- 11 Jun, 2015 17 commits
-
-
Alexey Kardashevskiy authored
Modern IBM POWERPC systems support multiple (currently two) TCE tables per IOMMU group (a.k.a. PE). This adds a iommu_table_group container for TCE tables. Right now just one table is supported. This defines iommu_table_group struct which stores pointers to iommu_group and iommu_table(s). This replaces iommu_table with iommu_table_group where iommu_table was used to identify a group: - iommu_register_group(); - iommudata of generic iommu_group; This removes @data from iommu_table as it_table_group provides same access to pnv_ioda_pe. For IODA, instead of embedding iommu_table, the new iommu_table_group keeps pointers to those. The iommu_table structs are allocated dynamically. For P5IOC2, both iommu_table_group and iommu_table are embedded into PE struct. As there is no EEH and SRIOV support for P5IOC2, iommu_free_table() should not be called on iommu_table struct pointers so we can keep it embedded in pnv_phb::p5ioc2. For pSeries, this replaces multiple calls of kzalloc_node() with a new iommu_pseries_alloc_group() helper and stores the table group struct pointer into the pci_dn struct. For release, a iommu_table_free_group() helper is added. This moves iommu_table struct allocation from SR-IOV code to the generic DMA initialization code in pnv_pci_ioda_setup_dma_pe and pnv_pci_ioda2_setup_dma_pe as this is where DMA is actually initialized. This change is here because those lines had to be changed anyway. This should cause no behavioural change. Signed-off-by:
Alexey Kardashevskiy <aik@ozlabs.ru> [aw: for the vfio related changes] Acked-by:
Alex Williamson <alex.williamson@redhat.com> Reviewed-by:
David Gibson <david@gibson.dropbear.id.au> Reviewed-by:
Gavin Shan <gwshan@linux.vnet.ibm.com> Signed-off-by:
Michael Ellerman <mpe@ellerman.id.au>
-
Alexey Kardashevskiy authored
The pnv_pci_ioda_tce_invalidate() helper invalidates TCE cache. It is supposed to be called on IODA1/2 and not called on p5ioc2. It receives start and end host addresses of TCE table. IODA2 actually needs PCI addresses to invalidate the cache. Those can be calculated from host addresses but since we are going to implement multi-level TCE tables, calculating PCI address from a host address might get either tricky or ugly as TCE table remains flat on PCI bus but not in RAM. This moves pnv_pci_ioda_tce_invalidate() from generic pnv_tce_build/ pnt_tce_free and defines IODA1/2-specific callbacks which call generic ones and do PHB-model-specific TCE cache invalidation. P5IOC2 keeps using generic callbacks as before. This changes pnv_pci_ioda2_tce_invalidate() to receives TCE index and number of pages which are PCI addresses shifted by IOMMU page shift. No change in behaviour is expected. Signed-off-by:
-
Alexey Kardashevskiy authored
This adds a iommu_table_ops struct and puts pointer to it into the iommu_table struct. This moves tce_build/tce_free/tce_get/tce_flush callbacks from ppc_md to the new struct where they really belong to. This adds the requirement for @it_ops to be initialized before calling iommu_init_table() to make sure that we do not leave any IOMMU table with iommu_table_ops uninitialized. This is not a parameter of iommu_init_table() though as there will be cases when iommu_init_table() will not be called on TCE tables, for example - VFIO. This does s/tce_build/set/, s/tce_free/clear/ and removes "tce_" redundant prefixes. This removes tce_xxx_rm handlers from ppc_md but does not add them to iommu_table_ops as this will be done later if we decide to support TCE hypercalls in real mode. This removes _vm callbacks as only virtual mode is supported by now so this also removes @rm parameter. For pSeries, this always uses tce_buildmulti_pSeriesLP/ tce_buildmulti_pSeriesLP. This changes multi callback to fall back to tce_build_pSeriesLP/tce_free_pSeriesLP if FW_FEATURE_MULTITCE is not present. The reason for this is we still have to support "multitce=off" boot parameter in disable_multitce() and we do not want to walk through all IOMMU tables in the system and replace "multi" callbacks with single ones. For powernv, this defines _ops per PHB type which are P5IOC2/IODA1/IODA2. This makes the callbacks for them public. Later patches will extend callbacks for IODA1/2. No change in behaviour is expected. Signed-off-by:
