| Age | Commit message (Collapse) | Author |
|---|
|
add HS400 mode and tune support
Signed-off-by: Chaotian Jing <chaotian.jing@mediatek.com>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
|
|
The mmc_execute_tuning() has already prepared the opcode,
there is no need to prepare it again at mmc_send_tuning(),
and, there is a BUG of mmc_send_tuning() to determine the opcode
by bus width, assume eMMC was running at HS200, 4bit mode,
then the mmc_send_tuning() will overwrite the opcode from CMD21
to CMD19, then got error.
in addition, extend an argument of "cmd_error" to allow getting
if there was cmd error when tune response.
Signed-off-by: Chaotian Jing <chaotian.jing@mediatek.com>
[Ulf: Rebased patch]
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
|
|
add implement of ops->hw_reset() for eMMC
Signed-off-by: Chaotian Jing <chaotian.jing@mediatek.com>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
|
|
Even if we only use one gpd, we need alloc 2 gpd and make
the gpd->next pointer to the second gpd, or may get gpd checksum
error, this was checked by hardware
Signed-off-by: Chaotian Jing <chaotian.jing@mediatek.com>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
|
|
use the ios->timing directly is better
It can reflect current timing and do settings by timing
Signed-off-by: Chaotian Jing <chaotian.jing@mediatek.com>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
|
|
cmd_ints_mask and data_ints_mask are constant value,
so make it to const
Signed-off-by: Chaotian Jing <chaotian.jing@mediatek.com>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
|
|
Add 400Mhz clock source for HS400 mode
Signed-off-by: Chaotian Jing <chaotian.jing@mediatek.com>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
|
|
Sometime only need set MMC_CAP_HW_RESET for one of MMC hosts,
So set it in device tree is better.
Signed-off-by: Chaotian Jing <chaotian.jing@mediatek.com>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
|
|
Jukka reported about the following warning:
"NOHZ: local_softirq_pending 08"
I remember this warning and we had a similar issue when using workqueues
and calling netif_rx. See commit 5ff3fec ("mac802154: fix NOHZ
local_softirq_pending 08 warning").
This warning occurs when calling "netif_rx" inside the wrong context
(non softirq context). The net core api offers "netif_rx_ni" to call
netif_rx inside the correct softirq context.
Reported-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
Signed-off-by: Alexander Aring <alex.aring@gmail.com>
Acked-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
|
|
Signed-off-by: Jubin John <jubin.john@intel.com>
Signed-off-by: Ira Weiny <ira.weiny@intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
|
|
Add additional rc traces to aid in debugging rc retry logic.
Reviewed-by: Dennis Dalessandro <dennis.dalessandro@intel.com>
Signed-off-by: Mike Marciniszyn <mike.marciniszyn@intel.com>
Signed-off-by: Ira Weiny <ira.weiny@intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
|
|
The name used to create the verbs txreq cache was not qualified with the unit
number. This causes a panic when destroying the cache on a dual HFI systems.
Create a unique name with the unit number with this patch
Reviewed-by: Mike Marciniszyn <mike.marciniszyn@intel.com>
Signed-off-by: Jubin John <jubin.john@intel.com>
Signed-off-by: Ira Weiny <ira.weiny@intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
|
|
Using fw_sbus_load to control SBus firmware load doesn't scale across multiple
HFI1 cards in a single system. This patch ensures that the SBus firmware is
loaded once per ASIC.
Reviewed-by: Dean Luick <dean.luick@intel.com>
Reviewed-by: Dennis Dalessandro <dennis.dalessandro@intel.com>
Signed-off-by: Easwar Hariharan <easwar.hariharan@intel.com>
Signed-off-by: Ira Weiny <ira.weiny@intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
|
|
When under heavy load, the receive interrupt handler can run too long with IRQs
disabled. Add a mixed-mode threading scheme. Initially process packets in the
handler for quick responses (latency). If there are too many packets to
process move to a thread to continue (bandwidth).
