Age | Commit message (Collapse) | Author | Files | Lines |
|
We need to add memory barrier so that the page table walk doesn't happen
before the cpumask is set and made visible to the other cpus. We need
to use a sync here instead of lwsync because lwsync is not sufficient for
store/load ordering.
We also need to add an if (mm) check so that we do the right thing when called
with a kernel context. For kernel context, we have mm = NULL. W.r.t kernel
address we can skip setting the mm cpumask.
Fixes: 0f4bc0932e ("powerpc/mm/cxl: Add the fault handling cpu to mm cpumask")
Cc: Andrew Donnellan <andrew.donnellan@au1.ibm.com>
Reported-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Reported-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: Aneesh Kumar K.V <aneesh.kumar@linux.vnet.ibm.com>
Acked-by: Andrew Donnellan <andrew.donnellan@au1.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
|
|
We use mm cpumask for serializing against lockless page table walk.
Anybody who is doing a lockless page table walk is expected to disable
irq and only cpus in mm cpumask is expected do the lockless walk. This
ensure that a THP split can send IPI to only cpus in the mm cpumask,
to make sure there are no parallel lockless page table walk.
Add the CAPI fault handling cpu to the mm cpumask so that we can do
the lockless page table walk while inserting hash page table entries.
Reviewed-by: Frederic Barrat <fbarrat@linux.vnet.ibm.com>
Signed-off-by: Aneesh Kumar K.V <aneesh.kumar@linux.vnet.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
|
|
This patch exports a in-kernel 'library' API which can be called by
other drivers to help interacting with an IBM XSL on a POWER9 system.
The XSL (Translation Service Layer) is a stripped down version of the
PSL (Power Service Layer) used in some cards such as the Mellanox CX5.
Like the PSL, it implements the CAIA architecture, but has a number
of differences, mostly in it's implementation dependent registers.
The XSL also uses a special DMA cxl mode, which uses a slightly
different init sequence for the CAPP and PHB.
Signed-off-by: Andrew Donnellan <andrew.donnellan@au1.ibm.com>
Signed-off-by: Christophe Lombard <clombard@linux.vnet.ibm.com>
Acked-by: Frederic Barrat <fbarrat@linux.vnet.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
|
|
A previous set of patches "cxl: Add support for Coherent Accelerator
Interface Architecture 2.0" has introduced a new support for the CAPI
cards. These patches have been tested on Simulation environment and
quite a bit of them have been tested on real hardware.
This patch brings new fixes after a series of tests carried out on new
equipment:
- Add POWER9 definition.
- Re-enable any masked interrupts when the AFU is not activated
after resetting the AFU.
- Remove the api cxl_is_psl8/9 which is no longer useful.
- Do not dump CAPI1 registers.
- Rewrite cxl_is_page_fault() function.
- Do not register slb callack on P9.
Fixes: f24be42aab37 ("cxl: Add psl9 specific code")
Signed-off-by: Christophe Lombard <clombard@linux.vnet.ibm.com>
Acked-by: Frederic Barrat <fbarrat@linux.vnet.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
|
|
The new Coherent Accelerator Interface Architecture, level 2, for the
IBM POWER9 brings new content and features:
- POWER9 Service Layer
- Registers
- Radix mode
- Process element entry
- Dedicated-Shared Process Programming Model
- Translation Fault Handling
- CAPP
- Memory Context ID
If a valid mm_struct is found the memory context id is used for each
transaction associated with the process handle. The PSL uses the
context ID to find the corresponding process element.
Signed-off-by: Christophe Lombard <clombard@linux.vnet.ibm.com>
Acked-by: Frederic Barrat <fbarrat@linux.vnet.ibm.com>
[mpe: Fixup comment formatting, unsplit long strings]
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
|
|
The mm_struct corresponding to the current task is acquired each time
an interrupt is raised. So to simplify the code, we only get the
mm_struct when attaching an AFU context to the process.
The mm_count reference is increased to ensure that the mm_struct can't
be freed. The mm_struct will be released when the context is detached.
A reference on mm_users is not kept to avoid a circular dependency if
the process mmaps its cxl mmio and forget to unmap before exiting.
The field glpid (pid of the group leader associated with the pid), of
the structure cxl_context, is removed because it's no longer useful.
