* [PATCH v2 0/6] iommu/vt-d: Reset DMAR_UNITS_SUPPORTED @ 2022-07-02 1:56 ` Lu Baolu 0 siblings, 0 replies; 18+ messages in thread From: Lu Baolu @ 2022-07-02 1:56 UTC (permalink / raw) To: Joerg Roedel, Steve Wahl, Kevin Tian Cc: David Woodhouse, Jerry Snitselaar, Mike Travis, Dimitri Sivanich, Russ Anderson, iommu, iommu, linux-kernel, Lu Baolu Hi folks, This is a follow-up series of changes proposed by this patch: https://lore.kernel.org/linux-iommu/20220615183650.32075-1-steve.wahl@hpe.com/ It removes several static arrays of size DMAR_UNITS_SUPPORTED and sets the DMAR_UNITS_SUPPORTED to 1024. This series is also available on github: https://github.com/LuBaolu/intel-iommu/commits/reset-DMAR_UNITS_SUPPORTED-v2 Please help review and suggest. Best regards, baolu Change log: v2: - Set the right @max for ida_alloc_range() and return the right error when ida_alloc_range() returns failure. - Replace xa_store() with xa_cmpxchg(). - Set domain->nid to NUMA_NO_NODE when domain is detached from an iommu. - Avoid adding a new VTD_FLAG_IOMMU_PROBED flag. Remove the duplicate check directly. v1: - https://lore.kernel.org/lkml/20220625125204.2199437-1-baolu.lu@linux.intel.com/ - Initial post. Lu Baolu (6): iommu/vt-d: Remove unused domain_get_iommu() iommu/vt-d: Use IDA interface to manage iommu sequence id iommu/vt-d: Refactor iommu information of each domain iommu/vt-d: Remove unnecessary check in intel_iommu_add() iommu/vt-d: Remove global g_iommus array iommu/vt-d: Make DMAR_UNITS_SUPPORTED default 1024 include/linux/dmar.h | 6 +- drivers/iommu/intel/iommu.h | 30 ++++-- drivers/iommu/intel/dmar.c | 35 ++----- drivers/iommu/intel/iommu.c | 189 ++++++++++++++---------------------- drivers/iommu/intel/pasid.c | 2 +- drivers/iommu/intel/svm.c | 2 +- 6 files changed, 103 insertions(+), 161 deletions(-) -- 2.25.1 ^ permalink raw reply [flat|nested] 18+ messages in thread
* [PATCH v2 0/6] iommu/vt-d: Reset DMAR_UNITS_SUPPORTED @ 2022-07-02 1:56 ` Lu Baolu 0 siblings, 0 replies; 18+ messages in thread From: Lu Baolu @ 2022-07-02 1:56 UTC (permalink / raw) To: Joerg Roedel, Steve Wahl, Kevin Tian Cc: Dimitri Sivanich, Russ Anderson, Mike Travis, Jerry Snitselaar, iommu, linux-kernel, iommu, David Woodhouse Hi folks, This is a follow-up series of changes proposed by this patch: https://lore.kernel.org/linux-iommu/20220615183650.32075-1-steve.wahl@hpe.com/ It removes several static arrays of size DMAR_UNITS_SUPPORTED and sets the DMAR_UNITS_SUPPORTED to 1024. This series is also available on github: https://github.com/LuBaolu/intel-iommu/commits/reset-DMAR_UNITS_SUPPORTED-v2 Please help review and suggest. Best regards, baolu Change log: v2: - Set the right @max for ida_alloc_range() and return the right error when ida_alloc_range() returns failure. - Replace xa_store() with xa_cmpxchg(). - Set domain->nid to NUMA_NO_NODE when domain is detached from an iommu. - Avoid adding a new VTD_FLAG_IOMMU_PROBED flag. Remove the duplicate check directly. v1: - https://lore.kernel.org/lkml/20220625125204.2199437-1-baolu.lu@linux.intel.com/ - Initial post. Lu Baolu (6): iommu/vt-d: Remove unused domain_get_iommu() iommu/vt-d: Use IDA interface to manage iommu sequence id iommu/vt-d: Refactor iommu information of each domain iommu/vt-d: Remove unnecessary check in intel_iommu_add() iommu/vt-d: Remove global g_iommus array iommu/vt-d: Make DMAR_UNITS_SUPPORTED default 1024 include/linux/dmar.h | 6 +- drivers/iommu/intel/iommu.h | 30 ++++-- drivers/iommu/intel/dmar.c | 35 ++----- drivers/iommu/intel/iommu.c | 189 ++++++++++++++---------------------- drivers/iommu/intel/pasid.c | 2 +- drivers/iommu/intel/svm.c | 2 +- 6 files changed, 103 insertions(+), 161 deletions(-) -- 2.25.1 _______________________________________________ iommu mailing list iommu@lists.linux-foundation.org https://lists.linuxfoundation.org/mailman/listinfo/iommu ^ permalink raw reply [flat|nested] 18+ messages in thread
* [PATCH v2 1/6] iommu/vt-d: Remove unused domain_get_iommu() 2022-07-02 1:56 ` Lu Baolu @ 2022-07-02 1:56 ` Lu Baolu -1 siblings, 0 replies; 18+ messages in thread From: Lu Baolu @ 2022-07-02 1:56 UTC (permalink / raw) To: Joerg Roedel, Steve Wahl, Kevin Tian Cc: David Woodhouse, Jerry Snitselaar, Mike Travis, Dimitri Sivanich, Russ Anderson, iommu, iommu, linux-kernel, Lu Baolu It is not used anywhere. Remove it to avoid dead code. Signed-off-by: Lu Baolu <baolu.lu@linux.intel.com> Reviewed-by: Kevin Tian <kevin.tian@intel.com> --- drivers/iommu/intel/iommu.h | 1 - drivers/iommu/intel/iommu.c | 18 ------------------ 2 files changed, 19 deletions(-) diff --git a/drivers/iommu/intel/iommu.h b/drivers/iommu/intel/iommu.h index df64d3d9c49a..56e0d8cd2102 100644 --- a/drivers/iommu/intel/iommu.h +++ b/drivers/iommu/intel/iommu.h @@ -725,7 +725,6 @@ extern int dmar_ir_support(void); void *alloc_pgtable_page(int node); void free_pgtable_page(void *vaddr); -struct intel_iommu *domain_get_iommu(struct dmar_domain *domain); void iommu_flush_write_buffer(struct intel_iommu *iommu); int intel_iommu_enable_pasid(struct intel_iommu *iommu, struct device *dev); struct intel_iommu *device_to_iommu(struct device *dev, u8 *bus, u8 *devfn); diff --git a/drivers/iommu/intel/iommu.c b/drivers/iommu/intel/iommu.c index da6cfea0f0d6..781e060352e6 100644 --- a/drivers/iommu/intel/iommu.c +++ b/drivers/iommu/intel/iommu.c @@ -445,24 +445,6 @@ int iommu_calculate_agaw(struct intel_iommu *iommu) return __iommu_calculate_agaw(iommu, DEFAULT_DOMAIN_ADDRESS_WIDTH); } -/* This functionin only returns single iommu in a domain */ -struct intel_iommu *domain_get_iommu(struct dmar_domain *domain) -{ - int iommu_id; - - /* si_domain and vm domain should not get here. */ - if (WARN_ON(!iommu_is_dma_domain(&domain->domain))) - return NULL; - - for_each_domain_iommu(iommu_id, domain) - break; - - if (iommu_id < 0 || iommu_id >= g_num_of_iommus) - return NULL; - - return g_iommus[iommu_id]; -} - static inline bool iommu_paging_structure_coherency(struct intel_iommu *iommu) { return sm_supported(iommu) ? -- 2.25.1 ^ permalink raw reply related [flat|nested] 18+ messages in thread