Reviewed-by: Mike Marciniszyn <mike.marciniszyn@intel.com>
Signed-off-by: Dean Luick <dean.luick@intel.com>
Signed-off-by: Ira Weiny <ira.weiny@intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
|
|
In preparation for threading the receive interrupt, add irqsaves in the packet
processing path.
When the receive interrupt is threaded, the packet processing path is no longer
guaranteed to have IRQs disabled. Add irqsaves where needed on several locks
in the packet processing path. Anything that did not have an obvious, "close"
irqsave in its caller is a candidate.
Reviewed-by: Mike Marciniszyn <mike.marciniszyn@intel.com>
Signed-off-by: Dean Luick <dean.luick@intel.com>
Signed-off-by: Ira Weiny <ira.weiny@intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
|
|
A value of 2048 increased PSM performance while not impacting verbs
performance.
Reviewed-by: Mike Marciniszyn <mike.marciniszyn@intel.com>
Reviewed-by: Mitko Haralanov <mitko.haralanov@intel.com>
Signed-off-by: Ignacio Hernandez <ignacio.hernandez@intel.com>
Signed-off-by: Ira Weiny <ira.weiny@intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
|
|
The SDMA engines were configured to generate progress interrupts every time they
processed N/2 descriptors (where N is the size of the descriptor queue). This
interval was too infrequent, leading to degraded performance.
This commit adds a module parameter, as well as a recommended default, which
allows for the tuning of the interrupt frequency.
Reviewed-by: Mike Marciniszyn <mike.marciniszyn@intel.com>
Signed-off-by: Mitko Haralanov <mitko.haralanov@intel.com>
Signed-off-by: Ira Weiny <ira.weiny@intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
|
|
device_ops.c should only contain functions that are called by ccw device
drivers. Move the cio internal functions that handle unconditional
reserve + release to device_pgid.c
Acked-by: Peter Oberparleiter <oberpar@linux.vnet.ibm.com>
Signed-off-by: Sebastian Ott <sebott@linux.vnet.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
|
|
device_ops.c should only contain functions that are called by ccw device
drivers. Move the cio internal function ccw_device_call_handler to
device_fsm.c where it's used. Remove some useless comments while at it.
Acked-by: Peter Oberparleiter <oberpar@linux.vnet.ibm.com>
Signed-off-by: Sebastian Ott <sebott@linux.vnet.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
|
|
Each per_cpu() invocation generates extra code. Since there are a lot
of similiar calls in the topology code we can avoid a lot of them.
Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
|
|
Change the flag fields within struct mcck_struct to simple bit fields
to reduce the size of the structure which is used as percpu variable.
Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
|
|
According to the architecture registers are validated and not
revalidated. So change comments and functions names to match.
Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
|
|
Remove all the casts to and from the machine check interruption code.
This patch changes struct mci to a union, which contains an anonymous
structure with the already known bits and in addition an unsigned
long field, which contains the raw machine check interruption code.
This allows to simply assign and decoce the interruption code value
without the need for all those casts we had all the time.
Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
|
|
s390_handle_damage() has character string parameter which was used as
a pointer to verbose error message. The hope was (a lot of years ago)
when analyzing dumps that register R2 would still contain the pointer
and therefore it would be rather easy to tell what went wrong.
However gcc optimizes the strings away since a long time. And even if
it wouldn't it is necessary to have a close look at the machine check
interruption code to tell what's wrong.
So remove the pointless error strings.
Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
|
|
The current disabled wait code stores register contents into their
save areas, however it is (at least) missing the new vector registers.
Given the fact that the whole exercise seems to be rather pointless
simply don't save any registers anymore.
In a "live" system it is always possible to inspect register contents,
and in case of a dump the register contents will be stored by the
dump mechanism.
Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
|
|
With the removal of 31 bit code we can always assume that the epsw
instruction is available. Therefore use the __extract_psw() function
to disable and enable machine checks.
Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
|
|
The address being mapped into a process's memory for notification events was
improperly calculated due to cast that was happening too early. dd->events is a
pointer and should have been casted to unsigned long after the pointer
arithmetic was done, not before.
