On 2019/04/19 11:05, Chao Yu wrote: > Hi Shinichiro, > > On 2019/4/19 9:18, Shinichiro Kawasaki wrote: >> >> I don't think so. The major scenario that read_file() returns NULL is fopen >> fails to read /sys/dev/block/%s/partition. It indicates that the target >> device >> is not a partition device. The sys_path and following code is still valid >> for >> the non-partition device. > > I guess you mean the case that ../partition is not existed, right? so we > should > check errno to confirm that? > >> >> The other minor read_file() NULL return scenario is strdup() failure because >> of >> memory shortage. Do you think f2fs-tools need strdup() return value check? >> If >> you suggest it, I will update the patch. > > Yes, I think we need to handle that as well.
Chao, I talked with Shinichiro. The original code is in fact mostly OK, modulo a change to get the correct path to the "zoned" sysfs attribute file for partition block devices. There is also one thing missing in the original patch, namely a check for ENOENT error on fopen(). With these 2 changes, every error case will be correctly covered. Shinichiro v2 patch will introduce these 2 fixes. Best regards. -- Damien Le Moal Western Digital Research _______________________________________________ Linux-f2fs-devel mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/linux-f2fs-devel
