[PATCHv2 6/6] Crashdump-Accepting-Active-IOMMU-Call-From-Mainline

From: Bill Sumner
Date: Thu Dec 19 2013 - 21:51:00 EST


At a high level, this code operates primarily during iommu initialization
and device-driver initialization

During intel-iommu hardware initialization:
In intel_iommu_init(void)
* If (This is the crash kernel)
. Set flag: crashdump_accepting_active_iommu (all changes below check this)
. Skip disabling the iommu hardware translations

In init_dmars()
* Duplicate the intel iommu translation tables from the old kernel
in the new kernel
. The root-entry table, all context-entry tables,
and all page-translation-entry tables
. The duplicate tables contain updated physical addresses to link them together.
. The duplicate tables are mapped into kernel virtual addresses
in the new kernel which allows most of the existing iommu code
to operate without change.
. Do some minimal sanity-checks during the copy
. Place the address of the new root-entry structure into "struct intel_iommu"

* Skip setting-up new domains for 'si', 'rmrr', 'isa'
. Translations for 'rmrr' and 'isa' ranges have been copied from the old kernel
. This patch has not yet been tested with iommu pass-through enabled

* Existing (unchanged) code near the end of dmar_init:
. Loads the address of the (now new) root-entry structure from
"struct intel_iommu" into the iommu hardware and does the hardware flushes.
This changes the active translation tables from the ones in the old kernel
to the copies in the new kernel.
. This is legal because the translations in the two sets of tables are
currently identical:
Virtualization Technology for Directed I/O. Architecture Specification,
February 2011, Rev. 1.3 (section 11.2, paragraph 2)

In iommu_init_domains()
* Mark as in-use all domain-id's from the old kernel
. In case the new kernel contains a device that was not in the old kernel
and a new, unused domain-id is actually needed, the bitmap will give us one.

When a new domain is created for a device:
* If (this device has a context in the old kernel)
. Get domain-id, address-width, and IOVA ranges from the old kernel context;
. Get address(page-entry-tables) from the copy in the new kernel;
. And apply all of the above values to the new domain structure.
* Else
. Create a new domain as normal

v1->v2:
Updated patch description

Signed-off-by: Bill Sumner <bill.sumner@xxxxxx>
---
drivers/iommu/intel-iommu.c | 272 +++++++++++++++++++++++++++++++++-----------
1 file changed, 204 insertions(+), 68 deletions(-)

diff --git a/drivers/iommu/intel-iommu.c b/drivers/iommu/intel-iommu.c
index 3b357e2..58f6d87 100644
--- a/drivers/iommu/intel-iommu.c
+++ b/drivers/iommu/intel-iommu.c
@@ -21,6 +21,8 @@
* Author: Fenghua Yu <fenghua.yu@xxxxxxxxx>
*/

+#define DEBUG 1 /* TEMPORARY */
+
#include <linux/init.h>
#include <linux/bitmap.h>
#include <linux/debugfs.h>
@@ -1357,6 +1359,12 @@ static int iommu_init_domains(struct intel_iommu *iommu)
*/
if (cap_caching_mode(iommu->cap))
set_bit(0, iommu->domain_ids);
+
+#ifdef CONFIG_CRASH_DUMP
+ if (crashdump_accepting_active_iommu)
+ intel_iommu_get_dids_from_old_kernel(iommu);
+#endif /* CONFIG_CRASH_DUMP */
+
return 0;
}

@@ -1430,7 +1438,8 @@ static struct dmar_domain *alloc_domain(void)
}

static int iommu_attach_domain(struct dmar_domain *domain,
- struct intel_iommu *iommu)
+ struct intel_iommu *iommu,
+ int domain_number)
{
int num;
unsigned long ndomains;
@@ -1440,12 +1449,15 @@ static int iommu_attach_domain(struct dmar_domain *domain,

spin_lock_irqsave(&iommu->lock, flags);