As a result, processes were looking at the wrong place and not seeing their
notification events.
Signed-off-by: Mitko Haralanov <mitko.haralanov@intel.com>
Signed-off-by: Ira Weiny <ira.weiny@intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
|
|
User SDMA keeps track of progress into the submitted IO vectors by tracking an
offset into the vectors when packets are submitted. This offset is updated
after a successful submission of a txreq to the SDMA engine.
The same offset was used when determining whether an IO vector should be
'freed' (pages unpinned) in the SDMA callback functions.
This was causing a silent data corruption in big jobs (> 2 nodes, 120 ranks
each) on the receive side because the send side was mistakenly unpinning the
vector pages before the HW has processed all descriptors referencing the
vector.
Reviewed-by: Mike Marciniszyn <mike.marciniszyn@intel.com>
Signed-off-by: Mitko Haralanov <mitko.haralanov@intel.com>
Signed-off-by: Ira Weiny <ira.weiny@intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
|
|
The DC firmware overwrites the enable_lane_tx register and does not update it
on a host request to go to Poll. This causes an infinite loop through the LNI
state machine if a link width downgrade occurs. This patch re-sets the
enable_lane_tx register to all 4 lanes.
Reviewed-by: Dean Luick <dean.luick@intel.com>
Signed-off-by: Easwar Hariharan <easwar.hariharan@intel.com>
Signed-off-by: Ira Weiny <ira.weiny@intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
|
|
When under heavy load, the send handler can run too long without allowing other
tasks to run. Add a conditional resched to break this up.
Reviewed-by: Mike Marciniszyn <mike.marciniszyn@intel.com>
Signed-off-by: Dean Luick <dean.luick@intel.com>
Signed-off-by: Ira Weiny <ira.weiny@intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
|
|
Reset firmware instead of reloading Sbus firmware if it's already done for this
ASIC. To work around thermal polling problem in firmware, don't reload Sbus
firmware, instead, reset the firmware on the initialization of the second HFI.
Reviewed-by: Easwar Hariharan <easwar.hariharan@intel.com>
Reviewed-by: Dennis Dalessandro <dennis.dalessandro@intel.com>
Reviewed-by: Dean Luick <dean.luick@intel.com>
Signed-off-by: Caz Yokoyama <caz.yokoyama@intel.com>
Signed-off-by: Ira Weiny <ira.weiny@intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
|
|
Driver does not verify userid for shared context assignments, allowing
malicious user access.
Reviewed by: Mike Marciniszyn <mike.marciniszyn@intel.com>
Signed-off-by: Jareer H Abdel-Qader <jareer.h.abdel-qader@intel.com>
Signed-off-by: Ira Weiny <ira.weiny@intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
|
|
This fixes transmit errors when the number of scatter gather elements in the
request is more that the number of per packet descriptors supported by the
hardware, allocate and coalesce the extra scatter gather elements into a single
buffer. The last descriptor is reserved and used for this coalesced buffer.
Verbs potentially need this support when transferring small data chunks
involving different memory regions.
Reviewed-by: Mike Marciniszyn <mike.marciniszyn@intel.com>
Reviewed-by: Mitko Haralanov <mitko.haralanov@intel.com>
Signed-off-by: Niranjana Vishwanathapura <niranjana.vishwanathapura@intel.com>
Signed-off-by: Ira Weiny <ira.weiny@intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
|
|
The QSFP interface code has been running without issues and the flag is
never set to off. This patch removes the QSFP_ENABLED bit from HFI1_CAP.
Reviewed-by: Mike Marciniszyn <mike.marciniszyn@intel.com>
Signed-off-by: Easwar Hariharan <easwar.hariharan@intel.com>
Signed-off-by: Ira Weiny <ira.weiny@intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
|
|
If packets stop egressing the hardware link, software can lock up.
Implement a timeout for send context halt recovery. This patch increases the
timeout for packet egress to 500 us and timer resets to zero if the packet
occupancy changes. Also we bounce the link on time out.