* [PATCH v2 1/6] iommu/vt-d: Remove unused domain_get_iommu() @ 2022-07-02 1:56 ` Lu Baolu 0 siblings, 0 replies; 18+ messages in thread From: Lu Baolu @ 2022-07-02 1:56 UTC (permalink / raw) To: Joerg Roedel, Steve Wahl, Kevin Tian Cc: Dimitri Sivanich, Russ Anderson, Mike Travis, Jerry Snitselaar, iommu, linux-kernel, iommu, David Woodhouse It is not used anywhere. Remove it to avoid dead code. Signed-off-by: Lu Baolu <baolu.lu@linux.intel.com> Reviewed-by: Kevin Tian <kevin.tian@intel.com> --- drivers/iommu/intel/iommu.h | 1 - drivers/iommu/intel/iommu.c | 18 ------------------ 2 files changed, 19 deletions(-) diff --git a/drivers/iommu/intel/iommu.h b/drivers/iommu/intel/iommu.h index df64d3d9c49a..56e0d8cd2102 100644 --- a/drivers/iommu/intel/iommu.h +++ b/drivers/iommu/intel/iommu.h @@ -725,7 +725,6 @@ extern int dmar_ir_support(void); void *alloc_pgtable_page(int node); void free_pgtable_page(void *vaddr); -struct intel_iommu *domain_get_iommu(struct dmar_domain *domain); void iommu_flush_write_buffer(struct intel_iommu *iommu); int intel_iommu_enable_pasid(struct intel_iommu *iommu, struct device *dev); struct intel_iommu *device_to_iommu(struct device *dev, u8 *bus, u8 *devfn); diff --git a/drivers/iommu/intel/iommu.c b/drivers/iommu/intel/iommu.c index da6cfea0f0d6..781e060352e6 100644 --- a/drivers/iommu/intel/iommu.c +++ b/drivers/iommu/intel/iommu.c @@ -445,24 +445,6 @@ int iommu_calculate_agaw(struct intel_iommu *iommu) return __iommu_calculate_agaw(iommu, DEFAULT_DOMAIN_ADDRESS_WIDTH); } -/* This functionin only returns single iommu in a domain */ -struct intel_iommu *domain_get_iommu(struct dmar_domain *domain) -{ - int iommu_id; - - /* si_domain and vm domain should not get here. */ - if (WARN_ON(!iommu_is_dma_domain(&domain->domain))) - return NULL; - - for_each_domain_iommu(iommu_id, domain) - break; - - if (iommu_id < 0 || iommu_id >= g_num_of_iommus) - return NULL; - - return g_iommus[iommu_id]; -} - static inline bool iommu_paging_structure_coherency(struct intel_iommu *iommu) { return sm_supported(iommu) ? -- 2.25.1 _______________________________________________ iommu mailing list iommu@lists.linux-foundation.org https://lists.linuxfoundation.org/mailman/listinfo/iommu ^ permalink raw reply related [flat|nested] 18+ messages in thread
* [PATCH v2 2/6] iommu/vt-d: Use IDA interface to manage iommu sequence id 2022-07-02 1:56 ` Lu Baolu @ 2022-07-02 1:56 ` Lu Baolu -1 siblings, 0 replies; 18+ messages in thread From: Lu Baolu @ 2022-07-02 1:56 UTC (permalink / raw) To: Joerg Roedel, Steve Wahl, Kevin Tian Cc: David Woodhouse, Jerry Snitselaar, Mike Travis, Dimitri Sivanich, Russ Anderson, iommu, iommu, linux-kernel, Lu Baolu Switch dmar unit sequence id allocation and release from bitmap to IDA interface. Signed-off-by: Lu Baolu <baolu.lu@linux.intel.com> --- drivers/iommu/intel/dmar.c | 35 ++++++++--------------------------- 1 file changed, 8 insertions(+), 27 deletions(-) diff --git a/drivers/iommu/intel/dmar.c b/drivers/iommu/intel/dmar.c index 2a5e0f91e647..6327b34f5aa7 100644 --- a/drivers/iommu/intel/dmar.c +++ b/drivers/iommu/intel/dmar.c @@ -60,7 +60,7 @@ LIST_HEAD(dmar_drhd_units); struct acpi_table_header * __initdata dmar_tbl; static int dmar_dev_scope_status = 1; -static unsigned long dmar_seq_ids[BITS_TO_LONGS(DMAR_UNITS_SUPPORTED)]; +static DEFINE_IDA(dmar_seq_ids); static int alloc_iommu(struct dmar_drhd_unit *drhd); static void free_iommu(struct intel_iommu *iommu); @@ -1023,28 +1023,6 @@ static int map_iommu(struct intel_iommu *iommu, u64 phys_addr) return err; } -static int dmar_alloc_seq_id(struct intel_iommu *iommu) -{ - iommu->seq_id = find_first_zero_bit(dmar_seq_ids, - DMAR_UNITS_SUPPORTED); - if (iommu->seq_id >= DMAR_UNITS_SUPPORTED) { - iommu->seq_id = -1; - } else { - set_bit(iommu->seq_id, dmar_seq_ids); - sprintf(iommu->name, "dmar%d", iommu->seq_id); - } - - return iommu->seq_id; -} - -static void dmar_free_seq_id(struct intel_iommu *iommu) -{ - if (iommu->seq_id >= 0) { - clear_bit(iommu->seq_id, dmar_seq_ids); - iommu->seq_id = -1; - } -} - static int alloc_iommu(struct dmar_drhd_unit *drhd) { struct intel_iommu *iommu; @@ -1062,11 +1040,14 @@ static int alloc_iommu(struct dmar_drhd_unit *drhd) if (!iommu) return -ENOMEM; - if (dmar_alloc_seq_id(iommu) < 0) { + iommu->seq_id = ida_alloc_range(&dmar_seq_ids, 0, + DMAR_UNITS_SUPPORTED - 1, GFP_KERNEL); + if (iommu->seq_id < 0) { pr_err("Failed to allocate seq_id\n"); - err = -ENOSPC; + err = iommu->seq_id; goto error; } + sprintf(iommu->name, "dmar%d", iommu->seq_id); err = map_iommu(iommu, drhd->reg_base_addr); if (err) { @@ -1150,7 +1131,7 @@ static int alloc_iommu(struct dmar_drhd_unit *drhd) err_unmap: unmap_iommu(iommu); error_free_seq_id: - dmar_free_seq_id(iommu); + ida_free(&dmar_seq_ids, iommu->seq_id); error: kfree(iommu); return err; @@ -1183,7 +1164,7 @@ static void free_iommu(struct intel_iommu *iommu) if (iommu->reg) unmap_iommu(iommu); - dmar_free_seq_id(iommu); + ida_free(&dmar_seq_ids, iommu->seq_id); kfree(iommu); } -- 2.25.1 ^ permalink raw reply related [flat|nested] 18+ messages in thread
* [PATCH v2 2/6] iommu/vt-d: Use IDA interface to manage iommu sequence id @ 2022-07-02 1:56 ` Lu Baolu 0 siblings, 0 replies; 18+ messages in thread From: Lu Baolu @ 2022-07-02 1:56 UTC (permalink / raw) To: Joerg Roedel, Steve Wahl, Kevin Tian Cc: Dimitri Sivanich, Russ Anderson, Mike Travis, Jerry Snitselaar, iommu, linux-kernel, iommu, David Woodhouse Switch dmar unit sequence id allocation and release from bitmap to IDA interface. Signed-off-by: Lu Baolu <baolu.lu@linux.intel.com> --- drivers/iommu/intel/dmar.c | 35 ++++++++--------------------------- 1 file changed, 8 insertions(+), 27 deletions(-) diff --git a/drivers/iommu/intel/dmar.c b/drivers/iommu/intel/dmar.c index 2a5e0f91e647..6327b34f5aa7 100644 --- a/drivers/iommu/intel/dmar.c +++ b/drivers/iommu/intel/dmar.c @@ -60,7 +60,7 @@ LIST_HEAD(dmar_drhd_units); struct acpi_table_header * __initdata dmar_tbl; static int dmar_dev_scope_status = 1; -static unsigned long dmar_seq_ids[BITS_TO_LONGS(DMAR_UNITS_SUPPORTED)]; +static DEFINE_IDA(dmar_seq_ids); static int alloc_iommu(struct dmar_drhd_unit *drhd); static void free_iommu(struct intel_iommu *iommu); @@ -1023,28 +1023,6 @@ static int map_iommu(struct intel_iommu *iommu, u64 phys_addr) return err; } -static int dmar_alloc_seq_id(struct intel_iommu *iommu) -{ - iommu->seq_id = find_first_zero_bit(dmar_seq_ids, - DMAR_UNITS_SUPPORTED); - if (iommu->seq_id >= DMAR_UNITS_SUPPORTED) { - iommu->seq_id = -1; - } else { - set_bit(iommu->seq_id, dmar_seq_ids); - sprintf(iommu->name, "dmar%d", iommu->seq_id); - } - - return iommu->seq_id; -} - -static void dmar_free_seq_id(struct intel_iommu *iommu) -{ - if (iommu->seq_id >= 0) { - clear_bit(iommu->seq_id, dmar_seq_ids); - iommu->seq_id = -1; - } -} - static int alloc_iommu(struct dmar_drhd_unit *drhd) { struct intel_iommu *iommu; @@ -1062,11 +1040,14 @@ static int alloc_iommu(struct dmar_drhd_unit *drhd) if (!iommu) return -ENOMEM; - if (dmar_alloc_seq_id(iommu) < 0) { + iommu->seq_id = ida_alloc_range(&dmar_seq_ids, 0, + DMAR_UNITS_SUPPORTED - 1, GFP_KERNEL); + if (iommu->seq_id < 0) { pr_err("Failed to allocate seq_id\n"); - err = -ENOSPC; + err = iommu->seq_id; goto error; } + sprintf(iommu->name, "dmar%d", iommu->seq_id); err = map_iommu(iommu, drhd->reg_base_addr); if (err) { @@ -1150,7 +1131,7 @@ static int alloc_iommu(struct dmar_drhd_unit *drhd) err_unmap: unmap_iommu(iommu); error_free_seq_id: - dmar_free_seq_id(iommu); + ida_free(&dmar_seq_ids, iommu->seq_id); error: kfree(iommu); return err; @@ -1183,7 +1164,7 @@ static void free_iommu(struct intel_iommu *iommu) if (iommu->reg) unmap_iommu(iommu); - dmar_free_seq_id(iommu); + ida_free(&dmar_seq_ids, iommu->seq_id); kfree(iommu); } -- 2.25.1 _______________________________________________ iommu mailing list iommu@lists.linux-foundation.org https://lists.linuxfoundation.org/mailman/listinfo/iommu ^ permalink raw reply related [flat|nested] 18+ messages in thread
