iommu: Enhance IOMMU default DMA mode build options
Subsystems (4)
v14 cover iommu: Enhance IOMMU default DMA mode build options
This is a reboot of Zhen Lei's series from a couple of years ago, which never made it across the line.
I still think that it has some value, so taking up the mantle.
Motivation:
Allow lazy mode be default mode for DMA domains for all ARCHs, and not only those who hardcode it (to be lazy). For ARM64, currently we must use a kernel command line parameter to use lazy mode, which is less than ideal.
I have now included the print for strict/lazy mode, which I originally sent in:
https://lore.kernel.org/linux-iommu/[email protected]/
There was some concern there about drivers and their custom prints conflicting with the print in that patch, but I think that it should be ok.
Based on next-20210611 + "iommu: Update "iommu.strict" documentation"
Differences to v13:
- Improve strict mode deprecation messages and cut out some kernel-parameters.txt legacy description
- Add tag in 1/6
- use pr_info_once() for vt-d message about VM and caching
Differences to v12:
- Add Robin's RB tags (thanks!)
- Add a patch to mark x86 strict cmdline params as deprecated
- Improve wording in Kconfig change and tweak iommu_dma_strict declaration
Differences to v11:
- Rebase to next-20210610
- Drop strict mode globals in Intel and AMD drivers
- Include patch to print strict vs lazy mode
- Include patch to remove argument from iommu_set_dma_strict()
John Garry (3):
iommu: Deprecate Intel and AMD cmdline methods to enable strict mode
iommu: Print strict or lazy mode at init time
iommu: Remove mode argument from iommu_set_dma_strict()
Zhen Lei (3):
iommu: Enhance IOMMU default DMA mode build options iommu/vt-d: Add support for IOMMU default DMA mode build options iommu/amd: Add support for IOMMU default DMA mode build options
.../admin-guide/kernel-parameters.txt | 12 ++---- drivers/iommu/Kconfig | 41 +++++++++++++++++++ drivers/iommu/amd/amd_iommu_types.h | 6 --- drivers/iommu/amd/init.c | 7 ++-- drivers/iommu/amd/iommu.c | 6 --- drivers/iommu/intel/iommu.c | 16 ++++---- drivers/iommu/iommu.c | 12 ++++-- include/linux/iommu.h | 2 +- 8 files changed, 65 insertions(+), 37 deletions(-)
Signature · 1 line
2.26.2
PATCH v14 1/6 iommu: Deprecate Intel and AMD cmdline methods to enable strict mode
Now that the x86 drivers support iommu.strict, deprecate the custom methods.
--- Documentation/admin-guide/kernel-parameters.txt | 9 ++------- drivers/iommu/amd/init.c | 4 +++- drivers/iommu/intel/iommu.c | 1 + 3 files changed, 6 insertions(+), 8 deletions(-)
Documentation/admin-guide/kernel-parameters.txt +2 -7
@@ -290,10 +290,7 @@ amd_iommu= [HW,X86-64] Pass parameters to the AMD IOMMU driver in the system. Possible values are: - fullflush - enable flushing of IO/TLB entries when - they are unmapped. Otherwise they are - flushed before they will be reused, which - is a lot of faster + fullflush - Deprecated, equivalent to iommu.strict=1 off - do not initialize any AMD IOMMU found in the system force_isolation - Force device isolation for all
@@ -1948,9 +1945,7 @@ this case, gfx device will use physical address for DMA. strict [Default Off] - With this option on every unmap_single operation will - result in a hardware IOTLB flush operation as opposed - to batching them for performance. + Deprecated, equivalent to iommu.strict=1. sp_off [Default Off] By default, super page will be supported if Intel IOMMU has the capability. With this option, super page will
drivers/iommu/amd/init.c +3 -1
@@ -3098,8 +3098,10 @@ static int __init parse_amd_iommu_intr(char *str)static int __init parse_amd_iommu_options (char *str) {for (; *str; ++str) { -if (strncmp (str,"fullflush" ,9 ) ==0 ) +if (strncmp (str,"fullflush" ,9 ) ==0 ) { +pr_warn ("amd_iommu=fullflush deprecated; use iommu.strict=1 instead\n" ); amd_iommu_unmap_flush = true; + }if (strncmp (str,"force_enable" ,12 ) ==0 ) amd_iommu_force_enable = true;if (strncmp (str,"off" ,3 ) ==0 )
drivers/iommu/intel/iommu.c +1 -0
@@ -454,6 +454,7 @@ static int __init intel_iommu_setup(char *str)pr_warn ("intel_iommu=forcedac deprecated; use iommu.forcedac instead\n" ); iommu_dma_forcedac = true; }else if (!strncmp (str,"strict" ,6 )) { +pr_warn ("intel_iommu=strict deprecated; use iommu.strict=1 instead\n" );pr_info ("Disable batched IOTLB flush\n" ); intel_iommu_strict =1 ; }else if (!strncmp (str,"sp_off" ,6 )) {
Signature · 1 line
2.26.2
Re: [PATCH v14 1/6] iommu: Deprecate Intel and AMD cmdline methods to enable strict mode
Quoted 60 lines · John Garry
Now that the x86 drivers support iommu.strict, deprecate the custom methods.
Signed-off-by John Garry <[email protected]>Acked-by Robin Murphy <[email protected]>--- Documentation/admin-guide/kernel-parameters.txt | 9 ++------- drivers/iommu/amd/init.c | 4 +++- drivers/iommu/intel/iommu.c | 1 + 3 files changed, 6 insertions(+), 8 deletions(-)Documentation/admin-guide/kernel-parameters.txt +2
-7@@ -290,10 +290,7 @@ amd_iommu= [HW,X86-64] Pass parameters to the AMD IOMMU driver in the system. Possible values are: - fullflush - enable flushing of IO/TLB entries when - they are unmapped. Otherwise they are - flushed before they will be reused, which - is a lot of faster + fullflush - Deprecated, equivalent to iommu.strict=1 off - do not initialize any AMD IOMMU found in the system force_isolation - Force device isolation for all@@ -1948,9 +1945,7 @@ this case, gfx device will use physical address for DMA. strict [Default Off] - With this option on every unmap_single operation will - result in a hardware IOTLB flush operation as opposed - to batching them for performance. + Deprecated, equivalent to iommu.strict=1. sp_off [Default Off] By default, super page will be supported if Intel IOMMU has the capability. With this option, super page willdrivers/iommu/amd/init.c +3
-1@@ -3098,8 +3098,10 @@ static int __init parse_amd_iommu_intr(char *str)static int __init parse_amd_iommu_options (char *str) {for (; *str; ++str) { -if (strncmp (str,"fullflush" ,9 ) ==0 ) +if (strncmp (str,"fullflush" ,9 ) ==0 ) { +pr_warn ("amd_iommu=fullflush deprecated; use iommu.strict=1 instead\n" ); amd_iommu_unmap_flush = true; + }if (strncmp (str,"force_enable" ,12 ) ==0 ) amd_iommu_force_enable = true;if (strncmp (str,"off" ,3 ) ==0 )drivers/iommu/intel/iommu.c +0
-0
drivers/iommu/intel/iommu.c static int __init intel_iommu_setup(char *str) · line 460
Quoted 7 lines
drivers/iommu/intel/iommu.c +1
-0@@ -454,6 +454,7 @@ static int __init intel_iommu_setup(char *str)pr_warn ("intel_iommu=forcedac deprecated; use iommu.forcedac instead\n" ); iommu_dma_forcedac = true; }else if (!strncmp (str,"strict" ,6 )) { +pr_warn ("intel_iommu=strict deprecated; use iommu.strict=1 instead\n" );pr_info ("Disable batched IOTLB flush\n" ); intel_iommu_strict =1 ; }else if (!strncmp (str,"sp_off" ,6 )) {
Best regards,
baolu
PATCH v14 2/6 iommu: Print strict or lazy mode at init time
In reply to John GarryAs well as the default domain type, it's useful to know whether strict or lazy for DMA domains, so add this info in a separate print.