Reviewed-by: Dean Luick <dean.luick@intel.com>
Signed-off-by: Vennila Megavannan <vennila.megavannan@intel.com>
Signed-off-by: Ira Weiny <ira.weiny@intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
|
|
The latest version of the 8051 firmware will wait longer
when bringing the link down. Extend the driver's timeout
to go with that.
Reviewed-by: Dennis Dalessandro <dennis.dalessandro@intel.com>
Signed-off-by: Dean Luick <dean.luick@intel.com>
Signed-off-by: Ira Weiny <ira.weiny@intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
|
|
The ASIC registers were not reset on FLR, and the code to
protect the ASIC block against multiple initializations by
peer HFIs did not extend to multiple ASICs in a system. This
patch addresses this gap.
Reviewed-by: Dean Luick <dean.luick@intel.com>
Signed-off-by: Easwar Hariharan <easwar.hariharan@intel.com>
Signed-off-by: Ira Weiny <ira.weiny@intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
|
|
The clear_ahg call is new in hfi1 vs qib. For small messages the progress
routine always builds one and clears out the ahg state when the queue has gone
to empty which is the predominant case for small messages.
Inline the routine and avoid the call to sdma_ahg_free to mitigate the
performance impact. Finally, move the routine to qp.h for scope reasons.
Reviewed-by: Dennis Dalessandro <dennis.dalessandro@intel.com>
Signed-off-by: Mike Marciniszyn <mike.marciniszyn@intel.com>
Signed-off-by: Ira Weiny <ira.weiny@intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
|
|
Use the timer API function setup_timer instead of init_timer, removing
the structure field assignments.
Signed-off-by: Muhammad Falak R Wani <falakreyaz@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
|
|
Use the timer API function setup_timer instead of init_timer, removing
the structure field assignments.
<smpl>
@timer@
expression e1,e2,e3,fn_ptr;
@@
-init_timer(&e1);
+setup_timer(&e1, fn_ptr, e2);
... when != fn_ptr = e3
-e1.function = fn_ptr;
-e1.data = e2;
</smpl>
Signed-off-by: Muhammad Falak R Wani <falakreyaz@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
|
|
The {x,y}LCD fields of struct sm750_dev are not used and can be removed.
Signed-off-by: Mike Rapoport <mike.rapoport@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
|
|
There are several occurencies of '#ifdef CAP_EXP<something>', spelled
differently each time. None of these is ever defined and therefore they
enclose dead code that can be removed.
Signed-off-by: Mike Rapoport <mike.rapoport@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
|
|
Both struct lynx_share and struct sm750_dev reprsent some parts of the
SM750 graphics adapter. There is no point to keep these parts in
different structures.
Signed-off-by: Mike Rapoport <mike.rapoport@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
|
|
The lynx_share field in lynx_cursor structure is never used and can be
removed.
Signed-off-by: Mike Rapoport <mike.rapoport@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
|
|
Fix wrong indentation of a line in lynxfb_set_fbinfo
Signed-off-by: Mike Rapoport <mike.rapoport@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
|
|
Use sm750_dev as private driver data for the PCI device
Signed-off-by: Mike Rapoport <mike.rapoport@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
|
|
Use proper 'struct sm750_dev *' type instead of 'void *'
Signed-off-by: Mike Rapoport <mike.rapoport@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
|
|
Use sm750_dev rather than lynx_share as parameter for hw_sm750_map,
hw_sm750_inithw and hw_sm750_initAccel functions.
Signed-off-by: Mike Rapoport <mike.rapoport@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
|
|
The hw_sm750_getVMSize wraps ddk750_getVMSize without any added
functionality. Call ddk750_getVMSize directly and remove
hw_sm750_getVMSize.
Signed-off-by: Mike Rapoport <mike.rapoport@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
|
|
Use sm750_dev instead of lynx_share in the private data of the
frambuffer device.
Signed-off-by: Mike Rapoport <mike.rapoport@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
|
