Previously we unloaded the interrupts and reloaded in order to work around
a channel reset bug that cleared the MSIX table. This approach just isn't
practical when a reset needs to happen in the error handler that just
happens to be running in interrupt context (bottom half). It looks like we
can work around the hardware issue by just storing a shadow copy of the
MSIX table and restore it after reset.
Signed-off-by: Dave Jiang <dave.jiang@intel.com>
Signed-off-by: Vinod Koul <vinod.koul@intel.com>
Future IOATDMA hardware will take advantage of descriptors residing in
contiguous memory. Setting the descriptor ring in max config DMA memory
of 2MB. Each channel will need 2 of these chunks. This should provide 64k
of 64B descriptors.
Signed-off-by: Dave Jiang <dave.jiang@intel.com>
Signed-off-by: Vinod Koul <vinod.koul@intel.com>
Moving to contingous memory backed descriptor rings. This makes is really
difficult and complex to do reshape. Going to remove this as I don't think
we need to do it anymore.
Signed-off-by: Dave Jiang <dave.jiang@intel.com>
Signed-off-by: Vinod Koul <vinod.koul@intel.com>
Converting old pci_pool_* calls to "new" dma_pool_* to make everything
uniform.
Signed-off-by: Dave Jiang <dave.jiang@intel.com>
Signed-off-by: Vinod Koul <vinod.koul@intel.com>
CC [M] drivers/dma/ioat/prep.o
drivers/dma/ioat/prep.c: In function 'ioat_prep_pqxor':
drivers/dma/ioat/prep.c:682:1: warning: the frame size of 1048 bytes is larger than 1024 bytes [-Wframe-larger-than=]
}
^
drivers/dma/ioat/prep.c: In function 'ioat_prep_pqxor_val':
drivers/dma/ioat/prep.c:714:1: warning: the frame size of 1048 bytes is larger than 1024 bytes [-Wframe-larger-than=]
}
gcc version 5.3.1 20151219 (Ubuntu 5.3.1-4ubuntu1)
Cc: Vinod Koul <vinod.koul@intel.com>
Cc: Dan Williams <dan.j.williams@intel.com>
Cc: Dave Jiang <dave.jiang@intel.com>
Cc: Prarit Bhargava <prarit@redhat.com>
Cc: Nicholas Mc Guire <der.herr@hofr.at>
Cc: Jarkko Nikula <jarkko.nikula@linux.intel.com>
Signed-off-by: Tim Gardner <tim.gardner@canonical.com>
Signed-off-by: Vinod Koul <vinod.koul@intel.com>
The dca_ops structure is never modified, so declare it as const.
Done with the help of Coccinelle.
Signed-off-by: Julia Lawall <Julia.Lawall@lip6.fr>
Acked-by: Dan Williams <dan.j.williams@intel.com>
Signed-off-by: Vinod Koul <vinod.koul@intel.com>
Remove pre-3.0 channel status reads. 3.0 and later chansts register
is 64bit and can be read 64bit. This was clarified with the hardware
architects and since the driver now only support 3.0+ we don't need the
legacy support
Signed-off-by: Dave Jiang <dave.jiang@intel.com>
Signed-off-by: Vinod Koul <vinod.koul@intel.com>
Adding AER handlers in order to handle any PCIe errors.
Signed-off-by: Dave Jiang <dave.jiang@intel.com>
Signed-off-by: Vinod Koul <vinod.koul@intel.com>
The ioatdma needs to be queisced and block all additional op submission
during reboots. When NET_DMA was used, this caused issue as ops were still
being sent to ioatdma during reboots even though PCI BME has been turned
off. Even though NET_DMA has been deprecated, we need to prevent similar
situations. The shutdown handler should address that.
Signed-off-by: Dave Jiang <dave.jiang@intel.com>
Signed-off-by: Vinod Koul <vinod.koul@intel.com>
Adding the Broadwell Xeon ioatdma PCI device IDs and
related bits. This is still IOATDMA 3.2 based hw.
Signed-off-by: Dave Jiang <dave.jiang@intel.com>
Signed-off-by: Vinod Koul <vinod.koul@intel.com>
Sparse reported:
drivers/dma/ioat/prep.c:637:27: sparse: Variable length array is used.