-
Alexey Kardashevskiy authored
Normally a bitmap from the iommu_table is used to track what TCE entry is in use. Since we are going to use iommu_table without its locks and do xchg() instead, it becomes essential not to put bits which are not implied in the direction flag as the old TCE value (more precisely - the permission bits) will be used to decide whether to put the page or not. This adds iommu_direction_to_tce_perm() (its counterpart is there already) and uses it for powernv's pnv_tce_build(). Signed-off-by:
-
Alexey Kardashevskiy authored
This is to make extended ownership and multiple groups support patches simpler for review. This should cause no behavioural change. Signed-off-by:
-
Alexey Kardashevskiy authored
This is a pretty mechanical patch to make next patches simpler. New tce_iommu_unuse_page() helper does put_page() now but it might skip that after the memory registering patch applied. As we are here, this removes unnecessary checks for a value returned by pfn_to_page() as it cannot possibly return NULL. This moves tce_iommu_disable() later to let tce_iommu_clear() know if the container has been enabled because if it has not been, then put_page() must not be called on TCEs from the TCE table. This situation is not yet possible but it will after KVM acceleration patchset is applied. This changes code to work with physical addresses rather than linear mapping addresses for better code readability. Following patches will add an xchg() callback for an IOMMU table which will accept/return physical addresses (unlike current tce_build()) which will eliminate redundant conversions. Signed-off-by:
-
Alexey Kardashevskiy authored
At the moment DMA map/unmap requests are handled irrespective to the container's state. This allows the user space to pin memory which it might not be allowed to pin. This adds checks to MAP/UNMAP that the container is enabled, otherwise -EPERM is returned. Signed-off-by:
-
Alexey Kardashevskiy authored
There moves locked pages accounting to helpers. Later they will be reused for Dynamic DMA windows (DDW). This reworks debug messages to show the current value and the limit. This stores the locked pages number in the container so when unlocking the iommu table pointer won't be needed. This does not have an effect now but it will with the multiple tables per container as then we will allow attaching/detaching groups on fly and we may end up having a container with no group attached but with the counter incremented. While we are here, update the comment explaining why RLIMIT_MEMLOCK might be required to be bigger than the guest RAM. This also prints pid of the current process in pr_warn/pr_debug. Signed-off-by:
-
Alexey Kardashevskiy authored
This makes use of the it_page_size from the iommu_table struct as page size can differ. This replaces missing IOMMU_PAGE_SHIFT macro in commented debug code as recently introduced IOMMU_PAGE_XXX macros do not include IOMMU_PAGE_SHIFT. Signed-off-by:
-
Alexey Kardashevskiy authored
This checks that the TCE table page size is not bigger that the size of a page we just pinned and going to put its physical address to the table. Otherwise the hardware gets unwanted access to physical memory between the end of the actual page and the end of the aligned up TCE page. Since compound_order() and compound_head() work correctly on non-huge pages, there is no need for additional check whether the page is huge. Signed-off-by:
-
Alexey Kardashevskiy authored
This moves page pinning (get_user_pages_fast()/put_page()) code out of the platform IOMMU code and puts it to VFIO IOMMU driver where it belongs to as the platform code does not deal with page pinning. This makes iommu_take_ownership()/iommu_release_ownership() deal with the IOMMU table bitmap only. This removes page unpinning from iommu_take_ownership() as the actual TCE table might contain garbage and doing put_page() on it is undefined behaviour. Besides the last part, the rest of the patch is mechanical. Signed-off-by:
-
Alexey Kardashevskiy authored
