Jeroen Domburg [Wed, 8 Feb 2017 04:19:34 +0000 (12:19 +0800)]
Merge branch 'bugfix/yield_other_core_prio_check_fix' into 'master'
Fix interrupting task on other CPU that has lower prio than current task on current CPU
Fix for a very subtle bug introduced somewhere in december. This bug would cause a cross-core interrupt to only be sent if the activated task has a higher priority than the task currently running on *this* cpu, instead of on the other CPU.
He Yin Ling [Thu, 2 Feb 2017 13:59:00 +0000 (21:59 +0800)]
CI: get test env config from gitlab
Previous design was put test env config on local runners. It's not easy
to manage as test runners count growing. Now we'll put config files for
test runners to a Gitlab repository. Test runners will get its config
from Gitlab every time before running.
Angus Gratton [Fri, 27 Jan 2017 00:33:16 +0000 (08:33 +0800)]
Merge branch 'bugfix/spi_multiple_cs_fix' into 'master'
Add missing variable initialisation to make SPI work for multiple CS
Ref https://github.com/espressif/esp-idf/pull/277
From PR: "I am not especially familiar with this platform or with esp-idf -- so review with care.
The spi-master driver would not work for me when the number of devices on a bus exceeded 1 (caused device hang). I believe this is because cur_cs is not properly set (anywhere), and yet is used. This is likely just not noticed because 0 is presumably the correct value when there's only 1 device (and I'm guessing multi-device setup is not tested as thoroughly).
Simply setting cur_cs to the correct slot value resolves the problem and multiple devices on a single bus behave correctly in my test environment."
Merge branch 'bugfix/fix_two_wifi_datapath_issues' into 'master'
esp32: udpate wifi lib to fix two datapath issues
1. fix ampdu<->mpdu<->ampdu switch may cause rx slow issue by put mpdu into ampdu reorder queue
2. fix each ac first sending always fail issue by adding retry
Liu Zhi Fu [Thu, 26 Jan 2017 07:35:04 +0000 (15:35 +0800)]
esp32: udpate wifi lib to fix two datapath issues
1. fix ampdu<->mpdu<->ampdu switch may cause rx slow issue by put mpdu into ampdu reorder queue
2. fix each ac first sending always fail issue by adding retry
Merge branch 'lwip/optimize_dhcpc_taken_time' into 'master'
lwip: optimize the dhcp client
1. modify the discover retry backoff time from (2,4,8,16,32,60,60)s to (500m,1,2,4,8,15,15)s.
2. add DHCP_DOES_ARP_CHECK to menuconfig for users to specify if do a ARP check on the offered address.
If enable, one more second will be taken.
qiyueixa [Sun, 22 Jan 2017 12:32:39 +0000 (20:32 +0800)]
lwip: optimize the dhcp client
1. modify the discover retry backoff time from (2,4,8,16,32,60,60)s to (500m,1,2,4,8,15,15)s.
2. add DHCP_DOES_ARP_CHECK to menuconfig for users to specify if do a ARP check on the offered address.
If enable, one more second will be taken in obtaining IP address.
3. update wifi libs
Angus Gratton [Thu, 5 Jan 2017 04:51:02 +0000 (15:51 +1100)]
spi_flash: Flush flash cache if flash_mmap()ing a written-to page
Without this, it's possible for stale information to be read from
cache via mmap, even if the MMU table entry had been invalidated
prior to writing flash (if the same MMU table entry was re-used after
writing flash.)
Ivan Grokhotkov [Fri, 20 Jan 2017 09:47:34 +0000 (17:47 +0800)]
Merge branch 'feature/docs_gh_link_roles' into 'master'
docs: use custom roles to generate GitHub links
This change replaces direct links to GitHub master branch with
auto-generated links using docutils custom roles.
These auto-generated links point to the tree or blob for the git commit
ID (or tag) of the repository. This is needed to ensure that links don’t
become broken when files in master branch are moved around or deleted.
The following roles are introduced:
```
- :idf:`path` - points to directory inside ESP-IDF
- :idf_blob:`path` - points to file inside ESP-IDF
- :idf_raw:`path` - points to raw view of the file inside ESP-IDF
- :component:`path` - points to directory inside ESP-IDF components dir
- :component_blob:`path` - points to file inside ESP-IDF components dir
- :component_raw:`path` - points to raw view of the file inside ESP-IDF
components dir
- :example:`path` - points to directory inside ESP-IDF examples dir
- :example_blob:`path` - points to file inside ESP-IDF examples dir
- :example_raw:`path` - points to raw view of the file inside ESP-IDF
examples dir
```
A check is added to the CI build script, which searches RST files for
presence of hard-coded links (identified by tree/master, blob/master,
or raw/master part of the URL).
