry->mode, __entry->i_blocks, - __entry->reserved_data_blocks, - __entry->reserved_meta_blocks) + __entry->reserved_data_blocks) ); TRACE_EVENT(ext4_da_release_space, @@ -1228,8 +1215,6 @@ TRACE_EVENT(ext4_da_release_space, __field( __u64, i_blocks ) __field( int, freed_blocks ) __field( int, reserved_data_blocks ) - __field( int, reserved_meta_blocks ) - __field( int, allocated_meta_blocks ) __field( __u16, mode ) ), @@ -1239,19 +1224,15 @@ TRACE_EVENT(ext4_da_release_space, __entry->i_blocks = inode->i_blocks; __entry->freed_blocks = freed_blocks; __entry->reserved_data_blocks = EXT4_I(inode)->i_reserved_data_blocks; - __entry->reserved_meta_blocks = EXT4_I(inode)->i_reserved_meta_blocks; - __entry->allocated_meta_blocks = EXT4_I(inode)->i_allocated_meta_blocks; __entry->mode = inode->i_mode; ), TP_printk("dev %d,%d ino %lu mode 0%o i_blocks %llu freed_blocks %d " - "reserved_data_blocks %d reserved_meta_blocks %d " - "allocated_meta_blocks %d", + "reserved_data_blocks %d", MAJOR(__entry->dev), MINOR(__entry->dev), (unsigned long) __entry->ino, __entry->mode, __entry->i_blocks, - __entry->freed_blocks, __entry->reserved_data_blocks, - __entry->reserved_meta_blocks, __entry->allocated_meta_blocks) + __entry->freed_blocks, __entry->reserved_data_blocks) ); DECLARE_EVENT_CLASS(ext4__bitmap_load, -- cgit v1.2.3 From e17e8969f5c59a10083af5e260bdad6026872203 Mon Sep 17 00:00:00 2001 From: Ilya Dryomov Date: Mon, 24 Jul 2017 16:43:49 +0200 Subject: libceph: fallback for when there isn't a pool-specific choose_arg There is now a fallback to a choose_arg index of -1 if there isn't a pool-specific choose_arg set. If you create a per-pool weight-set, that works for that pool. Otherwise we try the compat/default one. If that doesn't exist either, then we use the normal CRUSH weights. Signed-off-by: Ilya Dryomov Reviewed-by: Sage Weil --- include/linux/crush/crush.h | 2 +- net/ceph/osdmap.c | 12 +++++++++++- 2 files changed, 12 insertions(+), 2 deletions(-) (limited to 'include') diff --git a/include/linux/crush/crush.h b/include/linux/crush/crush.h index 92e165d417a6..07eed95e10c7 100644 --- a/include/linux/crush/crush.h +++ b/include/linux/crush/crush.h @@ -193,7 +193,7 @@ struct crush_choose_arg { struct crush_choose_arg_map { #ifdef __KERNEL__ struct rb_node node; - u64 choose_args_index; + s64 choose_args_index; #endif struct crush_choose_arg *args; /*!< replacement for each bucket in the crushmap */ diff --git a/net/ceph/osdmap.c b/net/ceph/osdmap.c index 64ae9f89773a..eb57a06373ca 100644 --- a/net/ceph/osdmap.c +++ b/net/ceph/osdmap.c @@ -2301,10 +2301,17 @@ static u32 raw_pg_to_pps(struct ceph_pg_pool_info *pi, } } +/* + * Magic value used for a "default" fallback choose_args, used if the + * crush_choose_arg_map passed to do_crush() does not exist. If this + * also doesn't exist, fall back to canonical weights. + */ +#define CEPH_DEFAULT_CHOOSE_ARGS -1 + static int do_crush(struct ceph_osdmap *map, int ruleno, int x, int *result, int result_max, const __u32 *weight, int weight_max, - u64 choose_args_index) + s64 choose_args_index) { struct crush_choose_arg_map *arg_map; int r; @@ -2313,6 +2320,9 @@ static int do_crush(struct ceph_osdmap *map, int ruleno, int x, arg_map = lookup_choose_arg_map(&map->crush->choose_args, choose_args_index); + if (!arg_map) + arg_map = lookup_choose_arg_map(&map->crush->choose_args, + CEPH_DEFAULT_CHOOSE_ARGS); mutex_lock(&map->crush_workspace_mutex); r = crush_do_rule(map->crush, ruleno, x, result, result_max, -- cgit v1.2.3 From ae78dd8139ce93a528beb7f3914531b7a7be9e30 Mon Sep 17 00:00:00 2001 From: Ilya Dryomov Date: Thu, 27 Jul 2017 17:59:14 +0200 Subject: libceph: make RECOVERY_DELETES feature create a new interval This is needed so that the OSDs can regenerate the missing set at the start of a new interval where support for recovery deletes changed. Signed-off-by: Ilya Dryomov Reviewed-by: Sage Weil --- include/linux/ceph/osd_client.h | 1 + include/linux/ceph/osdmap.h | 2 ++ include/linux/ceph/rados.h | 4 ++++ net/ceph/osd_client.c | 5 +++++ net/ceph/osdmap.c | 5 ++++- 5 files changed, 16 insertions(+), 1 deletion(-) (limited to 'include') diff --git a/include/linux/ceph/osd_client.h b/include/linux/ceph/osd_client.h index c6d96a5f46fd..adf670ecaf94 100644 --- a/include/linux/ceph/osd_client.h +++ b/include/linux/ceph/osd_client.h @@ -148,6 +148,7 @@ struct ceph_osd_request_target { int size; int min_size; bool sort_bitwise; + bool recovery_deletes; unsigned int flags; /* CEPH_OSD_FLAG_* */ bool paused; diff --git a/include/linux/ceph/osdmap.h b/include/linux/ceph/osdmap.h index a0996cb9faed..af3444a5bfdd 100644 --- a/include/linux/ceph/osdmap.h +++ b/include/linux/ceph/osdmap.h @@ -272,6 +272,8 @@ bool ceph_is_new_interval(const struct ceph_osds *old_acting, u32 new_pg_num, bool old_sort_bitwise, bool new_sort_bitwise, + bool old_recovery_deletes, + bool new_recovery_deletes, const struct ceph_pg *pgid); bool ceph_osds_changed(const struct ceph_osds *old_acting, const struct ceph_osds *new_acting, diff --git a/include/linux/ceph/rados.h b/include/linux/ceph/rados.h index 385db08bb8b2..b8281feda9c7 100644 --- a/include/linux/ceph/rados.h +++ b/include/linux/ceph/rados.h @@ -158,6 +158,10 @@ extern const char *ceph_osd_state_name(int s); #define CEPH_OSDMAP_NOTIERAGENT (1<<13) /* disable tiering agent */ #define CEPH_OSDMAP_NOREBALANCE (1<<14) /* block osd backfill unless pg is degraded */ #define CEPH_OSDMAP_SORTBITWISE (1<<15) /* use bitwise hobject_t sort */ +#define CEPH_OSDMAP_REQUIRE_JEWEL (1<<16) /* require jewel for booting osds */ +#define CEPH_OSDMAP_REQUIRE_KRAKEN (1<<17) /* require kraken for booting osds */ +#define CEPH_OSDMAP_REQUIRE_LUMINOUS (1<<18) /* require l for booting osds */ +#define CEPH_OSDMAP_RECOVERY_DELETES (1<<19) /* deletes performed during recovery instead of peering */ /* * The error code to return when an OSD can't handle a write diff --git a/net/ceph/osd_client.c b/net/ceph/osd_client.c index b5f016cb9569..dcfbdd74dfd1 100644 --- a/net/ceph/osd_client.c +++ b/net/ceph/osd_client.c @@ -1337,6 +1337,8 @@ static enum calc_target_result calc_target(struct ceph_osd_client *osdc, bool legacy_change; bool split = false; bool sort_bitwise = ceph_osdmap_flag(osdc, CEPH_OSDMAP_SORTBITWISE); + bool recovery_deletes = ceph_osdmap_flag(osdc, + CEPH_OSDMAP_RECOVERY_DELETES); enum calc_target_result ct_res; int ret; @@ -1399,6 +1401,8 @@ static enum calc_target_result calc_target(struct ceph_osd_client *osdc, pi->pg_num, t->sort_bitwise, sort_bitwise, + t->recovery_deletes, + recovery_deletes, &last_pgid)) force_resend = true; @@ -1421,6 +1425,7 @@ static enum calc_target_result calc_target(struct ceph_osd_client *osdc, t->pg_num = pi->pg_num; t->pg_num_mask = pi->pg_num_mask; t->sort_bitwise = sort_bitwise; + t->recovery_deletes = recovery_deletes; t->osd = acting.primary; } diff --git a/net/ceph/osdmap.c b/net/ceph/osdmap.c index 0bec71fa712e..f358d0bfa76b 100644 --- a/net/ceph/osdmap.c +++ b/net/ceph/osdmap.c @@ -2082,6 +2082,8 @@ bool ceph_is_new_interval(const struct ceph_osds *old_acting, u32 new_pg_num, bool old_sort_bitwise, bool new_sort_bitwise, + bool old_recovery_deletes, + bool new_recovery_deletes, const struct ceph_pg *pgid) { return !osds_equal(old_acting, new_acting) || @@ -2089,7 +2091,8 @@ bool ceph_is_new_interval(const struct ceph_osds *old_acting, old_size != new_size || old_min_size != new_min_size || ceph_pg_is_split(pgid, old_pg_num, new_pg_num) || - old_sort_bitwise != new_sort_bitwise; + old_sort_bitwise != new_sort_bitwise || + old_recovery_deletes != new_recovery_deletes; } static int calc_pg_rank(int osd, const struct ceph_osds *acting) -- cgit v1.2.3 From fd40559c8657418385e42f797e0b04bfc0add748 Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Tue, 1 Aug 2017 16:02:47 -0400 Subject: NFSv4: Fix EXCHANGE_ID corrupt verifier issue The verifier is allocated on the stack, but the EXCHANGE_ID RPC call was changed to be asynchronous by commit 8d89bd70bc939. If we interrrupt the call to rpc_wait_for_completion_task(), we can therefore end up transmitting random stack contents in lieu of the verifier. Fixes: 8d89bd70bc939 ("NFS setup async exchange_id") Cc: stable@vger.kernel.org # v4.9+ Signed-off-by: Trond Myklebust Signed-off-by: Anna Schumaker --- fs/nfs/nfs4proc.c | 11 ++++------- fs/nfs/nfs4xdr.c | 2 +- include/linux/nfs_xdr.h | 2 +- 3 files changed, 6 insertions(+), 9 deletions(-) (limited to 'include') diff --git a/fs/nfs/nfs4proc.c b/fs/nfs/nfs4proc.c index 93c1d7352238..c458a1e28b91 100644 --- a/fs/nfs/nfs4proc.c +++ b/fs/nfs/nfs4proc.c @@ -7460,7 +7460,7 @@ static void nfs4_exchange_id_done(struct rpc_task *task, void *data) cdata->res.server_scope = NULL; } /* Save the EXCHANGE_ID verifier session trunk tests */ - memcpy(clp->cl_confirm.data, cdata->args.verifier->data, + memcpy(clp->cl_confirm.data, cdata->args.verifier.data, sizeof(clp->cl_confirm.data)); } out: @@ -7497,7 +7497,6 @@ static const struct rpc_call_ops nfs4_exchange_id_call_ops = { static int _nfs4_proc_exchange_id(struct nfs_client *clp, struct rpc_cred *cred, u32 sp4_how, struct rpc_xprt *xprt) { - nfs4_verifier verifier; struct rpc_message msg = { .rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_EXCHANGE_ID], .rpc_cred = cred, @@ -7521,8 +7520,7 @@ static int _nfs4_proc_exchange_id(struct nfs_client *clp, struct rpc_cred *cred, return -ENOMEM; } - if (!xprt) - nfs4_init_boot_verifier(clp, &verifier); + nfs4_init_boot_verifier(clp, &calldata->args.verifier); status = nfs4_init_uniform_client_string(clp); if (status) @@ -7563,9 +7561,8 @@ static int _nfs4_proc_exchange_id(struct nfs_client *clp, struct rpc_cred *cred, task_setup_data.rpc_xprt = xprt; task_setup_data.flags = RPC_TASK_SOFT|RPC_TASK_SOFTCONN|RPC_TASK_ASYNC; - calldata->args.verifier = &clp->cl_confirm; - } else { - calldata->args.verifier = &verifier; + memcpy(calldata->args.verifier.data, clp->cl_confirm.data, + sizeof(calldata->args.verifier.data)); } calldata->args.client = clp; #ifdef CONFIG_NFS_V4_1_MIGRATION diff --git a/fs/nfs/nfs4xdr.c b/fs/nfs/nfs4xdr.c index fa3eb361d4f8..37c8af003275 100644 --- a/fs/nfs/nfs4xdr.c +++ b/fs/nfs/nfs4xdr.c @@ -1785,7 +1785,7 @@ static void encode_exchange_id(struct xdr_stream *xdr, int len = 0; encode_op_hdr(xdr, OP_EXCHANGE_ID, decode_exchange_id_maxsz, hdr); - encode_nfs4_verifier(xdr, args->verifier); + encode_nfs4_verifier(xdr, &args->verifier); encode_string(xdr, strlen(args->client->cl_owner_id), args->client->cl_owner_id); diff --git a/include/linux/nfs_xdr.h b/include/linux/nfs_xdr.h index ca3bcc4ed4e5..62cbcb842f99 100644 --- a/include/linux/nfs_xdr.h +++ b/include/linux/nfs_xdr.h @@ -1235,7 +1235,7 @@ struct nfs41_state_protection { struct nfs41_exchange_id_args { struct nfs_client *client; - nfs4_verifier *verifier; + nfs4_verifier verifier; u32 flags; struct nfs41_state_protection state_protect; }; -- cgit v1.2.3 From 3898da947bbaf9e7fd5816e825978d360028bba2 Mon Sep 17 00:00:00 2001 From: Paolo Bonzini Date: Wed, 2 Aug 2017 17:55:54 +0200 Subject: KVM: avoid using rcu_dereference_protected MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit During teardown, accesses to memslots and buses are using rcu_dereference_protected with an always-true condition because these accesses are done outside the usual mutexes. This is because the last reference is gone and there cannot be any concurrent modifications, but rcu_dereference_protected is ugly and unobvious. Instead, check the refcount in kvm_get_bus and __kvm_memslots. Signed-off-by: Paolo Bonzini Signed-off-by: Radim Krčmář --- include/linux/kvm_host.h | 6 ++++-- virt/kvm/kvm_main.c | 11 ++++------- 2 files changed, 8 insertions(+), 9 deletions(-) (limited to 'include') diff --git a/include/linux/kvm_host.h b/include/linux/kvm_host.h index 890b706d1943..21a6fd6c44af 100644 --- a/include/linux/kvm_host.h +++ b/include/linux/kvm_host.h @@ -477,7 +477,8 @@ struct kvm { static inline struct kvm_io_bus *kvm_get_bus(struct kvm *kvm, enum kvm_bus idx) { return srcu_dereference_check(kvm->buses[idx], &kvm->srcu, - lockdep_is_held(&kvm->slots_lock)); + lockdep_is_held(&kvm->slots_lock) || + !refcount_read(&kvm->users_count)); } static inline struct kvm_vcpu *kvm_get_vcpu(struct kvm *kvm, int i) @@ -570,7 +571,8 @@ void kvm_put_kvm(struct kvm *kvm); static inline struct kvm_memslots *__kvm_memslots(struct kvm *kvm, int as_id) { return srcu_dereference_check(kvm->memslots[as_id], &kvm->srcu, - lockdep_is_held(&kvm->slots_lock)); + lockdep_is_held(&kvm->slots_lock) || + !refcount_read(&kvm->users_count)); } static inline struct kvm_memslots *kvm_memslots(struct kvm *kvm) diff --git a/virt/kvm/kvm_main.c b/virt/kvm/kvm_main.c index f3f74271f1a9..15252d723b54 100644 --- a/virt/kvm/kvm_main.c +++ b/virt/kvm/kvm_main.c @@ -717,10 +717,9 @@ out_err_no_srcu: hardware_disable_all(); out_err_no_disable: for (i = 0; i < KVM_NR_BUSES; i++) - kfree(rcu_access_pointer(kvm->buses[i])); + kfree(kvm_get_bus(kvm, i)); for (i = 0; i < KVM_ADDRESS_SPACE_NUM; i++) - kvm_free_memslots(kvm, - rcu_dereference_protected(kvm->memslots[i], 1)); + kvm_free_memslots(kvm, __kvm_memslots(kvm, i)); kvm_arch_free_vm(kvm); mmdrop(current->mm); return ERR_PTR(r); @@ -754,9 +753,8 @@ static void kvm_destroy_vm(struct kvm *kvm) spin_unlock(&kvm_lock); kvm_free_irq_routing(kvm); for (i = 0; i < KVM_NR_BUSES; i++) { - struct kvm_io_bus *bus; + struct kvm_io_bus *bus = kvm_get_bus(kvm, i); - bus = rcu_dereference_protected(kvm->buses[i], 1); if (bus) kvm_io_bus_destroy(bus); kvm->buses[i] = NULL; @@ -770,8 +768,7 @@ static void kvm_destroy_vm(struct kvm *kvm) kvm_arch_destroy_vm(kvm); kvm_destroy_devices(kvm); for (i = 0; i < KVM_ADDRESS_SPACE_NUM; i++) - kvm_free_memslots(kvm, - rcu_dereference_protected(kvm->memslots[i], 1)); + kvm_free_memslots(kvm, __kvm_memslots(kvm, i)); cleanup_srcu_struct(&kvm->irq_srcu); cleanup_srcu_struct(&kvm->srcu); kvm_arch_free_vm(kvm); -- cgit v1.2.3 From 3ea277194daaeaa84ce75180ec7c7a2075027a68 Mon Sep 17 00:00:00 2001 From: Mel Gorman Date: Wed, 2 Aug 2017 13:31:52 -0700 Subject: mm, mprotect: flush TLB if potentially racing with a parallel reclaim leaving stale TLB entries Nadav Amit identified a theoritical race between page reclaim and mprotect due to TLB flushes being batched outside of the PTL being held. He described the race as follows: CPU0 CPU1 ---- ---- user accesses memory using RW PTE [PTE now cached in TLB] try_to_unmap_one() ==> ptep_get_and_clear() ==> set_tlb_ubc_flush_pending() mprotect(addr, PROT_READ) ==> change_pte_range() ==> [ PTE non-present - no flush ] user writes using cached RW PTE ... try_to_unmap_flush() The same type of race exists for reads when protecting for PROT_NONE and also exists for operations that can leave an old TLB entry behind such as munmap, mremap and madvise. For some operations like mprotect, it's not necessarily a data integrity issue but it is a correctness issue as there is a window where an mprotect that limits access still allows access. For munmap, it's potentially a data integrity issue although the race is massive as an munmap, mmap and return to userspace must all complete between the window when reclaim drops the PTL and flushes the TLB. However, it's theoritically possible so handle this issue by flushing the mm if reclaim is potentially currently batching TLB flushes. Other instances where a flush is required for a present pte should be ok as either the page lock is held preventing parallel reclaim or a page reference count is elevated preventing a parallel free leading to corruption. In the case of page_mkclean there isn't an obvious path that userspace could take advantage of without using the operations that are guarded by this patch. Other users such as gup as a race with reclaim looks just at PTEs. huge page variants should be ok as they don't race with reclaim. mincore only looks at PTEs. userfault also should be ok as if a parallel reclaim takes place, it will either fault the page back in or read some of the data before the flush occurs triggering a fault. Note that a variant of this patch was acked by Andy Lutomirski but this was for the x86 parts on top of his PCID work which didn't make the 4.13 merge window as expected. His ack is dropped from this version and there will be a follow-on patch on top of PCID that will include his ack. [akpm@linux-foundation.org: tweak comments] [akpm@linux-foundation.org: fix spello] Link: http://lkml.kernel.org/r/20170717155523.emckq2esjro6hf3z@suse.de Reported-by: Nadav Amit Signed-off-by: Mel Gorman Cc: Andy Lutomirski Cc: [v4.4+] Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- include/linux/mm_types.h | 4 ++++ mm/internal.h | 5 ++++- mm/madvise.c | 1 + mm/memory.c | 1 + mm/mprotect.c | 1 + mm/mremap.c | 1 + mm/rmap.c | 36 ++++++++++++++++++++++++++++++++++++ 7 files changed, 48 insertions(+), 1 deletion(-) (limited to 'include') diff --git a/include/linux/mm_types.h b/include/linux/mm_types.h index ff151814a02d..7f384bb62d8e 100644 --- a/include/linux/mm_types.h +++ b/include/linux/mm_types.h @@ -494,6 +494,10 @@ struct mm_struct { * PROT_NONE or PROT_NUMA mapped page. */ bool tlb_flush_pending; +#endif +#ifdef CONFIG_ARCH_WANT_BATCHED_UNMAP_TLB_FLUSH + /* See flush_tlb_batched_pending() */ + bool tlb_flush_batched; #endif struct uprobes_state uprobes_state; #ifdef CONFIG_HUGETLB_PAGE diff --git a/mm/internal.h b/mm/internal.h index 24d88f084705..4ef49fc55e58 100644 --- a/mm/internal.h +++ b/mm/internal.h @@ -498,6 +498,7 @@ extern struct workqueue_struct *mm_percpu_wq; #ifdef CONFIG_ARCH_WANT_BATCHED_UNMAP_TLB_FLUSH void try_to_unmap_flush(void); void try_to_unmap_flush_dirty(void); +void flush_tlb_batched_pending(struct mm_struct *mm); #else static inline void try_to_unmap_flush(void) { @@ -505,7 +506,9 @@ static inline void try_to_unmap_flush(void) static inline void try_to_unmap_flush_dirty(void) { } - +static inline void flush_tlb_batched_pending(struct mm_struct *mm) +{ +} #endif /* CONFIG_ARCH_WANT_BATCHED_UNMAP_TLB_FLUSH */ extern const struct trace_print_flags pageflag_names[]; diff --git a/mm/madvise.c b/mm/madvise.c index 9976852f1e1c..47d8d8a25eae 100644 --- a/mm/madvise.c +++ b/mm/madvise.c @@ -320,6 +320,7 @@ static int madvise_free_pte_range(pmd_t *pmd, unsigned long addr, tlb_remove_check_page_size_change(tlb, PAGE_SIZE); orig_pte = pte = pte_offset_map_lock(mm, pmd, addr, &ptl); + flush_tlb_batched_pending(mm); arch_enter_lazy_mmu_mode(); for (; addr != end; pte++, addr += PAGE_SIZE) { ptent = *pte; diff --git a/mm/memory.c b/mm/memory.c index 0e517be91a89..f65beaad319b 100644 --- a/mm/memory.c +++ b/mm/memory.c @@ -1197,6 +1197,7 @@ again: init_rss_vec(rss); start_pte = pte_offset_map_lock(mm, pmd, addr, &ptl); pte = start_pte; + flush_tlb_batched_pending(mm); arch_enter_lazy_mmu_mode(); do { pte_t ptent = *pte; diff --git a/mm/mprotect.c b/mm/mprotect.c index 1a8c9ca83e48..4180ad8cc9c5 