Signed-off-by: Christophe Lombard <clombard@linux.vnet.ibm.com>
Reviewed-by: Andrew Donnellan <andrew.donnellan@au1.ibm.com>
Acked-by: Frederic Barrat <fbarrat@linux.vnet.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
|
|
<linux/sched.h> into <linux/sched/signal.h>
Fix up affected files that include this signal functionality via sched.h.
Acked-by: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: linux-kernel@vger.kernel.org
Signed-off-by: Ingo Molnar <mingo@kernel.org>
|
|
<linux/sched/mm.h>
We are going to split <linux/sched/mm.h> out of <linux/sched.h>, which
will have to be picked up from other headers and a couple of .c files.
Create a trivial placeholder <linux/sched/mm.h> file that just
maps to <linux/sched.h> to make this patch obviously correct and
bisectable.
The APIs that are going to be moved first are:
mm_alloc()
__mmdrop()
mmdrop()
mmdrop_async_fn()
mmdrop_async()
mmget_not_zero()
mmput()
mmput_async()
get_task_mm()
mm_access()
mm_release()
Include the new header in the files that are going to need it.
Acked-by: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: linux-kernel@vger.kernel.org
Signed-off-by: Ingo Molnar <mingo@kernel.org>
|
|
_PAGE_PRIVILEGED means the page can be accessed only by the kernel. This
is done to keep pte bits similar to PowerISA 3.0 Radix PTE format. User
pages are now marked by clearing _PAGE_PRIVILEGED bit.
Previously we allowed the kernel to have a privileged page in the lower
address range (USER_REGION). With this patch such access is denied.
We also prevent a kernel access to a non-privileged page in higher
address range (ie, REGION_ID != 0).
Both the above access scenarios should never happen.
Cc: Arnd Bergmann <arnd@arndb.de>
Cc: Jeremy Kerr <jk@ozlabs.org>
Cc: Frederic Barrat <fbarrat@linux.vnet.ibm.com>
Acked-by: Ian Munsie <imunsie@au1.ibm.com>
Signed-off-by: Aneesh Kumar K.V <aneesh.kumar@linux.vnet.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
|
|
This splits the _PAGE_RW bit into _PAGE_READ and _PAGE_WRITE. It also
removes the dependency on _PAGE_USER for implying read only. Few things
to note here is that, we have read implied with write and execute
permission. Hence we should always find _PAGE_READ set on hash pte
fault.
We still can't switch PROT_NONE to !(_PAGE_RWX). Auto numa depends on
marking a prot none pte _PAGE_WRITE. (For more details look at
b191f9b106ea "mm: numa: preserve PTE write permissions across a NUMA
hinting fault")
Cc: Arnd Bergmann <arnd@arndb.de>
Cc: Jeremy Kerr <jk@ozlabs.org>
Cc: Frederic Barrat <fbarrat@linux.vnet.ibm.com>
Acked-by: Ian Munsie <imunsie@au1.ibm.com>
Signed-off-by: Aneesh Kumar K.V <aneesh.kumar@linux.vnet.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
|
|
The current code will set _PAGE_USER to the access flags for any
fault address, because the ~ operation will be true for all address we
take a fault on. But setting _PAGE_USER also means that the fault will
be handled only if the page table have _PAGE_USER set. Hence there is
no security hole with the current code.
Now if it is an user space access, then the change in this patch really
don't have an impact because we have (!ctx->kernel) set true
and we take the if condition true.
Now kernel context created fault on an address in the kernel range
will result in a fault loop because we will not insert the
hash pte due to access and pte permission mismatch. This patch fix
the above issue.
Fixes: f204e0b8cedd ("cxl: Driver code for powernv PCIe based cards for userspace access")
Reviewed-by: Andrew Donnellan <andrew.donnellan@au1.ibm.com>
Acked-by: Ian Munsie <imunsie@au1.ibm.com>
Signed-off-by: Aneesh Kumar K.V <aneesh.kumar@linux.vnet.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
|
|
A few functions are mostly common between bare-metal and guest and
just need minor tuning. To avoid crowding the backend API, introduce a
few 'if' based on the CPU being in HV mode.
