On Wed, Aug 19, 2020 at 5:20 PM Pranay Srivastava <pranay.srivastava@xxxxxxxxxxxx> wrote: > > Hello, > > I'm running a system where rootfs is ramfs. For kernel version 5.2.11 > > # cat /proc/self/mountinfo > 0 0 0:1 / / rw - rootfs rootfs rw > %<---snip>% > > while for kernel 5.4.58 > # cat /proc/self/mountinfo > 0 0 0:1 / / rw - rootfs none rw > %<---snip>% > > The reason for the above difference is because for kernel 5.2.11 the > parse_param for > rootfs was set to legacy_parse_param which handled the "source" param > instead of > ignoring it. > > With kernel 5.4.58 this is set to ramfs_parse_param which ignores any > parameters not > recognized and also returns 0 instead of -ENOPARAM. This causes > vfs_parse_fs_param > to not set the file context source(fc->source) which results in "none" from > alloc_vfs_mount(fc->source ? : "none") > > The commit which introduced the above change was > > commit f32356261d44d580649a7abce1156d15d49cf20f > Author: David Howells <dhowells@xxxxxxxxxx> > Date: Mon Mar 25 16:38:31 2019 +0000 > > vfs: Convert ramfs, shmem, tmpfs, devtmpfs, rootfs to use the new mount API > > I'm not sure if this is a regression? But if it is, do we handle it like > > diff --git a/fs/ramfs/inode.c b/fs/ramfs/inode.c > index ee179a81b3da..47a39baa0535 100644 > --- a/fs/ramfs/inode.c > +++ b/fs/ramfs/inode.c > > @@ -200,7 +200,7 @@ static int ramfs_parse_param(struct fs_context > *fc, struct fs_parameter *param) > * and as it is used as a !CONFIG_SHMEM simple substitute > * for tmpfs, better continue to ignore other mount options. > */ > - if (opt == -ENOPARAM) > + if (opt == -ENOPARAM && strcmp(param->key, "source")) > opt = 0; > return opt; > } > > so that mountinfo gives the same information as for earlier kernels. > > Thanks! > -- > Regards, > Pranay Srivastava Kindly let me know if the above is a regression and needs to be fixed? -- Regards, Pranay Srivastava