100644 --- a/mm/mprotect.c +++ b/mm/mprotect.c @@ -64,6 +64,7 @@ static unsigned long change_pte_range(struct vm_area_struct *vma, pmd_t *pmd, atomic_read(&vma->vm_mm->mm_users) == 1) target_node = numa_node_id(); + flush_tlb_batched_pending(vma->vm_mm); arch_enter_lazy_mmu_mode(); do { oldpte = *pte; diff --git a/mm/mremap.c b/mm/mremap.c index cd8a1b199ef9..6e3d857458de 100644 --- a/mm/mremap.c +++ b/mm/mremap.c @@ -152,6 +152,7 @@ static void move_ptes(struct vm_area_struct *vma, pmd_t *old_pmd, new_ptl = pte_lockptr(mm, new_pmd); if (new_ptl != old_ptl) spin_lock_nested(new_ptl, SINGLE_DEPTH_NESTING); + flush_tlb_batched_pending(vma->vm_mm); arch_enter_lazy_mmu_mode(); for (; old_addr < old_end; old_pte++, old_addr += PAGE_SIZE, diff --git a/mm/rmap.c b/mm/rmap.c index ced14f1af6dc..c8993c63eb25 100644 --- a/mm/rmap.c +++ b/mm/rmap.c @@ -604,6 +604,13 @@ static void set_tlb_ubc_flush_pending(struct mm_struct *mm, bool writable) arch_tlbbatch_add_mm(&tlb_ubc->arch, mm); tlb_ubc->flush_required = true; + /* + * Ensure compiler does not re-order the setting of tlb_flush_batched + * before the PTE is cleared. + */ + barrier(); + mm->tlb_flush_batched = true; + /* * If the PTE was dirty then it's best to assume it's writable. The * caller must use try_to_unmap_flush_dirty() or try_to_unmap_flush() @@ -631,6 +638,35 @@ static bool should_defer_flush(struct mm_struct *mm, enum ttu_flags flags) return should_defer; } + +/* + * Reclaim unmaps pages under the PTL but do not flush the TLB prior to + * releasing the PTL if TLB flushes are batched. It's possible for a parallel + * operation such as mprotect or munmap to race between reclaim unmapping + * the page and flushing the page. If this race occurs, it potentially allows + * access to data via a stale TLB entry. Tracking all mm's that have TLB + * batching in flight would be expensive during reclaim so instead track + * whether TLB batching occurred in the past and if so then do a flush here + * if required. This will cost one additional flush per reclaim cycle paid + * by the first operation at risk such as mprotect and mumap. + * + * This must be called under the PTL so that an access to tlb_flush_batched + * that is potentially a "reclaim vs mprotect/munmap/etc" race will synchronise + * via the PTL. + */ +void flush_tlb_batched_pending(struct mm_struct *mm) +{ + if (mm->tlb_flush_batched) { + flush_tlb_mm(mm); + + /* + * Do not allow the compiler to re-order the clearing of + * tlb_flush_batched before the tlb is flushed. + */ + barrier(); + mm->tlb_flush_batched = false; + } +} #else static void set_tlb_ubc_flush_pending(struct mm_struct *mm, bool writable) { -- cgit v1.2.3 From d16977f3a6cfbb5e9ce477f423a1bf343347c1ed Mon Sep 17 00:00:00 2001 From: Jonathan Corbet Date: Wed, 2 Aug 2017 13:32:01 -0700 Subject: kthread: fix documentation build warning The kerneldoc comment for kthread_create() had an incorrect argument name, leading to a warning in the docs build. Correct it, and make one more small step toward a warning-free build. Link: http://lkml.kernel.org/r/20170724135916.7f486c6f@lwn.net Signed-off-by: Jonathan Corbet Cc: Randy Dunlap Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- include/linux/kthread.h | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'include') diff --git a/include/linux/kthread.h b/include/linux/kthread.h index 4fec8b775895..82e197eeac91 100644 --- a/include/linux/kthread.h +++ b/include/linux/kthread.h @@ -15,7 +15,7 @@ struct task_struct *kthread_create_on_node(int (*threadfn)(void *data), * @threadfn: the function to run in the thread * @data: data pointer for @threadfn() * @namefmt: printf-style format string for the thread name - * @...: arguments for @namefmt. + * @arg...: arguments for @namefmt. * * This macro will create a kthread on the current node, leaving it in * the stopped state. This is just a helper for kthread_create_on_node(); -- cgit v1.2.3 From 89affbf5d9ebb15c6460596822e8857ea2f9e735 Mon Sep 17 00:00:00 2001 From: Dima Zavin Date: Wed, 2 Aug 2017 13:32:18 -0700 Subject: cpuset: fix a deadlock due to incomplete patching of cpusets_enabled() In codepaths that use the begin/retry interface for reading mems_allowed_seq with irqs disabled, there exists a race condition that stalls the patch process after only modifying a subset of the static_branch call sites. This problem manifested itself as a deadlock in the slub allocator, inside get_any_partial. The loop reads mems_allowed_seq value (via read_mems_allowed_begin), performs the defrag operation, and then verifies the consistency of mem_allowed via the read_mems_allowed_retry and the cookie returned by xxx_begin. The issue here is that both begin and retry first check if cpusets are enabled via cpusets_enabled() static branch. This branch can be rewritted dynamically (via cpuset_inc) if a new cpuset is created. The x86 jump label code fully synchronizes across all CPUs for every entry it rewrites. If it rewrites only one of the callsites (specifically the one in read_mems_allowed_retry) and then waits for the smp_call_function(do_sync_core) to complete while a CPU is inside the begin/retry section with IRQs off and the mems_allowed value is changed, we can hang. This is because begin() will always return 0 (since it wasn't patched yet) while retry() will test the 0 against the actual value of the seq counter. The fix is to use two different static keys: one for begin (pre_enable_key) and one for retry (enable_key). In cpuset_inc(), we first bump the pre_enable key to ensure that cpuset_mems_allowed_begin() always return a valid seqcount if are enabling cpusets. Similarly, when disabling cpusets via cpuset_dec(), we first ensure that callers of cpuset_mems_allowed_retry() will start ignoring the seqcount value before we let cpuset_mems_allowed_begin() return 0. The relevant stack traces of the two stuck threads: CPU: 1 PID: 1415 Comm: mkdir Tainted: G L 4.9.36-00104-g540c51286237 #4 Hardware name: Default string Default string/Hardware, BIOS 4.29.1-20170526215256 05/26/2017 task: ffff8817f9c28000 task.stack: ffffc9000ffa4000 RIP: smp_call_function_many+0x1f9/0x260 Call Trace: smp_call_function+0x3b/0x70 on_each_cpu+0x2f/0x90 text_poke_bp+0x87/0xd0 arch_jump_label_transform+0x93/0x100 __jump_label_update+0x77/0x90 jump_label_update+0xaa/0xc0 static_key_slow_inc+0x9e/0xb0 cpuset_css_online+0x70/0x2e0 online_css+0x2c/0xa0 cgroup_apply_control_enable+0x27f/0x3d0 cgroup_mkdir+0x2b7/0x420 kernfs_iop_mkdir+0x5a/0x80 vfs_mkdir+0xf6/0x1a0 SyS_mkdir+0xb7/0xe0 entry_SYSCALL_64_fastpath+0x18/0xad ... CPU: 2 PID: 1 Comm: init Tainted: G L 4.9.36-00104-g540c51286237 #4 Hardware name: Default string Default string/Hardware, BIOS 4.29.1-20170526215256 05/26/2017 task: ffff8818087c0000 task.stack: ffffc90000030000 RIP: int3+0x39/0x70 Call Trace: <#DB> ? ___slab_alloc+0x28b/0x5a0 ? copy_process.part.40+0xf7/0x1de0 __slab_alloc.isra.80+0x54/0x90 copy_process.part.40+0xf7/0x1de0 copy_process.part.40+0xf7/0x1de0 kmem_cache_alloc_node+0x8a/0x280 copy_process.part.40+0xf7/0x1de0 _do_fork+0xe7/0x6c0 _raw_spin_unlock_irq+0x2d/0x60 trace_hardirqs_on_caller+0x136/0x1d0 entry_SYSCALL_64_fastpath+0x5/0xad do_syscall_64+0x27/0x350 SyS_clone+0x19/0x20 do_syscall_64+0x60/0x350 entry_SYSCALL64_slow_path+0x25/0x25 Link: http://lkml.kernel.org/r/20170731040113.14197-1-dmitriyz@waymo.com Fixes: 46e700abc44c ("mm, page_alloc: remove unnecessary taking of a seqlock when cpusets are disabled") Signed-off-by: Dima Zavin Reported-by: Cliff Spradlin Acked-by: Vlastimil Babka Cc: Peter Zijlstra Cc: Christopher Lameter Cc: Li Zefan Cc: Pekka Enberg Cc: David Rientjes Cc: Joonsoo Kim Cc: Mel Gorman Cc: Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- include/linux/cpuset.h | 19 +++++++++++++++++-- kernel/cgroup/cpuset.c | 1 + 2 files changed, 18 insertions(+), 2 deletions(-) (limited to 'include') diff --git a/include/linux/cpuset.h b/include/linux/cpuset.h index 119a3f9604b0..898cfe2eeb42 100644 --- a/include/linux/cpuset.h +++ b/include/linux/cpuset.h @@ -18,6 +18,19 @@ #ifdef CONFIG_CPUSETS +/* + * Static branch rewrites can happen in an arbitrary order for a given + * key. In code paths where we need to loop with read_mems_allowed_begin() and + * read_mems_allowed_retry() to get a consistent view of mems_allowed, we need + * to ensure that begin() always gets rewritten before retry() in the + * disabled -> enabled transition. If not, then if local irqs are disabled + * around the loop, we can deadlock since retry() would always be + * comparing the latest value of the mems_allowed seqcount against 0 as + * begin() still would see cpusets_enabled() as false. The enabled -> disabled + * transition should happen in reverse order for the same reasons (want to stop + * looking at real value of mems_allowed.sequence in retry() first). + */ +extern struct static_key_false cpusets_pre_enable_key; extern struct static_key_false cpusets_enabled_key; static inline bool cpusets_enabled(void) { @@ -32,12 +45,14 @@ static inline int nr_cpusets(void) static inline void cpuset_inc(void) { + static_branch_inc(&cpusets_pre_enable_key); static_branch_inc(&cpusets_enabled_key); } static inline void cpuset_dec(void) { static_branch_dec(&cpusets_enabled_key); + static_branch_dec(&cpusets_pre_enable_key); } extern int cpuset_init(void); @@ -115,7 +130,7 @@ extern void cpuset_print_current_mems_allowed(void); */ static inline unsigned int read_mems_allowed_begin(void) { - if (!cpusets_enabled()) + if (!static_branch_unlikely(&cpusets_pre_enable_key)) return 0; return read_seqcount_begin(¤t->mems_allowed_seq); @@ -129,7 +144,7 @@ static inline unsigned int read_mems_allowed_begin(void) */ static inline bool read_mems_allowed_retry(unsigned int seq) { - if (!cpusets_enabled()) + if (!static_branch_unlikely(&cpusets_enabled_key)) return false; return read_seqcount_retry(¤t->mems_allowed_seq, seq); diff --git a/kernel/cgroup/cpuset.c b/kernel/cgroup/cpuset.c index ca8376e5008c..8d5151688504 100644 --- a/kernel/cgroup/cpuset.c +++ b/kernel/cgroup/cpuset.c @@ -63,6 +63,7 @@ #include #include +DEFINE_STATIC_KEY_FALSE(cpusets_pre_enable_key); DEFINE_STATIC_KEY_FALSE(cpusets_enabled_key); /* See "Frequency meter" comments, below. */ -- cgit v1.2.3 From 1ee1c3f5b5cff959e0ac95a125bd15eaf88cc638 Mon Sep 17 00:00:00 2001 From: Kan Liang Date: Wed, 2 Aug 2017 13:32:27 -0700 Subject: mm: allow page_cache_get_speculative in interrupt context Kernel panic when calling the IRQ-safe __get_user_pages_fast in NMI handler. The bug was introduced by commit 2947ba054a4d ("x86/mm/gup: Switch GUP to the generic get_user_page_fast() implementation"). The original x86 __get_user_page_fast used plain get_page() or page_ref_add(). However, the generic __get_user_page_fast uses page_cache_get_speculative(), which has VM_BUG_ON(in_interrupt()). There is no reason to prevent page_cache_get_speculative from using in interrupt context. According to the author, putting a BUG_ON there is just because the code is not verifying correctness of interrupt races. I did some tests in interrupt context. There is no issue found. Removing VM_BUG_ON(in_interrupt()) for page_cache_get_speculative(). Link: http://lkml.kernel.org/r/1501609146-59730-1-git-send-email-kan.liang@intel.com Fixes: 2947ba054a4d ("x86/mm/gup: Switch GUP to the generic get_user_page_fast() implementation") Signed-off-by: Kan Liang Cc: Jens Axboe Cc: Al Viro Cc: Kirill A. Shutemov Cc: Ying Huang Cc: Nicholas Piggin Cc: Ingo Molnar Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- include/linux/pagemap.h | 2 -- 1 file changed, 2 deletions(-) (limited to 'include') diff --git a/include/linux/pagemap.h b/include/linux/pagemap.h index baa9344dcd10..79b36f57c3ba 100644 --- a/include/linux/pagemap.h +++ b/include/linux/pagemap.h @@ -163,8 +163,6 @@ void release_pages(struct page **pages, int nr, bool cold); */ static inline int page_cache_get_speculative(struct page *page) { - VM_BUG_ON(in_interrupt()); - #ifdef CONFIG_TINY_RCU # ifdef CONFIG_PREEMPT_COUNT VM_BUG_ON(!in_atomic() && !irqs_disabled()); -- cgit v1.2.3 From eeb66cdb682678bfd1f02a4547e3649b38ffea7e Mon Sep 17 00:00:00 2001 From: Saeed Mahameed Date: Sun, 4 Jun 2017 23:11:55 +0300 Subject: net/mlx5: Separate between E-Switch and MPFS Multi-Physical Function Switch (MPFs) is required for when multi-PF configuration is enabled to allow passing user configured unicast MAC addresses to the requesting PF. Before this patch eswitch.c used to manage the HW MPFS l2 table, E-Switch always (regardless of sriov) enabled vport(0) (NIC PF) vport's contexts update on unicast mac address list changes, to populate the PF's MPFS L2 table accordingly. In downstream patch we would like to allow compiling the driver without E-Switch functionalities, for that we move MPFS l2 table logic out of eswitch.c into its own file, and provide Kconfig flag (MLX5_MPFS) to allow compiling out MPFS for those who don't want Multi-PF support. NIC PF netdevice will now directly update MPFS l2 table via the new MPFS API. VF netdevice has no access to MPFS L2 table, so E-Switch will remain responsible of updating its MPFS l2 table on behalf of its VFs. Due to this change we also don't require enabling vport(0) (PF vport) unicast mac changes events anymore, for when SRIOV is not enabled. Which means E-Switch is now activated only on SRIOV activation, and not required otherwise. Signed-off-by: Saeed Mahameed Cc: Jes Sorensen Cc: kernel-team@fb.com --- drivers/net/ethernet/mellanox/mlx5/core/Kconfig | 10 + drivers/net/ethernet/mellanox/mlx5/core/Makefile | 2 + drivers/net/ethernet/mellanox/mlx5/core/en_fs.c | 17 +- drivers/net/ethernet/mellanox/mlx5/core/eswitch.c | 190 ++++--------------- drivers/net/ethernet/mellanox/mlx5/core/eswitch.h | 55 +----- drivers/net/ethernet/mellanox/mlx5/core/lib/mpfs.c | 201 +++++++++++++++++++++ drivers/net/ethernet/mellanox/mlx5/core/lib/mpfs.h | 95 ++++++++++ drivers/net/ethernet/mellanox/mlx5/core/main.c | 26 ++- include/linux/mlx5/driver.h | 2 + 9 files changed, 377 insertions(+), 221 deletions(-) create mode 100644 drivers/net/ethernet/mellanox/mlx5/core/lib/mpfs.c create mode 100644 drivers/net/ethernet/mellanox/mlx5/core/lib/mpfs.h (limited to 'include') diff --git a/drivers/net/ethernet/mellanox/mlx5/core/Kconfig b/drivers/net/ethernet/mellanox/mlx5/core/Kconfig index 5aee05992f27..d7174295b6ef 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/Kconfig +++ b/drivers/net/ethernet/mellanox/mlx5/core/Kconfig @@ -34,6 +34,16 @@ config MLX5_CORE_EN ---help--- Ethernet support in Mellanox Technologies ConnectX-4 NIC. +config MLX5_MPFS + bool "Mellanox Technologies MLX5 MPFS support" + depends on MLX5_CORE_EN + default y + ---help--- + Mellanox Technologies Ethernet Multi-Physical Function Switch (MPFS) + support in ConnectX NIC. MPFs is required for when multi-PF configuration + is enabled to allow passing user configured unicast MAC addresses to the + requesting PF. + config MLX5_CORE_EN_DCB bool "Data Center Bridging (DCB) Support" default y diff --git a/drivers/net/ethernet/mellanox/mlx5/core/Makefile b/drivers/net/ethernet/mellanox/mlx5/core/Makefile index 9d17e4e76d3a..c867e48f8a4c 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/Makefile +++ b/drivers/net/ethernet/mellanox/mlx5/core/Makefile @@ -16,6 +16,8 @@ mlx5_core-$(CONFIG_MLX5_CORE_EN) += eswitch.o eswitch_offloads.o \ en_rx.o en_rx_am.o en_txrx.o en_clock.o vxlan.o \ en_tc.o en_arfs.o en_rep.o en_fs_ethtool.o en_selftest.o +mlx5_core-$(CONFIG_MLX5_MPFS) += lib/mpfs.o + mlx5_core-$(CONFIG_MLX5_CORE_EN_DCB) += en_dcbnl.o mlx5_core-$(CONFIG_MLX5_CORE_IPOIB) += ipoib/ipoib.o ipoib/ethtool.o diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_fs.c b/drivers/net/ethernet/mellanox/mlx5/core/en_fs.c index dfccb5305e9c..eecbc6d4f51f 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en_fs.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_fs.c @@ -36,6 +36,7 @@ #include #include #include "en.h" +#include "lib/mpfs.h" static int mlx5e_add_l2_flow_rule(struct mlx5e_priv *priv, struct mlx5e_l2_rule *ai, int type); @@ -65,6 +66,7 @@ struct mlx5e_l2_hash_node { struct hlist_node hlist; u8 action; struct mlx5e_l2_rule ai; + bool mpfs; }; static inline int mlx5e_hash_l2(u8 *addr) @@ -362,17 +364,30 @@ static void mlx5e_del_vlan_rules(struct mlx5e_priv *priv) static void mlx5e_execute_l2_action(struct mlx5e_priv *priv, struct mlx5e_l2_hash_node *hn) { - switch (hn->action) { + u8 action = hn->action; + int l2_err = 0; + + switch (action) { case MLX5E_ACTION_ADD: mlx5e_add_l2_flow_rule(priv, &hn->ai, MLX5E_FULLMATCH); + if (!is_multicast_ether_addr(hn->ai.addr)) { + l2_err = mlx5_mpfs_add_mac(priv->mdev, hn->ai.addr); + hn->mpfs = !l2_err; + } hn->action = MLX5E_ACTION_NONE; break; case MLX5E_ACTION_DEL: + if (!is_multicast_ether_addr(hn->ai.addr) && hn->mpfs) + l2_err = mlx5_mpfs_del_mac(priv->mdev, hn->ai.addr); mlx5e_del_l2_flow_rule(priv, &hn->ai); mlx5e_del_l2_from_hash(hn); break; } + + if (l2_err) + netdev_warn(priv->netdev, "MPFS, failed to %s mac %pM, err(%d)\n", + action == MLX5E_ACTION_ADD ? "add" : "del", hn->ai.addr, l2_err); } static void mlx5e_sync_netdev_addr(struct mlx5e_priv *priv) diff --git a/drivers/net/ethernet/mellanox/mlx5/core/eswitch.c b/drivers/net/ethernet/mellanox/mlx5/core/eswitch.c index 5c001b61d04a..fd51f0ea8df9 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/eswitch.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/eswitch.c @@ -46,19 +46,13 @@ enum { MLX5_ACTION_DEL = 2, }; -/* E-Switch UC L2 table hash node */ -struct esw_uc_addr { - struct l2addr_node node; - u32 table_index; - u32 vport; -}; - /* Vport UC/MC hash node */ struct vport_addr { struct l2addr_node node; u8 action; u32 vport; - struct mlx5_flow_handle *flow_rule; /* SRIOV only */ + struct mlx5_flow_handle *flow_rule; + bool mpfs; /* UC MAC was added to MPFs */ /* A flag indicating that mac was added due to mc promiscuous vport */ bool mc_promisc; }; @@ -154,81 +148,6 @@ static int modify_esw_vport_cvlan(struct mlx5_core_dev *dev, u32 vport, return modify_esw_vport_context_cmd(dev, vport, in, sizeof(in)); } -/* HW L2 Table (MPFS) management */ -static int set_l2_table_entry_cmd(struct mlx5_core_dev *dev, u32 index, - u8 *mac, u8 vlan_valid, u16 vlan) -{ - u32 in[MLX5_ST_SZ_DW(set_l2_table_entry_in)] = {0}; - u32 out[MLX5_ST_SZ_DW(set_l2_table_entry_out)] = {0}; - u8 *in_mac_addr; - - MLX5_SET(set_l2_table_entry_in, in, opcode, - MLX5_CMD_OP_SET_L2_TABLE_ENTRY); - MLX5_SET(set_l2_table_entry_in, in, table_index, index); - MLX5_SET(set_l2_table_entry_in, in, vlan_valid, vlan_valid); - MLX5_SET(set_l2_table_entry_in, in, vlan, vlan); - - in_mac_addr = MLX5_ADDR_OF(set_l2_table_entry_in, in, mac_address); - ether_addr_copy(&in_mac_addr[2], mac); - - return mlx5_cmd_exec(dev, in, sizeof(in), out, sizeof(out)); -} - -static int del_l2_table_entry_cmd(struct mlx5_core_dev *dev, u32 index) -{ - u32 