The (stict/lazy) mode may be also set via iommu.strict earlyparm, but this will be processed prior to iommu_subsys_init(), so that print will be accurate for drivers which don't set the mode via custom means.
For the drivers which set the mode via custom means - AMD and Intel drivers
- they maintain prints to inform a change in policy or that custom cmdline methods to change policy are deprecated.
--- drivers/iommu/iommu.c | 5 +++++ 1 file changed, 5 insertions(+)
drivers/iommu/iommu.c +5 -0
@@ -138,6 +138,11 @@ static int __init iommu_subsys_init(void) (iommu_cmd_line &IOMMU_CMD_LINE_DMA_API ) ? "(set via kernel command line)" : ""); +pr_info ("DMA domain TLB invalidation policy: %s mode %s\n" , + iommu_dma_strict ?"strict" :"lazy" , + (iommu_cmd_line &IOMMU_CMD_LINE_STRICT ) ? +"(set via kernel command line)" :"" ); +return 0 ; }subsys_initcall (iommu_subsys_init);
Signature · 1 line
2.26.2
Re: [PATCH v14 2/6] iommu: Print strict or lazy mode at init time
Quoted 24 lines · John Garry
As well as the default domain type, it's useful to know whether strict or lazy for DMA domains, so add this info in a separate print.
The (stict/lazy) mode may be also set via iommu.strict earlyparm, but this will be processed prior to
iommu_subsys_init(), so that print will be accurate for drivers which don't set the mode via custom means.For the drivers which set the mode via custom means - AMD and Intel drivers
- they maintain prints to inform a change in policy or that custom cmdline methods to change policy are deprecated.Signed-off-by John Garry <[email protected]>Reviewed-by Robin Murphy <[email protected]>--- drivers/iommu/iommu.c | 5 +++++ 1 file changed, 5 insertions(+)drivers/iommu/iommu.c +0
-0@@ -138,6 +138,11 @@ static int __init iommu_subsys_init(void) (iommu_cmd_line & IOMMU_CMD_LINE_DMA_API) ?
drivers/iommu/iommu.c static int __init iommu_subsys_init(void) · line 148
Quoted 10 lines
drivers/iommu/iommu.c +5
-0"(set via kernel command line)" : ""); +pr_info ("DMA domain TLB invalidation policy: %s mode %s\n" , + iommu_dma_strict ?"strict" :"lazy" , + (iommu_cmd_line &IOMMU_CMD_LINE_STRICT ) ? +"(set via kernel command line)" :"" ); +return 0 ; }subsys_initcall (iommu_subsys_init);
Best regards,
baolu
PATCH v14 3/6 iommu: Enhance IOMMU default DMA mode build options
In reply to John GarryFrom: Zhen Lei <[email protected]>
First, add build options IOMMU_DEFAULT_{LAZY|STRICT}, so that we have the opportunity to set {lazy|strict} mode as default at build time. Then put the two config options in an choice, as they are mutually exclusive.
--- .../admin-guide/kernel-parameters.txt | 3 +- drivers/iommu/Kconfig | 40 +++++++++++++++++++ drivers/iommu/iommu.c | 2 +- 3 files changed, 43 insertions(+), 2 deletions(-)
Documentation/admin-guide/kernel-parameters.txt +2 -1
@@ -2046,9 +2046,10 @@ throughput at the cost of reduced device isolation. Will fall back to strict mode if not supported by the relevant IOMMU driver. - 1 - Strict mode (default). + 1 - Strict mode. DMA unmap operations invalidate IOMMU hardware TLBs synchronously. + unset - Use value of CONFIG_IOMMU_DEFAULT_{LAZY,STRICT}. Note: on x86, the default behaviour depends on the equivalent driver-specific parameters, but a strict mode explicitly specified by either method takes
drivers/iommu/Kconfig +40 -0
@@ -90,6 +90,46 @@ config IOMMU_DEFAULT_PASSTHROUGH If unsure, say N here. +choice + prompt "IOMMU default DMA IOTLB invalidation mode" + depends on IOMMU_DMA + + default IOMMU_DEFAULT_STRICT + help + This option allows an IOMMU DMA IOTLB invalidation mode to be + chosen at build time, to override the default mode of each ARCH, + removing the need to pass in kernel parameters through command line. + It is still possible to provide common boot params to override this + config. + + If unsure, keep the default. + +config IOMMU_DEFAULT_STRICT + bool "strict" + help + For every IOMMU DMA unmap operation, the flush operation of IOTLB and + the free operation of IOVA are guaranteed to be done in the unmap + function. + +config IOMMU_DEFAULT_LAZY + bool "lazy" + help + Support lazy mode, where for every IOMMU DMA unmap operation, the + flush operation of IOTLB and the free operation of IOVA are deferred. + They are only guaranteed to be done before the related IOVA will be + reused. + + The isolation provided in this mode is not as secure as STRICT mode, + such that a vulnerable time window may be created between the DMA + unmap and the mappings cached in the IOMMU IOTLB or device TLB + finally being invalidated, where the device could still access the + memory which has already been unmapped by the device driver. + However this mode may provide better performance in high throughput + scenarios, and is still considerably more secure than passthrough + mode or no IOMMU. + +endchoice + config OF_IOMMU def_bool y depends on OF && IOMMU_API
drivers/iommu/iommu.c +1 -1
@@ -29,7 +29,7 @@ static struct kset *iommu_group_kset;static DEFINE_IDA (iommu_group_ida);static unsigned int iommu_def_domain_type__read_mostly ; -static bool iommu_dma_strict__read_mostly = true; +static bool iommu_dma_strict__read_mostly =IS_ENABLED (CONFIG_IOMMU_DEFAULT_STRICT );static u32 iommu_cmd_line__read_mostly ;struct iommu_group {
Signature · 1 line
2.26.2
Quoted 89 lines · John Garry
From: Zhen Lei <[email protected]>
First, add build options IOMMU_DEFAULT_{LAZY|STRICT}, so that we have the opportunity to set {lazy|strict} mode as default at build time. Then put the two config options in an choice, as they are mutually exclusive.
[jpg: Make choice between strict and lazy only (and not passthrough)]Signed-off-by Zhen Lei <[email protected]>Signed-off-by John Garry <[email protected]>Reviewed-by Robin Murphy <[email protected]>--- .../admin-guide/kernel-parameters.txt | 3 +- drivers/iommu/Kconfig | 40 +++++++++++++++++++ drivers/iommu/iommu.c | 2 +- 3 files changed, 43 insertions(+), 2 deletions(-)Documentation/admin-guide/kernel-parameters.txt +2
-1@@ -2046,9 +2046,10 @@ throughput at the cost of reduced device isolation. Will fall back to strict mode if not supported by the relevant IOMMU driver. - 1 - Strict mode (default). + 1 - Strict mode. DMA unmap operations invalidate IOMMU hardware TLBs synchronously. + unset - Use value of CONFIG_IOMMU_DEFAULT_{LAZY,STRICT}. Note: on x86, the default behaviour depends on the equivalent driver-specific parameters, but a strict mode explicitly specified by either method takesdrivers/iommu/Kconfig +40
-0@@ -90,6 +90,46 @@ config IOMMU_DEFAULT_PASSTHROUGH If unsure, say N here. +choice + prompt "IOMMU default DMA IOTLB invalidation mode" + depends on IOMMU_DMA + + default IOMMU_DEFAULT_STRICT + help + This option allows an IOMMU DMA IOTLB invalidation mode to be + chosen at build time, to override the default mode of each ARCH, + removing the need to pass in kernel parameters through command line. + It is still possible to provide common boot params to override this + config. + + If unsure, keep the default. + +config IOMMU_DEFAULT_STRICT + bool "strict" + help + For every IOMMU DMA unmap operation, the flush operation of IOTLB and + the free operation of IOVA are guaranteed to be done in the unmap + function. + +config IOMMU_DEFAULT_LAZY + bool "lazy" + help + Support lazy mode, where for every IOMMU DMA unmap operation, the + flush operation of IOTLB and the free operation of IOVA are deferred. + They are only guaranteed to be done before the related IOVA will be + reused. + + The isolation provided in this mode is not as secure as STRICT mode, + such that a vulnerable time window may be created between the DMA + unmap and the mappings cached in the IOMMU IOTLB or device TLB + finally being invalidated, where the device could still access the + memory which has already been unmapped by the device driver. + However this mode may provide better performance in high throughput + scenarios, and is still considerably more secure than passthrough + mode or no IOMMU. + +endchoice + config OF_IOMMU def_bool y depends on OF && IOMMU_APIdrivers/iommu/iommu.c +0
-0
drivers/iommu/iommu.c static struct kset *iommu_group_kset; · line 35
Quoted 8 lines
drivers/iommu/iommu.c +1
-1@@ -29,7 +29,7 @@ static struct kset *iommu_group_kset;static DEFINE_IDA (iommu_group_ida);static unsigned int iommu_def_domain_type__read_mostly ; -static bool iommu_dma_strict__read_mostly = true; +static bool iommu_dma_strict__read_mostly =IS_ENABLED (CONFIG_IOMMU_DEFAULT_STRICT );static u32 iommu_cmd_line__read_mostly ;struct iommu_group {
Best regards,
baolu
PATCH v14 4/6 iommu/vt-d: Add support for IOMMU default DMA mode build options
In reply to John GarryFrom: Zhen Lei <[email protected]>
Make IOMMU_DEFAULT_LAZY default for when INTEL_IOMMU config is set, as is current behaviour.