* RE: [PATCH v2 2/6] iommu/vt-d: Use IDA interface to manage iommu sequence id 2022-07-02 1:56 ` Lu Baolu (?) @ 2022-07-07 0:53 ` Tian, Kevin -1 siblings, 0 replies; 18+ messages in thread From: Tian, Kevin @ 2022-07-07 0:53 UTC (permalink / raw) To: Lu Baolu, Joerg Roedel, Steve Wahl Cc: David Woodhouse, Jerry Snitselaar, Mike Travis, Dimitri Sivanich, Anderson, Russ, iommu, iommu, linux-kernel > From: Lu Baolu <baolu.lu@linux.intel.com> > Sent: Saturday, July 2, 2022 9:56 AM > > Switch dmar unit sequence id allocation and release from bitmap to IDA > interface. > > Signed-off-by: Lu Baolu <baolu.lu@linux.intel.com> Reviewed-by: Kevin Tian <kevin.tian@intel.com> ^ permalink raw reply [flat|nested] 18+ messages in thread
* [PATCH v2 3/6] iommu/vt-d: Refactor iommu information of each domain 2022-07-02 1:56 ` Lu Baolu @ 2022-07-02 1:56 ` Lu Baolu -1 siblings, 0 replies; 18+ messages in thread From: Lu Baolu @ 2022-07-02 1:56 UTC (permalink / raw) To: Joerg Roedel, Steve Wahl, Kevin Tian Cc: David Woodhouse, Jerry Snitselaar, Mike Travis, Dimitri Sivanich, Russ Anderson, iommu, iommu, linux-kernel, Lu Baolu When a DMA domain is attached to a device, it needs to allocate a domain ID from its IOMMU. Currently, the domain ID information is stored in two static arrays embedded in the domain structure. This can lead to memory waste when the driver is running on a small platform. This optimizes these static arrays by replacing them with an xarray and consuming memory on demand. Signed-off-by: Lu Baolu <baolu.lu@linux.intel.com> --- drivers/iommu/intel/iommu.h | 29 ++++++--- drivers/iommu/intel/iommu.c | 124 +++++++++++++++++++++--------------- drivers/iommu/intel/pasid.c | 2 +- drivers/iommu/intel/svm.c | 2 +- 4 files changed, 94 insertions(+), 63 deletions(-) diff --git a/drivers/iommu/intel/iommu.h b/drivers/iommu/intel/iommu.h index 56e0d8cd2102..fae45bbb0c7f 100644 --- a/drivers/iommu/intel/iommu.h +++ b/drivers/iommu/intel/iommu.h @@ -21,6 +21,7 @@ #include <linux/dmar.h> #include <linux/ioasid.h> #include <linux/bitfield.h> +#include <linux/xarray.h> #include <asm/cacheflush.h> #include <asm/iommu.h> @@ -524,17 +525,17 @@ struct context_entry { */ #define DOMAIN_FLAG_USE_FIRST_LEVEL BIT(1) -struct dmar_domain { - int nid; /* node id */ - - unsigned int iommu_refcnt[DMAR_UNITS_SUPPORTED]; - /* Refcount of devices per iommu */ - - - u16 iommu_did[DMAR_UNITS_SUPPORTED]; - /* Domain ids per IOMMU. Use u16 since +struct iommu_domain_info { + struct intel_iommu *iommu; + unsigned int refcnt; /* Refcount of devices per iommu */ + u16 did; /* Domain ids per IOMMU. Use u16 since * domain ids are 16 bit wide according * to VT-d spec, section 9.3 */ +}; + +struct dmar_domain { + int nid; /* node id */ + struct xarray iommu_array; /* Attached IOMMU array */ u8 has_iotlb_device: 1; u8 iommu_coherency: 1; /* indicate coherency of iommu access */ @@ -640,6 +641,16 @@ static inline struct dmar_domain *to_dmar_domain(struct iommu_domain *dom) return container_of(dom, struct dmar_domain, domain); } +/* Retrieve the domain ID which has allocated to the domain */ +static inline u16 +domain_id_iommu(struct dmar_domain *domain, struct intel_iommu *iommu) +{ + struct iommu_domain_info *info = + xa_load(&domain->iommu_array, iommu->seq_id); + + return info->did; +} + /* * 0: readable * 1: writable diff --git a/drivers/iommu/intel/iommu.c b/drivers/iommu/intel/iommu.c index 781e060352e6..70408c234f5b 100644 --- a/drivers/iommu/intel/iommu.c +++ b/drivers/iommu/intel/iommu.c @@ -254,10 +254,6 @@ static inline void context_clear_entry(struct context_entry *context) static struct dmar_domain *si_domain; static int hw_pass_through = 1; -#define for_each_domain_iommu(idx, domain) \ - for (idx = 0; idx < g_num_of_iommus; idx++) \ - if (domain->iommu_refcnt[idx]) - struct dmar_rmrr_unit { struct list_head list; /* list of rmrr units */ struct acpi_dmar_header *hdr; /* ACPI header */ @@ -453,16 +449,16 @@ static inline bool iommu_paging_structure_coherency(struct intel_iommu *iommu) static void domain_update_iommu_coherency(struct dmar_domain *domain) { + struct iommu_domain_info *info; struct dmar_drhd_unit *drhd; struct intel_iommu *iommu; bool found = false; - int i; + unsigned long i; domain->iommu_coherency = true; - - for_each_domain_iommu(i, domain) { + xa_for_each(&domain->iommu_array, i, info) { found = true; - if (!iommu_paging_structure_coherency(g_iommus[i])) { + if (!iommu_paging_structure_coherency(info->iommu)) { domain->iommu_coherency = false; break; } @@ -1495,7 +1491,7 @@ static void iommu_flush_iotlb_psi(struct intel_iommu *iommu, unsigned int aligned_pages = __roundup_pow_of_two(pages); unsigned int mask = ilog2(aligned_pages); uint64_t addr = (uint64_t)pfn << VTD_PAGE_SHIFT; - u16 did = domain->iommu_did[iommu->seq_id]; + u16 did = domain_id_iommu(domain, iommu); BUG_ON(pages == 0); @@ -1565,11 +1561,12 @@ static inline void __mapping_notify_one(struct intel_iommu *iommu, static void intel_flush_iotlb_all(struct iommu_domain *domain) { struct dmar_domain *dmar_domain = to_dmar_domain(domain); - int idx; + struct iommu_domain_info *info; + unsigned long idx; - for_each_domain_iommu(idx, dmar_domain) { - struct intel_iommu *iommu = g_iommus[idx]; - u16 did = dmar_domain->iommu_did[iommu->seq_id]; + xa_for_each(&dmar_domain->iommu_array, idx, info) { + struct intel_iommu *iommu = info->iommu; + u16 did = domain_id_iommu(dmar_domain, iommu); if (domain_use_first_level(dmar_domain)) qi_flush_piotlb(iommu, did, PASID_RID2PASID, 0, -1, 0); @@ -1745,6 +1742,7 @@ static struct dmar_domain *alloc_domain(unsigned int type) domain->has_iotlb_device = false; INIT_LIST_HEAD(&domain->devices); spin_lock_init(&domain->lock); + xa_init(&domain->iommu_array); return domain; } @@ -1752,45 +1750,67 @@ static struct dmar_domain *alloc_domain(unsigned int type) static int domain_attach_iommu(struct dmar_domain *domain, struct intel_iommu *iommu) { + struct iommu_domain_info *info, *curr; unsigned long ndomains; - int num, ret = 0; + int num, ret = -ENOSPC; + + info = kzalloc(sizeof(*info), GFP_KERNEL); + if (!info) + return -ENOMEM; spin_lock(&iommu->lock); - domain->iommu_refcnt[iommu->seq_id] += 1; - if (domain->iommu_refcnt[iommu->seq_id] == 1) { - ndomains = cap_ndoms(iommu->cap); - num = find_first_zero_bit(iommu->domain_ids, ndomains); - - if (num >= ndomains) { - pr_err("%s: No free domain ids\n", iommu->name); - domain->iommu_refcnt[iommu->seq_id] -= 1; - ret = -ENOSPC; - goto out_unlock; - } + curr = xa_load(&domain->iommu_array, iommu->seq_id); + if (curr) { + curr->refcnt++; + spin_unlock(&iommu->lock); + kfree(info); + return 0; + } - set_bit(num, iommu->domain_ids); - domain->iommu_did[iommu->seq_id] = num; - domain->nid = iommu->node; - domain_update_iommu_cap(domain); + ndomains = cap_ndoms(iommu->cap); + num = find_first_zero_bit(iommu->domain_ids, ndomains); + if (num >= ndomains) { + pr_err("%s: No free domain ids\n", iommu->name); + goto