At the moment iommu_free_table() only releases memory if the table was initialized for the platform code use, i.e. it had it_map initialized (which purpose is to track DMA memory space use). With dynamic DMA windows, we will need to be able to release iommu_table even if it was used for VFIO in which case it_map is NULL so does the patch. Signed-off-by:
-
Alexey Kardashevskiy authored
So far an iommu_table lifetime was the same as PE. Dynamic DMA windows will change this and iommu_free_table() will not always require the group to be released. This moves iommu_group_put() out of iommu_free_table(). This adds a iommu_pseries_free_table() helper which does iommu_group_put() and iommu_free_table(). Later it will be changed to receive a table_group and we will have to change less lines then. This should cause no behavioural change. Signed-off-by:
-
Alexey Kardashevskiy authored
The existing code has 3 calls to iommu_register_group() and all 3 branches actually cover all possible cases. This replaces 3 calls with one and moves the registration earlier; the latter will make more sense when we add TCE table sharing. Signed-off-by:
-
Alexey Kardashevskiy authored
The set_iommu_table_base_and_group() name suggests that the function sets table base and add a device to an IOMMU group. The actual purpose for table base setting is to put some reference into a device so later iommu_add_device() can get the IOMMU group reference and the device to the group. At the moment a group cannot be explicitly passed to iommu_add_device() as we want it to work from the bus notifier, we can fix it later and remove confusing calls of set_iommu_table_base(). This replaces set_iommu_table_base_and_group() with a couple of set_iommu_table_base() + iommu_add_device() which makes reading the code easier. This adds few comments why set_iommu_table_base() and iommu_add_device() are called where they are called. For IODA1/2, this essentially removes iommu_add_device() call from the pnv_pci_ioda_dma_dev_setup() as it will always fail at this particular place: - for physical PE, the device is already attached by iommu_add_device() in pnv_pci_ioda_setup_dma_pe(); - for virtual PE, the sysfs entries are not ready to create all symlinks so actual adding is happening in tce_iommu_bus_notifier. Signed-off-by:
-
Alexey Kardashevskiy authored
This relies on the fact that a PCI device always has an IOMMU table which may not be the case when we get dynamic DMA windows so let's use more reliable check for IOMMU group here. As we do not rely on the table presence here, remove the workaround from pnv_pci_ioda2_set_bypass(); also remove the @add_to_iommu_group parameter from pnv_ioda_setup_bus_dma(). Signed-off-by:
-
Cyril Bur authored
Powerpc powernv platforms allow access to certain system flash devices through a firmwarwe interface. This change adds an mtd driver for these flash devices. Minor updates from Jeremy Kerr and Joel Stanley. Signed-off-by:
Cyril Bur <cyrilbur@gmail.com> Signed-off-by:
Joel Stanley <joel@jms.id.au> Signed-off-by:
Jeremy Kerr <jk@ozlabs.org> Reviewed-by:
Neelesh Gupta <neelegup@linux.vnet.ibm.com> Signed-off-by:
-
- 10 Jun, 2015 1 commit
-
-
Aneesh Kumar K.V authored
This enables us to understand how many hash fault we are taking when running benchmarks. For ex: -bash-4.2# ./perf stat -e powerpc:hash_fault -e page-faults /tmp/ebizzy.ppc64 -S 30 -P -n 1000 ... Performance counter stats for '/tmp/ebizzy.ppc64 -S 30 -P -n 1000': 1,10,04,075 powerpc:hash_fault 1,10,03,429 page-faults 30.865978991 seconds time elapsed NOTE: The impact of the tracepoint was not noticeable when running test. It was within the run-time variance of the test. For ex: without-patch: -------------- Performance counter stats for './a.out 3000 300': 643 page-faults # 0.089 M/sec 7.236562 task-clock (msec) # 0.928 CPUs utilized 2,179,213 stalled-cycles-frontend # 0.00% frontend cycles idle 17,174,367 stalled-cycles-backend # 0.00% backend cycles idle 0 context-switches # 0.000 K/sec 0.007794658 seconds time elapsed And with-patch: --------------- Performance counter stats for './a.out 3000 300': 643 page-faults # 0.089 M/sec 7.233746 task-clock (msec) # 0.921 CPUs utilized 0 context-switches # 0.000 K/sec 0.007854876 seconds time elapsed Performance counter stats for './a.out 3000 300': 643 page-faults # 0.087 M/sec 649 powerpc:hash_fault # 0.087 M/sec 7.430376 task-clock (msec) # 0.938 CPUs utilized 2,347,174 stalled-cycles-frontend # 0.00% frontend cycles idle 17,524,282 stalled-cycles-backend # 0.00% backend cycles idle 0 context-switches # 0.000 K/sec 0.007920284 seconds time elapsed Signed-off-by:Aneesh Kumar K.V <aneesh.kumar@linux.vnet.ibm.com> Signed-off-by:
-
- 07 Jun, 2015 15 commits
-
-
Anshuman Khandual authored
This patch adds .gitignore for all the newly added DSCR tests. Signed-off-by:
Anshuman Khandual <khandual@linux.vnet.ibm.com> Signed-off-by:
-
Michael Ellerman authored
This patch adds a test to update the system wide DSCR value repeatedly and then verifies that any thread on any given CPU on the system must be able to see the same DSCR value whether its is being read through the problem state based SPR or the privilege state based SPR. Signed-off-by:
-
Anshuman Khandual authored
This test continuously updates the system wide DSCR default value in the sysfs interface and makes sure that the same is reflected across all the sysfs interfaces for each individual CPUs present on the system. Acked-by:
Shuah Khan <shuahkh@osg.samsung.com> Signed-off-by:
-
Anshuman Khandual authored
This patch adds a test case to verify that the changed DSCR value inside any process would be inherited to it's child across the fork and exec system call. Acked-by:
Anton Blanchard <anton@samba.org> Signed-off-by:
-
Anshuman Khandual authored
This patch adds a test to verify that the changed DSCR value inside any process would be inherited to it's child process across the fork system call. Acked-by:
-
Anshuman Khandual authored
This patch adds a test which verifies that the DSCR privilege and problem state SPR read & write accesses while making sure that the results are always the same irrespective of which SPR number is being used. Acked-by:
-
Anshuman Khandual authored
This patch adds a test which modifies the DSCR using mtspr instruction and verifies the change using mfspr instruction. It uses both the privilege state SPR as well as the problem state SPR for the purpose. Acked-by:
-
Anshuman Khandual authored
This patch adds a test case for the system wide DSCR default value, which when changed through it's sysfs interface must be visible to all threads reading DSCR either through the privilege state SPR or the problem state SPR. The DSCR value change should be immediate as well. Acked-by:
-
Anshuman Khandual authored
This patch adds a new documentation file explaining the DSCR support on powerpc platforms. This explains DSCR related data structure, code paths and also available user interfaces. Any further functional changes to the DSCR support in the kernel should definitely update the documentation here. Signed-off-by:
-
Anshuman Khandual authored
This patch adds some in-code documentation to the DSCR related code to make it more readable without having any functional change to it. Signed-off-by:
-
Anshuman Khandual authored
PACA_DSCR offset macro tracks dscr_default element in the paca structure. Better change the name of this macro to match that of the data element it tracks. Makes the code more readable. Signed-off-by:
-
Anshuman Khandual authored
The process context switch code no longer uses dscr_default variable from the sysfs.c file. The variable became unused when we started storing the CPU specific DSCR value in the PACA structure instead. This patch just removes this extern declaration. It was originally added by the following commit. Signed-off-by:
-
Anshuman Khandual authored