Co-authored-by: Christophe Lombard <clombard@linux.vnet.ibm.com>
Signed-off-by: Frederic Barrat <fbarrat@linux.vnet.ibm.com>
Signed-off-by: Christophe Lombard <clombard@linux.vnet.ibm.com>
Reviewed-by: Manoj Kumar <manoj@linux.vnet.ibm.com>
Acked-by: Ian Munsie <imunsie@au1.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
|
|
The backend API (in cxl.h) lists some low-level functions whose
implementation is different on bare-metal and in a guest. Each
environment implements its own functions, and the common code uses
them through function pointers, defined in cxl_backend_ops
Co-authored-by: Christophe Lombard <clombard@linux.vnet.ibm.com>
Signed-off-by: Frederic Barrat <fbarrat@linux.vnet.ibm.com>
Signed-off-by: Christophe Lombard <clombard@linux.vnet.ibm.com>
Reviewed-by: Manoj Kumar <manoj@linux.vnet.ibm.com>
Acked-by: Ian Munsie <imunsie@au1.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
|
|
Presently when a user-space process issues CXL_IOCTL_START_WORK ioctl we
store the pid of the current task_struct and use it to get pointer to
the mm_struct of the process, while processing page or segment faults
from the capi card. However this causes issues when the thread that had
originally issued the start-work ioctl exits in which case the stored
pid is no more valid and the cxl driver is unable to handle faults as
the mm_struct corresponding to process is no more accessible.
This patch fixes this issue by using the mm_struct of the next alive
task in the thread group. This is done by iterating over all the tasks
in the thread group starting from thread group leader and calling
get_task_mm on each one of them. When a valid mm_struct is obtained the
pid of the associated task is stored in the context replacing the
exiting one for handling future faults.
The patch introduces a new function named get_mem_context that checks if
the current task pointed to by ctx->pid is dead? If yes it performs the
steps described above. Also a new variable cxl_context.glpid is
introduced which stores the pid of the thread group leader associated
with the context owning task.
Reported-by: Matthew R. Ochs <mrochs@linux.vnet.ibm.com>
Reported-by: Frank Haverkamp <HAVERKAM@de.ibm.com>
Suggested-by: Ian Munsie <imunsie@au1.ibm.com>
Signed-off-by: Vaibhav Jain <vaibhav@linux.vnet.ibm.com>
Acked-by: Ian Munsie <imunsie@au1.ibm.com>
Reviewed-by: Frederic Barrat <fbarrat@linux.vnet.ibm.com>
Reviewed-by: Matthew R. Ochs <mrochs@linux.vnet.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
|
|
We only need to check the pid attached to this context for userspace contexts.
Kernel contexts can skip this check.
Signed-off-by: Michael Neuling <mikey@neuling.org>
Acked-by: Ian Munsie <imunsie@au1.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
|
|
This patch adds tracepoints throughout the cxl driver, which can provide
insight into:
- Context lifetimes
- Commands sent to the PSL and AFU and their completion status
- Segment and page table misses and their resolution
- PSL and AFU interrupts
- slbia calls from the powerpc copro_fault code
These tracepoints are mostly intended to aid in debugging (particularly
for new AFU designs), and may be useful standalone or in conjunction
with hardware traces collected by the PSL (read out via the trace
interface in debugfs) and AFUs.
Signed-off-by: Ian Munsie <imunsie@au1.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
|
|
If a context is being detached and we get a translation fault for it
there is little point getting it's mm and handling the fault, so just
respond with an address error and return earlier.
Signed-off-by: Ian Munsie <imunsie@au1.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
|
|
upatepp can get called for a nohpte fault when we find from the linux
page table that the translation was hashed before. In that case
we are sure that there is no existing translation, hence we could
avoid doing tlbie.
We could possibly race with a parallel fault filling the TLB. But
that should be ok because updatepp is only ever relaxing permissions.
We also look at linux pte permission bits when filling hash pte
permission bits. We also hold the linux pte busy bits while
inserting/updating a hashpte entry, hence a paralle update of
linux pte is not possible. On the other hand mprotect involves
ptep_modify_prot_start which cause a hpte invalidate and not updatepp.
Performance number:
We use randbox_access_bench written by Anton.
Kernel with THP disabled and smaller hash page table size.