Assigning a static value for the array.
Signed-off-by: Dave Jiang <dave.jiang@intel.com>
Signed-off-by: Vinod Koul <vinod.koul@intel.com>
The prep lock gets acquired in ioat_check_space_lock and released in
ioat_tx_submit_unlock. Setting the annotations so sparse does not freak out.
drivers/dma/ioat/dma.c:273:30: sparse: context imbalance in 'ioat_tx_submit_unlock' - unexpected unlock
drivers/dma/ioat/dma.c:476:5: sparse: context imbalance in 'ioat_check_space_lock' - wrong count at exit
Signed-off-by: Dave Jiang <dave.jiang@intel.com>
Signed-off-by: Vinod Koul <vinod.koul@intel.com>
Simplifying the end return. This existed in the original code but was
flagged when refactoring of the code made it appear it's new.
coccinelle warnings: (new ones prefixed by >>)
>> drivers/dma/ioat/init.c:1018:1-3: WARNING: end returns can be simpified
Signed-off-by: Dave Jiang <dave.jiang@intel.com>
Signed-off-by: Vinod Koul <vinod.koul@intel.com>
IOAT_COMPLETION_PENDING flag was deprecated for v2 and v3 drivers but was
not cleaned up. Doing that now. The commit deprecated this flag was
4dec23d7 ioatdma: fix race between updating ioat->head and
IOAT_COMPLETION_PENDING.
Signed-off-by: Dave Jiang <dave.jiang@intel.com>
Signed-off-by: Vinod Koul <vinod.koul@intel.com>
./scripts/kerne-doc is reporting errors on dma.h. Clean up all reported
errors.
Signed-off-by: Dave Jiang <dave.jiang@intel.com>
Acked-by: Dan Williams <dan.j.williams@intel.com>
Signed-off-by: Vinod Koul <vinod.koul@intel.com>
Since we are a "single" device driver now we no longer require the function
pointers in ioatdma_device. Remove.
Signed-off-by: Dave Jiang <dave.jiang@intel.com>
Acked-by: Dan Williams <dan.j.williams@intel.com>
Signed-off-by: Vinod Koul <vinod.koul@intel.com>
Moving the relevant functions to their respective .c files and removal of
dma_v3.c file. Also removed various ioat3 references when appropriate.
Signed-off-by: Dave Jiang <dave.jiang@intel.com>
Acked-by: Dan Williams <dan.j.williams@intel.com>
Signed-off-by: Vinod Koul <vinod.koul@intel.com>
Move all DMA descriptor prepping functions to prep.c file. Fixup all
broken bits caused by the move.
Signed-off-by: Dave Jiang <dave.jiang@intel.com>
Acked-by: Dan Williams <dan.j.williams@intel.com>
Signed-off-by: Vinod Koul <vinod.koul@intel.com>
Moving all the init routines to init.c and fixup anything broken during
the move.
Signed-off-by: Dave Jiang <dave.jiang@intel.com>
Acked-by: Dan Williams <dan.j.williams@intel.com>
Signed-off-by: Vinod Koul <vinod.koul@intel.com>
Move and fixup all sysfs related bits to sysfs.c file.
Signed-off-by: Dave Jiang <dave.jiang@intel.com>
Acked-by: Dan Williams <dan.j.williams@intel.com>
Signed-off-by: Vinod Koul <vinod.koul@intel.com>
Clean out dma_v2 and remove ioat2 calls since we are moving everything
to just ioat.
Signed-off-by: Dave Jiang <dave.jiang@intel.com>
Acked-by: Dan Williams <dan.j.williams@intel.com>
Signed-off-by: Vinod Koul <vinod.koul@intel.com>
Changing the variable names for ioatdma_device to be consistently named
ioat_dma instead of device/dma in order to avoid confusion and distinct
from struct device. This will clearly indicate that it is an
ioatdma_device. This also make all the naming consistent that the dma
device is ioat_dma and all the channels are ioat_chan.
