...and all their callers.
With the upcoming switch from sizeof(time_t)==4 to sizeof(time_t)==8,
sizeof(struct stat) is also increasing.
A few newlib functions present in ROM allocate 'struct stat' on the
stack and call _fstat_r on this structure. The implementation of
fstat is provided in ESP-IDF. This implementation will often do
memset(st, 0, sizeof(*st)), where st is 'struct stat*', before setting
some fields of this structure. If IDF is built with sizeof(st)
different from sizeof(st) which ROM was built with, this will lead
to an out-of-bounds write and a stack corruption.
This commit removes problematic ROM functions from the linker script.
Here are the functions which allocate 'struct stat':
* _isatty_r (in ROM)
* __swhatbuf_r, called by __smakebuf_r, called by __swsetup_r and
__srefill_r (in ROM)
* _fseeko_r (not in ROM)
* glob2 (not in ROM)
* _gettemp (not in ROM)
As a result, these functions are used from libc.a, and use correct
size of 'stat' structure.
Closes https://github.com/espressif/esp-idf/issues/7980
This commit updates the USB Host Library as follows:
- usb_helpers.h
- Removed dependency on USB Host Library API
- Added function to print string descriptors
- usbh
- Fixed bug where an interface/endpoint could be claimed/allocated multiple times
- Removed redundant device ref_count change
- Added unit test for USB Host Library API usage
This commit updates the Hub Driver and USBH as follows:
- Updated communication mechanism between Hub Driver and USBH
- USBH notifies the Hub Driver via Hub request callback (usbh_hub_req_cb_t)
- Hub Driver notifies the USBH via usbh_hub_pass_event()
- Hub Driver now defers all event handling to hub_process.
- Hub Driver and USBH will now automatically recover the port after a disconnected
device has been closed.
- Fixed incorrect assert in usbh_dev_close()
Hub Driver is refactored as follows:
This commit update and refactors the Hub Driver as follows:
- Refactored enumeration state machine and stage functions
- Enumeration stage is now incremented
- Combined transfer stages of enumeration into common functions
- Comments updated
- Fixed usbh_hal_disable_debounce_lock() that would cause root_port_handle_events()
to fail the HCD_PORT_CMD_RESET call because the previous port connection interrupt
was not cleared.
The following features were added to the Hub Driver
- Enumeration config descriptor is now fetched in two separate stages
- Header is fetched first to determine the wTotalLength of the descriptor
- Fetching the full descriptor will request exactly wTotalLength bytes
- This works around some non-compliant devices that will babble/return zero
when requesting a length > wTotalLength
- Closes https://github.com/espressif/esp-idf/issues/7799
- Enumeration now stores string descriptors
- The Manufacturer, Product, and Serial Number string descriptors are
now read and stored during enumeration
- String descriptors are now part of usb_device_info_t
- Added unit test to test enumeration
This commit updates the USB Host stack to use the USB PHY driver. The
USB PHY and the OTG Controller should now both be setup/deleted using
usb_new_phy() and usb_del_phy() respectively.
- The hcd_install() now expects the USB PHY and OTG Contorller to be
already setup before it is called
- usb_host_install() now has an option to skip calling usb_del_phy() if
the user wants to setup their own USB PHY (e.g., in the case of using
and external PHY).
- CDC-ACM and MSC examples/test updated to use internal PHY
Closes https://github.com/espressif/esp-idf/issues/8061
This commit fixes the following bugs in the HCD
- HCD control pipes fill incorrect transfer size
_buffer_fill_ctrl() would fill the transfer descriptors length with
wLength instead of transfer->num_bytes. Therefore, the remaining length
would be incorrect when a control transfer requests more bytes than are
available.
- Fix USB_TRANSFER_FLAG_ZERO_PACK behavior
The previous behavior of USB_TRANSFER_FLAG_ZERO_PACK was incorrect, and did not
support interrupt pipes. A zero length packet can now be added to Bulk/Interrupt
OUT transfers where the length is a multiple of the endpoint's MPS.
- Fixed HCD port suspend and resume test case
Halting a control pipe mid control transfer can lead some test devices to get stuck
