linux-kernel - Re: [PATCH] efivarfs: expose used and total size

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: <ZGI81p+2SNjI4VOb@anisse-laptop>
Date:   Mon, 15 May 2023 16:08:22 +0200
From:   Anisse Astier <anisse@...ier.eu>
To:     Christian Brauner <brauner@...nel.org>
Cc:     Ard Biesheuvel <ardb@...nel.org>,
        Anisse Astier <an.astier@...teo.com>,
        Al Viro <viro@...iv.linux.org.uk>,
        Greg Kroah-Hartman <gregkh@...uxfoundation.org>,
        linux-efi@...r.kernel.org, Johan Hovold <johan+linaro@...nel.org>,
        Darren Hart <dvhart@...radead.org>,
        Andy Shevchenko <andy@...radead.org>,
        Thomas Gleixner <tglx@...utronix.de>,
        Ingo Molnar <mingo@...hat.com>, Borislav Petkov <bp@...en8.de>,
        Dave Hansen <dave.hansen@...ux.intel.com>, x86@...nel.org,
        "H. Peter Anvin" <hpa@...or.com>, Jeremy Kerr <jk@...abs.org>,
        platform-driver-x86@...r.kernel.org, linux-kernel@...r.kernel.org,
        lennart@...ttering.net
Subject: Re: [PATCH] efivarfs: expose used and total size

Hi Christian,

On Mon, May 15, 2023 at 10:59:49AM +0200, Christian Brauner wrote:
> On Wed, May 10, 2023 at 05:13:36PM +0200, Ard Biesheuvel wrote:
> > On Wed, 26 Apr 2023 at 21:59, Anisse Astier <anisse@...ier.eu> wrote:
> > >
> > > From: Anisse Astier <an.astier@...teo.com>
> > >
> > > When writing variables, one might get errors with no other message on
> > > why it fails.
> > >
> > > Being able to see how much is used by EFI variables helps analyzing such
> > > issues.
> > >
> > > Since this is not a conventionnal filesystem, block size is
> > > intentionnally set to 1 instead of PAGE_SIZE.
> > >
> > > x86 quirks of reserved size are taken into account and available and
> > > free size can be different, further helping debugging space issues.
> > >
> > 
> > I have no objections to this, but I'm not much of a user space/ VFS
> > person, so adding some other folks that can chime in if they want.
> > 
> > The point of this patch is that user space can query this information
> > using statvfs(), right?
> 
> Seems ok to me.
> 
> > 
> > 
> > 
> > > Signed-off-by: Anisse Astier <an.astier@...teo.com>
> > > ---
> > > Notes:
> > > Patch isn't split per subsystem intentionally, for better understanding
> > > of intent; split could be trivial in a later version.
> > >
> > > I'm not sure whether statfs(2) should return an error if the efi request
> > > fails; I think it could be ignored with maybe a WARN_ONCE; which would
> > > be close to the current behaviour.
> 
> Not sure. If you're not returning an error you would have to report made
> up/magic values that could end up confusing userspace. So better to be
> honest and report an error.

Ack.