Signed-off-by: Dave Jiang <dave.jiang@intel.com>
Acked-by: Dan Williams <dan.j.williams@intel.com>
Signed-off-by: Vinod Koul <vinod.koul@intel.com>
Kill the common ioatdma channel structure and everything that is not
dma_chan to be ioat_dma_chan. Since we don't have to worry about v1
and v2 ioatdma anymore this makes it much cleaner and obvious for
maintenance.
Signed-off-by: Dave Jiang <dave.jiang@intel.com>
Acked-by: Dan Williams <dan.j.williams@intel.com>
Signed-off-by: Vinod Koul <vinod.koul@intel.com>
Removal of support for ioatdma v2 device support.
Signed-off-by: Dave Jiang <dave.jiang@intel.com>
Acked-by: Dan Williams <dan.j.williams@intel.com>
Signed-off-by: Vinod Koul <vinod.koul@intel.com>
Cleaning up of ioat1 specific code as it is no longer supported
Signed-off-by: Dave Jiang <dave.jiang@intel.com>
Acked-by: Dan Williams <dan.j.williams@intel.com>
Signed-off-by: Vinod Koul <vinod.koul@intel.com>
Removal of any devices that are ioatdma pre-3.0. This is the first step
in attempting to clean up the ioatdma driver and remove hw no longer
supported.
Signed-off-by: Dave Jiang <dave.jiang@intel.com>
Acked-by: Dan Williams <dan.j.williams@intel.com>
Signed-off-by: Vinod Koul <vinod.koul@intel.com>
If the allocation order is 16, then the u16 count will overflow and wrap
to zero when assigned the value 1 << 16.
Change the type of 'total_descs' to int, so that it is large enough to
store a value equal or greater than 1 << 16.
Signed-off-by: Allen Hubbe <Allen.Hubbe@emc.com>
Acked-by: Dave Jiang <dave.jiang@intel.com>
Signed-off-by: Vinod Koul <vinod.koul@intel.com>
If the allocation order is 16, then the u16 index will overflow and wrap
to zero instead of being equal or greater than 1 << 16. The loop
condition will always be true, and the loop will run until all the
memory resources are depleted.
Change the type of index 'i' to u32, so that it is large enough to store
a value equal or greater than 1 << 16.
Signed-off-by: Allen Hubbe <Allen.Hubbe@emc.com>
Acked-by: Dave Jiang <dave.jiang@intel.com>
Signed-off-by: Vinod Koul <vinod.koul@intel.com>
The dmaengine core assumes that async DMA devices will only be removed
when they not used anymore, or it assumes dma_async_device_unregister()
will only be called by dma driver exit routines. But this assumption is
not true for the IOAT driver, which calls dma_async_device_unregister()
from ioat_remove(). So current IOAT driver doesn't support device
hot-removal because it may cause system crash to hot-remove an inuse
IOAT device.
To support CPU socket hot-removal, all PCI devices, including IOAT
devices embedded in the socket, will be hot-removed. The idea solution
is to enhance the dmaengine core and IOAT driver to support hot-removal,
but that's too hard.
This patch implements a hack to disable IOAT devices under hotplug-capable
CPU socket so it won't break socket hot-removal.
Signed-off-by: Jiang Liu <jiang.liu@linux.intel.com>
Signed-off-by: Vinod Koul <vinod.koul@intel.com>
This allows claiming of non-RAID channels as a private channel. This
prevents breakage of MDRAID using the IOATDMA channels via
async_tx but also allows agents such as NTB to claim channels
exclusively for its usages.
Signed-off-by: Dave Jiang <dave.jiang@intel.com>
Signed-off-by: Vinod Koul <vinod.koul@intel.com>
Free Software Foundation mailing address has been moved in the past and some
of the addresses here are outdated. Remove them from file headers since the
COPYING file in the kernel sources includes it.
Signed-off-by: Jarkko Nikula <jarkko.nikula@linux.intel.com>
Signed-off-by: Vinod Koul <vinod.koul@intel.com>
BDX-DE IOATDMA reports incorrect DMACAP register for PQ related
ops. Ignoring those bits.
