This enables 128 bit matching in ip6tables to help
optimize cases where there is no match required.
This shows a rough improvement of 40Mbps in a UDPv6
downlink test using iperf.
CRs-Fixed: 2034577
Change-Id: I093069b17f21cfbbe60b2bd941f8e2dfef562401
Signed-off-by: Subash Abhinov Kasiviswanathan <subashab@codeaurora.org>
iptables matching for ipv6 always has a mask if an address is
specified for match. Adding a check for mask prior to that
helps to improve performance as it avoids the masked comparison.
This is achieved by defining a new feature IP6_NF_IPTABLES_128
which allows to perform 128 bit arithmetic while matching.
CRs-Fixed: 2034577
Change-Id: I9ffd34e29d1b06ad74545d4751aa0e1ea06ded49
Signed-off-by: Subash Abhinov Kasiviswanathan <subashab@codeaurora.org>
The 4k display variant of the 8998 MTP has a new device subtype.
Creating a new device tree file with the default panel set to
4k command mode to support this MTP variant.
Change-Id: I1f1eee1baf2bfb1f452e724b55536674a9d6d973
Signed-off-by: Rajkumar Subbiah <rsubbia@codeaurora.org>
OLEDB driver needs to configure the PBS RAM sequence for
disabling the OLEDB pull down during the module disable.
Add it for PM660A.
Change-Id: I8a969125faa14b2011714ad5bd98a1f9a0f5b649
Signed-off-by: Kiran Gunda <kgunda@codeaurora.org>
On SSR scenario disable statuses for modem producer pipes.
Since modem is not up these statuses are redundant and are
not read by any EE.
Change-Id: I6e7a475d74902f7c5d1b9083f9f7935d3ed2c7c9
CRs-Fixed: 2034903
Signed-off-by: Amir Levy <alevy@codeaurora.org>
OLEDB_SPARE_CTL register bit is set by the PBS during the display enable
and is used during the display suspend to trigger the PBS event.
Hence, move clearing of the OLEDB_SPARE_CTL register to regulator disable.
Also enable the force_pd_ctl only for DRAX-A revisions <=2.0.
Change-Id: I339cefc91f94f772bc4f84e85f2471e1a73e4aed
Signed-off-by: Kiran Gunda <kgunda@codeaurora.org>
Use safer version of strcpy i.e. strlcpy instead of
strncpy for NULL terminated strings.
Change-Id: I5e2800d1d539545744a1b1231e1d589a7c92797a
Signed-off-by: Manu Gautam <mgautam@codeaurora.org>
During stability runs, CCI dump registers are
accessed after the cci release call causing NOC
errors.Enable dump registers if necessary
Change-Id: Ib80bed8de0bdb68dc0ad72d7be8bd5e973ed7f57
Signed-off-by: Shankar Ravi <rshankar@codeaurora.org>
This fix checks the validity of memory device session's
process descriptor before issuing a signal to it while
subsystem restart is performed. This fix avoids accessing
of cleaned-up process descriptor's fields.
CRs-Fixed: 2034816
Change-Id: I2725d3348fbeb8290c887cccea4bca6d41c5b184
Signed-off-by: Gopikrishna Mogasati <gmogas@codeaurora.org>
Register a notifier callback for LAB_VREG_NOT_OK, which is triggered
by the LABIBB driver when short circuit(SC) is detected on the LAB
rail. Disable the OLEDB module upon receiving the LAB_VREG_NOT_OK
notification.
Change-Id: Ia3cc19d9a3c2f32acc413677a9ea4b3100464cd2
Signed-off-by: Kiran Gunda <kgunda@codeaurora.org>
On certain PMICs, there is no short circuit (SC) detection available
in LAB module. Add support to detect it manually and disable LAB/IBB
module upon SC detection. In addition, notify it to OLEDB driver to
disable OLEDB module.
Change-Id: I2a9dc5b1ae8ece45c85b3992026c7f389fd4d135
Signed-off-by: Kiran Gunda <kgunda@codeaurora.org>
Restore default android binder domains. As 'binder,hwbinder,vndbinder'
is the default, this change enables vndbinder.
Change-Id: I48ce75fc6090ad3d9a5b66ca437acbde5a913751
Signed-off-by: Neeraj Upadhyay <neeraju@codeaurora.org>
Memory dump table doesn't support de registration, but
when ssr ramdump device creation failed, it frees
the buffer registered with Memory dump table also.
So re-order ramdump device creation to avoid freeing
memory dump table buffer.