err_unlock; } -out_unlock: + set_bit(num, iommu->domain_ids); + info->refcnt = 1; + info->did = num; + info->iommu = iommu; + domain->nid = iommu->node; + curr = xa_cmpxchg(&domain->iommu_array, iommu->seq_id, + NULL, info, GFP_ATOMIC); + if (curr) { + ret = xa_err(curr) ? : -EBUSY; + goto err_clear; + } + domain_update_iommu_cap(domain); + spin_unlock(&iommu->lock); + return 0; + +err_clear: + clear_bit(info->did, iommu->domain_ids); +err_unlock: + spin_unlock(&iommu->lock); + kfree(info); return ret; } static void domain_detach_iommu(struct dmar_domain *domain, struct intel_iommu *iommu) { - int num; + struct iommu_domain_info *info; spin_lock(&iommu->lock); - domain->iommu_refcnt[iommu->seq_id] -= 1; - if (domain->iommu_refcnt[iommu->seq_id] == 0) { - num = domain->iommu_did[iommu->seq_id]; - clear_bit(num, iommu->domain_ids); + info = xa_load(&domain->iommu_array, iommu->seq_id); + if (--info->refcnt == 0) { + clear_bit(info->did, iommu->domain_ids); + xa_erase(&domain->iommu_array, iommu->seq_id); + domain->nid = NUMA_NO_NODE; domain_update_iommu_cap(domain); - domain->iommu_did[iommu->seq_id] = 0; + kfree(info); } spin_unlock(&iommu->lock); } @@ -1880,7 +1900,7 @@ static int domain_context_mapping_one(struct dmar_domain *domain, { struct device_domain_info *info = iommu_support_dev_iotlb(domain, iommu, bus, devfn); - u16 did = domain->iommu_did[iommu->seq_id]; + u16 did = domain_id_iommu(domain, iommu); int translation = CONTEXT_TT_MULTI_LEVEL; struct context_entry *context; int ret; @@ -2130,8 +2150,9 @@ static void switch_to_super_page(struct dmar_domain *domain, unsigned long end_pfn, int level) { unsigned long lvl_pages = lvl_to_nr_pages(level); + struct iommu_domain_info *info; struct dma_pte *pte = NULL; - int i; + unsigned long i; while (start_pfn <= end_pfn) { if (!pte) @@ -2142,8 +2163,8 @@ static void switch_to_super_page(struct dmar_domain *domain, start_pfn + lvl_pages - 1, level + 1); - for_each_domain_iommu(i, domain) - iommu_flush_iotlb_psi(g_iommus[i], domain, + xa_for_each(&domain->iommu_array, i, info) + iommu_flush_iotlb_psi(info->iommu, domain, start_pfn, lvl_pages, 0, 0); } @@ -2273,7 +2294,7 @@ static void domain_context_clear_one(struct device_domain_info *info, u8 bus, u8 if (hw_pass_through && domain_type_is_si(info->domain)) did_old = FLPT_DEFAULT_DID; else - did_old = info->domain->iommu_did[iommu->seq_id]; + did_old = domain_id_iommu(info->domain, iommu); } else { did_old = context_domain_id(context); } @@ -2331,7 +2352,7 @@ static int domain_setup_first_level(struct intel_iommu *iommu, flags |= PASID_FLAG_PAGE_SNOOP; return intel_pasid_setup_first_level(iommu, dev, (pgd_t *)pgd, pasid, - domain->iommu_did[iommu->seq_id], + domain_id_iommu(domain, iommu), flags); } @@ -4369,15 +4390,16 @@ static void intel_iommu_tlb_sync(struct iommu_domain *domain, struct dmar_domain *dmar_domain = to_dmar_domain(domain); unsigned long iova_pfn = IOVA_PFN(gather->start); size_t size = gather->end - gather->start; + struct iommu_domain_info *info; unsigned long start_pfn; unsigned long nrpages; - int iommu_id; + unsigned long i; nrpages = aligned_nrpages(gather->start, size); start_pfn = mm_to_dma_pfn(iova_pfn); - for_each_domain_iommu(iommu_id, dmar_domain) - iommu_flush_iotlb_psi(g_iommus[iommu_id], dmar_domain, + xa_for_each(&dmar_domain->iommu_array, i, info) + iommu_flush_iotlb_psi(info->iommu, dmar_domain, start_pfn, nrpages, list_empty(&gather->freelist), 0); @@ -4622,7 +4644,7 @@ int intel_iommu_enable_pasid(struct intel_iommu *iommu, struct device *dev) context[0].lo = ctx_lo; wmb(); iommu->flush.flush_context(iommu, - domain->iommu_did[iommu->seq_id], + domain_id_iommu(domain, iommu), PCI_DEVID(info->bus, info->devfn), DMA_CCMD_MASK_NOBIT, DMA_CCMD_DEVICE_INVL); @@ -4759,13 +4781,11 @@ static void intel_iommu_iotlb_sync_map(struct iommu_domain *domain, struct dmar_domain *dmar_domain = to_dmar_domain(domain); unsigned long pages = aligned_nrpages(iova, size); unsigned long pfn = iova >> VTD_PAGE_SHIFT; - struct intel_iommu *iommu; - int iommu_id; + struct iommu_domain_info *info; + unsigned long i; - for_each_domain_iommu(iommu_id, dmar_domain) { - iommu = g_iommus[iommu_id]; - __mapping_notify_one(iommu, dmar_domain, pfn, pages); - } + xa_for_each(&dmar_domain->iommu_array, i, info) + __mapping_notify_one(info->iommu, dmar_domain, pfn, pages); } const struct iommu_ops intel_iommu_ops = { diff --git a/drivers/iommu/intel/pasid.c b/drivers/iommu/intel/pasid.c index c30de2339de5..006026dc18d4 100644 --- a/drivers/iommu/intel/pasid.c +++ b/drivers/iommu/intel/pasid.c @@ -626,7 +626,7 @@ int intel_pasid_setup_second_level(struct intel_iommu *iommu, } pgd_val = virt_to_phys(pgd); - did = domain->iommu_did[iommu->seq_id]; + did = domain_id_iommu(domain, iommu); spin_lock(&iommu->lock); pte = intel_pasid_get_entry(dev, pasid); diff --git a/drivers/iommu/intel/svm.c b/drivers/iommu/intel/svm.c index 64072e628bbd..8bcfb93dda56 100644 --- a/drivers/iommu/intel/svm.c +++ b/drivers/iommu/intel/svm.c @@ -541,7 +541,7 @@ static void intel_svm_drain_prq(struct device *dev, u32 pasid) domain = info->domain; pdev = to_pci_dev(dev); sid = PCI_DEVID(info->bus, info->devfn); - did = domain->iommu_did[iommu->seq_id]; + did = domain_id_iommu(domain, iommu); qdep = pci_ats_queue_depth(pdev); /* -- 2.25.1 ^ permalink raw reply related [flat|nested] 18+ messages in thread
* [PATCH v2 3/6] iommu/vt-d: Refactor iommu information of each domain @ 2022-07-02 1:56 ` Lu Baolu 0 siblings, 0 replies; 18+ messages in thread From: Lu Baolu @ 2022-07-02 1:56 UTC (permalink / raw) To: Joerg Roedel, Steve Wahl, Kevin Tian Cc: Dimitri Sivanich, Russ Anderson, Mike Travis, Jerry Snitselaar, iommu, linux-kernel, iommu, David Woodhouse When a DMA domain is attached to a device, it needs to allocate a domain ID from its IOMMU. Currently, the domain ID information is stored in two static arrays embedded in the domain structure. This can lead to memory waste when the driver is running on a small platform. This optimizes these static arrays by replacing them with an xarray and consuming memory on demand. Signed-off-by: Lu Baolu <baolu.lu@linux.intel.com> --- drivers/iommu/intel/iommu.h | 29 ++++++--- drivers/iommu/intel/iommu.c | 124 +++++++++++++++++++++--------------- drivers/iommu/intel/pasid.c | 2 +- drivers/iommu/intel/svm.c | 2 +- 4 files changed, 94 insertions(+), 63 deletions(-) diff --git a/drivers/iommu/intel/iommu.h b/drivers/iommu/intel/iommu.h index 56e0d8cd2102..fae45bbb0c7f 100644 --- a/drivers/iommu/intel/iommu.h +++ b/drivers/iommu/intel/iommu.h @@ -21,6 +21,7 @@ #include <linux/dmar.h> #include <linux/ioasid.h> #include <linux/bitfield.h> +#include <linux/xarray.h> #include <asm/cacheflush.h> #include <asm/iommu.h> @@ -524,17 +525,17 @@ struct context_entry { */ #define DOMAIN_FLAG_USE_FIRST_LEVEL BIT(1) -struct dmar_domain { - int nid; /* node id */ - - unsigned int iommu_refcnt[DMAR_UNITS_SUPPORTED]; - /* Refcount of devices per iommu */ - - - u16 iommu_did[DMAR_UNITS_SUPPORTED]; - /* Domain