Two instances of the IDE code setting the error register did not set it in the cdrom struct if drive is CD-ROM, now they do, fixes CD-ROM booting on the Intel Advanced/ATX and possibly other things;

Fixed IDE and ATAPI 8-bit data reads and writes by rewriting the IDE and CD-ROM PIO data read and write code so that the base code is now 8-bit;
Added some compiler directives that if set, will enable togglable logging and log breakpoints;
Empty IDE channels now always report a status of 0x10 rather than 0x20, and do it on port 3F6h/376h/36Eh/3EEh too, fixes OAKCDROM.SYS infinite loop when hitting an empty IDE channel.

src/disc_86f.c

@@ -1,3 +1,4 @@
+#include <stdarg.h>
 #include <stdint.h>
 #include <stdio.h>
 #include <stdlib.h>
@@ -203,7 +204,7 @@ static struct __attribute__((packed))
 	uint32_t dma_over;
 } d86f[FDD_NUM];
 
-int d86f_do_log = 1;
+int d86f_do_log = 0;
 
 void d86f_log(const char *format, ...)
 {