Change-Id: I0ac595259159f3d3fe2e5c864dae8a6a1d6b4078
Signed-off-by: Lingutla Chandrasekhar <clingutla@codeaurora.org>
Add support to dump dsi debug bus registers to memory
or print in xlog. This will be helpful in analyzing HW
behavior more precisely while debugging DSI stability
issues.
Change-Id: I7735897d9f24ca06e15015cf0d277bfe800b35bf
Signed-off-by: Sandeep Panda <spanda@codeaurora.org>
Invalid typecast in MAC address store sysfs entry method
causing device crash due to the buffer overflow.
To fix the above issue check the length of the userspace MAC
address before write. Update the datatype and parsing method
to store the userspace MAC address.
CRs-Fixed: 2034549
Change-Id: I0d4709c5b623c8333a99991c042552df1e7da923
Signed-off-by: Sarada Prasanna Garnayak <sgarna@codeaurora.org>
IPAv2 hardware works with 32 bit addressing,
so allocate a kernel memory using GFP_DMA flag
which is processed by IPA hardware.
Add a error check for return of dma_map_single
using dma_mapping_error API.
Change-Id: I08e3eec8e9b2d81f945283a72338c0a8f23e2ac9
Acked-by: Mohammed Javid <mjavid@qti.qualcomm.com>
Signed-off-by: Utkarsh Saxena <usaxena@codeaurora.org>
In the current implementation DBA driver always assumes
the TV connected is supporting audio. However if the TV
connected is in DVI mode or it's EDID does not contain
any audio block then it will not support audio data.
Hence after parsing EDID, check if the connected TV has
support for audio and accordingly configure and send HPD
events to audio framework.
Change-Id: Id1d175b9c0c1c07bcb4447eeaa907a17c15df098
Signed-off-by: Sandeep Panda <spanda@codeaurora.org>
Signed-off-by: Vishnuvardhan Prodduturi <vproddut@codeaurora.org>
Fix the security issue where mux channel name might
not be null-terminated in ipa wan driver.
Change-Id: I3ef440b62cf3861464fb60c1e7f65f2be5e39ed0
Acked-by: Shihuan Liu <shihuanl@qti.qualcomm.com>
Signed-off-by: Skylar Chang <chiaweic@codeaurora.org>
Allow clients to specify the IOVA range for fastmap clients
via the DOMAIN_ATTR_GEOMETRY domain attribute.
Presently fastmap only allocates page tables for the IOVA
range specified during the create mapping call. However
clients may want to use IOVA addresses outside this range,
such as for their calls to iommu_map.
So allow clients to extend the available IOVA space by setting
the DOMAIN_ATTR_GEOMETRY domain attribute's
iommu_domain_geometry.aperture_start to
the new start address of the IOVA space and by setting
iommu_domain_geometry.aperture_end to the new end address of the
IOVA space.
The new IOVA space created by iommu_domain_geometry.aperture_start
and iommu_domain_geometry.aperture_end will be a superset of the
IOVA range which was created through the create mapping call.
The DOMAIN_ATTR_GEOMETRY domain attribute can only be set before
attaching.
Calls to set the DOMAIN_ATTR_GEOMETRY domain attribute can only
be used to extend the IOVA space, it cannot shrink the range.
Note that extending the IOVA range will not change the range of
IOVA addresses which will be available to the DMA APIs.
Change-Id: Ib389e019a022d98417884002de08115fb0fc9384
Signed-off-by: Liam Mark <lmark@codeaurora.org>
On some conditions, such as reboot, the allocation context
for DMA task might be a user space processes with SIGKILL.
Pre allocate this buffer to avoid DMA allocation failure.
Change-Id: Ie0c5ecbb7aebbf03c42c2f8976aad40e9987fd4a
CRs-Fixed: 2032907
Acked-by: Ady Abraham <adya@qti.qualcomm.com>
Signed-off-by: Skylar Chang <chiaweic@codeaurora.org>
This reverts 'commit 9730348075 ("arm64: Increase the max
granular size") and also sets ARM_DMA_MINALIGN to 128.
ARCH_DMA_MINALIGN is dependent on L1_CACHE_SHIFT but it should
be set to maximum *known* cache line size on ARMv8 systems to
avoid DMA coherecy issues. So setting ARM_DMA_MINALIGN to 128.
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
Change-Id: Ie771d1b693789fce8793538a6efddfe68e2a0043
Patch-mainline: linux-kernel @ 21/03/16, 17:14:03
Signed-off-by: Imran Khan <kimran@codeaurora.org>
