From: Boris Brezillon <boris.brezillon@free-electrons.com>
To: David Woodhouse <dwmw2@infradead.org>,
Brian Norris <computersforpeace@gmail.com>,
Boris Brezillon <boris.brezillon@free-electrons.com>,
Marek Vasut <marek.vasut@gmail.com>,
Richard Weinberger <richard@nod.at>,
Cyrille Pitchen <cyrille.pitchen@wedev4u.fr>,
linux-mtd@lists.infradead.org
Cc: "Mario J. Rugiero" <mrugiero@gmail.com>
Subject: Re: [PATCH 1/2] mtd: Make sure MTD objects always have a valid debugfs dir
Date: Wed, 4 Oct 2017 15:53:42 +0200 [thread overview]
Message-ID: <20171004155342.32e096b8@bbrezillon> (raw)
In-Reply-To: <20171004135239.16316-1-boris.brezillon@free-electrons.com>
On Wed, 4 Oct 2017 15:52:38 +0200
Boris Brezillon <boris.brezillon@free-electrons.com> wrote:
> Master MTD devices are not registered to MTD subsystem if they are
> exposing MTD partitions unless the CONFIG_MTD_PARTITIONED_MASTER option
> is enabled.
>
> This lead to a weird situation where some MTD device drivers are trying
> to add debugfs files to the master MTD device, but this device has no
> valid debugfs directory.
>
> Rework the core logic to do most of the MTD registration steps
> (including debugfs dir creation) except the registration to the device
> model, so that master devices are never exposed to the outside world
> but are still able to expose debugfs entries.
>
> These devices will be exposed as mtd-hiddenX in the mtd debugfs dir.
> Note that X is unique within the hidden MTD device pool but can collide
> with ids of exposed MTD devs.
>
> This commit fixes a bug introduced by commit e8e3edb95ce6 ("mtd: create
> per-device and module-scope debugfs entries") which is preventing
> nandsim from loading when CONFIG_DEBUG_FS is enabled.
>
> Fixes: e8e3edb95ce6 ("mtd: create per-device and module-scope debugfs entries")
Oops, I forgot
Reported-by: Richard Weinberger <richard@nod.at>
> Signed-off-by: Boris Brezillon <boris.brezillon@free-electrons.com>
> ---
> drivers/mtd/mtdcore.c | 83 ++++++++++++++++++++++++++++++++-----------------
> include/linux/mtd/mtd.h | 1 +
> 2 files changed, 56 insertions(+), 28 deletions(-)
>
> diff --git a/drivers/mtd/mtdcore.c b/drivers/mtd/mtdcore.c
> index e7ea842ba3db..323e621f2eba 100644
> --- a/drivers/mtd/mtdcore.c
> +++ b/drivers/mtd/mtdcore.c
> @@ -79,6 +79,10 @@ static struct class mtd_class = {
> .pm = MTD_CLS_PM_OPS,
> };
>
> +#define MTD_MAX_IDS 0x10000
> +#define MTD_ID_START 0x0
> +#define MTD_HIDDEN_ID_START (MTD_ID_START + MTD_MAX_IDS)
> +
> static DEFINE_IDR(mtd_idr);
>
> /* These are exported solely for the purpose of mtd_blkdevs.c. You
> @@ -88,7 +92,13 @@ EXPORT_SYMBOL_GPL(mtd_table_mutex);
>
> struct mtd_info *__mtd_next_device(int i)
> {
> - return idr_get_next(&mtd_idr, &i);
> + struct mtd_info *mtd;
> +
> + mtd = idr_get_next(&mtd_idr, &i);
> + if (!mtd || i >= MTD_MAX_IDS)
> + return NULL;
> +
> + return mtd;
> }
> EXPORT_SYMBOL_GPL(__mtd_next_device);
>
> @@ -505,7 +515,13 @@ int add_mtd_device(struct mtd_info *mtd)
> BUG_ON(mtd->writesize == 0);
> mutex_lock(&mtd_table_mutex);
>
> - i = idr_alloc(&mtd_idr, mtd, 0, 0, GFP_KERNEL);
> + if (!mtd->hidden)
> + i = idr_alloc(&mtd_idr, mtd, MTD_ID_START, MTD_MAX_IDS,
> + GFP_KERNEL);
> + else
> + i = idr_alloc(&mtd_idr, mtd, MTD_HIDDEN_ID_START,
> + MTD_HIDDEN_ID_START + MTD_MAX_IDS,
> + GFP_KERNEL);
> if (i < 0) {
> error = i;
> goto fail_locked;
> @@ -545,15 +561,20 @@ int add_mtd_device(struct mtd_info *mtd)
> /* Caller should have set dev.parent to match the
> * physical device, if appropriate.
