The check for EROFS was wrong. For example, if you try to mount a filesystem appended to an executable that's being run (yes, I'm doing this) you get EPERM, but mounting readonly fixes it. Doing the fallback all the time shouldn't hurt, and is one less test.
This commit was genuinely broken at the time (not related to our build system): scripts/config/mkconfigs.c (2005-08-01 - 2006-10-05 arasi) uretilen bbconfigopts.h icine .config satirlarini kacissiz, ham newline'larla gömüyordu - gecerli C degildi, busybox'un kendi bb_mkdep araci bile parse edemiyordu (7d219aab7 ile duzeltildi, mkconfigs shell scripte cevrildi)
No binary (build failed)
—
(log yok)
| @@ -88,10 +88,9 @@ extern int set_loop(char **device, const char *file, int offset) | |
| int i, dfd, ffd, mode, rc=1; | |
| // Open the file. Barf if this doesn't work. | |
| - | if((ffd = open(file, mode=O_RDWR))<0) |
| - | if(errno!=EROFS || (ffd=open(file,mode=O_RDONLY))<0) |
| - | return errno; |
| - | |
| + | if((ffd = open(file, mode=O_RDWR))<0 && (ffd = open(file,mode=O_RDONLY))<0) |
| + | return errno; |
| + | |
| // Find a loop device | |
| for(i=0;rc;i++) { | |
| sprintf(dev, LOOP_FORMAT, i++); | |