> 
> > >
> > > Regards,
> > >
> > > Anisse
> > >
> > > ---
> > >  arch/x86/platform/efi/quirks.c |  8 ++++++++
> > >  drivers/firmware/efi/efi.c     |  1 +
> > >  drivers/firmware/efi/vars.c    | 12 ++++++++++++
> > >  fs/efivarfs/super.c            | 26 +++++++++++++++++++++++++-
> > >  include/linux/efi.h            | 10 ++++++++++
> > >  5 files changed, 56 insertions(+), 1 deletion(-)
> > >
> > > diff --git a/arch/x86/platform/efi/quirks.c b/arch/x86/platform/efi/quirks.c
> > > index b0b848d6933a..587fa51230e2 100644
> > > --- a/arch/x86/platform/efi/quirks.c
> > > +++ b/arch/x86/platform/efi/quirks.c
> > > @@ -114,6 +114,14 @@ void efi_delete_dummy_variable(void)
> > >                                      EFI_VARIABLE_RUNTIME_ACCESS, 0, NULL);
> > >  }
> > >
> > > +u64 efi_reserved_space(void)
> > > +{
> > > +       if (efi_no_storage_paranoia)
> > > +               return 0;
> > > +       return EFI_MIN_RESERVE;
> > > +}
> > > +EXPORT_SYMBOL_GPL(efi_reserved_space);
> > > +
> > >  /*
> > >   * In the nonblocking case we do not attempt to perform garbage
> > >   * collection if we do not have enough free space. Rather, we do the
> > > diff --git a/drivers/firmware/efi/efi.c b/drivers/firmware/efi/efi.c
> > > index abeff7dc0b58..d0dfa007bffc 100644
> > > --- a/drivers/firmware/efi/efi.c
> > > +++ b/drivers/firmware/efi/efi.c
> > > @@ -211,6 +211,7 @@ static int generic_ops_register(void)
> > >         generic_ops.get_variable = efi.get_variable;
> > >         generic_ops.get_next_variable = efi.get_next_variable;
> > >         generic_ops.query_variable_store = efi_query_variable_store;
> > > +       generic_ops.query_variable_info = efi.query_variable_info;
> > >
> > >         if (efi_rt_services_supported(EFI_RT_SUPPORTED_SET_VARIABLE)) {
> > >                 generic_ops.set_variable = efi.set_variable;
> > > diff --git a/drivers/firmware/efi/vars.c b/drivers/firmware/efi/vars.c
> > > index bd75b87f5fc1..c5382d5c3073 100644
> > > --- a/drivers/firmware/efi/vars.c
> > > +++ b/drivers/firmware/efi/vars.c
> > > @@ -245,3 +245,15 @@ efi_status_t efivar_set_variable(efi_char16_t *name, efi_guid_t *vendor,
> > >         return status;
> > >  }
> > >  EXPORT_SYMBOL_NS_GPL(efivar_set_variable, EFIVAR);
> > > +
> > > +efi_status_t efivar_query_variable_info(u32 attr,
> > > +                                       u64 *storage_space,
> > > +                                       u64 *remaining_space,
> > > +                                       u64 *max_variable_size)
> > > +{
> > > +       if (!__efivars->ops->query_variable_info)
> > > +               return EFI_UNSUPPORTED;
> > > +       return __efivars->ops->query_variable_info(attr, storage_space,
> > > +                       remaining_space, max_variable_size);
> > > +}
> > > +EXPORT_SYMBOL_NS_GPL(efivar_query_variable_info, EFIVAR);
> > > diff --git a/fs/efivarfs/super.c b/fs/efivarfs/super.c
> > > index 482d612b716b..064bfc0243c9 100644
> > > --- a/fs/efivarfs/super.c
> > > +++ b/fs/efivarfs/super.c
> > > @@ -13,6 +13,7 @@
> > >  #include <linux/ucs2_string.h>
> > >  #include <linux/slab.h>
> > >  #include <linux/magic.h>
> > > +#include <linux/statfs.h>
> > >
> > >  #include "internal.h"
> > >
> > > @@ -23,8 +24,31 @@ static void efivarfs_evict_inode(struct inode *inode)
> > >         clear_inode(inode);
> > >  }
> > >
> > > +static int efivarfs_statfs(struct dentry *dentry, struct kstatfs *buf)
> > > +{
> > > +       u64 storage_space, remaining_space, max_variable_size;
> > > +       efi_status_t status;
> > > +       const u32 attr = (EFI_VARIABLE_NON_VOLATILE | EFI_VARIABLE_BOOTSERVICE_ACCESS |
> > > +        EFI_VARIABLE_RUNTIME_ACCESS);
> > > +
> > > +       buf->f_type = dentry->d_sb->s_magic;
> > > +       buf->f_bsize = 1;
> > > +       buf->f_namelen = NAME_MAX;
> > > +
> > > +       status = efivar_query_variable_info(attr, &storage_space, &remaining_space,
> > > +                                           &max_variable_size);
> > > +       if (status != EFI_SUCCESS)
> > > +               return efi_status_to_err(status);
> > > +       buf->f_blocks = storage_space;
> 
> I have no idea about efivarfs specifically but I would add comments
> why f_bsize is set to what it is and clarify the relationship between
> f_bsize and f_blocks. Even if just for the sake of userspace to be able
> to interpret this.

Ack.

Thanks for taking the time to review this patch, I'll send a new version
soon.

Regards,

Anisse