- num = find_first_zero_bit(iommu->domain_ids, ndomains);
- if (num >= ndomains) {
- spin_unlock_irqrestore(&iommu->lock, flags);
- printk(KERN_ERR "IOMMU: no free domain ids\n");
- return -ENOMEM;
- }
+ if (domain_number < 0) {
+ num = find_first_zero_bit(iommu->domain_ids, ndomains);
+ if (num >= ndomains) {
+ spin_unlock_irqrestore(&iommu->lock, flags);
+ printk(KERN_ERR "IOMMU: no free domain ids\n");
+ return -ENOMEM;
+ }
+ } else
+ num = domain_number;

domain->id = num;
set_bit(num, iommu->domain_ids);
@@ -2056,8 +2068,17 @@ static struct dmar_domain *get_domain_for_dev(struct pci_dev *pdev, int gaw)
int bus = 0, devfn = 0;
int segment;
int ret;
+ int did = -1; /* Default to "no domain_id supplied" */

domain = find_domain(pdev);
+
+#ifdef CONFIG_CRASH_DUMP
+ if (domain)
+ if (pr_dbg.in_crashdump && crashdump_accepting_active_iommu)
+ pr_debug("IOMMU: Found domain (%d) for device %s\n",
+ domain->id, pci_name(pdev));
+#endif /* CONFIG_CRASH_DUMP */
+
if (domain)
return domain;

@@ -2088,6 +2109,12 @@ static struct dmar_domain *get_domain_for_dev(struct pci_dev *pdev, int gaw)
}
}

+#ifdef CONFIG_CRASH_DUMP
+ if (pr_dbg.in_crashdump && crashdump_accepting_active_iommu)
+ pr_debug("IOMMU: Allocating new domain for device %s\n",
+ pci_name(pdev));
+#endif /* CONFIG_CRASH_DUMP */
+
domain = alloc_domain();
if (!domain)
goto error;
@@ -2102,7 +2129,26 @@ static struct dmar_domain *get_domain_for_dev(struct pci_dev *pdev, int gaw)
}
iommu = drhd->iommu;

- ret = iommu_attach_domain(domain, iommu);
+#ifdef CONFIG_CRASH_DUMP
+ /* See if this device had a did & gaw in the old kernel */
+ if (crashdump_accepting_active_iommu) {
+ did = domain_get_did_from_old_kernel(iommu, pdev);
+ if (did > 0 || (did == 0 && !cap_caching_mode(iommu->cap))) {
+ ret = domain_get_gaw_from_old_kernel(iommu, pdev);
+ if (ret > 0)
+ gaw = ret;
+ else
+ did = -1;
+ } else
+ did = -1;
+ }
+
+ if (pr_dbg.in_crashdump && crashdump_accepting_active_iommu)
+ pr_debug("IOMMU: new domain for device %s: gaw(%d) did(%d)\n",
+ pci_name(pdev), gaw, did);
+#endif /* CONFIG_CRASH_DUMP */
+
+ ret = iommu_attach_domain(domain, iommu, did);
if (ret) {
free_domain_mem(domain);
goto error;
@@ -2113,6 +2159,23 @@ static struct dmar_domain *get_domain_for_dev(struct pci_dev *pdev, int gaw)
goto error;
}

+#ifdef CONFIG_CRASH_DUMP
+ if (crashdump_accepting_active_iommu && did >= 0) {
+ u64 temp_pgd; /* Top page-translation-table */
+
+ domain_get_ranges_from_old_kernel(domain, iommu, pdev);
+
+ temp_pgd = domain_get_pgd_from_old_kernel(iommu, pdev);
+ if (temp_pgd) {
+ if (domain->pgd)
+ free_pgtable_page(domain->pgd);
+ domain->pgd = (struct dma_pte *)temp_pgd;
+ }
+ pr_debug("IOMMU: New Domain for device %s Did:%d Pgd: 0x%12.12llx\n",
+ pci_name(pdev), did, temp_pgd);
+ }
+#endif /* CONFIG_CRASH_DUMP */
+
/* register pcie-to-pci device */
if (dev_tmp) {
info = alloc_devinfo_mem();
@@ -2323,7 +2386,7 @@ static int __init si_domain_init(int hw)
pr_debug("Identity mapping domain is domain %d\n", si_domain->id);

for_each_active_iommu(iommu, drhd) {
- ret = iommu_attach_domain(si_domain, iommu);
+ ret = iommu_attach_domain(si_domain, iommu, (int) -1);
if (ret) {
domain_exit(si_domain);
return -EFAULT;
@@ -2531,6 +2594,10 @@ static int __init init_dmars(void)
struct pci_dev *pdev;
struct intel_iommu *iommu;
int i, ret;
+#ifdef CONFIG_CRASH_DUMP
+ struct root_entry *root_old_phys;
+ struct root_entry *root_new_virt;
+#endif /* CONFIG_CRASH_DUMP */