Also delete global flag intel_iommu_strict:
- In intel_iommu_setup(), call iommu_set_dma_strict(true) directly. Also remove the print, as iommu_subsys_init() prints the mode and we have already marked this param as deprecated.
- For cap_caching_mode() check in intel_iommu_setup(), call iommu_set_dma_strict(true) directly; also reword the accompanying print with a level downgrade and also add the missing '\n'.
- For Ironlake GPU, again call iommu_set_dma_strict(true) directly and keep the accompanying print.
--- drivers/iommu/Kconfig | 1 + drivers/iommu/intel/iommu.c | 15 ++++++--------- 2 files changed, 7 insertions(+), 9 deletions(-)
drivers/iommu/Kconfig +1 -0
@@ -94,6 +94,7 @@ choice prompt "IOMMU default DMA IOTLB invalidation mode" depends on IOMMU_DMA + default IOMMU_DEFAULT_LAZY if INTEL_IOMMU default IOMMU_DEFAULT_STRICT help This option allows an IOMMU DMA IOTLB invalidation mode to be
drivers/iommu/intel/iommu.c +6 -9
@@ -361,7 +361,6 @@ int intel_iommu_enabled = 0;EXPORT_SYMBOL_GPL (intel_iommu_enabled);static int dmar_map_gfx =1 ; -static int intel_iommu_strict;static int intel_iommu_superpage =1 ;static int iommu_identity_mapping;static int iommu_skip_te_disable;
@@ -455,8 +454,7 @@ static int __init intel_iommu_setup(char *str) iommu_dma_forcedac = true; }else if (!strncmp (str,"strict" ,6 )) {pr_warn ("intel_iommu=strict deprecated; use iommu.strict=1 instead\n" ); -pr_info ("Disable batched IOTLB flush\n" ); - intel_iommu_strict =1 ; +iommu_set_dma_strict (true); }else if (!strncmp (str,"sp_off" ,6 )) {pr_info ("Disable supported super page\n" ); intel_iommu_superpage =0 ;
@@ -4382,9 +4380,9 @@ int __init intel_iommu_init(void)* is likely to be much lower than the overhead of synchronizing * the virtual and physical IOMMU page-tables. */ -if (!intel_iommu_strict &&cap_caching_mode (iommu->cap)) { -pr_warn ("IOMMU batching is disabled due to virtualization" ); - intel_iommu_strict =1 ; +if (cap_caching_mode (iommu->cap)) { +pr_info_once ("IOMMU batching disallowed due to virtualization\n" ); +iommu_set_dma_strict (true); } iommu_device_sysfs_add(&iommu->iommu, NULL, intel_iommu_groups,
@@ -4393,7 +4391,6 @@ int __init intel_iommu_init(void) }up_read (&dmar_global_lock); -iommu_set_dma_strict (intel_iommu_strict);bus_set_iommu (&pci_bus_type, &intel_iommu_ops);if (si_domain && !hw_pass_through)register_memory_notifier (&intel_iommu_memory_nb);
@@ -5702,8 +5699,8 @@ static void quirk_calpella_no_shadow_gtt(struct pci_dev *dev) }else if (dmar_map_gfx) {/* we have to ensure the gfx device is idle before we flush */ pci_info (dev,"Disabling batched IOTLB flush on Ironlake\n" ); - intel_iommu_strict =1 ; - } +iommu_set_dma_strict (true); + } }DECLARE_PCI_FIXUP_HEADER (PCI_VENDOR_ID_INTEL ,0x0040 , quirk_calpella_no_shadow_gtt);DECLARE_PCI_FIXUP_HEADER (PCI_VENDOR_ID_INTEL ,0x0044 , quirk_calpella_no_shadow_gtt);
Signature · 1 line
2.26.2
Re: [PATCH v14 4/6] iommu/vt-d: Add support for IOMMU default DMA mode build options
Quoted 82 lines · John Garry
From: Zhen Lei <[email protected]>
Make
IOMMU_DEFAULT_LAZYdefault for whenINTEL_IOMMUconfig is set, as is current behaviour.Also delete global flag
intel_iommu_strict:
- Inintel_iommu_setup(), calliommu_set_dma_strict(true)directly. Also remove the print, asiommu_subsys_init()prints the mode and we have already marked this param as deprecated.- For
cap_caching_mode()check inintel_iommu_setup(), calliommu_set_dma_strict(true)directly; also reword the accompanying print with a level downgrade and also add the missing '\n'.- For Ironlake GPU, again call
iommu_set_dma_strict(true)directly and keep the accompanying print.[jpg: Remove intel_iommu_strict]Signed-off-by Zhen Lei <[email protected]>Signed-off-by John Garry <[email protected]>--- drivers/iommu/Kconfig | 1 + drivers/iommu/intel/iommu.c | 15 ++++++--------- 2 files changed, 7 insertions(+), 9 deletions(-)drivers/iommu/Kconfig +1
-0@@ -94,6 +94,7 @@ choice prompt "IOMMU default DMA IOTLB invalidation mode" depends on IOMMU_DMA + default IOMMU_DEFAULT_LAZY if INTEL_IOMMU default IOMMU_DEFAULT_STRICT help This option allows an IOMMU DMA IOTLB invalidation mode to bedrivers/iommu/intel/iommu.c +4
-7@@ -361,7 +361,6 @@ int intel_iommu_enabled = 0;EXPORT_SYMBOL_GPL (intel_iommu_enabled);static int dmar_map_gfx =1 ; -static int intel_iommu_strict;static int intel_iommu_superpage =1 ;static int iommu_identity_mapping;static int iommu_skip_te_disable;@@ -455,8 +454,7 @@ static int __init intel_iommu_setup(char *str) iommu_dma_forcedac = true; }else if (!strncmp (str,"strict" ,6 )) {pr_warn ("intel_iommu=strict deprecated; use iommu.strict=1 instead\n" ); -pr_info ("Disable batched IOTLB flush\n" ); - intel_iommu_strict =1 ; +iommu_set_dma_strict (true); }else if (!strncmp (str,"sp_off" ,6 )) {pr_info ("Disable supported super page\n" ); intel_iommu_superpage =0 ;@@ -4382,9 +4380,9 @@ int __init intel_iommu_init(void)* is likely to be much lower than the overhead of synchronizing * the virtual and physical IOMMU page-tables. */ -if (!intel_iommu_strict &&cap_caching_mode (iommu->cap)) { -pr_warn ("IOMMU batching is disabled due to virtualization" ); - intel_iommu_strict =1 ; +if (cap_caching_mode (iommu->cap)) { +pr_info_once ("IOMMU batching disallowed due to virtualization\n" ); +iommu_set_dma_strict (true); } iommu_device_sysfs_add(&iommu->iommu, NULL, intel_iommu_groups,@@ -4393,7 +4391,6 @@ int __init intel_iommu_init(void) }up_read (&dmar_global_lock); -iommu_set_dma_strict (intel_iommu_strict);bus_set_iommu (&pci_bus_type, &intel_iommu_ops);if (si_domain && !hw_pass_through)register_memory_notifier (&intel_iommu_memory_nb);
drivers/iommu/intel/iommu.c static void quirk_calpella_no_shadow_gtt(struct pci_dev *dev) · line 5706
Quoted 10 lines
drivers/iommu/intel/iommu.c +2
-2@@ -5702,8 +5699,8 @@ static void quirk_calpella_no_shadow_gtt(struct pci_dev *dev) }else if (dmar_map_gfx) {/* we have to ensure the gfx device is idle before we flush */ pci_info (dev,"Disabling batched IOTLB flush on Ironlake\n" ); - intel_iommu_strict =1 ; - } +iommu_set_dma_strict (true); + } }DECLARE_PCI_FIXUP_HEADER (PCI_VENDOR_ID_INTEL ,0x0040 , quirk_calpella_no_shadow_gtt);DECLARE_PCI_FIXUP_HEADER (PCI_VENDOR_ID_INTEL ,0x0044 , quirk_calpella_no_shadow_gtt);
Best regards,
baolu
PATCH v14 5/6 iommu/amd: Add support for IOMMU default DMA mode build options
In reply to John GarryFrom: Zhen Lei <[email protected]>
Make IOMMU_DEFAULT_LAZY default for when AMD_IOMMU config is set, which matches current behaviour.