> */
> - mtd->dev.type = &mtd_devtype;
> - mtd->dev.class = &mtd_class;
> - mtd->dev.devt = MTD_DEVT(i);
> - dev_set_name(&mtd->dev, "mtd%d", i);
> dev_set_drvdata(&mtd->dev, mtd);
> of_node_get(mtd_get_of_node(mtd));
> - error = device_register(&mtd->dev);
> - if (error)
> - goto fail_added;
> + if (!mtd->hidden) {
> + mtd->dev.type = &mtd_devtype;
> + mtd->dev.class = &mtd_class;
> + mtd->dev.devt = MTD_DEVT(i);
> + dev_set_name(&mtd->dev, "mtd%d", i);
> + error = device_register(&mtd->dev);
> + if (error)
> + goto fail_added;
> + } else {
> + dev_set_name(&mtd->dev, "mtdmaster%d",
> + i - MTD_HIDDEN_ID_START);
> + }
>
> if (!IS_ERR_OR_NULL(dfs_dir_mtd)) {
> mtd->dbg.dfs_dir = debugfs_create_dir(dev_name(&mtd->dev), dfs_dir_mtd);
> @@ -563,14 +584,16 @@ int add_mtd_device(struct mtd_info *mtd)
> }
> }
>
> - device_create(&mtd_class, mtd->dev.parent, MTD_DEVT(i) + 1, NULL,
> - "mtd%dro", i);
> + if (!mtd->hidden) {
> + device_create(&mtd_class, mtd->dev.parent, MTD_DEVT(i) + 1,
> + NULL, "mtd%dro", i);
>
> - pr_debug("mtd: Giving out device %d to %s\n", i, mtd->name);
> - /* No need to get a refcount on the module containing
> - the notifier, since we hold the mtd_table_mutex */
> - list_for_each_entry(not, &mtd_notifiers, list)
> - not->add(mtd);
> + pr_debug("mtd: Giving out device %d to %s\n", i, mtd->name);
> + /* No need to get a refcount on the module containing
> + the notifier, since we hold the mtd_table_mutex */
> + list_for_each_entry(not, &mtd_notifiers, list)
> + not->add(mtd);
> + }
>
> mutex_unlock(&mtd_table_mutex);
> /* We _know_ we aren't being removed, because
> @@ -612,17 +635,20 @@ int del_mtd_device(struct mtd_info *mtd)
> goto out_error;
> }
>
> - /* No need to get a refcount on the module containing
> - the notifier, since we hold the mtd_table_mutex */
> - list_for_each_entry(not, &mtd_notifiers, list)
> - not->remove(mtd);
> + if (!mtd->hidden) {
> + /* No need to get a refcount on the module containing
> + the notifier, since we hold the mtd_table_mutex */
> + list_for_each_entry(not, &mtd_notifiers, list)
> + not->remove(mtd);
> + }
>
> if (mtd->usecount) {
> printk(KERN_NOTICE "Removing MTD device #%d (%s) with use count %d\n",
> mtd->index, mtd->name, mtd->usecount);
> ret = -EBUSY;
> } else {
> - device_unregister(&mtd->dev);
> + if (!mtd->hidden)
> + device_unregister(&mtd->dev);
>
> idr_remove(&mtd_idr, mtd->index);
> of_node_put(mtd_get_of_node(mtd));
> @@ -643,15 +669,16 @@ static int mtd_add_device_partitions(struct mtd_info *mtd,
> int nbparts = parts->nr_parts;
> int ret;
>
> - if (nbparts == 0 || IS_ENABLED(CONFIG_MTD_PARTITIONED_MASTER)) {
> - ret = add_mtd_device(mtd);
> - if (ret)
> - return ret;
> - }
> + if (nbparts && !IS_ENABLED(CONFIG_MTD_PARTITIONED_MASTER))
> + mtd->hidden = true;
> +
> + ret = add_mtd_device(mtd);
> + if (ret)
> + return ret;
>
> if (nbparts > 0) {
> ret = add_mtd_partitions(mtd, real_parts, nbparts);
> - if (ret && IS_ENABLED(CONFIG_MTD_PARTITIONED_MASTER))
> + if (ret)
> del_mtd_device(mtd);
> return ret;
> }
> @@ -857,7 +884,7 @@ struct mtd_info *get_mtd_device(struct mtd_info *mtd, int num)
> break;
> }
> }
> - } else if (num >= 0) {
> + } else if (num >= MTD_ID_START && num < MTD_MAX_IDS) {
> ret = idr_find(&mtd_idr, num);
> if (mtd && mtd != ret)
> ret = NULL;
> diff --git a/include/linux/mtd/mtd.h b/include/linux/mtd/mtd.h
> index 6cd0f6b7658b..382c996b51df 100644
> --- a/include/linux/mtd/mtd.h
> +++ b/include/linux/mtd/mtd.h
> @@ -268,6 +268,7 @@ struct mtd_info {
> unsigned int bitflip_threshold;
>
> // Kernel-only stuff starts here.
> + bool hidden;
> const char *name;
> int index;
>
next prev parent reply other threads:[~2017-10-04 13:54 UTC|newest]
Thread overview: 5+ messages / expand[flat|nested] mbox.gz Atom feed top
2017-10-04 13:52 [PATCH 1/2] mtd: Make sure MTD objects always have a valid debugfs dir Boris Brezillon
2017-10-04 13:52 ` [PATCH 2/2] mtd: mtdpart: Create a symlink to the master " Boris Brezillon
2017-10-04 13:53 ` Boris Brezillon [this message]
2017-10-04 13:56 ` [PATCH 1/2] mtd: Make sure MTD objects always have a valid " Boris Brezillon
2017-10-04 14:26 ` Richard Weinberger
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20171004155342.32e096b8@bbrezillon \
--to=boris.brezillon@free-electrons.com \
--cc=computersforpeace@gmail.com \
--cc=cyrille.pitchen@wedev4u.fr \
--cc=dwmw2@infradead.org \
--cc=linux-mtd@lists.infradead.org \
--cc=marek.vasut@gmail.com \
--cc=mrugiero@gmail.com \
--cc=richard@nod.at \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox