tree 024fb96ec7f6df755c6d3927a85bf9fdbe999d9a
parent d02ab775f321e5f3b59e5a3c609ea6c06c0199c1
author Jeff Tenney <jeff.tenney@gmail.com> 1690433533 -0700
committer GitHub <noreply@github.com> 1690433533 +0530
gpgsig -----BEGIN PGP SIGNATURE-----
 
 wsBcBAABCAAQBQJkwff9CRBK7hj4Ov3rIwAAof8IAJ6JiVTY9A0M5x6yhbXEjsI+
 //wSS0aA/39rlD/jtdsyvYshTsALCnDcPLdTwGvhQvciLZGWNtdF/UPhgaZSwhLV
 zksm8qaZ6AgUgzUqqnGQ69U4nZyaTvH2OIe9oLXeFhlom7AMU6l0zNj8rPV+Pcx8
 aKJGuxWx+qj6blyO6zmXE4daah+TqAT3P8mrTvBa88tGGj1TU8BQu41P/RXbiLul
 mGBO113p/lMdKmqb4kdJShRGgJexGR96YadngOA51CaCXcxJDP8rCgyFd2Aw8Mc3
 aw+vJqSFF/ebuFGiIiceRUPGDWn/tK1M8CPvVdIEgBriHCO8l8/KT2HFQ6qTAtA=
 =Etio
 -----END PGP SIGNATURE-----
 

Work around SysTick bug for QEMU ARMv8-M (#724)

* Set SysTick CLKSOURCE bit before enabling SysTick

* Use portNVIC_SYSTICK_CLK_BIT_CONFIG

The workaround now uses portNVIC_SYSTICK_CLK_BIT_CONFIG instead of
portNVIC_SYSTICK_CLK_BIT, which saves us from having to explain in the
comments why it's OK to temporarily set the CLKSOURCE bit even if the
user's FreeRTOS configuration clears the CLKSOURCE bit.

Using portNVIC_SYSTICK_CLK_BIT_CONFIG here still correctly prevents the
firmware from triggering the QEMU bug.