Signed-off-by: Dave Jiang <dave.jiang@intel.com>
Acked-by: Dan Williams <dan.j.williams@intel.com>
Signed-off-by: Vinod Koul <vinod.koul@intel.com>
wait_for_completion_timeout reaching timeout was being ignored,
fail the self-test if timeout condition occurs.
v2: fixup of coding style issues.
Signed-off-by: Nicholas Mc Guire <der.herr@hofr.at>
Acked-by: Dave Jiang <dave.jiang@intel.com>
Reviewed-by: Prarit Bhargava <prarit@redhat.com>
Signed-off-by: Vinod Koul <vinod.koul@intel.com>
Adding PCI device IDs and hooks in workarounds for Broadwell DE ioatdma.
Signed-off-by: Dave Jiang <dave.jiang@intel.com>
Signed-off-by: Vinod Koul <vinod.koul@intel.com>
The err completion callback is missing from the error handler. Two
reasons we never hit this. On Xeon because the hw err workaround, the
completion happens on a NULL descriptor so we don't do callback on the
PQ descriptor. On Atom we have DWBES support and thus the callback already
happened or we don't halt on error, so that was take cared of. But this code
needs to be corrected for future error handlers.
Signed-off-by: Dave Jiang <dave.jiang@intel.com>
Acked-by Dan Williams <dan.j.williams@intel.com>
Signed-off-by: Vinod Koul <vinod.koul@intel.com>
Per commit "77873803363c net_dma: mark broken" net_dma is no longer used
and there is no plan to fix it.
This is the mechanical removal of bits in CONFIG_NET_DMA ifdef guards.
Reverting the remainder of the net_dma induced changes is deferred to
subsequent patches.
Marked for stable due to Roman's report of a memory leak in
dma_pin_iovec_pages():
https://lkml.org/lkml/2014/9/3/177
Cc: Dave Jiang <dave.jiang@intel.com>
Cc: Vinod Koul <vinod.koul@intel.com>
Cc: David Whipple <whipple@securedatainnovations.ch>
Cc: Alexander Duyck <alexander.h.duyck@intel.com>
Cc: <stable@vger.kernel.org>
Reported-by: Roman Gushchin <klamm@yandex-team.ru>
Acked-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
To be future-proof and for better readability the time comparisons are modified
to use time_before_jiffies() instead of plain, error-prone math.
Signed-off-by: Manuel Schölling <manuel.schoelling@gmx.de>
[djbw: use time_before_jiffies() to make argument order more clear]
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
As result of deprecation of MSI-X/MSI enablement functions
pci_enable_msix() and pci_enable_msi_block() all drivers
using these two interfaces need to be updated to use the
new pci_enable_msi_range() or pci_enable_msi_exact()
and pci_enable_msix_range() or pci_enable_msix_exact()
interfaces.
Function pci_enable_msix() returns a tri-state value while
pci_enable_msi_exact() is a canonical zero/-errno variant.
The former is being phased out in favor of the latter.
In case of 'ioat' there (should be) no difference.
Cc: Vinod Koul <vinod.koul@intel.com>
Signed-off-by: Alexander Gordeev <agordeev@redhat.com>
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
Includes an appropriate header file dma_v2.h in ioat/dca.c because
functions ioat2_dca_init() and ioat3_dca_init() have their function
declarations in dma_v2.h.
This eliminates the following warning in ioat/dca.c:
drivers/dma/ioat/dca.c:410:22: warning: no previous prototype for ‘ioat2_dca_init’ [-Wmissing-prototypes]
drivers/dma/ioat/dca.c:624:22: warning: no previous prototype for ‘ioat3_dca_init’ [-Wmissing-prototypes]
Signed-off-by: Rashika Kheria <rashika.kheria@gmail.com>
Reviewed-by: Josh Triplett <josh@joshtriplett.org>
Acked-by: Vinod Koul <vinod.koul@intel.com>
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
Mark the functions ioat3_prep_xor_val(), ioat3_prep_pq_val() and
ioat3_prep_pqxor_val() as static in dma_v3.c because they are not used
outside this file.