For "fullflush" param, just call iommu_set_dma_strict(true) directly.
Since we get a strict vs lazy mode print already in iommu_subsys_init(), and maintain a deprecation print when "fullflush" param is passed, drop the prints in amd_iommu_init_dma_ops().
Finally drop global flag amd_iommu_unmap_flush, as it has no longer has any purpose.
--- drivers/iommu/Kconfig | 2 +- drivers/iommu/amd/amd_iommu_types.h | 6 ------ drivers/iommu/amd/init.c | 3 +-- drivers/iommu/amd/iommu.c | 6 ------ 4 files changed, 2 insertions(+), 15 deletions(-)
drivers/iommu/Kconfig +1 -1
@@ -94,7 +94,7 @@ choice prompt "IOMMU default DMA IOTLB invalidation mode" depends on IOMMU_DMA - default IOMMU_DEFAULT_LAZY if INTEL_IOMMU + default IOMMU_DEFAULT_LAZY if (AMD_IOMMU || INTEL_IOMMU) default IOMMU_DEFAULT_STRICT help This option allows an IOMMU DMA IOTLB invalidation mode to be
drivers/iommu/amd/amd_iommu_types.h +0 -6
@@ -779,12 +779,6 @@ extern u16 amd_iommu_last_bdf;/* allocation bitmap for domain ids */ extern unsigned long *amd_iommu_pd_alloc_bitmap; -/* -* If true, the addresses will be flushed on unmap time, not when -* they are reused -*/ -extern bool amd_iommu_unmap_flush; -/* Smallest max PASID supported by any IOMMU in the system */ extern u32 amd_iommu_max_pasid;
drivers/iommu/amd/init.c +1 -2
@@ -161,7 +161,6 @@ u16 amd_iommu_last_bdf; /* largest PCI device id we haveto handle */ LIST_HEAD (amd_iommu_unity_map);/* a list of required unity mappings we find in ACPI */ -bool amd_iommu_unmap_flush;/* if true, flush on every unmap */ LIST_HEAD (amd_iommu_list);/* list of all AMD IOMMUs in the system */
@@ -3100,7 +3099,7 @@ static int __init parse_amd_iommu_options(char *str)for (; *str; ++str) {if (strncmp (str,"fullflush" ,9 ) ==0 ) {pr_warn ("amd_iommu=fullflush deprecated; use iommu.strict=1 instead\n" ); - amd_iommu_unmap_flush = true; +iommu_set_dma_strict (true); }if (strncmp (str,"force_enable" ,12 ) ==0 ) amd_iommu_force_enable = true;
drivers/iommu/amd/iommu.c +0 -6
@@ -1775,12 +1775,6 @@ void amd_iommu_domain_update(struct protection_domain *domain)static void __init amd_iommu_init_dma_ops (void ) { swiotlb = (iommu_default_passthrough () || sme_me_mask) ?1 :0 ; - -if (amd_iommu_unmap_flush) -pr_info ("IO/TLB flush on unmap enabled\n" ); -else -pr_info ("Lazy IO/TLB flushing enabled\n" ); -iommu_set_dma_strict (amd_iommu_unmap_flush); }int __init amd_iommu_init_api (void )
Signature · 6 lines
2.26.2 _______________________________________________ iommu mailing list [email protected] https://lists.linuxfoundation.org/mailman/listinfo/iommu
PATCH v14 6/6 iommu: Remove mode argument from iommu_set_dma_strict()
In reply to John GarryWe only ever now set strict mode enabled in iommu_set_dma_strict(), so just remove the argument.
--- drivers/iommu/amd/init.c | 2 +- drivers/iommu/intel/iommu.c | 6 +++--- drivers/iommu/iommu.c | 5 ++--- include/linux/iommu.h | 2 +- 4 files changed, 7 insertions(+), 8 deletions(-)
drivers/iommu/amd/init.c +1 -1
@@ -3099,7 +3099,7 @@ static int __init parse_amd_iommu_options(char *str)for (; *str; ++str) {if (strncmp (str,"fullflush" ,9 ) ==0 ) {pr_warn ("amd_iommu=fullflush deprecated; use iommu.strict=1 instead\n" ); -iommu_set_dma_strict (true); +iommu_set_dma_strict (); }if (strncmp (str,"force_enable" ,12 ) ==0 ) amd_iommu_force_enable = true;
drivers/iommu/intel/iommu.c +3 -3
@@ -454,7 +454,7 @@ static int __init intel_iommu_setup(char *str) iommu_dma_forcedac = true; }else if (!strncmp (str,"strict" ,6 )) {pr_warn ("intel_iommu=strict deprecated; use iommu.strict=1 instead\n" ); -iommu_set_dma_strict (true); +iommu_set_dma_strict (); }else if (!strncmp (str,"sp_off" ,6 )) {pr_info ("Disable supported super page\n" ); intel_iommu_superpage =0 ;
@@ -4382,7 +4382,7 @@ int __init intel_iommu_init(void)*/ if (cap_caching_mode (iommu->cap)) {pr_info_once ("IOMMU batching disallowed due to virtualization\n" ); -iommu_set_dma_strict (true); +iommu_set_dma_strict (); } iommu_device_sysfs_add(&iommu->iommu, NULL, intel_iommu_groups,
@@ -5699,7 +5699,7 @@ static void quirk_calpella_no_shadow_gtt(struct pci_dev *dev) }else if (dmar_map_gfx) {/* we have to ensure the gfx device is idle before we flush */ pci_info (dev,"Disabling batched IOTLB flush on Ironlake\n" ); -iommu_set_dma_strict (true); +iommu_set_dma_strict (); } }DECLARE_PCI_FIXUP_HEADER (PCI_VENDOR_ID_INTEL ,0x0040 , quirk_calpella_no_shadow_gtt);
drivers/iommu/iommu.c +2 -3
@@ -349,10 +349,9 @@ static int __init iommu_dma_setup(char *str) }early_param ("iommu.strict" , iommu_dma_setup); -void iommu_set_dma_strict (bool strict) +void iommu_set_dma_strict (void ) { -if (strict || !(iommu_cmd_line &IOMMU_CMD_LINE_STRICT )) - iommu_dma_strict = strict; + iommu_dma_strict = true; }bool iommu_get_dma_strict (struct iommu_domain *domain)
include/linux/iommu.h +1 -1
@@ -476,7 +476,7 @@ int iommu_enable_nesting(struct iommu_domain *domain);int iommu_set_pgtable_quirks (struct iommu_domain *domain,unsigned long quirks); -void iommu_set_dma_strict (bool val); +void iommu_set_dma_strict (void );bool iommu_get_dma_strict (struct iommu_domain *domain); extern int report_iommu_fault(struct iommu_domain *domain, struct device *dev,
Signature · 1 line
2.26.2
Re: [PATCH v14 6/6] iommu: Remove mode argument from iommu_set_dma_strict()
Quoted 79 lines · John Garry
We only ever now set strict mode enabled in
iommu_set_dma_strict(), so just remove the argument.Signed-off-by John Garry <[email protected]>Reviewed-by Robin Murphy <[email protected]>--- drivers/iommu/amd/init.c | 2 +- drivers/iommu/intel/iommu.c | 6 +++--- drivers/iommu/iommu.c | 5 ++--- include/linux/iommu.h | 2 +- 4 files changed, 7 insertions(+), 8 deletions(-)drivers/iommu/amd/init.c +1
-1@@ -3099,7 +3099,7 @@ static int __init parse_amd_iommu_options(char *str)for (; *str; ++str) {if (strncmp (str,"fullflush" ,9 ) ==0 ) {pr_warn ("amd_iommu=fullflush deprecated; use iommu.strict=1 instead\n" ); -iommu_set_dma_strict (true); +iommu_set_dma_strict (); }if (strncmp (str,"force_enable" ,12 ) ==0 ) amd_iommu_force_enable = true;drivers/iommu/intel/iommu.c +3
