From 68deb9972079c9904fe714c049a7f08bd997a9ee Mon Sep 17 00:00:00 2001 From: Dan Williams Date: Tue, 16 Jan 2024 13:17:17 -0800 Subject: [PATCH 1/4] tools/testing/cxl: Disable "missing prototypes / declarations" warnings MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Prevent warnings of the form: tools/testing/cxl/test/mock.c:44:6: error: no previous prototype for ‘__wrap_is_acpi_device_node’ [-Werror=missing-prototypes] tools/testing/cxl/test/mock.c:63:5: error: no previous prototype for ‘__wrap_acpi_table_parse_cedt’ [-Werror=missing-prototypes] tools/testing/cxl/test/mock.c:81:13: error: no previous prototype for ‘__wrap_acpi_evaluate_integer’ [-Werror=missing-prototypes] ...by locally disabling some warnings. It turns out that: Commit 0fcb70851fbf ("Makefile.extrawarn: turn on missing-prototypes globally") ...in addition to expanding in-tree coverage, also impacts out-of-tree module builds like those in tools/testing/cxl/. Filter out the warning options on unit test code that does not effect mainline builds. Reviewed-by: Alison Schofield Link: https://lore.kernel.org/r/170543983780.460832.10920261849128601697.stgit@dwillia2-xfh.jf.intel.com Signed-off-by: Dan Williams --- tools/testing/cxl/Kbuild | 2 ++ tools/testing/cxl/test/Kbuild | 2 ++ 2 files changed, 4 insertions(+) diff --git a/tools/testing/cxl/Kbuild b/tools/testing/cxl/Kbuild index 0b12c36902d8..caff3834671f 100644 --- a/tools/testing/cxl/Kbuild +++ b/tools/testing/cxl/Kbuild @@ -65,4 +65,6 @@ cxl_core-y += config_check.o cxl_core-y += cxl_core_test.o cxl_core-y += cxl_core_exports.o +KBUILD_CFLAGS := $(filter-out -Wmissing-prototypes -Wmissing-declarations, $(KBUILD_CFLAGS)) + obj-m += test/ diff --git a/tools/testing/cxl/test/Kbuild b/tools/testing/cxl/test/Kbuild index 61d5f7bcddf9..6b1927897856 100644 --- a/tools/testing/cxl/test/Kbuild +++ b/tools/testing/cxl/test/Kbuild @@ -8,3 +8,5 @@ obj-m += cxl_mock_mem.o cxl_test-y := cxl.o cxl_mock-y := mock.o cxl_mock_mem-y := mem.o + +KBUILD_CFLAGS := $(filter-out -Wmissing-prototypes -Wmissing-declarations, $(KBUILD_CFLAGS)) From c97dac57c804b53eab492ca0230d86356729d633 Mon Sep 17 00:00:00 2001 From: Dan Williams Date: Tue, 16 Jan 2024 13:17:23 -0800 Subject: [PATCH 2/4] tools/testing/nvdimm: Disable "missing prototypes / declarations" warnings MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Prevent warnings of the form: tools/testing/nvdimm/config_check.c:4:6: error: no previous prototype for ‘check’ [-Werror=missing-prototypes] ...by locally disabling some warnings. It turns out that: Commit 0fcb70851fbf ("Makefile.extrawarn: turn on missing-prototypes globally") ...in addition to expanding in-tree coverage, also impacts out-of-tree module builds like those in tools/testing/nvdimm/. Filter out the warning options on unit test code that does not effect mainline builds. Reviewed-by: Alison Schofield Link: https://lore.kernel.org/r/170543984331.460832.1780246477583036191.stgit@dwillia2-xfh.jf.intel.com Signed-off-by: Dan Williams --- tools/testing/nvdimm/Kbuild | 2 ++ 1 file changed, 2 insertions(+) diff --git a/tools/testing/nvdimm/Kbuild b/tools/testing/nvdimm/Kbuild index 8153251ea389..91a3627f301a 100644 --- a/tools/testing/nvdimm/Kbuild +++ b/tools/testing/nvdimm/Kbuild @@ -82,4 +82,6 @@ libnvdimm-$(CONFIG_NVDIMM_KEYS) += $(NVDIMM_SRC)/security.o libnvdimm-y += libnvdimm_test.o