in[MLX5_ST_SZ_DW(delete_l2_table_entry_in)] = {0}; - u32 out[MLX5_ST_SZ_DW(delete_l2_table_entry_out)] = {0}; - - MLX5_SET(delete_l2_table_entry_in, in, opcode, - MLX5_CMD_OP_DELETE_L2_TABLE_ENTRY); - MLX5_SET(delete_l2_table_entry_in, in, table_index, index); - return mlx5_cmd_exec(dev, in, sizeof(in), out, sizeof(out)); -} - -static int alloc_l2_table_index(struct mlx5_l2_table *l2_table, u32 *ix) -{ - int err = 0; - - *ix = find_first_zero_bit(l2_table->bitmap, l2_table->size); - if (*ix >= l2_table->size) - err = -ENOSPC; - else - __set_bit(*ix, l2_table->bitmap); - - return err; -} - -static void free_l2_table_index(struct mlx5_l2_table *l2_table, u32 ix) -{ - __clear_bit(ix, l2_table->bitmap); -} - -static int set_l2_table_entry(struct mlx5_core_dev *dev, u8 *mac, - u8 vlan_valid, u16 vlan, - u32 *index) -{ - struct mlx5_l2_table *l2_table = &dev->priv.eswitch->l2_table; - int err; - - err = alloc_l2_table_index(l2_table, index); - if (err) - return err; - - err = set_l2_table_entry_cmd(dev, *index, mac, vlan_valid, vlan); - if (err) - free_l2_table_index(l2_table, *index); - - return err; -} - -static void del_l2_table_entry(struct mlx5_core_dev *dev, u32 index) -{ - struct mlx5_l2_table *l2_table = &dev->priv.eswitch->l2_table; - - del_l2_table_entry_cmd(dev, index); - free_l2_table_index(l2_table, index); -} - /* E-Switch FDB */ static struct mlx5_flow_handle * __esw_fdb_set_vport_rule(struct mlx5_eswitch *esw, u32 vport, bool rx_rule, @@ -455,65 +374,60 @@ typedef int (*vport_addr_action)(struct mlx5_eswitch *esw, static int esw_add_uc_addr(struct mlx5_eswitch *esw, struct vport_addr *vaddr) { - struct hlist_head *hash = esw->l2_table.l2_hash; - struct esw_uc_addr *esw_uc; u8 *mac = vaddr->node.addr; u32 vport = vaddr->vport; int err; - esw_uc = l2addr_hash_find(hash, mac, struct esw_uc_addr); - if (esw_uc) { + /* Skip mlx5_mpfs_add_mac for PFs, + * it is already done by the PF netdev in mlx5e_execute_l2_action + */ + if (!vport) + goto fdb_add; + + err = mlx5_mpfs_add_mac(esw->dev, mac); + if (err) { esw_warn(esw->dev, - "Failed to set L2 mac(%pM) for vport(%d), mac is already in use by vport(%d)\n", - mac, vport, esw_uc->vport); - return -EEXIST; + "Failed to add L2 table mac(%pM) for vport(%d), err(%d)\n", + mac, vport, err); + return err; } + vaddr->mpfs = true; - esw_uc = l2addr_hash_add(hash, mac, struct esw_uc_addr, GFP_KERNEL); - if (!esw_uc) - return -ENOMEM; - esw_uc->vport = vport; - - err = set_l2_table_entry(esw->dev, mac, 0, 0, &esw_uc->table_index); - if (err) - goto abort; - +fdb_add: /* SRIOV is enabled: Forward UC MAC to vport */ if (esw->fdb_table.fdb && esw->mode == SRIOV_LEGACY) vaddr->flow_rule = esw_fdb_set_vport_rule(esw, mac, vport); - esw_debug(esw->dev, "\tADDED UC MAC: vport[%d] %pM index:%d fr(%p)\n", - vport, mac, esw_uc->table_index, vaddr->flow_rule); - return err; -abort: - l2addr_hash_del(esw_uc); + esw_debug(esw->dev, "\tADDED UC MAC: vport[%d] %pM fr(%p)\n", + vport, mac, vaddr->flow_rule); + return err; } static int esw_del_uc_addr(struct mlx5_eswitch *esw, struct vport_addr *vaddr) { - struct hlist_head *hash = esw->l2_table.l2_hash; - struct esw_uc_addr *esw_uc; u8 *mac = vaddr->node.addr; u32 vport = vaddr->vport; + int err = 0; - esw_uc = l2addr_hash_find(hash, mac, struct esw_uc_addr); - if (!esw_uc || esw_uc->vport != vport) { - esw_debug(esw->dev, - "MAC(%pM) doesn't belong to vport (%d)\n", - mac, vport); - return -EINVAL; - } - esw_debug(esw->dev, "\tDELETE UC MAC: vport[%d] %pM index:%d fr(%p)\n", - vport, mac, esw_uc->table_index, vaddr->flow_rule); + /* Skip mlx5_mpfs_del_mac for PFs, + * it is already done by the PF netdev in mlx5e_execute_l2_action + */ + if (!vport || !vaddr->mpfs) + goto fdb_del; - del_l2_table_entry(esw->dev, esw_uc->table_index); + err = mlx5_mpfs_del_mac(esw->dev, mac); + if (err) + esw_warn(esw->dev, + "Failed to del L2 table mac(%pM) for vport(%d), err(%d)\n", + mac, vport, err); + vaddr->mpfs = false; +fdb_del: if (vaddr->flow_rule) mlx5_del_flow_rules(vaddr->flow_rule); vaddr->flow_rule = NULL; - l2addr_hash_del(esw_uc); return 0; } @@ -1635,7 +1549,6 @@ int mlx5_eswitch_enable_sriov(struct mlx5_eswitch *esw, int nvfs, int mode) esw_info(esw->dev, "E-Switch enable SRIOV: nvfs(%d) mode (%d)\n", nvfs, mode); esw->mode = mode; - esw_disable_vport(esw, 0); if (mode == SRIOV_LEGACY) err = esw_create_legacy_fdb_table(esw, nvfs + 1); @@ -1648,7 +1561,11 @@ int mlx5_eswitch_enable_sriov(struct mlx5_eswitch *esw, int nvfs, int mode) if (err) esw_warn(esw->dev, "Failed to create eswitch TSAR"); - enabled_events = (mode == SRIOV_LEGACY) ? SRIOV_VPORT_EVENTS : UC_ADDR_CHANGE; + /* Don't enable vport events when in SRIOV_OFFLOADS mode, since: + * 1. L2 table (MPFS) is programmed by PF/VF representors netdevs set_rx_mode + * 2. FDB/Eswitch is programmed by user space tools + */ + enabled_events = (mode == SRIOV_LEGACY) ? SRIOV_VPORT_EVENTS : 0; for (i = 0; i <= nvfs; i++) esw_enable_vport(esw, i, enabled_events); @@ -1657,7 +1574,6 @@ int mlx5_eswitch_enable_sriov(struct mlx5_eswitch *esw, int nvfs, int mode) return 0; abort: - esw_enable_vport(esw, 0, UC_ADDR_CHANGE); esw->mode = SRIOV_NONE; return err; } @@ -1691,30 +1607,10 @@ void mlx5_eswitch_disable_sriov(struct mlx5_eswitch *esw) esw_offloads_cleanup(esw, nvports); esw->mode = SRIOV_NONE; - /* VPORT 0 (PF) must be enabled back with non-sriov configuration */ - esw_enable_vport(esw, 0, UC_ADDR_CHANGE); -} - -void mlx5_eswitch_attach(struct mlx5_eswitch *esw) -{ - if (!ESW_ALLOWED(esw)) - return; - - esw_enable_vport(esw, 0, UC_ADDR_CHANGE); - /* VF Vports will be enabled when SRIOV is enabled */ -} - -void mlx5_eswitch_detach(struct mlx5_eswitch *esw) -{ - if (!ESW_ALLOWED(esw)) - return; - - esw_disable_vport(esw, 0); } int mlx5_eswitch_init(struct mlx5_core_dev *dev) { - int l2_table_size = 1 << MLX5_CAP_GEN(dev, log_max_l2_table); int total_vports = MLX5_TOTAL_VPORTS(dev); struct mlx5_eswitch *esw; int vport_num; @@ -1724,8 +1620,8 @@ int mlx5_eswitch_init(struct mlx5_core_dev *dev) return 0; esw_info(dev, - "Total vports %d, l2 table size(%d), per vport: max uc(%d) max mc(%d)\n", - total_vports, l2_table_size, + "Total vports %d, per vport: max uc(%d) max mc(%d)\n", + total_vports, MLX5_MAX_UC_PER_VPORT(dev), MLX5_MAX_MC_PER_VPORT(dev)); @@ -1735,14 +1631,6 @@ int mlx5_eswitch_init(struct mlx5_core_dev *dev) esw->dev = dev; - esw->l2_table.bitmap = kcalloc(BITS_TO_LONGS(l2_table_size), - sizeof(uintptr_t), GFP_KERNEL); - if (!esw->l2_table.bitmap) { - err = -ENOMEM; - goto abort; - } - esw->l2_table.size = l2_table_size; - esw->work_queue = create_singlethread_workqueue("mlx5_esw_wq"); if (!esw->work_queue) { err = -ENOMEM; @@ -1793,7 +1681,6 @@ int mlx5_eswitch_init(struct mlx5_core_dev *dev) abort: if (esw->work_queue) destroy_workqueue(esw->work_queue); - kfree(esw->l2_table.bitmap); kfree(esw->vports); kfree(esw->offloads.vport_reps); kfree(esw); @@ -1809,7 +1696,6 @@ void mlx5_eswitch_cleanup(struct mlx5_eswitch *esw) esw->dev->priv.eswitch = NULL; destroy_workqueue(esw->work_queue); - kfree(esw->l2_table.bitmap); kfree(esw->offloads.vport_reps); kfree(esw->vports); kfree(esw); diff --git a/drivers/net/ethernet/mellanox/mlx5/core/eswitch.h b/drivers/net/ethernet/mellanox/mlx5/core/eswitch.h index 834a33050969..701d228de4ad 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/eswitch.h +++ b/drivers/net/ethernet/mellanox/mlx5/core/eswitch.h @@ -37,6 +37,7 @@ #include #include #include +#include "lib/mpfs.h" #define MLX5_MAX_UC_PER_VPORT(dev) \ (1 << MLX5_CAP_GEN(dev, log_max_current_uc_list)) @@ -44,9 +45,6 @@ #define MLX5_MAX_MC_PER_VPORT(dev) \ (1 << MLX5_CAP_GEN(dev, log_max_current_mc_list)) -#define MLX5_L2_ADDR_HASH_SIZE (BIT(BITS_PER_BYTE)) -#define MLX5_L2_ADDR_HASH(addr) (addr[5]) - #define FDB_UPLINK_VPORT 0xffff #define MLX5_MIN_BW_SHARE 1 @@ -54,48 +52,6 @@ #define MLX5_RATE_TO_BW_SHARE(rate, divider, limit) \ min_t(u32, max_t(u32, (rate) / (divider), MLX5_MIN_BW_SHARE), limit) -/* L2 -mac address based- hash helpers */ -struct l2addr_node { - struct hlist_node hlist; - u8 addr[ETH_ALEN]; -}; - -#define for_each_l2hash_node(hn, tmp, hash, i) \ - for (i = 0; i < MLX5_L2_ADDR_HASH_SIZE; i++) \ - hlist_for_each_entry_safe(hn, tmp, &hash[i], hlist) - -#define l2addr_hash_find(hash, mac, type) ({ \ - int ix = MLX5_L2_ADDR_HASH(mac); \ - bool found = false; \ - type *ptr = NULL; \ - \ - hlist_for_each_entry(ptr, &hash[ix], node.hlist) \ - if (ether_addr_equal(ptr->node.addr, mac)) {\ - found = true; \ - break; \ - } \ - if (!found) \ - ptr = NULL; \ - ptr; \ -}) - -#define l2addr_hash_add(hash, mac, type, gfp) ({ \ - int ix = MLX5_L2_ADDR_HASH(mac); \ - type *ptr = NULL; \ - \ - ptr = kzalloc(sizeof(type), gfp); \ - if (ptr) { \ - ether_addr_copy(ptr->node.addr, mac); \ - hlist_add_head(&ptr->node.hlist, &hash[ix]);\ - } \ - ptr; \ -}) - -#define l2addr_hash_del(ptr) ({ \ - hlist_del(&ptr->node.hlist); \ - kfree(ptr); \ -}) - struct vport_ingress { struct mlx5_flow_table *acl; struct mlx5_flow_group *allow_untagged_spoofchk_grp; @@ -150,12 +106,6 @@ struct mlx5_vport { u16 enabled_events; }; -struct mlx5_l2_table { - struct hlist_head l2_hash[MLX5_L2_ADDR_HASH_SIZE]; - u32 size; - unsigned long *bitmap; -}; - struct mlx5_eswitch_fdb { void *fdb; union { @@ -222,7 +172,6 @@ struct esw_mc_addr { /* SRIOV only */ struct mlx5_eswitch { struct mlx5_core_dev *dev; - struct mlx5_l2_table l2_table; struct mlx5_eswitch_fdb fdb_table; struct