This eliminates the following warnings in dma_v3.c:
drivers/dma/ioat/dma_v3.c:741:1: warning: no previous prototype for ‘ioat3_prep_xor_val’ [-Wmissing-prototypes]
drivers/dma/ioat/dma_v3.c:1092:1: warning: no previous prototype for ‘ioat3_prep_pq_val’ [-Wmissing-prototypes]
drivers/dma/ioat/dma_v3.c:1134:1: warning: no previous prototype for ‘ioat3_prep_pqxor_val’ [-Wmissing-prototypes]
Signed-off-by: Rashika Kheria <rashika.kheria@gmail.com>
Reviewed-by: Josh Triplett <josh@joshtriplett.org>
Acked-by: Vinod Koul <vinod.koul@intel.com>
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
Use PCI standard marco dev_is_pci() instead of directly compare
pci_bus_type to check whether it is pci device.
Signed-off-by: Yijing Wang <wangyijing@huawei.com>
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
Since commit 7787380336 "net_dma: mark broken" we no longer pin dma
engines active for the network-receive-offload use case. As a result
the ->free_chan_resources() that occurs after the driver self test no
longer has a NET_DMA induced ->alloc_chan_resources() to back it up. A
late firing irq can lead to ksoftirqd spinning indefinitely due to the
tasklet_disable() performed by ->free_chan_resources(). Only
->alloc_chan_resources() can clear this condition in affected kernels.
This problem has been present since commit 3e037454bc "I/OAT: Add
support for MSI and MSI-X" in 2.6.24, but is now exposed. Given the
NET_DMA use case is deprecated we can revisit moving the driver to use
threaded irqs. For now, just tear down the irq and tasklet properly by:
1/ Disable the irq from triggering the tasklet
2/ Disable the irq from re-arming
3/ Flush inflight interrupts
4/ Flush the timer
5/ Flush inflight tasklets
References:
https://lkml.org/lkml/2014/1/27/282https://lkml.org/lkml/2014/2/19/672
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Steven Rostedt <rostedt@goodmis.org>
Cc: <stable@vger.kernel.org>
Reported-by: Mike Galbraith <bitbucket@online.de>
Reported-by: Stanislav Fomichev <stfomichev@yandex-team.ru>
Tested-by: Mike Galbraith <bitbucket@online.de>
Tested-by: Stanislav Fomichev <stfomichev@yandex-team.ru>
Reviewed-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
Pull dmaengine changes from Dan
1/ Bartlomiej and Dan finalized a rework of the dma address unmap
implementation.
2/ In the course of testing 1/ a collection of enhancements to dmatest
fell out. Notably basic performance statistics, and fixed / enhanced
test control through new module parameters 'run', 'wait', 'noverify',
and 'verbose'. Thanks to Andriy and Linus for their review.
3/ Testing the raid related corner cases of 1/ triggered bugs in the
recently added 16-source operation support in the ioatdma driver.
4/ Some minor fixes / cleanups to mv_xor and ioatdma.
Conflicts:
drivers/dma/dmatest.c
Signed-off-by: Vinod Koul <vinod.koul@intel.com>
The implementation of ioat3_irq_reinit has two bugs:
1/ The mode is incorrectly set to MSIX for the MSI case
2/ The 'dev_id' parameter to free_irq is the ioatdma_device not the channel in
the msi and intx case
Include a small cleanup to clarify that ioat3_irq_reinit is only for bwd
hardware
Cc: Dave Jiang <dave.jiang@intel.com>
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
Once we have determined that we will not have all of our desired msix
vectors there is no point in attempting a single msix allocation. The
driver will already need to read registers to determine the source of
the interrupt the fact that it is msix is moot. Fallback directly to
msi.
Reported-by: Alexander Gordeev <agordeev@redhat.com>
Cc: Dave Jiang <dave.jiang@intel.com>
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
Use a single cache for all sed allocations. No need to make it per
channel. This also avoids the slub_debug warnings for multiple caches
with the same name.
Switching to dmam_pool_create() to fix leaking the dma pools on
initialization failure and lets us kill ioat3_dma_remove().
Cc: Dave Jiang <dave.jiang@intel.com>
Acked-by: Dave Jiang <dave.jiang@intel.com>
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