86.60% random_access_b [kernel.kallsyms] [k] .native_hpte_updatepp
2.10% random_access_b random_access_bench [.] doit
1.99% random_access_b [kernel.kallsyms] [k] .do_raw_spin_lock
1.85% random_access_b [kernel.kallsyms] [k] .native_hpte_insert
1.26% random_access_b [kernel.kallsyms] [k] .native_flush_hash_range
1.18% random_access_b [kernel.kallsyms] [k] .__delay
0.69% random_access_b [kernel.kallsyms] [k] .native_hpte_remove
0.37% random_access_b [kernel.kallsyms] [k] .clear_user_page
0.34% random_access_b [kernel.kallsyms] [k] .__hash_page_64K
0.32% random_access_b [kernel.kallsyms] [k] fast_exception_return
0.30% random_access_b [kernel.kallsyms] [k] .hash_page_mm
With Fix:
27.54% random_access_b random_access_bench [.] doit
22.90% random_access_b [kernel.kallsyms] [k] .native_hpte_insert
5.76% random_access_b [kernel.kallsyms] [k] .native_hpte_remove
5.20% random_access_b [kernel.kallsyms] [k] fast_exception_return
5.12% random_access_b [kernel.kallsyms] [k] .__hash_page_64K
4.80% random_access_b [kernel.kallsyms] [k] .hash_page_mm
3.31% random_access_b [kernel.kallsyms] [k] data_access_common
1.84% random_access_b [kernel.kallsyms] [k] .trace_hardirqs_on_caller
Signed-off-by: Aneesh Kumar K.V <aneesh.kumar@linux.vnet.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
|
|
In certain circumstances the PSL (Power Service Layer, which provides
translation services for CXL hardware) can send an interrupt for a
segment miss that the kernel has already handled. This can happen if
multiple translations for the same segment are queued in the PSL before
the kernel has restarted the first translation.
The CXL driver does not expect this situation and does not check if a
segment had already been handled. This could cause a duplicate segment
table entry which in turn caused a PSL error taking down the card.
This patch fixes the issue by checking for existing entries in the
segment table that match the segment we are trying to insert, so as to
avoid inserting duplicate entries.
Signed-off-by: Ian Munsie <imunsie@au1.ibm.com>
Reviewed-by: Aneesh Kumar K.V <aneesh.kumar@linux.vnet.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
|
|
This moves the segment table hash calculation from cxl_load_segment()
into find_free_sste() since that is the only place it is actually used.
Signed-off-by: Ian Munsie <imunsie@au1.ibm.com>
Reviewed-by: Aneesh Kumar K.V <aneesh.kumar@linux.vnet.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
|
|
This patch simplifies the process of finding a free segment table entry
by disabling the secondary hash. This reduces the number of possible
entries in the segment table for a given address from 16 to 8.
Due to the large segment sizes we use it is extremely unlikely that the
secondary hash would ever have been used in practice, so this should not
have any negative impacts and may even improve performance due to the
reduced number of comparisons that software & hardware need to perform.
This patch clears the SC bit in the hardware's state register
(CXL_PSL_SR_An) to disable the secondary hash in the hardware since we
can no longer fill out entries using it.
Signed-off-by: Ian Munsie <imunsie@au1.ibm.com>
Reviewed-by: Aneesh Kumar K.V <aneesh.kumar@linux.vnet.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
|
|
This is the core of the cxl driver.
It adds support for using cxl cards in the powernv environment only (ie POWER8
bare metal). It allows access to cxl accelerators by userspace using the
/dev/cxl/afuM.N char devices.
The kernel driver has no knowledge of the function implemented by the
accelerator. It provides services to userspace via the /dev/cxl/afuM.N
devices. When a program opens this device and runs the start work IOCTL, the
accelerator will have coherent access to that processes memory using the same
virtual addresses. That process may mmap the device to access any MMIO space
the accelerator provides. Also, reads on the device will allow interrupts to
be received. These services are further documented in a later patch in
Documentation/powerpc/cxl.txt.
Documentation of the cxl hardware architecture and userspace API is provided in
subsequent patches.
Signed-off-by: Ian Munsie <imunsie@au1.ibm.com>
Signed-off-by: Michael Neuling <mikey@neuling.org>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
|