This check can be run manually: cd docs && make gh-linkcheck
Additionally, Sphinx linkcheck build type is used to create new CI test,
which check for broken links. This test has to be triggered explicitly,
because including it in normal build process (when the commit is not yet
deployed to Github) will not work. It can be triggered in a regular
fashion using a combination of cron and Curl, similar to stress tests.
Ivan Grokhotkov [Thu, 19 Jan 2017 08:16:06 +0000 (16:16 +0800)]
docs: use custom roles to generate GitHub links
This change replaces direct links to GitHub master branch with
auto-generated links using docutils custom roles.
These auto-generated links point to the tree or blob for the git commit
ID (or tag) of the repository. This is needed to ensure that links don’t
become broken when files in master branch are moved around or deleted.
The following roles are introduced:
- :idf:`path` - points to directory inside ESP-IDF
- :idf_blob:`path` - points to file inside ESP-IDF
- :idf_raw:`path` - points to raw view of the file inside ESP-IDF
- :component:`path` - points to directory inside ESP-IDF components dir
- :component_blob:`path` - points to file inside ESP-IDF components dir
- :component_raw:`path` - points to raw view of the file inside ESP-IDF
components dir
- :example:`path` - points to directory inside ESP-IDF examples dir
- :example_blob:`path` - points to file inside ESP-IDF examples dir
- :example_raw:`path` - points to raw view of the file inside ESP-IDF
examples dir
A check is added to the CI build script, which searches RST files for
presence of hard-coded links (identified by tree/master, blob/master,
or raw/master part of the URL).
This check can be run manually: cd docs && make gh-linkcheck
Additionally, Sphinx linkcheck build type is used to create new CI test,
which check for broken links. This test has to be triggered explicitly,
because including it in normal build process (when the commit is not yet
deployed to Github) will not work. It can be triggered in a regular
fashion using a combination of cron and Curl, similar to stress tests.
Angus Gratton [Fri, 20 Jan 2017 07:45:03 +0000 (15:45 +0800)]
Merge branch 'feature/esptool_v20_beta' into 'master'
Update esptool.py to v2.0-beta1
* Minor bug fixes, no major functionality changes from the esp-idf perspective.
* Now tracks master branch of esptool repo (finally!)
* Supports Python 3, although our other esp-idf Python tools don't yet support Python 3.
Mike Ryan [Thu, 12 Jan 2017 00:40:37 +0000 (16:40 -0800)]
gatts_demo: re-enter advertising after client disconnects
In the existing tree, after a connected client disconnects the device
will never re-enter the advertising state. Additional clients will not
be able to discover or connect to the device. This patch forces the
device back into the advertising state upon disconnection.
Jan Schmidt [Fri, 20 Jan 2017 03:17:03 +0000 (14:17 +1100)]
netconn_gethostbyname: Fix race reporting success
If the DNS request is dispatched and performed very quickly,
then it can complete before tcpip_callback() actually returns,
in which case we'll destroy the actual err_t error value passed
in the message. Use a local variable for the tcpip_callback
error code so that can't happen.
Ivan Grokhotkov [Thu, 19 Jan 2017 02:58:09 +0000 (10:58 +0800)]
kconfig: define ESP32_PHY_MAX_TX_POWER if PHY_ENABLED
ESP32_PHY_MAX_TX_POWER option is only meaningful for WiFi, so previous
change made it depend on WIFI_ENABLED. However if WiFi is not enabled,
but BT is, this option becomes undefined which breaks phy_init_data
generation.
This change turns ESP32_PHY_MAX_TX_POWER into a hidden parameter, which
depends on PHY_ENABLED. New user-visible parameter,
ESP32_PHY_MAX_WIFI_TX_POWER is introduced which depends on WIFI_ENABLED
and is used as default value for ESP32_PHY_MAX_TX_POWER if WIFI_ENABLED
is set. Otherwise ESP32_PHY_MAX_WIFI_TX_POWER is set to 20.
Ivan Grokhotkov [Thu, 19 Jan 2017 02:19:06 +0000 (10:19 +0800)]
Merge branch 'feature/category_readme' into 'master'
Readme and links to example category folders
- Added README.md to example category folders, for BT ref. https://github.com/espressif/esp-idf/issues/262
- Added links from api category TOC to example category folders
Ivan Grokhotkov [Thu, 19 Jan 2017 02:17:36 +0000 (10:17 +0800)]
Merge branch 'bugfix/frc1_timer_clear' into 'master'
newlib: fix register used for DPORT/RTC bug workaround
While there was no register at DR_REG_FRC_TIMER_BASE + 0x60, due to
peripheral address space wraparound this write actually affected one of
FRC2 registers, which is used by WiFi stack to implement legacy
ets_timer APIs.
This change uses FRC_TIMER_LOAD_REG(0) instead, which can be set to
known value safely.