/*
* for each drhd
@@ -2578,16 +2645,41 @@ static int __init init_dmars(void)
if (ret)
goto error;

- /*
- * TBD:
- * we could share the same root & context tables
- * among all IOMMU's. Need to Split it later.
- */
- ret = iommu_alloc_root_entry(iommu);
- if (ret) {
- printk(KERN_ERR "IOMMU: allocate root entry failed\n");
- goto error;
+#ifdef CONFIG_CRASH_DUMP
+ if (crashdump_accepting_active_iommu) {
+ print_intel_iommu_registers(drhd);
+
+ pr_debug("Calling copy_intel_iommu_translation_tables\n");
+ pr_debug("(lists tables in OLD KERNEL during copy)\n");
+ ret = copy_intel_iommu_translation_tables(drhd,
+ &root_old_phys, &root_new_virt);
+ if (ret) {
+ pr_err("IOMMU: Copy translate tables failed\n");
+
+ /* Best to stop trying */
+ crashdump_accepting_active_iommu = false;
+ goto error;
+ }
+ iommu->root_entry = root_new_virt;
+ pr_debug("IOMMU: root_new_virt:0x%12.12llx phys:0x%12.12llx\n",
+ (u64)root_new_virt,
+ virt_to_phys(root_new_virt));
+ } else {
+#endif /* CONFIG_CRASH_DUMP */
+ /*
+ * TBD:
+ * we could share the same root & context tables
+ * among all IOMMU's. Need to Split it later.
+ */
+ ret = iommu_alloc_root_entry(iommu);
+ if (ret) {
+ printk(KERN_ERR "IOMMU: allocate root entry failed\n");
+ goto error;
+ }
+#ifdef CONFIG_CRASH_DUMP
}
+#endif /* CONFIG_CRASH_DUMP */
+
if (!ecap_pass_through(iommu->ecap))
hw_pass_through = 0;
}
@@ -2656,50 +2748,69 @@ static int __init init_dmars(void)

check_tylersburg_isoch();

- /*
- * If pass through is not set or not enabled, setup context entries for
- * identity mappings for rmrr, gfx, and isa and may fall back to static
- * identity mapping if iommu_identity_mapping is set.
- */
- if (iommu_identity_mapping) {
- ret = iommu_prepare_static_identity_mapping(hw_pass_through);
- if (ret) {
- printk(KERN_CRIT "Failed to setup IOMMU pass-through\n");
- goto error;
+#ifdef CONFIG_CRASH_DUMP
+ if (!crashdump_accepting_active_iommu) {
+ /* Skip setting-up new domains for si, rmrr, and the isa bus
+ * on the expectation that these translations
+ * were copied from the old kernel.
+ *
+ * NOTE: Indented the existing code below because it is now
+ * conditional upon the 'if' statement above.
+ * This pushed many of the lines over 80 characters.
+ * Chose to leave them and live with the 'checkpatch' warnings
+ * about "over 80 characters".
+ */
+#endif /* CONFIG_CRASH_DUMP */
+ /*
+ * If pass through is not set or not enabled, setup context entries for
+ * identity mappings for rmrr, gfx, and isa and may fall back to static
+ * identity mapping if iommu_identity_mapping is set.
+ */
+ if (iommu_identity_mapping) {
+ ret = iommu_prepare_static_identity_mapping(hw_pass_through);
+ if (ret) {
+ printk(KERN_CRIT "Failed to setup IOMMU pass-through\n");
+ goto error;
+ }
}
- }
- /*
- * For each rmrr
- * for each dev attached to rmrr
- * do
- * locate drhd for dev, alloc domain for dev
- * allocate free domain
- * allocate page table entries for rmrr
- * if context not allocated for bus
- * allocate and init context
- * set present in root table for this bus
- * init context with domain, translation etc
- * endfor
- * endfor
- */
- printk(KERN_INFO "IOMMU: Setting RMRR:\n");
- for_each_rmrr_units(rmrr) {
- for (i = 0; i < rmrr->devices_cnt; i++) {
- pdev = rmrr->devices[i];
- /*
- * some BIOS lists non-exist devices in DMAR
- * table.
- */
- if (!pdev)
- continue;
- ret = iommu_prepare_rmrr_dev(rmrr, pdev);
- if (ret)
- printk(KERN_ERR
- "IOMMU: mapping reserved region failed\n");
+ /*
+ * For each rmrr
+ * for each dev attached to rmrr
+ * do
+ * locate drhd for dev, alloc domain for dev
+ * allocate free domain
+ * allocate page table entries for rmrr
+ * if context not allocated for bus
+ * allocate and init context
+ * set present in root table for this bus
+ * init context with domain, translation etc
+ * endfor
+ * endfor
+ */
+ printk(KERN_INFO "IOMMU: Setting RMRR:\n");
+ for_each_rmrr_units(rmrr) {
+ for (i = 0; i < rmrr->devices_cnt; i++) {
+ pdev = rmrr->devices[i];
+ /*
+ * some BIOS lists non-exist devices in DMAR
+ * table.
+ */
+ if (!pdev)
+ continue;
+ ret = iommu_prepare_rmrr_dev(rmrr, pdev);
+ if (ret)
+ printk(KERN_ERR
+ "IOMMU: mapping reserved region failed\n");
+ }
}
- }

