linux-kernel - Re: [RFC 3/3] efi: add capsule update capability via sysfs

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

Message-ID: <1430350592.2189.50.camel@HansenPartnership.com>
Date:	Wed, 29 Apr 2015 16:36:32 -0700
From:	James Bottomley <James.Bottomley@...senPartnership.com>
To:	Andy Lutomirski <luto@...capital.net>
Cc:	"linux-efi@...r.kernel.org" <linux-efi@...r.kernel.org>,
	"Kweh, Hock Leong" <hock.leong.kweh@...el.com>,
	LKML <linux-kernel@...r.kernel.org>,
	Greg Kroah-Hartman <gregkh@...uxfoundation.org>,
	Peter Jones <pjones@...hat.com>
Subject: Re: [RFC 3/3] efi: add capsule update capability via sysfs

On Wed, 2015-04-29 at 16:25 -0700, Andy Lutomirski wrote:
> On Wed, Apr 29, 2015 at 4:12 PM, James Bottomley
> <James.Bottomley@...senpartnership.com> wrote:
> > From: James Bottomley <JBottomley@...n.com>
> >
> > The firmware update should be applied simply by doing
> >
> > cat fw_file > /sys/firmware/capsule/update
> >
> > With a properly formatted fw_file.  Any error will be returned on close of
> > stdout.  util-linux returns errors correctly from closing stdout, but firmware
> > shippers should check whatever utilities package they use correctly captures
> > the error return on close.
> 
> s/util-linux/coreutils/
> 
> This still makes my API sense itch.  It's kind of an abuse of
> open/write/close.

It works ... and according to Alan, NFS is already doing it.  I suppose
we can have a do over of the whole debate again ...

> >
> > Signed-off-by: James Bottomley <JBottomley@...n.com>
> > ---
> >  drivers/firmware/efi/Makefile  |  2 +-
> >  drivers/firmware/efi/capsule.c | 78 ++++++++++++++++++++++++++++++++++++++++++
> >  drivers/firmware/efi/capsule.h |  2 ++
> >  drivers/firmware/efi/efi.c     |  8 +++++
> >  4 files changed, 89 insertions(+), 1 deletion(-)
> >  create mode 100644 drivers/firmware/efi/capsule.c
> >  create mode 100644 drivers/firmware/efi/capsule.h
> >
> > diff --git a/drivers/firmware/efi/Makefile b/drivers/firmware/efi/Makefile
> > index d8be608..698846e 100644
> > --- a/drivers/firmware/efi/Makefile
> > +++ b/drivers/firmware/efi/Makefile
> > @@ -1,7 +1,7 @@
> >  #
> >  # Makefile for linux kernel
> >  #
> > -obj-$(CONFIG_EFI)                      += efi.o vars.o reboot.o
> > +obj-$(CONFIG_EFI)                      += efi.o vars.o reboot.o capsule.o
> >  obj-$(CONFIG_EFI_VARS)                 += efivars.o
> >  obj-$(CONFIG_EFI_VARS_PSTORE)          += efi-pstore.o
> >  obj-$(CONFIG_UEFI_CPER)                        += cper.o
> > diff --git a/drivers/firmware/efi/capsule.c b/drivers/firmware/efi/capsule.c
> > new file mode 100644
> > index 0000000..1fd78e7
> > --- /dev/null
> > +++ b/drivers/firmware/efi/capsule.c
> > @@ -0,0 +1,78 @@
> > +#include <linux/efi.h>
> > +#include <linux/slab.h>
> > +#include <linux/transaction_helper.h>
> > +
> > +#include "capsule.h"
> > +
> > +static struct kset *capsule_kset;
> > +static struct transaction_buf *capsule_buf;
> > +
> > +static int capsule_data_write(struct file *file, struct kobject *kobj,
> > +                             struct bin_attribute *attr,
> > +                             char *buffer, loff_t offset, size_t count)
> > +{
> > +       if (!capsule_buf) {
> > +               capsule_buf = kmalloc(sizeof(*capsule_buf), GFP_KERNEL);
> > +               if (!capsule_buf)
> > +                       return -ENOMEM;
> > +               transaction_init(capsule_buf);
> > +       }
> > +
> > +       return transaction_write(capsule_buf, buffer, offset, count);
> > +}
> 
> This seems unlikely to have good effects if two struct files are
> active at once.

I thought of threading ->open and using that to make it exclusive.  But
then I thought caveat emptor.

I think for multiple files, I need a mutex in the transaction code just
to ensure orderly access.

> Also, I think you crash if you open and close without calling write,

yes there should be an if (!capsule_buf) return -EINVAL in flush

> and I don't know what whether there can be spurious flushes (fsync?).

It turns out that the bdi flusher and the fop->flush() operation are
totally different things.  ->flush() is used mostly just to do stuff on
close (NFS uses it to tidy up for instance).

James


--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/