Ivan Grokhotkov [Thu, 19 Jan 2017 02:14:34 +0000 (10:14 +0800)]
Merge branch 'bugfix/flash_op_unpinned_task' into 'master'
fixes for issues observed when using spi_flash
This MR fixes three unrelated issues:
- Race condition in spi_flash_enable_interrupts_caches_and_other_cpu
when operations on unpinned tasks are performed.
The issue is reported in https://github.com/espressif/esp-idf/pull/258
- esp_intr_noniram_disable doesn’t disable interrupts when compiled in
release mode. This issue manifested itself with an illegal instruction
exception when task WDT ISR was called at the time when flash was
disabled.
Fixes https://github.com/espressif/esp-idf/issues/263.
- Tick hooks on CPU1 were not called if CPU0 scheduler was disabled for
significant amount of time (which could happen when doing flash erase).
The issue manifested itself as “INT WDT timeout on core 1” error.
Fixes https://github.com/espressif/esp-idf/issues/219.
Ivan Grokhotkov [Wed, 18 Jan 2017 14:49:41 +0000 (22:49 +0800)]
newlib: fix register used for DPORT/RTC bug workaround
While there was no register at DR_REG_FRC_TIMER_BASE + 0x60, due to
peripheral address space wraparound this write actually affected one of
FRC2 registers, which is used by WiFi stack to implement legacy
ets_timer APIs.
This change uses FRC_TIMER_LOAD_REG(0) instead, which can be set to
known value safely.
Ivan Grokhotkov [Wed, 18 Jan 2017 16:33:31 +0000 (00:33 +0800)]
Merge branch 'bugfix/docs_cleanup' into 'master'
Organize examples and API documentation - supplement to !430
This supplement to !430, both intended to replace bulky !421.
1. Fixed broken links
2. Added links to available examples in API documentation, where missing
3. Updated links to point directly to example folder (to show README.md with example description)
4. Added README.md in docs folder info to point reader to ReadTheDocs instead of doc sources, ref. https://github.com/espressif/esp-idf/issues/243
5. Fixed some typo errors / Sphinx warnings
6. Fixed memory leak in example, ref. https://github.com/espressif/esp-idf/issues/209
Ivan Grokhotkov [Wed, 18 Jan 2017 15:14:29 +0000 (23:14 +0800)]
freertos: call tick hooks on CPU1 even if CPU0 scheduler is suspended
The block which dispatches ticks on CPU1 was a copy of the code block
for the normal path (CPU0). It used to check uxPendedTicks, with the
logic that uxPendedTicks can be 0 iff the scheduler is not suspended.
On CPU1 however, uxPendedTicks is not related to the state of the
scheduler (as uxPendedTicks is updated on CPU0). Due to this, if CPU0
scheduler was suspended, and uxPendedTicks happened to be nonzero,
tick hooks on CPU1 didn’t run, even though CPU1 scheduler was working.
This change removes the check for uxPendedTicks in CPU1 code path,
so that the tick hooks on CPU1 always get called (as for the CPU0 code
path).
Ivan Grokhotkov [Wed, 18 Jan 2017 10:31:06 +0000 (18:31 +0800)]
intr_alloc: mark inline asm operand as earlyclobber
When compiling in release mode, compiler was choosing same register for
oldint and intmask variables, so INTENABLE was never modified.
This effectively broke disabling of non-IRAM interrupts during flash
operations, observed in the existing tests if task watchdog is enabled.
This change adds an extra constraint tells the compiler that output
operand should not be placed into the same register as an input one.
Ivan Grokhotkov [Wed, 18 Jan 2017 09:23:04 +0000 (17:23 +0800)]
Merge branch 'feature/parse_unit_test_cases_from_test_files' into 'master'
Parse unit test cases from test files
As of now, we need to sync three different places when adding new unit test cases. The plan is to add a python script in tools/unit-test-app. This script would parse unit test cases from the C files with tests and generate the needed YAML files in components/idf_test/unit_test. CI would run this script in build stage to create the files and unit-test stage would use the generated files to run tests.
To get the needed test metadata, we need to add some more tags in the test case functions. For example:
```
TEST_CASE("SYS_LIB_0102 test time functions", "[stdlib][fails][env=UT_T1_APC]")
{
//...
}
```
Reason for this kind of implementation is that we need the YAML files for generating documents.
Ivan Grokhotkov [Wed, 18 Jan 2017 07:07:27 +0000 (15:07 +0800)]
spi_flash: fix race condition when doing operations in unpinned tasks
spi_flash_enable_interrupts_caches_and_other_cpu function used to enable
non-IRAM interrupts after giving up flash operation lock, which would
cause problems if another task was waiting on the lock to start a flash
operation. In fact, non-IRAM interrupts should be re-enabled before the
task scheduler is resumed. Otherwise non-pinned task can be moved to the
other CPU due to preemption, causing esp_intr_noniram_enable to be
called on the other CPU, causing an abort to be triggered.
Fixes the issue reported in
https://github.com/espressif/esp-idf/pull/258
projects / esp-idf / log