[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20241101071208.GA2962282@thelio-3990X>
Date: Fri, 1 Nov 2024 00:12:08 -0700
From: Nathan Chancellor <nathan@...nel.org>
To: André Almeida <andrealmeid@...lia.com>
Cc: Hugh Dickins <hughd@...gle.com>,
Andrew Morton <akpm@...ux-foundation.org>,
Alexander Viro <viro@...iv.linux.org.uk>,
Christian Brauner <brauner@...nel.org>, Jan Kara <jack@...e.cz>,
krisman@...nel.org, Stephen Rothwell <sfr@...b.auug.org.au>,
linux-mm@...ck.org, linux-kernel@...r.kernel.org,
linux-fsdevel@...r.kernel.org, kernel-dev@...lia.com,
Theodore Ts'o <tytso@....edu>
Subject: Re: [PATCH 2/3] tmpfs: Fix type for sysfs' casefold attribute
On Thu, Oct 31, 2024 at 10:37:40PM -0300, André Almeida wrote:
> DEVICE_STRING_ATTR_RO should be only used by device drivers since it
> relies on `struct device` to use device_show_string() function. Using
> this with non device code led to a kCFI violation:
>
> > cat /sys/fs/tmpfs/features/casefold
> [ 70.558496] CFI failure at kobj_attr_show+0x2c/0x4c (target: device_show_string+0x0/0x38; expected type: 0xc527b809)
>
> Like the other filesystems, fix this by manually declaring the attribute
> using kobj_attribute() and writing a proper show() function.
>
> Also, leave macros for anyone that need to expand tmpfs sysfs' with
> more attributes (as seen in fs/btrfs/sysfs.c).
>
> Fixes: 5132f08bd332 ("tmpfs: Expose filesystem features via sysfs")
> Reported-by: Nathan Chancellor <nathan@...nel.org>
> Closes: https://lore.kernel.org/lkml/20241031051822.GA2947788@thelio-3990X/
> Signed-off-by: André Almeida <andrealmeid@...lia.com>
Reviewed-by: Nathan Chancellor <nathan@...nel.org>
Tested-by: Nathan Chancellor <nathan@...nel.org>
> ---
> mm/shmem.c | 29 +++++++++++++++++++++++++++--
> 1 file changed, 27 insertions(+), 2 deletions(-)
>
> diff --git a/mm/shmem.c b/mm/shmem.c
> index b86f526a1cb1..6038e1d11987 100644
> --- a/mm/shmem.c
> +++ b/mm/shmem.c
> @@ -5548,13 +5548,38 @@ struct page *shmem_read_mapping_page_gfp(struct address_space *mapping,
> EXPORT_SYMBOL_GPL(shmem_read_mapping_page_gfp);
>
> #if defined(CONFIG_SYSFS) && defined(CONFIG_TMPFS)
> +
> +#define __INIT_KOBJ_ATTR(_name, _mode, _show, _store) \
> +{ \
> + .attr = { .name = __stringify(_name), .mode = _mode }, \
> + .show = _show, \
> + .store = _store, \
> +}
> +
> +#define TMPFS_ATTR_W(_name, _store) \
> + static struct kobj_attribute tmpfs_attr_##_name = \
> + __INIT_KOBJ_ATTR(_name, 0200, NULL, _store)
> +
> +#define TMPFS_ATTR_RW(_name, _show, _store) \
> + static struct kobj_attribute tmpfs_attr_##_name = \
> + __INIT_KOBJ_ATTR(_name, 0644, _show, _store)
> +
> +#define TMPFS_ATTR_RO(_name, _show) \
> + static struct kobj_attribute tmpfs_attr_##_name = \
> + __INIT_KOBJ_ATTR(_name, 0444, _show, NULL)
> +
> #if IS_ENABLED(CONFIG_UNICODE)
> -static DEVICE_STRING_ATTR_RO(casefold, 0444, "supported");
> +static ssize_t casefold_show(struct kobject *kobj, struct kobj_attribute *a,
> + char *buf)
> +{
> + return sysfs_emit(buf, "supported\n");
Small nit, I think this might be overindented?
> +}
> +TMPFS_ATTR_RO(casefold, casefold_show);
> #endif
>
> static struct attribute *tmpfs_attributes[] = {
> #if IS_ENABLED(CONFIG_UNICODE)
> - &dev_attr_casefold.attr.attr,
> + &tmpfs_attr_casefold.attr,
> #endif
> NULL
> };
> --
> 2.47.0
>
Powered by blists - more mailing lists