- iommu_prepare_isa();
+ iommu_prepare_isa();
+#ifdef CONFIG_CRASH_DUMP
+ } else {
+ intel_iommu_translation_tables_are_mapped = true;
+ pr_debug("intel_iommu_translation_tables_are_mapped = true\n");
+ }
+#endif /* CONFIG_CRASH_DUMP */

/*
* for each drhd
@@ -2893,6 +3004,12 @@ static dma_addr_t __intel_map_single(struct device *hwdev, phys_addr_t paddr,

BUG_ON(dir == DMA_NONE);

+#ifdef CONFIG_CRASH_DUMP
+ if (pr_dbg.in_crashdump && crashdump_accepting_active_iommu)
+ pr_debug("ENTER %s paddr(0x%12.12llx) size(0x%12.12lx)\n",
+ __func__, paddr, size);
+#endif /* CONFIG_CRASH_DUMP */
+
if (iommu_no_mapping(hwdev))
return paddr;

@@ -2935,6 +3052,12 @@ static dma_addr_t __intel_map_single(struct device *hwdev, phys_addr_t paddr,

start_paddr = (phys_addr_t)iova->pfn_lo << PAGE_SHIFT;
start_paddr += paddr & ~PAGE_MASK;
+
+#ifdef CONFIG_CRASH_DUMP
+ if (pr_dbg.in_crashdump && crashdump_accepting_active_iommu)
+ pr_debug("LEAVE %s dma_addr_t(0x%16.16llx)\n",
+ __func__, start_paddr);
+#endif /* CONFIG_CRASH_DUMP */
return start_paddr;

error:
@@ -3754,19 +3877,32 @@ int __init intel_iommu_init(void)
return -ENODEV;
}

+#ifdef CONFIG_CRASH_DUMP
/*
- * Disable translation if already enabled prior to OS handover.
+ * If (This is the crash kernel)
+ * Set: copy iommu translate tables from old kernel
+ * Skip disabling the iommu hardware translations
*/
- for_each_drhd_unit(drhd) {
- struct intel_iommu *iommu;
+ if (is_kdump_kernel()) {
+ crashdump_accepting_active_iommu = true;
+ pr_info("IOMMU crashdump_accepting_active_iommu = true\n");
+ pr_info("IOMMU Skip disabling iommu hardware translations\n");
+ } else
+#endif /* CONFIG_CRASH_DUMP */
+ /*
+ * Disable translation if already enabled prior to OS handover.
+ */
+ for_each_drhd_unit(drhd) {
+ struct intel_iommu *iommu;

- if (drhd->ignored)
- continue;
+ if (drhd->ignored)
+ continue;
+
+ iommu = drhd->iommu;
+ if (iommu->gcmd & DMA_GCMD_TE)
+ iommu_disable_translation(iommu);
+ }

- iommu = drhd->iommu;
- if (iommu->gcmd & DMA_GCMD_TE)
- iommu_disable_translation(iommu);
- }

if (dmar_dev_scope_init() < 0) {
if (force_on)
--
Bill Sumner <bill.sumner@xxxxxx>

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/