(Cross-posted from http://e2e.ti.com/support/microcontrollers/msp430/f/166/t/348108.aspx since it's not clear how TI wants to receive bug reports related to the new toolchain.)
Back when I added intrinsics to mspgcc I got confirmation that CCS and IAR both implicitly added the required NOP after a DINT instruction generated with _disable_interrupts(). (The MSP430 instruction set explains that this is necessary to eliminate the possibility of interrupting a subsequent pipelined instruction.) The implementation of _disable_interrupts() in in430.h from TI's GCC_RH_20140508.zip does not include this required instruction, making it inconsistent with all other MSP430 toolchains. (Fortunately, msp430-elf does generate a warning in this situation.) Interestingly, the implementation of _enable_interrupts() unconditionally adds a NOP after EINT, even though that workaround is arguably required only on certain devices with a specific erratum, and this behavior is AFAIK also inconsistent with other toolchains. msp430-elf should behave as close to CCS and IAR as possible so developers don't have to work around toolchain-specific inconsistencies in how these operations are implemented. TI, please fix this for the next header release. Please also add a monotonically increasing version number to in430.h or iomacros.h so applications can work around bugs like this. Thanks. Peter ------------------------------------------------------------------------------ HPCC Systems Open Source Big Data Platform from LexisNexis Risk Solutions Find What Matters Most in Your Big Data with HPCC Systems Open Source. Fast. Scalable. Simple. Ideal for Dirty Data. Leverages Graph Analysis for Fast Processing & Easy Data Exploration http://p.sf.net/sfu/hpccsystems _______________________________________________ Mspgcc-users mailing list Mspgcc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/mspgcc-users