Currently DSCR (Data Stream Control Register) can be accessed with mfspr or mtspr instructions inside a thread via two different SPR numbers. One being the user accessible problem state SPR number 0x03 and the other being the privilege state SPR number 0x11. All access through the privilege state SPR number get emulated through illegal instruction exception. Any access through the problem state SPR number raises one facility unavailable exception which sets the thread based dscr_inherit bit and enables DSCR facility through FSCR register thus allowing direct access to DSCR without going through this exception in the future. We set the thread.dscr_inherit bit whether the access was with mfspr or mtspr instruction which is neither correct nor does it match the behaviour through the instruction emulation code path driven from privilege state SPR number. User currently observes two different kind of behaviour when accessing the DSCR through these two SPR numbers. This problem can be observed through these two test cases by replacing the privilege state SPR number with the problem state SPR number. (1) http://ozlabs.org/~anton/junkcode/dscr_default_test.c (2) http://ozlabs.org/~anton/junkcode/dscr_explicit_test.c This patch fixes the problem by making sure that the behaviour visible to the user remains the same irrespective of which SPR number is being used. Inside facility unavailable exception, we check whether it was cuased by a mfspr or a mtspr isntrucction. In case of mfspr instruction, just emulate the instruction. In case of mtspr instruction, set the thread based dscr_inherit bit and also enable the facility through FSCR. All user SPR based mfspr instruction will be emulated till one user SPR based mtspr has been executed. Signed-off-by:
-
Michael Neuling authored
When we release the device, we should also invalidate the default context. With this cxl_get_context() will return null after removal. Signed-off-by:
Michael Neuling <mikey@neuling.org> Signed-off-by:
-
David Gibson authored
Commit 28158cd1 "powerpc/eeh: Enhance pcibios_set_pcie_reset_state()" introduced a fix for a problem where certain configurations could lead to pci_reset_function() destroying the state of PCI devices other than the one specified. Unfortunately, the fix has a trivial bug - it calls pci_save_state() again, when it should be calling pci_restore_state(). This corrects the problem. Cc: Gavin Shan <gwshan@au1.ibm.com> Signed-off-by:
-
- 04 Jun, 2015 5 commits
-
-
Jeremy Kerr authored
This change adds a char device to access the "PRD" (processor runtime diagnostics) channel to OPAL firmware. Includes contributions from Vaidyanathan Srinivasan, Neelesh Gupta & Vishal Kulkarni. Signed-off-by:
Stewart Smith <stewart@linux.vnet.ibm.com> Signed-off-by:
-
Jeremy Kerr authored
The (upcoming) opal-prd driver needs to access the message notifier and xscom code, so add EXPORT_SYMBOL_GPL macros for these. Signed-off-by:
-
Jeremy Kerr authored
opal_ipmi_init and opal_flash_init are equivalent, except for the compatbile string. Merge these two into a common opal_pdev_init function. Signed-off-by:
-
Anton Blanchard authored
Signed-off-by:
-
Cédric Le Goater authored
The opal_{get,set}_param calls return internal error codes which need to be translated in errnos in Linux. Signed-off-by:Cédric Le Goater <clg@fr.ibm.com> Signed-off-by:
-
- 03 Jun, 2015 2 commits
-
-
Michael Neuling authored
This patch does two things. Firstly it presents the Accelerator Function Unit (AFUs) behind the POWER Service Layer (PSL) as PCI devices on a virtual PCI Host Bridge (vPHB). This in in addition to the PSL being a PCI device itself. As part of the Coherent Accelerator Interface Architecture (CAIA) AFUs can provide an AFU configuration. This AFU configuration recored is architected to be the same as a PCI config space. This patch sets discovers the AFU configuration records, provides AFU config space read/write functions to these configuration records. It then enumerates the PCI bus. It also hooks in PCI ops where appropriate. It also destroys the vPHB when the physical card is removed. Secondly, it add an in kernel API for AFU to use CXL. AFUs must present a driver that firstly binds as a PCI device. This PCI device can then be using to do CXL specific operations (that can't sit in the PCI ops) using this API. Signed-off-by:
Ian Munsie <imunsie@au1.ibm.com> Signed-off-by:
-
Michael Neuling authored
The cxl kernel API will allow drivers other than cxl to export a file descriptor which has the same userspace API. These file descriptors will be able to be used against libcxl. This exports those file ops for use by other drivers. Signed-off-by:
-