ids per IOMMU. Use u16 since +struct iommu_domain_info { + struct intel_iommu *iommu; + unsigned int refcnt; /* Refcount of devices per iommu */ + u16 did; /* Domain ids per IOMMU. Use u16 since * domain ids are 16 bit wide according * to VT-d spec, section 9.3 */ +}; + +struct dmar_domain { + int nid; /* node id */ + struct xarray iommu_array; /* Attached IOMMU array */ u8 has_iotlb_device: 1; u8 iommu_coherency: 1; /* indicate coherency of iommu access */ @@ -640,6 +641,16 @@ static inline struct dmar_domain *to_dmar_domain(struct iommu_domain *dom) return container_of(dom, struct dmar_domain, domain); } +/* Retrieve the domain ID which has allocated to the domain */ +static inline u16 +domain_id_iommu(struct dmar_domain *domain, struct intel_iommu *iommu) +{ + struct iommu_domain_info *info = + xa_load(&domain->iommu_array, iommu->seq_id); + + return info->did; +} + /* * 0: readable * 1: writable diff --git a/drivers/iommu/intel/iommu.c b/drivers/iommu/intel/iommu.c index 781e060352e6..70408c234f5b 100644 --- a/drivers/iommu/intel/iommu.c +++ b/drivers/iommu/intel/iommu.c @@ -254,10 +254,6 @@ static inline void context_clear_entry(struct context_entry *context) static struct dmar_domain *si_domain; static int hw_pass_through = 1; -#define for_each_domain_iommu(idx, domain) \ - for (idx = 0; idx < g_num_of_iommus; idx++) \ - if (domain->iommu_refcnt[idx]) - struct dmar_rmrr_unit { struct list_head list; /* list of rmrr units */ struct acpi_dmar_header *hdr; /* ACPI header */ @@ -453,16 +449,16 @@ static inline bool iommu_paging_structure_coherency(struct intel_iommu *iommu) static void domain_update_iommu_coherency(struct dmar_domain *domain) { + struct iommu_domain_info *info; struct dmar_drhd_unit *drhd; struct intel_iommu *iommu; bool found = false; - int i; + unsigned long i; domain->iommu_coherency = true; - - for_each_domain_iommu(i, domain) { + xa_for_each(&domain->iommu_array, i, info) { found = true; - if (!iommu_paging_structure_coherency(g_iommus[i])) { + if (!iommu_paging_structure_coherency(info->iommu)) { domain->iommu_coherency = false; break; } @@ -1495,7 +1491,7 @@ static void iommu_flush_iotlb_psi(struct intel_iommu *iommu, unsigned int aligned_pages = __roundup_pow_of_two(pages); unsigned int mask = ilog2(aligned_pages); uint64_t addr = (uint64_t)pfn << VTD_PAGE_SHIFT; - u16 did = domain->iommu_did[iommu->seq_id]; + u16 did = domain_id_iommu(domain, iommu); BUG_ON(pages == 0); @@ -1565,11 +1561,12 @@ static inline void __mapping_notify_one(struct intel_iommu *iommu, static void intel_flush_iotlb_all(struct iommu_domain *domain) { struct dmar_domain *dmar_domain = to_dmar_domain(domain); - int idx; + struct iommu_domain_info *info; + unsigned long idx; - for_each_domain_iommu(idx, dmar_domain) { - struct intel_iommu *iommu = g_iommus[idx]; - u16 did = dmar_domain->iommu_did[iommu->seq_id]; + xa_for_each(&dmar_domain->iommu_array, idx, info) { + struct intel_iommu *iommu = info->iommu; + u16 did = domain_id_iommu(dmar_domain, iommu); if (domain_use_first_level(dmar_domain)) qi_flush_piotlb(iommu, did, PASID_RID2PASID, 0, -1, 0); @@ -1745,6 +1742,7 @@ static struct dmar_domain *alloc_domain(unsigned int type) domain->has_iotlb_device = false; INIT_LIST_HEAD(&domain->devices); spin_lock_init(&domain->lock); + xa_init(&domain->iommu_array); return domain; } @@ -1752,45 +1750,67 @@ static struct dmar_domain *alloc_domain(unsigned int type) static int domain_attach_iommu(struct dmar_domain *domain, struct intel_iommu *iommu) { + struct iommu_domain_info *info, *curr; unsigned long ndomains; - int num, ret = 0; + int num, ret = -ENOSPC; + + info = kzalloc(sizeof(*info), GFP_KERNEL); + if (!info) + return -ENOMEM; spin_lock(&iommu->lock); - domain->iommu_refcnt[iommu->seq_id] += 1; - if (domain->iommu_refcnt[iommu->seq_id] == 1) { - ndomains = cap_ndoms(iommu->cap); - num = find_first_zero_bit(iommu->domain_ids, ndomains); - - if (num >= ndomains) { - pr_err("%s: No free domain ids\n", iommu->name); - domain->iommu_refcnt[iommu->seq_id] -= 1; - ret = -ENOSPC; - goto out_unlock; - } + curr = xa_load(&domain->iommu_array, iommu->seq_id); + if (curr) { + curr->refcnt++; + spin_unlock(&iommu->lock); + kfree(info); + return 0; + } - set_bit(num, iommu->domain_ids); - domain->iommu_did[iommu->seq_id] = num; - domain->nid = iommu->node; - domain_update_iommu_cap(domain); + ndomains = cap_ndoms(iommu->cap); + num = find_first_zero_bit(iommu->domain_ids, ndomains); + if (num >= ndomains) { + pr_err("%s: No free domain ids\n", iommu->name); + goto err_unlock; } -out_unlock: + set_bit(num, iommu->domain_ids); + info->refcnt = 1; + info->did = num; + info->iommu = iommu; + domain->nid = iommu->node; + curr = xa_cmpxchg(&domain->iommu_array, iommu->seq_id, + NULL, info, GFP_ATOMIC); + if (curr) { + ret = xa_err(curr) ? : -EBUSY; + goto err_clear; + } + domain_update_iommu_cap(domain); + spin_unlock(&iommu->lock); + return 0; + +err_clear: + clear_bit(info->did, iommu->domain_ids); +err_unlock: + spin_unlock(&iommu->lock); + kfree(info); return ret; } static void domain_detach_iommu(struct dmar_domain *domain, struct intel_iommu *iommu) { - int num; + struct iommu_domain_info *info; spin_lock(&iommu->lock); - domain->iommu_refcnt[iommu->seq_id] -= 1; - if (domain->iommu_refcnt[iommu->seq_id] == 0) { - num = domain->iommu_did[iommu->seq_id]; - clear_bit(num, iommu->domain_ids); + info = xa_load(&domain->iommu_array, iommu->seq_id); + if (--info->refcnt == 0) { + clear_bit(info->did, iommu->domain_ids); + xa_erase(&domain->iommu_array, iommu->seq_id); + domain->nid = NUMA_NO_NODE; domain_update_iommu_cap(domain); - domain->iommu_did[iommu->seq_id] = 0; + kfree(info); } spin_unlock(&iommu->lock); } @@ -1880,7 +1900,7 @@ static int domain_context_mapping_one(struct dmar_domain *domain, { struct device_domain_info *info = iommu_support_dev_iotlb(domain, iommu, bus, devfn); - u16 did = domain->iommu_did[iommu->seq_id]; + u16 did = domain_id_iommu(domain, iommu); int translation = CONTEXT_TT_MULTI_LEVEL; struct context_entry *context; int ret; @@ -2130,8 +2150,9 @@ static void switch_to_super_page(struct dmar_domain *domain, unsigned long end_pfn, int level) { unsigned long lvl_pages = lvl_to_nr_pages(level); + struct iommu_domain_info *info; struct dma_pte *pte = NULL; - int i; + unsigned long i; while (start_pfn <= end_pfn) { if (!pte) @@ -2142,8 +2163,8 @@ static void switch_to_super_page(struct dmar_domain *domain, start_pfn + lvl_pages - 1, level + 1); - for_each_domain_iommu(i, domain) - iommu_flush_iotlb_psi(g_iommus[i], domain, + xa_for_each(&domain->iommu_array, i, info) + iommu_flush_iotlb_psi(info->iommu, domain, start_pfn, lvl_pages, 0, 0); } @@ -2273,7 +2294,7 @@ static void domain_context_clear_one(struct device_domain_info *info, u8 bus, u8 if (hw_pass_through && domain_type_is_si(info->domain)) did_old = FLPT_DEFAULT_DID; else - did_old = info->domain->iommu_did[iommu->seq_id]; + did_old = domain_id_iommu(info->domain, iommu); } else { did_old = context_domain_id(context); } @@ -2331,7 +2352,7 @@ static int domain_setup_first_level(struct intel_iommu *iommu, flags |= PASID_FLAG_PAGE_SNOOP; return intel_pasid_setup_first_level(iommu, dev, (pgd_t *)pgd, pasid, - domain->iommu_did[iommu->seq_id], + domain_id_iommu(domain, iommu), flags); } @@ -4369,15 +4390,16 @@ static void intel_iommu_tlb_sync(struct iommu_domain *domain, struct dmar_domain *dmar_domain = to_dmar_domain(domain); unsigned long iova_pfn = IOVA_PFN(gather->start); size_t size = gather->end - gather->start; + struct iommu_domain_info *info; unsigned long start_pfn; unsigned long nrpages; - int iommu_id; + unsigned long i; nrpages = aligned_nrpages(gather->start, size); start_pfn = mm_to_dma_pfn(iova_pfn); - for_each_domain_iommu(iommu_id, dmar_domain) - iommu_flush_iotlb_psi(g_iommus[iommu_id], dmar_domain, + xa_for_each(&dmar_domain->iommu_array, i, info) + iommu_flush_iotlb_psi(info->iommu, dmar_domain, start_pfn, nrpages, list_empty(&gather->freelist), 0); @@ -4622,7 +4644,7 @@ int intel_iommu_enable_pasid(struct intel_iommu *iommu, struct device *dev) context[0].lo = ctx_lo; wmb(); iommu->flush.flush_context(iommu, - domain->iommu_did[iommu->seq_id], + domain_id_iommu(domain, iommu), PCI_DEVID(info->bus, info->devfn), DMA_CCMD_MASK_NOBIT, DMA_CCMD_DEVICE_INVL); @@ -4759,13 +4781,11 @@ static void intel_iommu_iotlb_sync_map(struct iommu_domain *domain, struct dmar_domain *dmar_domain = to_dmar_domain(domain); unsigned long pages = aligned_nrpages(iova, size); unsigned long pfn = iova >> VTD_PAGE_SHIFT; - struct intel_iommu *iommu; - int iommu_id; + struct iommu_domain_info *info; + unsigned long i; - for_each_domain_iommu(iommu_id, dmar_domain) { - iommu = g_iommus[iommu_id]; - __mapping_notify_one(iommu, dmar_domain, pfn, pages); - } + xa_for_each(&dmar_domain->iommu_array, i, info) + __mapping_notify_one(info->iommu, dmar_domain, pfn, pages); } const struct iommu_ops intel_iommu_ops = { diff --git a/drivers/iommu/intel/pasid.c b/drivers/iommu/intel/pasid.c index c30de2339de5..006026dc18d4 100644 --- a/drivers/iommu/intel/pasid.c +++ b/drivers/iommu/intel/pasid.c @@ -626,7 +626,7 @@ int intel_pasid_setup_second_level(struct intel_iommu *iommu, } pgd_val = virt_to_phys(pgd); - did = domain->iommu_did[iommu->seq_id]; + did = domain_id_iommu(domain, iommu); spin_lock(&iommu->lock); pte = intel_pasid_get_entry(dev, pasid); diff --git a/drivers/iommu/intel/svm.c b/drivers/iommu/intel/svm.c index 64072e628bbd..8bcfb93dda56 100644 --- a/drivers/iommu/intel/svm.c +++ b/drivers/iommu/intel/svm.c @@ -541,7 +541,7 @@ static void intel_svm_drain_prq(struct device *dev, u32 pasid) domain = info->domain; pdev = to_pci_dev(dev); sid = PCI_DEVID(info->bus, info->devfn); - did = domain->iommu_did[iommu->seq_id]; + did = domain_id_iommu(domain, iommu); qdep = pci_ats_queue_depth(pdev); /* -- 2.25.1 _______________________________________________ iommu mailing list iommu@lists.linux-foundation.org https://lists.linuxfoundation.org/mailman/listinfo/iommu ^ permalink raw reply related [flat|nested] 18+ messages in thread
* RE: [PATCH v2 3/6] iommu/vt-d: Refactor iommu information of each domain 2022-07-02 1:56 ` Lu Baolu (?) @ 2022-07-07 1:01 ` Tian, Kevin 2022-07-07 2:22 ` Baolu Lu -1 siblings, 1 reply; 18+ messages in thread From: Tian, Kevin @ 2022-07-07 1:01 UTC (permalink / raw) To: Lu Baolu, Joerg Roedel, Steve Wahl Cc: David Woodhouse, Jerry Snitselaar, Mike Travis, Dimitri Sivanich, Anderson, Russ, iommu, iommu, linux-kernel > From: Lu Baolu <baolu.lu@linux.intel.com> > Sent: Saturday, July 2, 2022 9:56 AM > > -out_unlock: > + set_bit(num, iommu->domain_ids); > + info->refcnt = 1; > + info->did = num; > + info->iommu = iommu; > + domain->nid = iommu->node; One nit. this line should be removed as it's incorrect to blindly update domain->nid and we should just leave to domain_update_iommu_cap() to decide the right node. Otherwise this causes a policy conflict as here it is the last attached device deciding the node which is different from domain_update_iommu_cap() which picks the node of the first attached device. Otherwise, Reviewed-by: Kevin Tian <kevin.tian@intel.com> ^ permalink raw reply [flat|nested] 18+ messages in thread
* Re: [PATCH v2 3/6] iommu/vt-d: Refactor iommu information of each domain 2022-07-07 1:01 ` Tian, Kevin @ 2022-07-07 2:22 ` Baolu Lu 0 siblings, 0 replies; 18+ messages in thread From: Baolu Lu @ 2022-07-07 2:22 UTC (permalink / raw) To: Tian, Kevin, Joerg Roedel, Steve Wahl Cc: baolu.lu, David Woodhouse, Jerry Snitselaar, Mike Travis, Dimitri Sivanich, Anderson, Russ, iommu, iommu, linux-kernel On 2022/7/7 09:01, Tian, Kevin wrote: >> From: Lu Baolu <baolu.lu@linux.intel.com> >> Sent: Saturday, July 2, 2022 9:56 AM >> >> -out_unlock: >> + set_bit(num, iommu->domain_ids); >> + info->refcnt = 1; >> + info->did = num; >> + info->iommu = iommu; >> + domain->nid = iommu->node; > > One nit. this line should be removed as it's incorrect to blindly update > domain->nid and we should just leave to domain_update_iommu_cap() > to decide the right node. Otherwise this causes a policy conflict as > here it is the last attached device deciding the node which is different > from domain_update_iommu_cap() which picks the node of the first > attached device. Agreed and updated. Thank you! > > Otherwise, > > Reviewed-by: Kevin Tian <kevin.tian@intel.com> Best regards, baolu ^ permalink raw reply [flat|nested] 18+ messages in thread
* [PATCH v2 4/6] iommu/vt-d: Remove unnecessary check in intel_iommu_add() 2022-07-02 1:56 ` Lu Baolu @ 2022-07-02 1:56 ` Lu Baolu -1 siblings, 0 replies; 18+ messages in thread From: Lu Baolu @ 2022-07-02 1:56 UTC (permalink / raw) To: Joerg Roedel, Steve Wahl, Kevin Tian Cc: David Woodhouse, Jerry Snitselaar, Mike Travis, Dimitri Sivanich, Russ Anderson, iommu, iommu, linux-kernel, Lu Baolu The Intel IOMMU hot-add process starts from dmar_device_hotplug(). It uses the global dmar_global_lock to synchronize all the hot-add and hot-remove paths. In the hot-add path, the new IOMMU data structures are allocated firstly by dmar_parse_one_drhd() and then initialized by dmar_hp_add_drhd(). All the IOMMU units are allocated and initialized in the same synchronized path. There is no case where any IOMMU unit is created and then initialized for multiple times. This removes the unnecessary check in intel_iommu_add() which is the last reference place of the global IOMMU array. Signed-off-by: Lu Baolu <baolu.lu@linux.intel.com> --- drivers/iommu/intel/iommu.c | 3 --- 1 file changed, 3 deletions(-) diff --git a/drivers/iommu/intel/iommu.c b/drivers/iommu/intel/iommu.c index 70408c234f5b..d79c48c5fc8c 100644 --- a/drivers/iommu/intel/iommu.c +++ b/drivers/iommu/intel/iommu.c @@ -3463,9 +3463,6 @@ static int intel_iommu_add(struct dmar_drhd_unit *dmaru) int sp, ret; struct intel_iommu *iommu = dmaru->iommu; - if (g_iommus[iommu->seq_id]) - return 0; - ret = intel_cap_audit(CAP_AUDIT_HOTPLUG_DMAR, iommu); if (ret) goto out; -- 2.25.1 ^ permalink raw reply related [flat|nested] 18+ messages in thread