-3@@ -454,7 +454,7 @@ static int __init intel_iommu_setup(char *str) iommu_dma_forcedac = true; }else if (!strncmp (str,"strict" ,6 )) {pr_warn ("intel_iommu=strict deprecated; use iommu.strict=1 instead\n" ); -iommu_set_dma_strict (true); +iommu_set_dma_strict (); }else if (!strncmp (str,"sp_off" ,6 )) {pr_info ("Disable supported super page\n" ); intel_iommu_superpage =0 ;@@ -4382,7 +4382,7 @@ int __init intel_iommu_init(void)*/ if (cap_caching_mode (iommu->cap)) {pr_info_once ("IOMMU batching disallowed due to virtualization\n" ); -iommu_set_dma_strict (true); +iommu_set_dma_strict (); } iommu_device_sysfs_add(&iommu->iommu, NULL, intel_iommu_groups,@@ -5699,7 +5699,7 @@ static void quirk_calpella_no_shadow_gtt(struct pci_dev *dev) }else if (dmar_map_gfx) {/* we have to ensure the gfx device is idle before we flush */ pci_info (dev,"Disabling batched IOTLB flush on Ironlake\n" ); -iommu_set_dma_strict (true); +iommu_set_dma_strict (); } }DECLARE_PCI_FIXUP_HEADER (PCI_VENDOR_ID_INTEL ,0x0040 , quirk_calpella_no_shadow_gtt);drivers/iommu/iommu.c +2
-3@@ -349,10 +349,9 @@ static int __init iommu_dma_setup(char *str) }early_param ("iommu.strict" , iommu_dma_setup); -void iommu_set_dma_strict (bool strict) +void iommu_set_dma_strict (void ) { -if (strict || !(iommu_cmd_line &IOMMU_CMD_LINE_STRICT )) - iommu_dma_strict = strict; + iommu_dma_strict = true; }bool iommu_get_dma_strict (struct iommu_domain *domain)include/linux/iommu.h +0
-0
include/linux/iommu.h int iommu_enable_nesting(struct iommu_domain *domain); · line 482
Quoted 8 lines
include/linux/iommu.h +1
-1@@ -476,7 +476,7 @@ int iommu_enable_nesting(struct iommu_domain *domain);int iommu_set_pgtable_quirks (struct iommu_domain *domain,unsigned long quirks); -void iommu_set_dma_strict (bool val); +void iommu_set_dma_strict (void );bool iommu_get_dma_strict (struct iommu_domain *domain); extern int report_iommu_fault(struct iommu_domain *domain, struct device *dev,
Best regards,
baolu
Lu Baolu
maintainerRe: [PATCH v14 6/6] iommu: Remove mode argument from iommu_set_dma_strict()
In reply to John GarryQuoted 5 lines · John Garry
drivers/iommu/iommu.c +0
-0
drivers/iommu/iommu.c static int __init iommu_dma_setup(char *str) · line 355
Quoted 10 lines
drivers/iommu/iommu.c +2
-3@@ -349,10 +349,9 @@ static int __init iommu_dma_setup(char *str) }early_param ("iommu.strict" , iommu_dma_setup); -void iommu_set_dma_strict (bool strict) +void iommu_set_dma_strict (void ) { -if (strict || !(iommu_cmd_line &IOMMU_CMD_LINE_STRICT )) - iommu_dma_strict = strict; + iommu_dma_strict = true; }
Sorry for this late comment.
Normally the cache invalidation policy should come from the user. We have pre-build kernel option and also a kernel boot command iommu.strict to override it. These seem reasonable.
We also have a helper (iommu_set_dma_strict()) so that the vendor iommu driver could squeeze in and change the previous settings mostly due to:
a) vendor iommu driver specific kernel boot command. (We are about to deprecate those.)
b) quirky hardware.
c) kernel optimization (e.x. strict mode in VM environment).
a) and b) are mandatory, while c) is optional. In any instance should c) override the flush mode specified by the user. Hence, probably we should also have another helper like:
void iommu_set_dma_strict_optional()
{
if (!(iommu_cmd_line & IOMMU_CMD_LINE_STRICT))
iommu_dma_strict = true;
}Any thoughts?
Best regards,
baolu
Re: [PATCH v14 6/6] iommu: Remove mode argument from iommu_set_dma_strict()
Quoted 16 lines · Lu Baolu
Quoted 5 lines · John Garry
drivers/iommu/iommu.c +0
-0drivers/iommu/iommu.c static int __init iommu_dma_setup(char *str) · line 354
Quoted 9 lines
drivers/iommu/iommu.c +2
-3@@ -349,10 +349,9 @@ static int __init iommu_dma_setup(char *str) } early_param("iommu.strict", iommu_dma_setup); -void iommu_set_dma_strict (bool strict) +void iommu_set_dma_strict (void ) { - if (strict || !(iommu_cmd_line & IOMMU_CMD_LINE_STRICT)) - iommu_dma_strict = strict; + iommu_dma_strict = true; }
Hi baolu,
Quoted 19 lines
Sorry for this late comment.
Quoted 1 line
Normally the cache invalidation policy should come from the user. We
have pre-build kernel option and also a kernel boot command iommu.strict to override it. These seem reasonable.
We also have a helper (
iommu_set_dma_strict()) so that the vendor iommu driver could squeeze in and change the previous settings mostly due to:a) vendor iommu driver specific kernel boot command. (We are about to deprecate those.)
b) quirky hardware.
c) kernel optimization (e.x. strict mode in VM environment).
a) and b) are mandatory, while c) is optional. In any instance should c) override the flush mode specified by the user. Hence, probably we should also have another helper like:
Quoted 7 lines
void
iommu_set_dma_strict_optional()
{
if (!(iommu_cmd_line &IOMMU_CMD_LINE_STRICT))iommu_dma_strict= true;
}Any thoughts?
What you are suggesting is a change in policy from mainline code. Currently for c) we always set strict enabled, regardless of any user cmdline input. But now you are saying that you want iommu.strict to override in particular scenario, right?
In that case I would think it's better to rework the current API, like adding an option to "force" strict mode:
void iommu_set_dma_strict(bool force)
{
if (force == true)
iommu_dma_strict = true;
else if (!(iommu_cmd_line & IOMMU_CMD_LINE_STRICT))
iommu_dma_strict = true;
}So we would use iommu_set_dma_strict(true) for a) and b), but iommu_set_dma_strict(false) for c).
Then I am not sure what you want to do with the accompanying print for c). It was:
"IOMMU batching is disabled due to virtualization"
And now is from this series:
"IOMMU batching disallowed due to virtualization"
Using iommu_get_dma_strict(domain) is not appropriate here to know the current mode (so we know whether to print).
Note that this change would mean that the current series would require non-trivial rework, which would be unfortunate so late in the cycle.
Thanks,
John
Lu Baolu
maintainerRe: [PATCH v14 6/6] iommu: Remove mode argument from iommu_set_dma_strict()
Quoted 18 lines · John Garry
Quoted 16 lines · Lu Baolu
Quoted 5 lines · John Garry
drivers/iommu/iommu.c +0
-0drivers/iommu/iommu.c static int __init iommu_dma_setup(char *str) · line 354
Quoted 9 lines
drivers/iommu/iommu.c +2
-3@@ -349,10 +349,9 @@ static int __init iommu_dma_setup(char *str) } early_param("iommu.strict", iommu_dma_setup); -void iommu_set_dma_strict (bool strict) +void iommu_set_dma_strict (void ) { - if (strict || !(iommu_cmd_line & IOMMU_CMD_LINE_STRICT)) - iommu_dma_strict = strict; + iommu_dma_strict = true; }
Quoted 1 line
Hi baolu,
Hi John,
Quoted 40 lines
Quoted 20 lines
Sorry for this late comment.
> Normally the cache invalidation policy should come from the user. We have pre-build kernel option and also a kernel boot command iommu.strict to override it. These seem reasonable.We also have a helper (
iommu_set_dma_strict()) so that the vendor iommu driver could squeeze in and change the previous settings mostly due to:a) vendor iommu driver specific kernel boot command. (We are about to deprecate those.)
b) quirky hardware.
c) kernel optimization (e.x. strict mode in VM environment).
a) and b) are mandatory, while c) is optional. In any instance should c) override the flush mode specified by the user. Hence, probably we should also have another helper like:
Quoted 7 lines
void
iommu_set_dma_strict_optional()
{
if (!(iommu_cmd_line &IOMMU_CMD_LINE_STRICT))iommu_dma_strict= true;
}Any thoughts?
What you are suggesting is a change in policy from mainline code. Currently for c) we always set strict enabled, regardless of any user cmdline input. But now you are saying that you want iommu.strict to override in particular scenario, right?
In that case I would think it's better to rework the current API, like adding an option to "force" strict mode:
void
iommu_set_dma_strict(bool force)
{
if (force == true)
iommu_dma_strict= true;
else if (!(iommu_cmd_line &IOMMU_CMD_LINE_STRICT))iommu_dma_strict= true;
}
Quoted 1 line
So we would use
iommu_set_dma_strict(true)for a) and b), butiommu_set_dma_strict(false)for c).
Yes. We need to distinguish the "must" and "nice-to-have" cases of setting strict mode.
Quoted 10 lines
Then I am not sure what you want to do with the accompanying print for c). It was:
"IOMMU batching is disabled due to virtualization"And now is from this series:
"IOMMU batching disallowed due to virtualization"Using
iommu_get_dma_strict(domain)is not appropriate here to know the current mode (so we know whether to print).Note that this change would mean that the current series would require non-trivial rework, which would be unfortunate so late in the cycle.
This patch series looks good to me and I have added by reviewed-by. Probably we could make another patch series to improve it so that the kernel optimization should not override the user setting.
Best regards,
baolu
Re: [PATCH v14 6/6] iommu: Remove mode argument from iommu_set_dma_strict()
Quoted 20 lines · Lu Baolu
Quoted 9 lines
void
iommu_set_dma_strict(bool force)
{
if (force == true)
iommu_dma_strict= true;
else if (!(iommu_cmd_line &IOMMU_CMD_LINE_STRICT))iommu_dma_strict= true;
}So we would use
iommu_set_dma_strict(true)for a) and b), butiommu_set_dma_strict(false)for c).Yes. We need to distinguish the "must" and "nice-to-have" cases of setting strict mode.
Quoted 7 lines
Then I am not sure what you want to do with the accompanying print for c). It was:
"IOMMU batching is disabled due to virtualization"And now is from this series:
"IOMMU batching disallowed due to virtualization"
Quoted 7 lines
Quoted 3 lines
Using
iommu_get_dma_strict(domain)is not appropriate here to know the current mode (so we know whether to print).Note that this change would mean that the current series would require non-trivial rework, which would be unfortunate so late in the cycle.
This patch series looks good to me and I have added by reviewed-by. Probably we could make another patch series to improve it so that the kernel optimization should not override the user setting.
On a personal level I would be happy with that approach, but I think it's better to not start changing things right away in a follow-up series.
So how about we add this patch (which replaces 6/6 "iommu: Remove mode argument from iommu_set_dma_strict()")?
Robin, any opinion?
------->8---------
[PATCH] iommu/vt-d: Make "iommu.strict" override batching due to
virtualization
As a change in policy, make iommu.strict cmdline argument override whether we disable batching due to virtualization.
The API of iommu_set_dma_strict() is changed to accept a "force" argument, which means that we always set iommu_dma_strict true, regardless of whether we already set via cmdline. Also return a boolean, to tell whether iommu_dma_strict was set or not.
Note that in all pre-existing callsites of iommu_set_dma_strict(), argument strict was true, so this argument is dropped.
drivers/iommu/intel/iommu.c +1 -3
@@ -4380,10 +4380,8 @@ int __init intel_iommu_init(void)* is likely to be much lower than the overhead of synchronizing * the virtual and physical IOMMU page-tables. */ -if (cap_caching_mode (iommu->cap)) { +if (cap_caching_mode (iommu->cap) &&iommu_set_dma_strict (false))pr_info_once ("IOMMU batching disallowed due to virtualization\n" ); -iommu_set_dma_strict (true); - }iommu_device_sysfs_add (&iommu->iommu,NULL , intel_iommu_groups,"%s" , iommu->name);
drivers/iommu/iommu.c +7 -3
@@ -349,10 +349,14 @@ static int __init iommu_dma_setup(char *str) }early_param ("iommu.strict" , iommu_dma_setup); -void iommu_set_dma_strict (bool strict) +/* Return true if we set iommu_dma_strict */ +bool iommu_set_dma_strict (bool force) { -if (strict || !(iommu_cmd_line &IOMMU_CMD_LINE_STRICT )) - iommu_dma_strict = strict; +if (force || !(iommu_cmd_line &IOMMU_CMD_LINE_STRICT )) { + iommu_dma_strict = true; +return true; + } +return false; }bool iommu_get_dma_strict (struct iommu_domain *domain)
include/linux/iommu.h +1 -1
@@ -476,7 +476,7 @@ int iommu_enable_nesting(struct iommu_domain *domain);int iommu_set_pgtable_quirks (struct iommu_domain *domain,unsigned long quirks); -void iommu_set_dma_strict (bool val); +bool iommu_set_dma_strict (bool force);bool iommu_get_dma_strict (struct iommu_domain *domain); extern int report_iommu_fault(struct iommu_domain *domain, struct
device *dev,
Signature · 1 line
2.26.2
Robin Murphy
reviewerRe: [PATCH v14 6/6] iommu: Remove mode argument from iommu_set_dma_strict()
Quoted 29 lines · John Garry
Quoted 20 lines · Lu Baolu
Quoted 9 lines
void
iommu_set_dma_strict(bool force)
{
if (force == true)
iommu_dma_strict= true;
else if (!(iommu_cmd_line &IOMMU_CMD_LINE_STRICT))iommu_dma_strict= true;
}So we would use
iommu_set_dma_strict(true)for a) and b), butiommu_set_dma_strict(false)for c).Yes. We need to distinguish the "must" and "nice-to-have" cases of setting strict mode.
Quoted 7 lines
Then I am not sure what you want to do with the accompanying print for c). It was:
"IOMMU batching is disabled due to virtualization"And now is from this series:
"IOMMU batching disallowed due to virtualization"Quoted 7 lines
Quoted 3 lines
Using
iommu_get_dma_strict(domain)is not appropriate here to know the current mode (so we know whether to print).Note that this change would mean that the current series would require non-trivial rework, which would be unfortunate so late in the cycle.
This patch series looks good to me and I have added by reviewed-by. Probably we could make another patch series to improve it so that the kernel optimization should not override the user setting.
Quoted 5 lines
On a personal level I would be happy with that approach, but I think it's better to not start changing things right away in a follow-up series.
