[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20231009021644.GC800259@ZenIV>
Date: Mon, 9 Oct 2023 03:16:44 +0100
From: Al Viro <viro@...iv.linux.org.uk>
To: Paolo Bonzini <pbonzini@...hat.com>
Cc: linux-kernel@...r.kernel.org, kvm@...r.kernel.org,
seanjc@...gle.com
Subject: Re: [PATCH gmem FIXUP] kvm: guestmem: do not use a file system
On Thu, Sep 28, 2023 at 02:06:51PM -0400, Paolo Bonzini wrote:
> Use a run-of-the-mill anonymous inode, there is nothing useful
> being provided by kvm_gmem_fs.
> - inode = alloc_anon_inode(mnt->mnt_sb);
> - if (IS_ERR(inode))
> - return PTR_ERR(inode);
> + fd = get_unused_fd_flags(0);
> + if (fd < 0)
> + return fd;
>
> - err = security_inode_init_security_anon(inode, &qname, NULL);
> - if (err)
> - goto err_inode;
> + gmem = kzalloc(sizeof(*gmem), GFP_KERNEL);
> + if (!gmem) {
> + err = -ENOMEM;
> + goto err_fd;
> + }
> +
> + file = anon_inode_getfile(anon_name, &kvm_gmem_fops, gmem,
> + O_RDWR);
> + inode = file->f_inode;
> + WARN_ON(file->f_mapping != inode->i_mapping);
>
> inode->i_private = (void *)(unsigned long)flags;
> inode->i_op = &kvm_gmem_iops;
That's very badly broken. The whole point of anon_inode_getfile() is
that *ALL* resulting files share the same inode. You are not allowed
to modify the damn thing.
Powered by blists - more mailing lists