For a BADD device, the audioformat structure was directly
being populated without initialising the format type
descriptor. This can lead to a crash later when the
format type and rates are being parsed but the NULL fmt
desc is dereferenced. Fix this by allocating a dummy copy
of fmt desc for BADD 3.0 devices and populating
necessary fields.
Change-Id: I80f33b0e400a9c522a800e989228da134100bb55
Signed-off-by: Ajay Agarwal <ajaya@codeaurora.org>
PCIe enumerate sysfs initialization should be
done after all other resources are setup.
Change-Id: I33b43e39453eeee9ac9e22bd1a5463af9eb1fd05
Signed-off-by: Tony Truong <truong@codeaurora.org>
This fix checks for valid process descriptor of a
memory device client before exporting information
from diag driver to memory device client's read buffer
for reliable data transfer.
CRs-Fixed: 2016396
Change-Id: I45aeb8fc9e2f6a678d48bbfcbb77c501adbbfce0
Signed-off-by: Gopikrishna Mogasati <gmogas@codeaurora.org>
For the status IPA_HW_2_CPU_WDI_RX_FSM_TRANSITION_ERROR,
receive from the uC, add a logic to retry sending the same
command for maximum of 10 retires.
Change-Id: I4b241311e0699aea0e24b3a279622a8a1aaa737d
Acked-by: Mohammed Javid <mjavid@qti.qualcomm.com>
Signed-off-by: Utkarsh Saxena <usaxena@codeaurora.org>
For A2DP/SCO Rx, SCO/FM Tx use cases, we need to program
multi channel registers. Without this there would be
glitches and random silence during A2DP playback.
While at it, add the function name to the debug print macro.
CRs-Fixed: 2027268
Change-Id: Ia8e5fde48bb94e97346fe323e92dc32ed60b0e65
Signed-off-by: Satish Kodishala <skodisha@codeaurora.org>
There are race condition B/T ion_client_destroy and debugfs callbacks.
Let's use a mutex to synchronize them.
Change-Id: I3373dc1dbb551b615105a485cc2d3c4bcc0e5e99
Signed-off-by: Neil Zhang <neilzhang1123@hotmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Git-commit: 948c4db4ee10d85fe78ed3755dcaeb85cd37a148
Git-repo: git://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git
[guptap@codeaurora.org: resolve trivial merge conflicts and
change usage of %p with %pK]
Signed-off-by: Prakash Gupta <guptap@codeaurora.org>
Without is_enabled function support, regualtor framework treats
it as an always on regulator and does not invoke the regulator enable
function call. Hence, add support for it.
Also, define a parent supply ("vin") for the regulator. This parent
can be used to pin-control the regulator. While at it, change the initcall
level to subsys_init. so that the driver registers after the parent supply
("vin" - provided by fixed-regualtor).
Change-Id: I93e870d2e4b89f7874c3e17be91bc417dadf1777
Signed-off-by: Kiran Gunda <kgunda@codeaurora.org>
Fix uninitialized local variable error which might have lead to
crash.
CRs-Fixed: 2030137
Change-Id: I3fd95cb343c3175e4190c8ebfe209399db0602a6
Signed-off-by: Abir Ghosh <abirg@codeaurora.org>
Terminate the string, coming from userspace and containing the name
of fingerprint trusted app, with null character, to make sure kernel
memory does not leak into logs
Change-Id: I1668a64fcb6747ce3ef3b1ee6321fa5fa4a1798a
CRs-Fixed: 2029409
Signed-off-by: Abir Ghosh <abirg@codeaurora.org>
Validate the name of the client app before passing it to
qseecom_start_app.
CRs-Fixed: 2006695
Change-Id: I9c6b16050d4f6fc94827021c7b0f2ab292452f60
Signed-off-by: Abir Ghosh <abirg@codeaurora.org>
Fix the below potential race between these two contexts -
mb_cache_entry_get() and mb_cache_shrink_scan(), which results
into use after free issue.
task a:
mb_cache_shrink_scan()
|--if(!list_empty(&mb_cache_lru_list))
|--get the ce entry
|--list_del_init(&ce->e_lru_list);
|--check ce->e_used, ce->e_queued,
ce->e_refcnt and continue
-> gets prempted here
task b:
ext4_xattr_release_block()
|--mb_cache_entry_get()
|--get ce from hlist_bl_for_each_entry()
|--increment ce->e_used and
list_del_init(&ce->e_lru_list)
|--mb_cache_entry_free()
|--hlist_bl_lock(ce->e_block_hash_p);
-> results into use after free
Also, fix similar potential race between mb_cache_entry_alloc() and
mb_cache_entry_get() in case if cache->c_max_entries is reached.
Change-Id: I01049bae5d914cfb8494ab299ec2e068745d1110
Signed-off-by: Sahitya Tummala <stummala@codeaurora.org>
Configure battery thermal coefficients for QRD660/630 to make sure the
device could read the correct battery temperatures.
CRs-Fixed: 2038922
Change-Id: I21d0eedfec0a0529d8864d11b703659f4ba43b6a
Signed-off-by: Yingwei Zhao <cyizhao@codeaurora.org>
This reverts commit 566b44214c (smb138x:
introduce support for smb1355).
This is not required since smb1355 support is moved to its own driver.
Change-Id: Ibad8595c09a865e273adb6f30d6a75f931220131
Signed-off-by: Abhijeet Dharmapurikar <adharmap@codeaurora.org>
This reverts commit 0f0e36134d ("smb138x:
use chg src bit to disable parallel charger)
This is not required, since smb1355 support is moved to its own driver.
Change-Id: I1bd454030d29195f03be132d167ec8294169b9d5
Signed-off-by: Abhijeet Dharmapurikar <adharmap@codeaurora.org>
Support faking input current limited state. This is very helpful
in testing/debugging the software based vbus changing algorithms.
Change-Id: I7cf548c94397ee680c9d65343c49ec7b16873e88
Signed-off-by: Abhijeet Dharmapurikar <adharmap@codeaurora.org>
Add POWER_SUPPLY_PROP_REAL_TYPE property for usb_psy to present its real
charger type. POWER_SUPPLY_PROP_TYPE in usb_psy is always set to
POWER_SUPPLY_TYPE_USB_PD for healthd to recognize it as an AC charger.
Also add usb_port_psy with POWER_SUPPLY_TYPE_USB type is added for healthd
to recognize it as an USB host. Their ONLINE properties will be updated
according to the VBUS status, type-c mode and real charger type.
With this type being set statically, update the usb phy and pd policy
engine code to look at real type.
Change-Id: I90aa69325cc82b09dfb513c0eeecbc61e092a57f
Signed-off-by: Fenglin Wu <fenglinw@codeaurora.org>
Signed-off-by: Abhijeet Dharmapurikar <adharmap@codeaurora.org>