So how about we add this patch (which replaces 6/6 "iommu: Remove mode argument from
iommu_set_dma_strict()")?Robin, any opinion?
For me it boils down to whether there are any realistic workloads where non-strict mode *would* still perform better under virtualisation. The only reason for the user to explicitly pass "iommu.strict=0" is because they expect it to increase unmap performance; if it's only ever going to lead to an unexpected performance loss, I don't see any value in overriding the kernel's decision purely for the sake of subservience.
If there *are* certain valid cases for allowing it for people who really know what they're doing, then we should arguably also log a counterpart message to say "we're honouring your override but beware it may have the opposite effect to what you expect" for the benefit of other users who assume it's a generic go-faster knob. At that point it starts getting non-trivial enough that I'd want to know for sure it's worthwhile.
The other reason this might be better to revisit later is that an AMD equivalent is still in flight[1], and there might be more that can eventually be factored out. I think both series are pretty much good to merge for 5.14, but time's already tight to sort out the conflicts which exist as-is, without making them any worse.
Robin.
[1] https://lore.kernel.org/linux-iommu/[email protected]/
Quoted 67 lines
------->8---------
[PATCH] iommu/vt-d: Make "iommu.strict" override batching due to
virtualizationAs a change in policy, make iommu.strict cmdline argument override whether we disable batching due to virtualization.
The API of
iommu_set_dma_strict()is changed to accept a "force" argument, which means that we always setiommu_dma_stricttrue, regardless of whether we already set via cmdline. Also return a boolean, to tell whetheriommu_dma_strictwas set or not.Note that in all pre-existing callsites of
iommu_set_dma_strict(), argument strict was true, so this argument is dropped.Signed-off-by John Garry <[email protected]>drivers/iommu/intel/iommu.c +1
-3@@ -4380,10 +4380,8 @@ int __init intel_iommu_init(void)* is likely to be much lower than the overhead of synchronizing * the virtual and physical IOMMU page-tables. */ - if (cap_caching_mode(iommu->cap)) { + if (cap_caching_mode(iommu->cap) && iommu_set_dma_strict(false)) pr_info_once("IOMMU batching disallowed due to virtualization\n"); - iommu_set_dma_strict(true); - } iommu_device_sysfs_add(&iommu->iommu, NULL, intel_iommu_groups,"%s",
iommu->name);drivers/iommu/iommu.c +7
-3@@ -349,10 +349,14 @@ static int __init iommu_dma_setup(char *str) } early_param("iommu.strict", iommu_dma_setup); -void iommu_set_dma_strict (bool strict) +/* Return true if we set iommu_dma_strict */ +bool iommu_set_dma_strict (bool force) { - if (strict || !(iommu_cmd_line & IOMMU_CMD_LINE_STRICT)) - iommu_dma_strict = strict; + if (force || !(iommu_cmd_line & IOMMU_CMD_LINE_STRICT)) { + iommu_dma_strict = true; + return true; + } + return false; } bool iommu_get_dma_strict(struct iommu_domain *domain)include/linux/iommu.h +1
-1@@ -476,7 +476,7 @@ int iommu_enable_nesting(struct iommu_domain *domain); int iommu_set_pgtable_quirks(struct iommu_domain *domain, unsigned long quirks); -void iommu_set_dma_strict (bool val); +bool iommu_set_dma_strict (bool force); bool iommu_get_dma_strict(struct iommu_domain *domain); extern int report_iommu_fault(struct iommu_domain *domain, structdevice *dev,
Re: [PATCH v14 6/6] iommu: Remove mode argument from iommu_set_dma_strict()
On 21/06/2021 12:59, Robin Murphy wrote:
+ Nadav
Quoted 21 lines
Quoted 7 lines
On a personal level I would be happy with that approach, but I think it's better to not start changing things right away in a follow-up series.
So how about we add this patch (which replaces 6/6 "iommu: Remove mode argument from
iommu_set_dma_strict()")?Robin, any opinion?
For me it boils down to whether there are any realistic workloads where non-strict mode*would* still perform better under virtualisation. The only reason for the user to explicitly pass "iommu.strict=0" is because they expect it to increase unmap performance; if it's only ever going to lead to an unexpected performance loss, I don't see any value in overriding the kernel's decision purely for the sake of subservience.
If there*are* certain valid cases for allowing it for people who really know what they're doing, then we should arguably also log a counterpart message to say "we're honouring your override but beware it may have the opposite effect to what you expect" for the benefit of other users who assume it's a generic go-faster knob. At that point it starts getting non-trivial enough that I'd want to know for sure it's worthwhile.
Quoted 5 lines
The other reason this might be better to revisit later is that an AMD equivalent is still in flight[1], and there might be more that can eventually be factored out. I think both series are pretty much good to merge for 5.14, but time's already tight to sort out the conflicts which exist as-is, without making them any worse.
ok, fine. Can revisit.
As for getting these merged, I'll dry-run merging both of those series to see the conflicts. It doesn't look too problematic from a glance.
Cheers,
John
Quoted 77 lines
Robin.
[1]
https://lore.kernel.org/linux-iommu/[email protected]/Quoted 71 lines
------->8---------
[PATCH] iommu/vt-d: Make "iommu.strict" override batching due to
virtualizationAs a change in policy, make iommu.strict cmdline argument override whether we disable batching due to virtualization.
The API of
iommu_set_dma_strict()is changed to accept a "force" argument, which means that we always setiommu_dma_stricttrue, regardless of whether we already set via cmdline. Also return a boolean, to tell whetheriommu_dma_strictwas set or not.Note that in all pre-existing callsites of
iommu_set_dma_strict(), argument strict was true, so this argument is dropped.Signed-off-by John Garry<[email protected]>drivers/iommu/intel/iommu.c +1
-3@@ -4380,10 +4380,8 @@ int __init intel_iommu_init(void)* is likely to be much lower than the overhead of synchronizing * the virtual and physical IOMMU page-tables. */ - if (cap_caching_mode(iommu->cap)) { + if (cap_caching_mode(iommu->cap) && iommu_set_dma_strict(false)) pr_info_once("IOMMU batching disallowed due to virtualization\n"); - iommu_set_dma_strict(true); - } iommu_device_sysfs_add(&iommu->iommu, NULL, intel_iommu_groups,"%s",
iommu->name);drivers/iommu/iommu.c +7
-3@@ -349,10 +349,14 @@ static int __init iommu_dma_setup(char *str) } early_param("iommu.strict", iommu_dma_setup); -void iommu_set_dma_strict (bool strict) +/* Return true if we set iommu_dma_strict */ +bool iommu_set_dma_strict (bool force) { - if (strict || !(iommu_cmd_line & IOMMU_CMD_LINE_STRICT)) - iommu_dma_strict = strict; + if (force || !(iommu_cmd_line & IOMMU_CMD_LINE_STRICT)) { + iommu_dma_strict = true; + return true; + } + return false; } bool iommu_get_dma_strict(struct iommu_domain *domain)include/linux/iommu.h +1
-1@@ -476,7 +476,7 @@ int iommu_enable_nesting(struct iommu_domain *domain); int iommu_set_pgtable_quirks(struct iommu_domain *domain, unsigned long quirks); -void iommu_set_dma_strict (bool val); +bool iommu_set_dma_strict (bool force); bool iommu_get_dma_strict(struct iommu_domain *domain); extern int report_iommu_fault(struct iommu_domain *domain, structdevice *dev,
Lu Baolu
maintainerRe: [PATCH v14 6/6] iommu: Remove mode argument from iommu_set_dma_strict()
In reply to Robin MurphyHi Robin,
Quoted 30 lines · Robin Murphy
Quoted 29 lines · John Garry
Quoted 27 lines · Lu Baolu
Quoted 9 lines
void
iommu_set_dma_strict(bool force)
{
if (force == true)
iommu_dma_strict= true;
else if (!(iommu_cmd_line &IOMMU_CMD_LINE_STRICT))iommu_dma_strict= true;
}So we would use
iommu_set_dma_strict(true)for a) and b), butiommu_set_dma_strict(false)for c).Yes. We need to distinguish the "must" and "nice-to-have" cases of setting strict mode.
