diff options
| author | tturney <tturney@google.com> | 2015-08-19 10:38:28 -0700 |
|---|---|---|
| committer | tturney <tturney@google.com> | 2015-08-31 15:41:47 -0700 |
| commit | b7b7af83922fc5918d2632402daa3824364806fb (patch) | |
| tree | ccfa5677a63a0ec066272d5eab655416af0eb262 /src/hardware.c | |
| parent | 955f8769d470fb2bbb8a1d44a39ce8e104085d9a (diff) | |
| download | android_hardware_broadcom_libbt-b7b7af83922fc5918d2632402daa3824364806fb.tar.gz android_hardware_broadcom_libbt-b7b7af83922fc5918d2632402daa3824364806fb.tar.bz2 android_hardware_broadcom_libbt-b7b7af83922fc5918d2632402daa3824364806fb.zip | |
Fix btwrite timer timeout for Angler (1/2)
1. For vnd_angler.txt:
Set PROC_BTWRITE_TIMER_TIMEOUT_MS to 0
to avoid upio start and stop timer when LMP
is enabled and BT_WAKE is asserted and
de-asserted. This time is useless and would
set btwrite to 0 by mistake even when host
is sending data to UART, and this would
de-assert BT_WAKE by mistake.
2. For hardware.c:
For chipset 4358 we need to set LPM_HC_IDLE_THRESHOLD
to 24, so we need to reduce the multiplier
to make timeout_ms a reasonable value. Without
modification, the timeout_ms would be 36000ms
which is too large.
3. For upio.c:
Need to set lmp_proc_cb.btwrite_active even though
PROC_BTWRITE_TIMER_TIMEOUT_MS is set to 0.
Otherwise, btwrite would be written very frequently.
Bug: 23248798
Change-Id: Ie55c346785679e51a771d5e46270c0858e6082ee
Diffstat (limited to 'src/hardware.c')
| -rwxr-xr-x | src/hardware.c | 3 |
1 files changed, 2 insertions, 1 deletions
diff --git a/src/hardware.c b/src/hardware.c index 02103ca..10bf22b 100755 --- a/src/hardware.c +++ b/src/hardware.c @@ -1211,9 +1211,10 @@ uint32_t hw_lpm_get_idle_timeout(void) */ timeout_ms = (uint32_t)lpm_param.host_stack_idle_threshold \ * LPM_IDLE_TIMEOUT_MULTIPLE; - if (strstr(hw_cfg_cb.local_chip_name, "BCM4325") != NULL) timeout_ms *= 25; // 12.5 or 25 ? + else if (strstr(hw_cfg_cb.local_chip_name, "BCM4358") != NULL) + timeout_ms *= 50; else timeout_ms *= 300; |