libnvdimm-y += config_check.o +KBUILD_CFLAGS := $(filter-out -Wmissing-prototypes -Wmissing-declarations, $(KBUILD_CFLAGS)) + obj-m += test/ From d72a4caf685989e353dff0a97d7376ee10edbf87 Mon Sep 17 00:00:00 2001 From: Ira Weiny Date: Wed, 17 Jan 2024 17:24:01 -0800 Subject: [PATCH 3/4] cxl/pci: Skip irq features if MSI/MSI-X are not supported CXL 3.1 Section 3.1.1 states: "A Function on a CXL device must not generate INTx messages if that Function participates in CXL.cache protocol or CXL.mem protocols." The generic CXL memory driver only supports devices which use the CXL.mem protocol. The current driver attempts to allocate MSI/MSI-X vectors in anticipation of their need for mailbox interrupts or event processing. However, the above requirement does not require a device to support interrupts, only that they use MSI/MSI-X. For example, a device may disable mailbox interrupts and either be configured for firmware first or skip event processing and function. Dave Larsen reported that the following Intel / Agilex card does not support interrupts on function 0. CXL: Intel Corporation Device 0ddb (rev 01) (prog-if 10 [CXL Memory Device (CXL 2.x)]) Rather than fail device probe if interrupts are not supported; flag that irqs are not enabled and avoid features which require interrupts. Emit messages appropriate for the situation to aid in debugging should device behavior be unexpected due to a failure to allocate vectors. Note that it is possible for a device to have host based event processing through polling. However, the driver does not support polling and it is not anticipated to be generally required. Leave that functionality to a future patch if such a device comes along. Reported-by: Dave Larsen Reviewed-by: Dave Jiang Reviewed-by: Fan Ni Signed-off-by: Ira Weiny Reviewed-and-tested-by: Davidlohr Bueso Link: https://lore.kernel.org/r/20240117-dont-fail-irq-v2-1-f33f26b0e365@intel.com Signed-off-by: Dan Williams --- drivers/cxl/pci.c | 26 +++++++++++++++----------- 1 file changed, 15 insertions(+), 11 deletions(-) diff --git a/drivers/cxl/pci.c b/drivers/cxl/pci.c index 4fd1f207c84e..233e7c42c161 100644 --- a/drivers/cxl/pci.c +++ b/drivers/cxl/pci.c @@ -382,7 +382,7 @@ static int cxl_pci_mbox_send(struct cxl_memdev_state *mds, return rc; } -static int cxl_pci_setup_mailbox(struct cxl_memdev_state *mds) +static int cxl_pci_setup_mailbox(struct cxl_memdev_state *mds, bool irq_avail) { struct cxl_dev_state *cxlds = &mds->cxlds; const int cap = readl(cxlds->regs.mbox + CXLDEV_MBOX_CAPS_OFFSET); @@ -441,7 +441,7 @@ static int cxl_pci_setup_mailbox(struct cxl_memdev_state *mds) INIT_DELAYED_WORK(&mds->security.poll_dwork, cxl_mbox_sanitize_work); /* background command interrupts are optional */ - if (!(cap & CXLDEV_MBOX_CAP_BG_CMD_IRQ)) + if (!(cap & CXLDEV_MBOX_CAP_BG_CMD_IRQ) || !irq_avail) return 0; msgnum = FIELD_GET(CXLDEV_MBOX_CAP_IRQ_MSGNUM_MASK, cap); @@ -588,7 +588,7 @@ static int cxl_mem_alloc_event_buf(struct cxl_memdev_state *mds) return devm_add_action_or_reset(mds->cxlds.dev, free_event_buf, buf); } -static int cxl_alloc_irq_vectors(struct pci_dev *pdev) +static bool cxl_alloc_irq_vectors(struct pci_dev *pdev) { int nvecs; @@ -605,9 +605,9 @@ static int cxl_alloc_irq_vectors(struct pci_dev *pdev) PCI_IRQ_MSIX | PCI_IRQ_MSI); if (nvecs < 1) { dev_dbg(&pdev->dev, "Failed to alloc irq