hlist_head mc_table[MLX5_L2_ADDR_HASH_SIZE]; struct workqueue_struct *work_queue; @@ -250,8 +199,6 @@ int esw_offloads_init(struct mlx5_eswitch *esw, int nvports); /* E-Switch API */ int mlx5_eswitch_init(struct mlx5_core_dev *dev); void mlx5_eswitch_cleanup(struct mlx5_eswitch *esw); -void mlx5_eswitch_attach(struct mlx5_eswitch *esw); -void mlx5_eswitch_detach(struct mlx5_eswitch *esw); void mlx5_eswitch_vport_event(struct mlx5_eswitch *esw, struct mlx5_eqe *eqe); int mlx5_eswitch_enable_sriov(struct mlx5_eswitch *esw, int nvfs, int mode); void mlx5_eswitch_disable_sriov(struct mlx5_eswitch *esw); diff --git a/drivers/net/ethernet/mellanox/mlx5/core/lib/mpfs.c b/drivers/net/ethernet/mellanox/mlx5/core/lib/mpfs.c new file mode 100644 index 000000000000..7cb67122e8b5 --- /dev/null +++ b/drivers/net/ethernet/mellanox/mlx5/core/lib/mpfs.c @@ -0,0 +1,201 @@ +/* + * Copyright (c) 2017, Mellanox Technologies. All rights reserved. + * + * This software is available to you under a choice of one of two + * licenses. You may choose to be licensed under the terms of the GNU + * General Public License (GPL) Version 2, available from the file + * COPYING in the main directory of this source tree, or the + * OpenIB.org BSD license below: + * + * Redistribution and use in source and binary forms, with or + * without modification, are permitted provided that the following + * conditions are met: + * + * - Redistributions of source code must retain the above + * copyright notice, this list of conditions and the following + * disclaimer. + * + * - Redistributions in binary form must reproduce the above + * copyright notice, this list of conditions and the following + * disclaimer in the documentation and/or other materials + * provided with the distribution. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, + * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF + * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND + * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS + * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN + * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN + * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE + * SOFTWARE. + */ + +#include +#include +#include +#include "mlx5_core.h" +#include "lib/mpfs.h" + +/* HW L2 Table (MPFS) management */ +static int set_l2table_entry_cmd(struct mlx5_core_dev *dev, u32 index, u8 *mac) +{ + u32 in[MLX5_ST_SZ_DW(set_l2_table_entry_in)] = {0}; + u32 out[MLX5_ST_SZ_DW(set_l2_table_entry_out)] = {0}; + u8 *in_mac_addr; + + MLX5_SET(set_l2_table_entry_in, in, opcode, MLX5_CMD_OP_SET_L2_TABLE_ENTRY); + MLX5_SET(set_l2_table_entry_in, in, table_index, index); + + in_mac_addr = MLX5_ADDR_OF(set_l2_table_entry_in, in, mac_address); + ether_addr_copy(&in_mac_addr[2], mac); + + return mlx5_cmd_exec(dev, in, sizeof(in), out, sizeof(out)); +} + +static int del_l2table_entry_cmd(struct mlx5_core_dev *dev, u32 index) +{ + u32 in[MLX5_ST_SZ_DW(delete_l2_table_entry_in)] = {0}; + u32 out[MLX5_ST_SZ_DW(delete_l2_table_entry_out)] = {0}; + + MLX5_SET(delete_l2_table_entry_in, in, opcode, MLX5_CMD_OP_DELETE_L2_TABLE_ENTRY); + MLX5_SET(delete_l2_table_entry_in, in, table_index, index); + return mlx5_cmd_exec(dev, in, sizeof(in), out, sizeof(out)); +} + +/* UC L2 table hash node */ +struct l2table_node { + struct l2addr_node node; + u32 index; /* index in HW l2 table */ +}; + +struct mlx5_mpfs { + struct hlist_head hash[MLX5_L2_ADDR_HASH_SIZE]; + struct mutex lock; /* Synchronize l2 table access */ + u32 size; + unsigned long *bitmap; +}; + +static int alloc_l2table_index(struct mlx5_mpfs *l2table, u32 *ix) +{ + int err = 0; + + *ix = find_first_zero_bit(l2table->bitmap, l2table->size); + if (*ix >= l2table->size) + err = -ENOSPC; + else + __set_bit(*ix, l2table->bitmap); + + return err; +} + +static void free_l2table_index(struct mlx5_mpfs *l2table, u32 ix) +{ + __clear_bit(ix, l2table->bitmap); +} + +int mlx5_mpfs_init(struct mlx5_core_dev *dev) +{ + int l2table_size = 1 << MLX5_CAP_GEN(dev, log_max_l2_table); + struct mlx5_mpfs *mpfs; + + if (!MLX5_VPORT_MANAGER(dev)) + return 0; + + mpfs = kzalloc(sizeof(*mpfs), GFP_KERNEL); + if (!mpfs) + return -ENOMEM; + + mutex_init(&mpfs->lock); + mpfs->size = l2table_size; + mpfs->bitmap = kcalloc(BITS_TO_LONGS(l2table_size), + sizeof(uintptr_t), GFP_KERNEL); + if (!mpfs->bitmap) { + kfree(mpfs); + return -ENOMEM; + } + + dev->priv.mpfs = mpfs; + return 0; +} + +void mlx5_mpfs_cleanup(struct mlx5_core_dev *dev) +{ + struct mlx5_mpfs *mpfs = dev->priv.mpfs; + + if (!MLX5_VPORT_MANAGER(dev)) + return; + + WARN_ON(!hlist_empty(mpfs->hash)); + kfree(mpfs->bitmap); + kfree(mpfs); +} + +int mlx5_mpfs_add_mac(struct mlx5_core_dev *dev, u8 *mac) +{ + struct mlx5_mpfs *mpfs = dev->priv.mpfs; + struct l2table_node *l2addr; + u32 index; + int err; + + if (!MLX5_VPORT_MANAGER(dev)) + return 0; + + mutex_lock(&mpfs->lock); + + l2addr = l2addr_hash_find(mpfs->hash, mac, struct l2table_node); + if (l2addr) { + err = -EEXIST; + goto abort; + } + + err = alloc_l2table_index(mpfs, &index); + if (err) + goto abort; + + l2addr = l2addr_hash_add(mpfs->hash, mac, struct l2table_node, GFP_KERNEL); + if (!l2addr) { + free_l2table_index(mpfs, index); + err = -ENOMEM; + goto abort; + } + + l2addr->index = index; + err = set_l2table_entry_cmd(dev, index, mac); + if (err) { + l2addr_hash_del(l2addr); + free_l2table_index(mpfs, index); + } + + mlx5_core_dbg(dev, "MPFS mac added %pM, index (%d)\n", mac, index); +abort: + mutex_unlock(&mpfs->lock); + return err; +} + +int mlx5_mpfs_del_mac(struct mlx5_core_dev *dev, u8 *mac) +{ + struct mlx5_mpfs *mpfs = dev->priv.mpfs; + struct l2table_node *l2addr; + int err = 0; + u32 index; + + if (!MLX5_VPORT_MANAGER(dev)) + return 0; + + mutex_lock(&mpfs->lock); + + l2addr = l2addr_hash_find(mpfs->hash, mac, struct l2table_node); + if (!l2addr) { + err = -ENOENT; + goto unlock; + } + + index = l2addr->index; + del_l2table_entry_cmd(dev, index); + l2addr_hash_del(l2addr); + free_l2table_index(mpfs, index); + mlx5_core_dbg(dev, "MPFS mac deleted %pM, index (%d)\n", mac, index); +unlock: + mutex_unlock(&mpfs->lock); + return err; +} diff --git a/drivers/net/ethernet/mellanox/mlx5/core/lib/mpfs.h b/drivers/net/ethernet/mellanox/mlx5/core/lib/mpfs.h new file mode 100644 index 000000000000..4a7b2c3203a7 --- /dev/null +++ b/drivers/net/ethernet/mellanox/mlx5/core/lib/mpfs.h @@ -0,0 +1,95 @@ +/* + * Copyright (c) 2017, Mellanox Technologies, Ltd. All rights reserved. + * + * This software is available to you under a choice of one of two + * licenses. You may choose to be licensed under the terms of the GNU + * General Public License (GPL) Version 2, available from the file + * COPYING in the main directory of this source tree, or the + * OpenIB.org BSD license below: + * + * Redistribution and use in source and binary forms, with or + * without modification, are permitted provided that the following + * conditions are met: + * + * - Redistributions of source code must retain the above + * copyright notice, this list of conditions and the following + * disclaimer. + * + * - Redistributions in binary form must reproduce the above + * copyright notice, this list of conditions and the following + * disclaimer in the documentation and/or other materials + * provided with the distribution. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, + * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF + * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND + * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS + * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN + * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN + * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE + * SOFTWARE. + */ + +#ifndef __MLX5_MPFS_H__ +#define __MLX5_MPFS_H__ + +#include +#include + +/* L2 -mac address based- hash helpers */ +#define MLX5_L2_ADDR_HASH_SIZE (BIT(BITS_PER_BYTE)) +#define MLX5_L2_ADDR_HASH(addr) (addr[5]) + +struct l2addr_node { + struct hlist_node hlist; + u8 addr[ETH_ALEN]; +}; + +#define for_each_l2hash_node(hn, tmp, hash, i) \ + for (i = 0; i < MLX5_L2_ADDR_HASH_SIZE; i++) \ + hlist_for_each_entry_safe(hn, tmp, &(hash)[i], hlist) + +#define l2addr_hash_find(hash, mac, type) ({ \ + int ix = MLX5_L2_ADDR_HASH(mac); \ + bool found = false; \ + type *ptr = NULL; \ + \ + hlist_for_each_entry(ptr, &(hash)[ix], node.hlist) \ + if (ether_addr_equal(ptr->node.addr, mac)) {\ + found = true; \ + break; \ + } \ + if (!found) \ + ptr = NULL; \ + ptr; \ +}) + +#define l2addr_hash_add(hash, mac, type, gfp) ({ \ + int ix = MLX5_L2_ADDR_HASH(mac); \ + type *ptr = NULL; \ + \ + ptr = kzalloc(sizeof(type), gfp); \ + if (ptr) { \ + ether_addr_copy(ptr->node.addr, mac); \ + hlist_add_head(&ptr->node.hlist, &(hash)[ix]);\ + } \ + ptr; \ +}) + +#define l2addr_hash_del(ptr) ({ \ + hlist_del(&(ptr)->node.hlist); \ + kfree(ptr); \ +}) + +#ifdef CONFIG_MLX5_MPFS +int mlx5_mpfs_init(struct mlx5_core_dev *dev); +void mlx5_mpfs_cleanup(struct mlx5_core_dev *dev); +int mlx5_mpfs_add_mac(struct mlx5_core_dev *dev, u8 *mac); +int mlx5_mpfs_del_mac(struct mlx5_core_dev *dev, u8 *mac); +#else /* #ifndef CONFIG_MLX5_MPFS */ +static inline int mlx5_mpfs_init(struct mlx5_core_dev *dev) { return 0; } +static inline void mlx5_mpfs_cleanup(struct mlx5_core_dev *dev) {} +static inline int mlx5_mpfs_add_mac(struct mlx5_core_dev *dev, u8 *mac) { return 0; } +static inline int mlx5_mpfs_del_mac(struct mlx5_core_dev *dev, u8 *mac) { return 0; } +#endif +#endif diff --git a/drivers/net/ethernet/mellanox/mlx5/core/main.c b/drivers/net/ethernet/mellanox/mlx5/core/main.c index c065132b956d..d4a9c9b7b6a2 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/main.