On 1 January 2014 13:56, Stefan Weil <s...@weilnetz.de> wrote: > hw_error is not available for user mode, so this patch fixes Darwin > (which only provides system emulation), but it breaks linking of > m68k-linux-user on Linux hosts.
Maybe we need to come up with some better guidelines about when to use the various choices of API we have for fatal and otherwise unexpected errors: * hw_error * qemu_log_mask(LOG_UNIMP, ...) and continue * assert * g_assert and friends * error* * qerror* * abort * tcg_abort * probably others I forget about... At the moment there's a bit of a mix in existing code, and I don't really know what the "best practices" are for when I'm reviewing new code. PS: does anybody actually find the full register dump hw_error() spits out useful? thanks -- PMM