* [PATCH v2 4/6] iommu/vt-d: Remove unnecessary check in intel_iommu_add() @ 2022-07-02 1:56 ` Lu Baolu 0 siblings, 0 replies; 18+ messages in thread From: Lu Baolu @ 2022-07-02 1:56 UTC (permalink / raw) To: Joerg Roedel, Steve Wahl, Kevin Tian Cc: Dimitri Sivanich, Russ Anderson, Mike Travis, Jerry Snitselaar, iommu, linux-kernel, iommu, David Woodhouse The Intel IOMMU hot-add process starts from dmar_device_hotplug(). It uses the global dmar_global_lock to synchronize all the hot-add and hot-remove paths. In the hot-add path, the new IOMMU data structures are allocated firstly by dmar_parse_one_drhd() and then initialized by dmar_hp_add_drhd(). All the IOMMU units are allocated and initialized in the same synchronized path. There is no case where any IOMMU unit is created and then initialized for multiple times. This removes the unnecessary check in intel_iommu_add() which is the last reference place of the global IOMMU array. Signed-off-by: Lu Baolu <baolu.lu@linux.intel.com> --- drivers/iommu/intel/iommu.c | 3 --- 1 file changed, 3 deletions(-) diff --git a/drivers/iommu/intel/iommu.c b/drivers/iommu/intel/iommu.c index 70408c234f5b..d79c48c5fc8c 100644 --- a/drivers/iommu/intel/iommu.c +++ b/drivers/iommu/intel/iommu.c @@ -3463,9 +3463,6 @@ static int intel_iommu_add(struct dmar_drhd_unit *dmaru) int sp, ret; struct intel_iommu *iommu = dmaru->iommu; - if (g_iommus[iommu->seq_id]) - return 0; - ret = intel_cap_audit(CAP_AUDIT_HOTPLUG_DMAR, iommu); if (ret) goto out; -- 2.25.1 _______________________________________________ iommu mailing list iommu@lists.linux-foundation.org https://lists.linuxfoundation.org/mailman/listinfo/iommu ^ permalink raw reply related [flat|nested] 18+ messages in thread
* RE: [PATCH v2 4/6] iommu/vt-d: Remove unnecessary check in intel_iommu_add() 2022-07-02 1:56 ` Lu Baolu (?) @ 2022-07-07 1:01 ` Tian, Kevin -1 siblings, 0 replies; 18+ messages in thread From: Tian, Kevin @ 2022-07-07 1:01 UTC (permalink / raw) To: Lu Baolu, Joerg Roedel, Steve Wahl Cc: David Woodhouse, Jerry Snitselaar, Mike Travis, Dimitri Sivanich, Anderson, Russ, iommu, iommu, linux-kernel > From: Lu Baolu <baolu.lu@linux.intel.com> > Sent: Saturday, July 2, 2022 9:56 AM > > The Intel IOMMU hot-add process starts from dmar_device_hotplug(). It > uses the global dmar_global_lock to synchronize all the hot-add and > hot-remove paths. In the hot-add path, the new IOMMU data structures > are allocated firstly by dmar_parse_one_drhd() and then initialized by > dmar_hp_add_drhd(). All the IOMMU units are allocated and initialized > in the same synchronized path. There is no case where any IOMMU unit > is created and then initialized for multiple times. > > This removes the unnecessary check in intel_iommu_add() which is the > last reference place of the global IOMMU array. > > Signed-off-by: Lu Baolu <baolu.lu@linux.intel.com> Reviewed-by: Kevin Tian <kevin.tian@intel.com> ^ permalink raw reply [flat|nested] 18+ messages in thread
* [PATCH v2 5/6] iommu/vt-d: Remove global g_iommus array 2022-07-02 1:56 ` Lu Baolu @ 2022-07-02 1:56 ` Lu Baolu -1 siblings, 0 replies; 18+ messages in thread From: Lu Baolu @ 2022-07-02 1:56 UTC (permalink / raw) To: Joerg Roedel, Steve Wahl, Kevin Tian Cc: David Woodhouse, Jerry Snitselaar, Mike Travis, Dimitri Sivanich, Russ Anderson, iommu, iommu, linux-kernel, Lu Baolu The g_iommus and g_num_of_iommus is not used anywhere. Remove them to avoid dead code. Signed-off-by: Lu Baolu <baolu.lu@linux.intel.com> Reviewed-by: Kevin Tian <kevin.tian@intel.com> --- drivers/iommu/intel/iommu.c | 44 ------------------------------------- 1 file changed, 44 deletions(-) diff --git a/drivers/iommu/intel/iommu.c b/drivers/iommu/intel/iommu.c index d79c48c5fc8c..73a48e2d4fbe 100644 --- a/drivers/iommu/intel/iommu.c +++ b/drivers/iommu/intel/iommu.c @@ -126,9 +126,6 @@ static inline unsigned long virt_to_dma_pfn(void *p) return page_to_dma_pfn(virt_to_page(p)); } -/* global iommu list, set NULL for ignored DMAR units */ -static struct intel_iommu **g_iommus; - static void __init check_tylersburg_isoch(void); static int rwbf_quirk; @@ -287,9 +284,6 @@ static LIST_HEAD(dmar_satc_units); #define for_each_rmrr_units(rmrr) \ list_for_each_entry(rmrr, &dmar_rmrr_units, list) -/* bitmap for indexing intel_iommus */ -static int g_num_of_iommus; - static void dmar_remove_one_dev_info(struct device *dev); int dmar_disabled = !IS_ENABLED(CONFIG_INTEL_IOMMU_DEFAULT_ON); @@ -1694,8 +1688,6 @@ static void free_dmar_iommu(struct intel_iommu *iommu) iommu->domain_ids = NULL; } - g_iommus[iommu->seq_id] = NULL; - /* free context mapping */ free_context_table(iommu); @@ -2901,36 +2893,6 @@ static int __init init_dmars(void) struct intel_iommu *iommu; int ret; - /* - * for each drhd - * allocate root - * initialize and program root entry to not present - * endfor - */ - for_each_drhd_unit(drhd) { - /* - * lock not needed as this is only incremented in the single - * threaded kernel __init code path all other access are read - * only - */ - if (g_num_of_iommus < DMAR_UNITS_SUPPORTED) { - g_num_of_iommus++; - continue; - } - pr_err_once("Exceeded %d IOMMUs\n", DMAR_UNITS_SUPPORTED); - } - - /* Preallocate enough resources for IOMMU hot-addition */ - if (g_num_of_iommus < DMAR_UNITS_SUPPORTED) - g_num_of_iommus = DMAR_UNITS_SUPPORTED; - - g_iommus = kcalloc(g_num_of_iommus, sizeof(struct intel_iommu *), - GFP_KERNEL); - if (!g_iommus) { - ret = -ENOMEM; - goto error; - } - ret = intel_cap_audit(CAP_AUDIT_STATIC_DMAR, NULL); if (ret) goto free_iommu; @@ -2953,8 +2915,6 @@ static int __init init_dmars(void) intel_pasid_max_id); } - g_iommus[iommu->seq_id] = iommu; - intel_iommu_init_qi(iommu); ret = iommu_init_domains(iommu); @@ -3080,9 +3040,6 @@ static int __init init_dmars(void) free_dmar_iommu(iommu); } - kfree(g_iommus); - -error: return ret; } @@ -3486,7 +3443,6 @@ static int intel_iommu_add(struct dmar_drhd_unit *dmaru) if (iommu->gcmd & DMA_GCMD_TE) iommu_disable_translation(iommu); - g_iommus[iommu->seq_id] = iommu; ret = iommu_init_domains(iommu); if (ret == 0) ret = iommu_alloc_root_entry(iommu); -- 2.25.1 ^ permalink raw reply related [flat|nested] 18+ messages in thread