vectors: %d\n", nvecs); - return -ENXIO; + return false; } - return 0; + return true; } static irqreturn_t cxl_event_thread(int irq, void *id) @@ -743,7 +743,7 @@ static bool cxl_event_int_is_fw(u8 setting) } static int cxl_event_config(struct pci_host_bridge *host_bridge, - struct cxl_memdev_state *mds) + struct cxl_memdev_state *mds, bool irq_avail) { struct cxl_event_interrupt_policy policy; int rc; @@ -755,6 +755,11 @@ static int cxl_event_config(struct pci_host_bridge *host_bridge, if (!host_bridge->native_cxl_error) return 0; + if (!irq_avail) { + dev_info(mds->cxlds.dev, "No interrupt support, disable event processing.\n"); + return 0; + } + rc = cxl_mem_alloc_event_buf(mds); if (rc) return rc; @@ -789,6 +794,7 @@ static int cxl_pci_probe(struct pci_dev *pdev, const struct pci_device_id *id) struct cxl_register_map map; struct cxl_memdev *cxlmd; int i, rc, pmu_count; + bool irq_avail; /* * Double check the anonymous union trickery in struct cxl_regs @@ -846,11 +852,9 @@ static int cxl_pci_probe(struct pci_dev *pdev, const struct pci_device_id *id) else dev_warn(&pdev->dev, "Media not active (%d)\n", rc); - rc = cxl_alloc_irq_vectors(pdev); - if (rc) - return rc; + irq_avail = cxl_alloc_irq_vectors(pdev); - rc = cxl_pci_setup_mailbox(mds); + rc = cxl_pci_setup_mailbox(mds, irq_avail); if (rc) return rc; @@ -909,7 +913,7 @@ static int cxl_pci_probe(struct pci_dev *pdev, const struct pci_device_id *id) } } - rc = cxl_event_config(host_bridge, mds); + rc = cxl_event_config(host_bridge, mds, irq_avail); if (rc) return rc; From d76779dd3681c01a4c6c3cae4d0627c9083e0ee6 Mon Sep 17 00:00:00 2001 From: Quanquan Cao Date: Wed, 24 Jan 2024 17:15:26 +0800 Subject: [PATCH 4/4] =?UTF-8?q?cxl/region=EF=BC=9AFix=20overflow=20issue?= =?UTF-8?q?=20in=20alloc=5Fhpa()?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Creating a region with 16 memory devices caused a problem. The div_u64_rem function, used for dividing an unsigned 64-bit number by a 32-bit one, faced an issue when SZ_256M * p->interleave_ways. The result surpassed the maximum limit of the 32-bit divisor (4G), leading to an overflow and a remainder of 0. note: At this point, p->interleave_ways is 16, meaning 16 * 256M = 4G To fix this issue, I replaced the div_u64_rem function with div64_u64_rem and adjusted the type of the remainder. Signed-off-by: Quanquan Cao Reviewed-by: Dave Jiang Fixes: 23a22cd1c98b ("cxl/region: Allocate HPA capacity to regions") Cc: Signed-off-by: Dan Williams --- drivers/cxl/core/region.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/drivers/cxl/core/region.c b/drivers/cxl/core/region.c index 0f05692bfec3..ce0e2d82bb2b 100644 --- a/drivers/cxl/core/region.c +++ b/drivers/cxl/core/region.c @@ -525,7 +525,7 @@ static int alloc_hpa(struct cxl_region *cxlr, resource_size_t size) struct cxl_root_decoder *cxlrd = to_cxl_root_decoder(cxlr->dev.parent); struct cxl_region_params *p = &cxlr->params; struct resource *res; - u32 remainder = 0; + u64 remainder = 0; lockdep_assert_held_write(&cxl_region_rwsem); @@ -545,7 +545,7 @@ static int alloc_hpa(struct cxl_region *cxlr, resource_size_t size) (cxlr->mode == CXL_DECODER_PMEM && uuid_is_null(&p->uuid))) return -ENXIO; - div_u64_rem(size, SZ_256M * p->interleave_ways, &remainder); + div64_u64_rem(size, (u64)SZ_256M * p->interleave_ways, &remainder); if (remainder) return -EINVAL;