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/main.c @@ -53,6 +53,7 @@ #include #include "mlx5_core.h" #include "fs_core.h" +#include "lib/mpfs.h" #ifdef CONFIG_MLX5_CORE_EN #include "eswitch.h" #endif @@ -946,11 +947,17 @@ static int mlx5_init_once(struct mlx5_core_dev *dev, struct mlx5_priv *priv) goto err_tables_cleanup; } + err = mlx5_mpfs_init(dev); + if (err) { + dev_err(&pdev->dev, "Failed to init l2 table %d\n", err); + goto err_rl_cleanup; + } + #ifdef CONFIG_MLX5_CORE_EN err = mlx5_eswitch_init(dev); if (err) { dev_err(&pdev->dev, "Failed to init eswitch %d\n", err); - goto err_rl_cleanup; + goto err_mpfs_cleanup; } #endif @@ -973,11 +980,11 @@ err_sriov_cleanup: err_eswitch_cleanup: #ifdef CONFIG_MLX5_CORE_EN mlx5_eswitch_cleanup(dev->priv.eswitch); - -err_rl_cleanup: +err_mpfs_cleanup: #endif + mlx5_mpfs_cleanup(dev); +err_rl_cleanup: mlx5_cleanup_rl_table(dev); - err_tables_cleanup: mlx5_cleanup_mkey_table(dev); mlx5_cleanup_srq_table(dev); @@ -998,6 +1005,7 @@ static void mlx5_cleanup_once(struct mlx5_core_dev *dev) #ifdef CONFIG_MLX5_CORE_EN mlx5_eswitch_cleanup(dev->priv.eswitch); #endif + mlx5_mpfs_cleanup(dev); mlx5_cleanup_rl_table(dev); mlx5_cleanup_reserved_gids(dev); mlx5_cleanup_mkey_table(dev); @@ -1155,10 +1163,6 @@ static int mlx5_load_one(struct mlx5_core_dev *dev, struct mlx5_priv *priv, goto err_fs; } -#ifdef CONFIG_MLX5_CORE_EN - mlx5_eswitch_attach(dev->priv.eswitch); -#endif - err = mlx5_sriov_attach(dev); if (err) { dev_err(&pdev->dev, "sriov init failed %d\n", err); @@ -1202,9 +1206,6 @@ err_fpga_start: mlx5_sriov_detach(dev); err_sriov: -#ifdef CONFIG_MLX5_CORE_EN - mlx5_eswitch_detach(dev->priv.eswitch); -#endif mlx5_cleanup_fs(dev); err_fs: @@ -1279,9 +1280,6 @@ static int mlx5_unload_one(struct mlx5_core_dev *dev, struct mlx5_priv *priv, mlx5_fpga_device_stop(dev); mlx5_sriov_detach(dev); -#ifdef CONFIG_MLX5_CORE_EN - mlx5_eswitch_detach(dev->priv.eswitch); -#endif mlx5_cleanup_fs(dev); mlx5_irq_clear_affinity_hints(dev); free_comp_eqs(dev); diff --git a/include/linux/mlx5/driver.h b/include/linux/mlx5/driver.h index df6ce59a1f95..88d6eb5b3a76 100644 --- a/include/linux/mlx5/driver.h +++ b/include/linux/mlx5/driver.h @@ -550,6 +550,7 @@ struct mlx5_fc_stats { unsigned long sampling_interval; /* jiffies */ }; +struct mlx5_mpfs; struct mlx5_eswitch; struct mlx5_lag; struct mlx5_pagefault; @@ -647,6 +648,7 @@ struct mlx5_priv { spinlock_t ctx_lock; struct mlx5_flow_steering *steering; + struct mlx5_mpfs *mpfs; struct mlx5_eswitch *eswitch; struct mlx5_core_sriov sriov; struct mlx5_lag *lag; -- cgit v1.2.3 From 97834eba7c194659a72c5bb0f8c19c7055bb69ea Mon Sep 17 00:00:00 2001 From: Erez Shitrit Date: Wed, 7 Jun 2017 12:14:24 +0300 Subject: net/mlx5: Delay events till ib registration ends When mlx5_ib registers itself to mlx5_core as an interface, it will call mlx5_add_device which will call mlx5_ib interface add callback, in case the latter successfully returns, only then mlx5_core will add it to the interface list and async events will be forwarded to mlx5_ib. Between mlx5_ib interface add callback and mlx5_core adding the mlx5_ib interface to its devices list, arriving mlx5_core events can be missed by the new mlx5_ib registering interface. In other words: thread 1: mlx5_ib: mlx5_register_interface(dev) thread 1: mlx5_core: mlx5_add_device(dev) thread 1: mlx5_core: ctx = dev->add => (mlx5_ib)->mlx5_ib_add thread 2: mlx5_core_event: **new event arrives, forward to dev_list thread 1: mlx5_core: add_ctx_to_dev_list(ctx) /* previous event was missed by the new interface.*/ It is ok to miss events before dev->add (mlx5_ib)->mlx5_ib_add_device but not after. We fix this race by accumulating the events that come between the ib_register_device (inside mlx5_add_device->(dev->add)) till the adding to the list completes and fire them to the new registering interface after that. Fixes: f1ee87fe55c8 ("net/mlx5: Organize device list API in one place") Signed-off-by: Erez Shitrit Signed-off-by: Saeed Mahameed --- drivers/net/ethernet/mellanox/mlx5/core/dev.c | 73 ++++++++++++++++++++++++++ drivers/net/ethernet/mellanox/mlx5/core/main.c | 3 ++ include/linux/mlx5/driver.h | 3 ++ 3 files changed, 79 insertions(+) (limited to 'include') diff --git a/drivers/net/ethernet/mellanox/mlx5/core/dev.c b/drivers/net/ethernet/mellanox/mlx5/core/dev.c index a62f4b6a21a5..ff60cf7342ca 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/dev.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/dev.c @@ -45,11 +45,70 @@ struct mlx5_device_context { unsigned long state; }; +struct mlx5_delayed_event { + struct list_head list; + struct mlx5_core_dev *dev; + enum mlx5_dev_event event; + unsigned long param; +}; + enum { MLX5_INTERFACE_ADDED, MLX5_INTERFACE_ATTACHED, }; +static void add_delayed_event(struct mlx5_priv *priv, + struct mlx5_core_dev *dev, + enum mlx5_dev_event event, + unsigned long param) +{ + struct mlx5_delayed_event *delayed_event; + + delayed_event = kzalloc(sizeof(*delayed_event), GFP_ATOMIC); + if (!delayed_event) { + mlx5_core_err(dev, "event %d is missed\n", event); + return; + } + + mlx5_core_dbg(dev, "Accumulating event %d\n", event); + delayed_event->dev = dev; + delayed_event->event = event; + delayed_event->param = param; + list_add_tail(&delayed_event->list, &priv->waiting_events_list); +} + +static void fire_delayed_event_locked(struct mlx5_device_context *dev_ctx, + struct mlx5_core_dev *dev, + struct mlx5_priv *priv) +{ + struct mlx5_delayed_event *de; + struct mlx5_delayed_event *n; + + /* stop delaying events */ + priv->is_accum_events = false; + + /* fire all accumulated events before new event comes */ + list_for_each_entry_safe(de, n, &priv->waiting_events_list, list) { + dev_ctx->intf->event(dev, dev_ctx->context, de->event, de->param); + list_del(&de->list); + kfree(de); + } +} + +static void cleanup_delayed_evets(struct mlx5_priv *priv) +{ + struct mlx5_delayed_event *de; + struct mlx5_delayed_event *n; + + spin_lock_irq(&priv->ctx_lock); + priv->is_accum_events = false; + list_for_each_entry_safe(de, n, &priv->waiting_events_list, list) { + list_del(&de->list); + kfree(de); + } + spin_unlock_irq(&priv->ctx_lock); +} + void mlx5_add_device(struct mlx5_interface *intf, struct mlx5_priv *priv) { struct mlx5_device_context *dev_ctx; @@ -63,6 +122,12 @@ void mlx5_add_device(struct mlx5_interface *intf, struct mlx5_priv *priv) return; dev_ctx->intf = intf; + /* accumulating events that can come after mlx5_ib calls to + * ib_register_device, till adding that interface to the events list. + */ + + priv->is_accum_events = true; + dev_ctx->context = intf->add(dev); set_bit(MLX5_INTERFACE_ADDED, &dev_ctx->state); if (intf->attach) @@ -71,6 +136,9 @@ void mlx5_add_device(struct mlx5_interface *intf, struct mlx5_priv *priv) if (dev_ctx->context) { spin_lock_irq(&priv->ctx_lock); list_add_tail(&dev_ctx->list, &priv->ctx_list); + + fire_delayed_event_locked(dev_ctx, dev, priv); + #ifdef CONFIG_INFINIBAND_ON_DEMAND_PAGING if (dev_ctx->intf->pfault) { if (priv->pfault) { @@ -84,6 +152,8 @@ void mlx5_add_device(struct mlx5_interface *intf, struct mlx5_priv *priv) spin_unlock_irq(&priv->ctx_lock); } else { kfree(dev_ctx); + /* delete all accumulated events */ + cleanup_delayed_evets(priv); } } @@ -341,6 +411,9 @@ void mlx5_core_event(struct mlx5_core_dev *dev, enum mlx5_dev_event event, spin_lock_irqsave(&priv->ctx_lock, flags); + if (priv->is_accum_events) + add_delayed_event(priv, dev, event, param); + list_for_each_entry(dev_ctx, &priv->ctx_list, list) if (dev_ctx->intf->event) dev_ctx->intf->event(dev, dev_ctx->context, event, param); diff --git a/drivers/net/ethernet/mellanox/mlx5/core/main.c b/drivers/net/ethernet/mellanox/mlx5/core/main.c index 124c7c3c3a00..6dbd637b4e66 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/main.