linux-kernel - Re: [PATCH RESEND] VMCI: remove unused ioctl definitions

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Message-ID: <jgdstohy4ocnsyeafdwlzdbjp3mrss4cpkxy2a66tgwaafaj6o@d5vubomi7osa>
Date: Fri, 20 Dec 2024 12:52:42 +0100
From: Alyssa Ross <hi@...ssa.is>
To: Greg Kroah-Hartman <gregkh@...uxfoundation.org>
Cc: Arnd Bergmann <arnd@...db.de>, 
	"open list:VMWARE VMCI DRIVER" <linux-kernel@...r.kernel.org>, Bryan Tan <bryan-bt.tan@...adcom.com>, 
	Vishnu Dasa <vishnu.dasa@...adcom.com>, 
	Broadcom internal kernel review list <bcm-kernel-feedback-list@...adcom.com>
Subject: Re: [PATCH RESEND] VMCI: remove unused ioctl definitions

On Fri, Dec 20, 2024 at 08:17:23AM +0100, Greg Kroah-Hartman wrote:
> On Fri, Dec 20, 2024 at 12:29:57AM +0100, Alyssa Ross wrote:
> > IOCTL_VMCI_SOCKETS_VERSION and IOCTL_VMCI_SOCKETS_GET_AF_VALUE were
> > never implemented, because VSOCK ended up being implemented as a
> > generic mechanism with a static AF value.  Likewise,
> > IOCTL_VMCI_SOCKETS_GET_LOCAL_CID ended up being implemented as
> > IOCTL_VM_SOCKETS_GET_LOCAL_CID.
> >
> > This isn't a UAPI header, so it should be fine to remove the unused
> > values.  I've left a comment noting IOCTL_VM_SOCKETS_GET_LOCAL_CID is
> > in the VMCI range to avoid unintentional reuse.
> >
> > Signed-off-by: Alyssa Ross <hi@...ssa.is>
> > ---
> > Hi misc maintainers!  I'm sending this version of the patch to you
> > because the VMWare VMCI maintainers don't seem to be around any more.
> > The last message on lore from either of them was in May.  Since then,
> > I've submitted this patch twice, and haven't heard anything in response.
> >
> >  include/linux/vmw_vmci_defs.h | 4 +---
> >  1 file changed, 1 insertion(+), 3 deletions(-)
> >
> > diff --git a/include/linux/vmw_vmci_defs.h b/include/linux/vmw_vmci_defs.h
> > index 6fb663b36f72..c2df94696593 100644
> > --- a/include/linux/vmw_vmci_defs.h
> > +++ b/include/linux/vmw_vmci_defs.h
> > @@ -453,9 +453,7 @@ enum {
> >  #define IOCTL_VMCI_CTX_GET_CPT_STATE		_IO(7, 0xb1)
> >  #define IOCTL_VMCI_CTX_SET_CPT_STATE		_IO(7, 0xb2)
> >  #define IOCTL_VMCI_GET_CONTEXT_ID		_IO(7, 0xb3)
> > -#define IOCTL_VMCI_SOCKETS_VERSION		_IO(7, 0xb4)
> > -#define IOCTL_VMCI_SOCKETS_GET_AF_VALUE		_IO(7, 0xb8)
> > -#define IOCTL_VMCI_SOCKETS_GET_LOCAL_CID	_IO(7, 0xb9)
> > +/*IOCTL_VM_SOCKETS_GET_LOCAL_CID		_IO(7, 0xb9)*/
>
> Why is just this one left as a comment?  Shoudn't you just delet them
> all?

I tried to explain this in the patch description, but I guess I didn't
communicate it clearly enuogh.  IOCTL_VMCI_SOCKETS_GET_LOCAL_CID ended
up being renamed to IOCTL_VM_SOCKETS_GET_LOCAL_CID when it was added to
the UAPI, because the VMCI sockets mechanism ended up being made
hypervisor-independent, not VMCI-specific.  However, the ioctl number
for it is still inside VMCI's range, so the comment here is to
communicate that that number is in use by an ioctl not defined here, and
so shouldn't be reused if new VMCI ioctls are added.

Download attachment "signature.asc" of type "application/pgp-signature" (229 bytes)