* [PATCH v2 5/6] iommu/vt-d: Remove global g_iommus array @ 2022-07-02 1:56 ` Lu Baolu 0 siblings, 0 replies; 18+ messages in thread From: Lu Baolu @ 2022-07-02 1:56 UTC (permalink / raw) To: Joerg Roedel, Steve Wahl, Kevin Tian Cc: Dimitri Sivanich, Russ Anderson, Mike Travis, Jerry Snitselaar, iommu, linux-kernel, iommu, David Woodhouse The g_iommus and g_num_of_iommus is not used anywhere. Remove them to avoid dead code. Signed-off-by: Lu Baolu <baolu.lu@linux.intel.com> Reviewed-by: Kevin Tian <kevin.tian@intel.com> --- drivers/iommu/intel/iommu.c | 44 ------------------------------------- 1 file changed, 44 deletions(-) diff --git a/drivers/iommu/intel/iommu.c b/drivers/iommu/intel/iommu.c index d79c48c5fc8c..73a48e2d4fbe 100644 --- a/drivers/iommu/intel/iommu.c +++ b/drivers/iommu/intel/iommu.c @@ -126,9 +126,6 @@ static inline unsigned long virt_to_dma_pfn(void *p) return page_to_dma_pfn(virt_to_page(p)); } -/* global iommu list, set NULL for ignored DMAR units */ -static struct intel_iommu **g_iommus; - static void __init check_tylersburg_isoch(void); static int rwbf_quirk; @@ -287,9 +284,6 @@ static LIST_HEAD(dmar_satc_units); #define for_each_rmrr_units(rmrr) \ list_for_each_entry(rmrr, &dmar_rmrr_units, list) -/* bitmap for indexing intel_iommus */ -static int g_num_of_iommus; - static void dmar_remove_one_dev_info(struct device *dev); int dmar_disabled = !IS_ENABLED(CONFIG_INTEL_IOMMU_DEFAULT_ON); @@ -1694,8 +1688,6 @@ static void free_dmar_iommu(struct intel_iommu *iommu) iommu->domain_ids = NULL; } - g_iommus[iommu->seq_id] = NULL; - /* free context mapping */ free_context_table(iommu); @@ -2901,36 +2893,6 @@ static int __init init_dmars(void) struct intel_iommu *iommu; int ret; - /* - * for each drhd - * allocate root - * initialize and program root entry to not present - * endfor - */ - for_each_drhd_unit(drhd) { - /* - * lock not needed as this is only incremented in the single - * threaded kernel __init code path all other access are read - * only - */ - if (g_num_of_iommus < DMAR_UNITS_SUPPORTED) { - g_num_of_iommus++; - continue; - } - pr_err_once("Exceeded %d IOMMUs\n", DMAR_UNITS_SUPPORTED); - } - - /* Preallocate enough resources for IOMMU hot-addition */ - if (g_num_of_iommus < DMAR_UNITS_SUPPORTED) - g_num_of_iommus = DMAR_UNITS_SUPPORTED; - - g_iommus = kcalloc(g_num_of_iommus, sizeof(struct intel_iommu *), - GFP_KERNEL); - if (!g_iommus) { - ret = -ENOMEM; - goto error; - } - ret = intel_cap_audit(CAP_AUDIT_STATIC_DMAR, NULL); if (ret) goto free_iommu; @@ -2953,8 +2915,6 @@ static int __init init_dmars(void) intel_pasid_max_id); } - g_iommus[iommu->seq_id] = iommu; - intel_iommu_init_qi(iommu); ret = iommu_init_domains(iommu); @@ -3080,9 +3040,6 @@ static int __init init_dmars(void) free_dmar_iommu(iommu); } - kfree(g_iommus); - -error: return ret; } @@ -3486,7 +3443,6 @@ static int intel_iommu_add(struct dmar_drhd_unit *dmaru) if (iommu->gcmd & DMA_GCMD_TE) iommu_disable_translation(iommu); - g_iommus[iommu->seq_id] = iommu; ret = iommu_init_domains(iommu); if (ret == 0) ret = iommu_alloc_root_entry(iommu); -- 2.25.1 _______________________________________________ iommu mailing list iommu@lists.linux-foundation.org https://lists.linuxfoundation.org/mailman/listinfo/iommu ^ permalink raw reply related [flat|nested] 18+ messages in thread
* [PATCH v2 6/6] iommu/vt-d: Make DMAR_UNITS_SUPPORTED default 1024 2022-07-02 1:56 ` Lu Baolu @ 2022-07-02 1:56 ` Lu Baolu -1 siblings, 0 replies; 18+ messages in thread From: Lu Baolu @ 2022-07-02 1:56 UTC (permalink / raw) To: Joerg Roedel, Steve Wahl, Kevin Tian Cc: David Woodhouse, Jerry Snitselaar, Mike Travis, Dimitri Sivanich, Russ Anderson, iommu, iommu, linux-kernel, Lu Baolu If the available hardware exceeds DMAR_UNITS_SUPPORTED (previously set to MAX_IO_APICS, or 128), it causes these messages: "DMAR: Failed to allocate seq_id", "DMAR: Parse DMAR table failure.", and "x2apic: IRQ remapping doesn't support X2APIC mode x2apic disabled"; and the system fails to boot properly. To support up to 64 sockets with 10 DMAR units each (640), make the value of DMAR_UNITS_SUPPORTED default 1024. Signed-off-by: Steve Wahl<steve.wahl@hpe.com> Link: https://lore.kernel.org/linux-iommu/20220615183650.32075-1-steve.wahl@hpe.com/ Signed-off-by: Lu Baolu <baolu.lu@linux.intel.com> Reviewed-by: Kevin Tian <kevin.tian@intel.com> --- include/linux/dmar.h | 6 +----- 1 file changed, 1 insertion(+), 5 deletions(-) diff --git a/include/linux/dmar.h b/include/linux/dmar.h index cbd714a198a0..d81a51978d01 100644 --- a/include/linux/dmar.h +++ b/include/linux/dmar.h @@ -18,11 +18,7 @@ struct acpi_dmar_header; -#ifdef CONFIG_X86 -# define DMAR_UNITS_SUPPORTED MAX_IO_APICS -#else -# define DMAR_UNITS_SUPPORTED 64 -#endif +#define DMAR_UNITS_SUPPORTED 1024 /* DMAR Flags */ #define DMAR_INTR_REMAP 0x1 -- 2.25.1 ^ permalink raw reply related [flat|nested] 18+ messages in thread
* [PATCH v2 6/6] iommu/vt-d: Make DMAR_UNITS_SUPPORTED default 1024 @ 2022-07-02 1:56 ` Lu Baolu 0 siblings, 0 replies; 18+ messages in thread From: Lu Baolu @ 2022-07-02 1:56 UTC (permalink / raw) To: Joerg Roedel, Steve Wahl, Kevin Tian Cc: Dimitri Sivanich, Russ Anderson, Mike Travis, Jerry Snitselaar, iommu, linux-kernel, iommu, David Woodhouse If the available hardware exceeds DMAR_UNITS_SUPPORTED (previously set to MAX_IO_APICS, or 128), it causes these messages: "DMAR: Failed to allocate seq_id", "DMAR: Parse DMAR table failure.", and "x2apic: IRQ remapping doesn't support X2APIC mode x2apic disabled"; and the system fails to boot properly. To support up to 64 sockets with 10 DMAR units each (640), make the value of DMAR_UNITS_SUPPORTED default 1024. Signed-off-by: Steve Wahl<steve.wahl@hpe.com> Link: https://lore.kernel.org/linux-iommu/20220615183650.32075-1-steve.wahl@hpe.com/ Signed-off-by: Lu Baolu <baolu.lu@linux.intel.com> Reviewed-by: Kevin Tian <kevin.tian@intel.com> --- include/linux/dmar.h | 6 +----- 1 file changed, 1 insertion(+), 5 deletions(-) diff --git a/include/linux/dmar.h b/include/linux/dmar.h index cbd714a198a0..d81a51978d01 100644 --- a/include/linux/dmar.h +++ b/include/linux/dmar.h @@ -18,11 +18,7 @@ struct acpi_dmar_header; -#ifdef CONFIG_X86 -# define DMAR_UNITS_SUPPORTED MAX_IO_APICS -#else -# define DMAR_UNITS_SUPPORTED 64 -#endif +#define DMAR_UNITS_SUPPORTED 1024 /* DMAR Flags */ #define DMAR_INTR_REMAP 0x1 -- 2.25.1 _______________________________________________ iommu mailing list iommu@lists.linux-foundation.org https://lists.linuxfoundation.org/mailman/listinfo/iommu ^ permalink raw reply related [flat|nested] 18+ messages in thread
end of thread, other threads:[~2022-07-07 2:22 UTC | newest] Thread overview: 18+ messages (download: mbox.gz / follow: Atom feed) -- links below jump to the message on this page -- 2022-07-02 1:56 [PATCH v2 0/6] iommu/vt-d: Reset DMAR_UNITS_SUPPORTED Lu Baolu 2022-07-02 1:56 ` Lu Baolu 2022-07-02 1:56 ` [PATCH v2 1/6] iommu/vt-d: Remove unused domain_get_iommu() Lu Baolu 2022-07-02 1:56 ` Lu Baolu 2022-07-02 1:56 ` [PATCH v2 2/6] iommu/vt-d: Use IDA interface to manage iommu sequence id Lu Baolu 2022-07-02 1:56 ` Lu Baolu 2022-07-07 0:53 ` Tian, Kevin 2022-07-02 1:56 ` [PATCH v2 3/6] iommu/vt-d: Refactor iommu information of each domain Lu Baolu 2022-07-02 1:56 ` Lu Baolu 2022-07-07 1:01 ` Tian, Kevin 2022-07-07 2:22 ` Baolu Lu 2022-07-02 1:56 ` [PATCH v2 4/6] iommu/vt-d: Remove unnecessary check in intel_iommu_add() Lu Baolu 2022-07-02 1:56 ` Lu Baolu 2022-07-07 1:01 ` Tian, Kevin 2022-07-02 1:56 ` [PATCH v2 5/6] iommu/vt-d: Remove global g_iommus array Lu Baolu 2022-07-02 1:56 ` Lu Baolu 2022-07-02 1:56 ` [PATCH v2 6/6] iommu/vt-d: Make DMAR_UNITS_SUPPORTED default 1024 Lu Baolu 2022-07-02 1:56 ` Lu Baolu
This is an external index of several public inboxes, see mirroring instructions on how to clone and mirror all data and code used by this external index.