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/main.c @@ -1343,6 +1343,9 @@ static int init_one(struct pci_dev *pdev, mutex_init(&dev->pci_status_mutex); mutex_init(&dev->intf_state_mutex); + INIT_LIST_HEAD(&priv->waiting_events_list); + priv->is_accum_events = false; + #ifdef CONFIG_INFINIBAND_ON_DEMAND_PAGING err = init_srcu_struct(&priv->pfault_srcu); if (err) { diff --git a/include/linux/mlx5/driver.h b/include/linux/mlx5/driver.h index 88d6eb5b3a76..d26f18b39c4a 100644 --- a/include/linux/mlx5/driver.h +++ b/include/linux/mlx5/driver.h @@ -647,6 +647,9 @@ struct mlx5_priv { struct list_head ctx_list; spinlock_t ctx_lock; + struct list_head waiting_events_list; + bool is_accum_events; + struct mlx5_flow_steering *steering; struct mlx5_mpfs *mpfs; struct mlx5_eswitch *eswitch; -- cgit v1.2.3 From 61690e09c3b4a40401bdaa89ee11522cc0dc4b11 Mon Sep 17 00:00:00 2001 From: Rabie Loulou Date: Mon, 10 Jul 2017 14:35:10 +0300 Subject: net/mlx5: Fix counter list hardware structure The counter list hardware structure doesn't contain a clear and num_of_counters fields, remove them. These wrong fields were never used by the driver hence no other driver changes. Fixes: a351a1b03bf1 ("net/mlx5: Introduce bulk reading of flow counters") Signed-off-by: Rabie Loulou Reviewed-by: Or Gerlitz Signed-off-by: Saeed Mahameed --- include/linux/mlx5/mlx5_ifc.h | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-) (limited to 'include') diff --git a/include/linux/mlx5/mlx5_ifc.h b/include/linux/mlx5/mlx5_ifc.h index 3030121b4746..f847a3a57913 100644 --- a/include/linux/mlx5/mlx5_ifc.h +++ b/include/linux/mlx5/mlx5_ifc.h @@ -1071,8 +1071,7 @@ struct mlx5_ifc_dest_format_struct_bits { }; struct mlx5_ifc_flow_counter_list_bits { - u8 clear[0x1]; - u8 num_of_counters[0xf]; + u8 reserved_at_0[0x10]; u8 flow_counter_id[0x10]; u8 reserved_at_20[0x20]; -- cgit v1.2.3 From a8ffcc741acb3c7f3dcf4c7d001209aa0995a5f1 Mon Sep 17 00:00:00 2001 From: Rabie Loulou Date: Sun, 9 Jul 2017 13:39:30 +0300 Subject: net/mlx5: Increase the maximum flow counters supported Read new NIC capability field which represnts 16 MSBs of the max flow counters number supported (max_flow_counter_31_16). Backward compatibility with older firmware is preserved, the modified driver reads max_flow_counter_31_16 as 0 from the older firmware and uses up to 64K counters. Changed flow counter id from 16 bits to 32 bits. Backward compatibility with older firmware is preserved as we kept the 16 LSBs of the counter id in place and added 16 MSBs from reserved field. Changed the background bulk reading of flow counters to work in chunks of at most 32K counters, to make sure we don't attempt to allocate very large buffers. Signed-off-by: Rabie Loulou Reviewed-by: Or Gerlitz Signed-off-by: Saeed Mahameed --- .../net/ethernet/mellanox/mlx5/core/eswitch_offloads.c | 6 ++++-- drivers/net/ethernet/mellanox/mlx5/core/fs_cmd.c | 12 ++++++------ drivers/net/ethernet/mellanox/mlx5/core/fs_cmd.h | 10 +++++----- drivers/net/ethernet/mellanox/mlx5/core/fs_core.h | 2 +- drivers/net/ethernet/mellanox/mlx5/core/fs_counters.c | 13 ++++++++++--- include/linux/mlx5/mlx5_ifc.h | 16 ++++++---------- 6 files changed, 32 insertions(+), 27 deletions(-) (limited to 'include') diff --git a/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c b/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c index 95b64025ce36..e7c186b58579 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c @@ -433,6 +433,8 @@ static int esw_create_offloads_fast_fdb_table(struct mlx5_eswitch *esw) struct mlx5_flow_table *fdb = NULL; int esw_size, err = 0; u32 flags = 0; + u32 max_flow_counter = (MLX5_CAP_GEN(dev, max_flow_counter_31_16) << 16) | + MLX5_CAP_GEN(dev, max_flow_counter_15_0); root_ns = mlx5_get_flow_namespace(dev, MLX5_FLOW_NAMESPACE_FDB); if (!root_ns) { @@ -443,9 +445,9 @@ static int esw_create_offloads_fast_fdb_table(struct mlx5_eswitch *esw) esw_debug(dev, "Create offloads FDB table, min (max esw size(2^%d), max counters(%d)*groups(%d))\n", MLX5_CAP_ESW_FLOWTABLE_FDB(dev, log_max_ft_size), - MLX5_CAP_GEN(dev, max_flow_counter), ESW_OFFLOADS_NUM_GROUPS); + max_flow_counter, ESW_OFFLOADS_NUM_GROUPS); - esw_size = min_t(int, MLX5_CAP_GEN(dev, max_flow_counter) * ESW_OFFLOADS_NUM_GROUPS, + esw_size = min_t(int, max_flow_counter * ESW_OFFLOADS_NUM_GROUPS, 1 << MLX5_CAP_ESW_FLOWTABLE_FDB(dev, log_max_ft_size)); if (esw->offloads.encap != DEVLINK_ESWITCH_ENCAP_MODE_NONE) diff --git a/drivers/net/ethernet/mellanox/mlx5/core/fs_cmd.c b/drivers/net/ethernet/mellanox/mlx5/core/fs_cmd.c index e750f07793b8..16b32f31d691 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/fs_cmd.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/fs_cmd.c @@ -359,7 +359,7 @@ int mlx5_cmd_delete_fte(struct mlx5_core_dev *dev, return mlx5_cmd_exec(dev, in, sizeof(in), out, sizeof(out)); } -int mlx5_cmd_fc_alloc(struct mlx5_core_dev *dev, u16 *id) +int mlx5_cmd_fc_alloc(struct mlx5_core_dev *dev, u32 *id) { u32 in[MLX5_ST_SZ_DW(alloc_flow_counter_in)] = {0}; u32 out[MLX5_ST_SZ_DW(alloc_flow_counter_out)] = {0}; @@ -374,7 +374,7 @@ int mlx5_cmd_fc_alloc(struct mlx5_core_dev *dev, u16 *id) return err; } -int mlx5_cmd_fc_free(struct mlx5_core_dev *dev, u16 id) +int mlx5_cmd_fc_free(struct mlx5_core_dev *dev, u32 id) { u32 in[MLX5_ST_SZ_DW(dealloc_flow_counter_in)] = {0}; u32 out[MLX5_ST_SZ_DW(dealloc_flow_counter_out)] = {0}; @@ -385,7 +385,7 @@ int mlx5_cmd_fc_free(struct mlx5_core_dev *dev, u16 id) return mlx5_cmd_exec(dev, in, sizeof(in), out, sizeof(out)); } -int mlx5_cmd_fc_query(struct mlx5_core_dev *dev, u16 id, +int mlx5_cmd_fc_query(struct mlx5_core_dev *dev, u32 id, u64 *packets, u64 *bytes) { u32 out[MLX5_ST_SZ_BYTES(query_flow_counter_out) + @@ -409,14 +409,14 @@ int mlx5_cmd_fc_query(struct mlx5_core_dev *dev, u16 id, } struct mlx5_cmd_fc_bulk { - u16 id; + u32 id; int num; int outlen; u32 out[0]; }; struct mlx5_cmd_fc_bulk * -mlx5_cmd_fc_bulk_alloc(struct mlx5_core_dev *dev, u16 id, int num) +mlx5_cmd_fc_bulk_alloc(struct mlx5_core_dev *dev, u32 id, int num) { struct mlx5_cmd_fc_bulk *b; int outlen = @@ -453,7 +453,7 @@ mlx5_cmd_fc_bulk_query(struct mlx5_core_dev *dev, struct mlx5_cmd_fc_bulk *b) } void mlx5_cmd_fc_bulk_get(struct mlx5_core_dev *dev, - struct mlx5_cmd_fc_bulk *b, u16 id, + struct mlx5_cmd_fc_bulk *b, u32 id, u64 *packets, u64 *bytes) { int index = id - b->id; diff --git a/drivers/net/ethernet/mellanox/mlx5/core/fs_cmd.h b/drivers/net/ethernet/mellanox/mlx5/core/fs_cmd.h index 0f98a7cf4877..c6d7bdf255b6 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/fs_cmd.h +++ b/drivers/net/ethernet/mellanox/mlx5/core/fs_cmd.h @@ -74,20 +74,20 @@ int mlx5_cmd_update_root_ft(struct mlx5_core_dev *dev, struct mlx5_flow_table *ft, u32 underlay_qpn); -int mlx5_cmd_fc_alloc(struct mlx5_core_dev *dev, u16 *id); -int mlx5_cmd_fc_free(struct mlx5_core_dev *dev, u16 id); -int mlx5_cmd_fc_query(struct mlx5_core_dev *dev, u16 id, +int mlx5_cmd_fc_alloc(struct mlx5_core_dev *dev, u32 *id); +int mlx5_cmd_fc_free(struct mlx5_core_dev *dev, u32 id); +int mlx5_cmd_fc_query(struct mlx5_core_dev *dev, u32 id, u64 *packets, u64 *bytes); struct mlx5_cmd_fc_bulk; struct mlx5_cmd_fc_bulk * -mlx5_cmd_fc_bulk_alloc(struct mlx5_core_dev *dev, u16 id, int num); +mlx5_cmd_fc_bulk_alloc(struct mlx5_core_dev *dev, u32 id, int num); void mlx5_cmd_fc_bulk_free(struct mlx5_cmd_fc_bulk *b); int mlx5_cmd_fc_bulk_query(struct mlx5_core_dev *dev, struct mlx5_cmd_fc_bulk *b); void mlx5_cmd_fc_bulk_get(struct mlx5_core_dev *dev, - struct mlx5_cmd_fc_bulk *b, u16 id, + struct mlx5_cmd_fc_bulk *b, u32 id, u64 *packets, u64 *bytes); #endif diff --git a/drivers/net/ethernet/mellanox/mlx5/core/fs_core.h b/drivers/net/ethernet/mellanox/mlx5/core/fs_core.h index 990acee6fb09..9fb5a333df52 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/fs_core.h +++ b/drivers/net/ethernet/mellanox/mlx5/core/fs_core.h @@ -136,7 +136,7 @@ struct mlx5_fc { u64 lastpackets; u64 lastbytes; - u16 id; + u32 id; bool deleted; bool aging; diff --git a/drivers/net/ethernet/mellanox/mlx5/core/fs_counters.c b/drivers/net/ethernet/mellanox/mlx5/core/fs_counters.c index 6507d8acc54d..89d1f8650033 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/fs_counters.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/fs_counters.c @@ -38,6 +38,8 @@ #include "fs_cmd.h" #define MLX5_FC_STATS_PERIOD msecs_to_jiffies(1000) +/* Max number of counters to query in bulk read is 32K */ +#define MLX5_SW_MAX_COUNTERS_BULK BIT(15) /* locking scheme: * @@ -90,16 +92,21 @@ static void mlx5_fc_stats_insert(struct rb_root *root, struct mlx5_fc *counter) rb_insert_color(&counter->node, root); } +/* The function returns the last node that was queried so the caller + * function can continue calling it till all counters are queried. + */ static struct rb_node *mlx5_fc_stats_query(struct mlx5_core_dev *dev, struct mlx5_fc *first, - u16 last_id) + u32 last_id) { struct mlx5_cmd_fc_bulk *b; struct rb_node *node = NULL; - u16 afirst_id; + u32 afirst_id; int num; int err; - int max_bulk = 1 << MLX5_CAP_GEN(dev, log_max_flow_counter_bulk); + + int max_bulk = min_t(int, MLX5_SW_MAX_COUNTERS_BULK, + (1 << MLX5_CAP_GEN(dev, log_max_flow_counter_bulk))); /* first id must be aligned to 4 when using bulk query */ afirst_id = first->id & ~0x3; diff --git a/include/linux/mlx5/mlx5_ifc.h b/include/linux/mlx5/mlx5_ifc.h index f847a3a57913..c99daffc3c3c 100644 --- a/include/linux/mlx5/mlx5_ifc.h +++ b/include/linux/mlx5/mlx5_ifc.h @@ -963,7 +963,7 @@ struct mlx5_ifc_cmd_hca_cap_bits { u8 reserved_at_2a0[0x10]; u8 max_wqe_sz_rq[0x10]; - u8 reserved_at_2c0[0x10]; + u8 max_flow_counter_31_16[0x10]; u8 max_wqe_sz_sq_dc[0x10]; u8 reserved_at_2e0[0x7]; @@ -981,7 +981,7 @@ struct mlx5_ifc_cmd_hca_cap_bits { u8 reserved_at_340[0x8]; u8 log_max_flow_counter_bulk[0x8]; - u8 max_flow_counter[0x10]; + u8 max_flow_counter_15_0[0x10]; u8 reserved_at_360[0x3]; @@ -1071,8 +1071,7 @@ struct mlx5_ifc_dest_format_struct_bits { }; struct mlx5_ifc_flow_counter_list_bits { - u8 reserved_at_0[0x10]; - u8 flow_counter_id[0x10]; + u8 flow_counter_id[0x20]; u8 reserved_at_20[0x20]; }; @@ -4402,8 +4401,7 @@ struct mlx5_ifc_query_flow_counter_in_bits { u8 reserved_at_c1[0xf]; u8 num_of_counters[0x10]; - u8 reserved_at_e0[0x10]; - u8 flow_counter_id[0x10]; + u8 flow_counter_id[0x20]; }; struct mlx5_ifc_query_esw_vport_context_out_bits { @@ -6271,8 +6269,7 @@ struct mlx5_ifc_dealloc_flow_counter_in_bits { u8 reserved_at_20[0x10]; u8 op_mod[0x10]; - u8 reserved_at_40[0x10]; - u8 flow_counter_id[0x10]; + u8 flow_counter_id[0x20]; u8 reserved_at_60[0x20]; }; @@ -7097,8 +7094,7 @@ struct mlx5_ifc_alloc_flow_counter_out_bits { u8 syndrome[0x20]; - u8 reserved_at_40[0x10]; - u8 flow_counter_id[0x10]; + u8 flow_counter_id[0x20]; u8 reserved_at_60[0x20]; }; -- cgit v1.2.3