Quoted 10 lines
Then I am not sure what you want to do with the accompanying print for c). It was:
"IOMMU batching is disabled due to virtualization"And now is from this series:
"IOMMU batching disallowed due to virtualization"Using
iommu_get_dma_strict(domain)is not appropriate here to know the current mode (so we know whether to print).Note that this change would mean that the current series would require non-trivial rework, which would be unfortunate so late in the cycle.
This patch series looks good to me and I have added by reviewed-by. Probably we could make another patch series to improve it so that the kernel optimization should not override the user setting.
Quoted 7 lines
Quoted 5 lines
On a personal level I would be happy with that approach, but I think it's better to not start changing things right away in a follow-up series.
So how about we add this patch (which replaces 6/6 "iommu: Remove mode argument from
iommu_set_dma_strict()")?Robin, any opinion?
For me it boils down to whether there are any realistic workloads where non-strict mode *would* still perform better under virtualisation. The
At present, we see that strict mode has better performance in the virtualization environment because it will make the shadow page table management more efficient. When the hardware supports nested translation, we may have to re-evaluate this since there's no need for a shadowing page table anymore.
Quoted 5 lines
only reason for the user to explicitly pass "iommu.strict=0" is because they expect it to increase unmap performance; if it's only ever going to lead to an unexpected performance loss, I don't see any value in overriding the kernel's decision purely for the sake of subservience.
If there *are* certain valid cases for allowing it for people who really know what they're doing, then we should arguably also log a counterpart message to say "we're honouring your override but beware it may have the opposite effect to what you expect" for the benefit of other users who assume it's a generic go-faster knob. At that point it starts getting non-trivial enough that I'd want to know for sure it's worthwhile.
The other reason this might be better to revisit later is that an AMD equivalent is still in flight[1], and there might be more that can eventually be factored out. I think both series are pretty much good to merge for 5.14, but time's already tight to sort out the conflicts which exist as-is, without making them any worse.
Agreed. We could revisit it later.
Best regards,
baolu
Robin Murphy
reviewerRe: [PATCH v14 6/6] iommu: Remove mode argument from iommu_set_dma_strict()
Quoted 39 lines · Lu Baolu
Hi Robin,
Quoted 36 lines · Robin Murphy
Quoted 34 lines · John Garry
Quoted 20 lines · Lu Baolu
Quoted 9 lines
void
iommu_set_dma_strict(bool force)
{
if (force == true)
iommu_dma_strict= true;
else if (!(iommu_cmd_line &IOMMU_CMD_LINE_STRICT))iommu_dma_strict= true;
}So we would use
iommu_set_dma_strict(true)for a) and b), butiommu_set_dma_strict(false)for c).Yes. We need to distinguish the "must" and "nice-to-have" cases of setting strict mode.
Quoted 7 lines
Then I am not sure what you want to do with the accompanying print for c). It was:
"IOMMU batching is disabled due to virtualization"And now is from this series:
"IOMMU batching disallowed due to virtualization"Quoted 7 lines
Quoted 3 lines
Using
iommu_get_dma_strict(domain)is not appropriate here to know the current mode (so we know whether to print).Note that this change would mean that the current series would require non-trivial rework, which would be unfortunate so late in the cycle.
This patch series looks good to me and I have added by reviewed-by. Probably we could make another patch series to improve it so that the kernel optimization should not override the user setting.
On a personal level I would be happy with that approach, but I think it's better to not start changing things right away in a follow-up series.
So how about we add this patch (which replaces 6/6 "iommu: Remove mode argument from
iommu_set_dma_strict()")?Robin, any opinion?
Quoted 7 lines
Quoted 1 line
For me it boils down to whether there are any realistic workloads where non-strict mode *would* still perform better under virtualisation. The
At present, we see that strict mode has better performance in the virtualization environment because it will make the shadow page table management more efficient. When the hardware supports nested translation, we may have to re-evaluate this since there's no need for a shadowing page table anymore.
I guess I was assuming that in most cases, proper nested mode could look distinct enough that we'd be able to treat it differently in the first place. For instance, if it's handing guest tables directly to the hardware, would the host have any reason to still set the "caching mode" ID bit?
Robin.
Lu Baolu
maintainerRe: [PATCH v14 6/6] iommu: Remove mode argument from iommu_set_dma_strict()
Quoted 42 lines · Robin Murphy
Quoted 41 lines · Lu Baolu
Hi Robin,
Quoted 37 lines · Robin Murphy
Quoted 29 lines · John Garry
Quoted 20 lines · Lu Baolu
Quoted 9 lines
void
iommu_set_dma_strict(bool force)
{
if (force == true)
iommu_dma_strict= true;
else if (!(iommu_cmd_line &IOMMU_CMD_LINE_STRICT))iommu_dma_strict= true;
}So we would use
iommu_set_dma_strict(true)for a) and b), butiommu_set_dma_strict(false)for c).Yes. We need to distinguish the "must" and "nice-to-have" cases of setting strict mode.
Quoted 7 lines
Then I am not sure what you want to do with the accompanying print for c). It was:
"IOMMU batching is disabled due to virtualization"And now is from this series:
"IOMMU batching disallowed due to virtualization"Quoted 7 lines
Quoted 3 lines
Using
iommu_get_dma_strict(domain)is not appropriate here to know the current mode (so we know whether to print).Note that this change would mean that the current series would require non-trivial rework, which would be unfortunate so late in the cycle.
This patch series looks good to me and I have added by reviewed-by. Probably we could make another patch series to improve it so that the kernel optimization should not override the user setting.
Quoted 5 lines
On a personal level I would be happy with that approach, but I think it's better to not start changing things right away in a follow-up series.
So how about we add this patch (which replaces 6/6 "iommu: Remove mode argument from
iommu_set_dma_strict()")?Robin, any opinion?
For me it boils down to whether there are any realistic workloads where non-strict mode *would* still perform better under virtualisation. The
Quoted 7 lines
Quoted 5 lines
At present, we see that strict mode has better performance in the virtualization environment because it will make the shadow page table management more efficient. When the hardware supports nested translation, we may have to re-evaluate this since there's no need for a shadowing page table anymore.
I guess I was assuming that in most cases, proper nested mode could look distinct enough that we'd be able to treat it differently in the first place. For instance, if it's handing guest tables directly to the hardware, would the host have any reason to still set the "caching mode" ID bit?
For Intel VT-d, yes, simply for compatible purpose. The guest kernel may use page tables that are not compatible with the first level translation. In this case, we must roll back to shadow page table.
Quoted 2 lines
Robin.
Best regards,
baolu
Quoted 15 lines · John Garry
This is a reboot of Zhen Lei's series from a couple of years ago, which never made it across the line.
I still think that it has some value, so taking up the mantle.
Motivation:
Allow lazy mode be default mode for DMA domains for all ARCHs, and not only those who hardcode it (to be lazy). For ARM64, currently we must use a kernel command line parameter to use lazy mode, which is less than ideal.I have now included the print for strict/lazy mode, which I originally sent in:
https://lore.kernel.org/linux-iommu/[email protected]/
Quoted 5 lines
There was some concern there about drivers and their custom prints conflicting with the print in that patch, but I think that it should be ok.
Based on next-20210611 + "iommu: Update "iommu.strict" documentation"
Hi Joerg, Will,
We think that this series is ready to go.
There would be a build conflict with the following:
https://lore.kernel.org/linux-iommu/[email protected]/
So please let us know where you stand on it, so that could be resolved.
Robin and Baolu have kindly reviewed all the patches, apart from the AMD one.
Thanks,
John
[email protected]
maintainerHi John,
Quoted 9 lines · John Garry
We think that this series is ready to go.
There would be a build conflict with the following:
https://lore.kernel.org/linux-iommu/[email protected]/So please let us know where you stand on it, so that could be resolved.
Robin and Baolu have kindly reviewed all the patches, apart from the AMD one.
The AMD one also looks good to me, please re-send after the merge window closes and I will take care of it then. Note that I usually start merging new stuff after -rc3 is out.
Regards,
Joerg