From: Alison Schofield <alison.schofield@intel.com>
To: John Groves <john@jagalactic.com>
Cc: John Groves <John@groves.net>, John Groves <jgroves@fastmail.com>,
"Dan Williams" <djbw@kernel.org>,
John Groves <jgroves@micron.com>,
Vishal Verma <vishal.l.verma@intel.com>,
Dave Jiang <dave.jiang@intel.com>,
"Jonathan Cameron" <Jonathan.Cameron@huawei.com>,
Aravind Ramesh <arramesh@micron.com>,
Ajay Joshi <ajayjoshi@micron.com>,
"venkataravis@micron.com" <venkataravis@micron.com>,
"dev.srinivasulu@gmail.com" <dev.srinivasulu@gmail.com>,
"linux-kernel@vger.kernel.org" <linux-kernel@vger.kernel.org>,
"nvdimm@lists.linux.dev" <nvdimm@lists.linux.dev>,
"linux-cxl@vger.kernel.org" <linux-cxl@vger.kernel.org>
Subject: Re: [PATCH V5 1/2] daxctl: Add support for famfs mode
Date: Tue, 12 May 2026 22:32:12 -0700 [thread overview]
Message-ID: <agQM3KXNMvtbpb_0@aschofie-mobl2.lan> (raw)
In-Reply-To: <0100019ddf06b207-eaf8cb8a-066e-4642-8947-effdb4848c20-000000@email.amazonses.com>
On Thu, Apr 30, 2026 at 03:34:11PM +0000, John Groves wrote:
> From: John Groves <John@Groves.net>
Hi John,
Please add an intro paragraph saying what famfs is, preferably
with links to reference material.
This patch doesn't update any docs. Please take a look at what is
needed. I think at least this one:
Documentation/daxctl/daxctl-reconfigure-device.txt needs to add
famfs as a 'mode' option.
> Putting a daxdev in famfs mode means binding it to fsdev_dax.ko
> (drivers/dax/fsdev.c). Finding a daxdev bound to fsdev_dax means
> it is in famfs mode.
>
> A test for this functionality is added in the next commit.
>
> With devdax, famfs, and system-ram modes, the previous logic that assumed
> 'not in mode X means in mode Y' needed to get slightly more complicated.
>
> Add explicit mode detection functions:
> - daxctl_dev_is_famfs_mode(): check if bound to fsdev_dax driver
> - daxctl_dev_is_devdax_mode(): check if bound to device_dax driver
> Both delegate to a shared static helper daxctl_dev_bound_to_module() to
> avoid duplicating the driver-symlink lookup, as does the pre-existing
> daxctl_dev_is_system_ram_capable().
>
> Update mode transition logic in device.c:
> - disable_devdax_device(): verify device is actually in devdax mode
> - disable_famfs_device(): verify device is actually in famfs mode
> - All reconfig_mode_*() functions explicitly check each mode
> - Handle unrecognized mode with an error instead of wrong assumption
>
> Update json.c to report fsdev_dax-bound devices as 'famfs' mode. An
> unbound device continues to be reported as 'devdax' (the legacy default
> when no driver is bound), to preserve existing behavior.
>
> Signed-off-by: John Groves <john@groves.net>
> ---
> daxctl/device.c | 132 +++++++++++++++++++++++++++++++++++++----
> daxctl/json.c | 13 +++-
> daxctl/lib/libdaxctl-private.h | 2 +
> daxctl/lib/libdaxctl.c | 39 ++++++++++--
> daxctl/lib/libdaxctl.sym | 7 +++
> daxctl/libdaxctl.h | 3 +
> 6 files changed, 181 insertions(+), 15 deletions(-)
>
> diff --git a/daxctl/device.c b/daxctl/device.c
> index a4e36b130a09..003609e4abba 100644
> --- a/daxctl/device.c
> +++ b/daxctl/device.c
> @@ -42,6 +42,7 @@ enum dev_mode {
> DAXCTL_DEV_MODE_UNKNOWN,
> DAXCTL_DEV_MODE_DEVDAX,
> DAXCTL_DEV_MODE_RAM,
> + DAXCTL_DEV_MODE_FAMFS,
> };
The above enum dev_mode in device.c shares enum names with enum
daxctl_dev_mode later in this patch (in libdaxctl-private.h) but assigns
different numeric values. The overlap predates this patch, so not
something you introduced, but adding FAMFS to both is a good moment to
fix it before it gets more entrenched.
Two enums with identical member names but different values is confusing
and risks silent cross-assignment bugs with no compiler warning.
Suggest renaming this local enum to `enum reconfig_mode` with members
`RECONFIG_MODE_{UNKNOWN,DEVDAX,RAM,FAMFS}`. That confines it to the reconfig
path. Then the library enum gets unambiguous ownership of `DAXCTL_DEV_MODE_*`.
After that accidental cross-assignments are obvious type mismatches rather
than a silent wrong-value bug.
>
> struct mapping {
> unsigned long long start, end, pgoff;
> @@ -471,6 +472,13 @@ static const char *parse_device_options(int argc, const char **argv,
> "--no-online is incompatible with --mode=devdax\n");
> rc = -EINVAL;
> }
> + } else if (strcmp(param.mode, "famfs") == 0) {
> + reconfig_mode = DAXCTL_DEV_MODE_FAMFS;
> + if (param.no_online) {
> + fprintf(stderr,
> + "--no-online is incompatible with --mode=famfs\n");
> + rc = -EINVAL;
rm extra whitespace after =
> + }
> }
> break;
> case ACTION_CREATE:
> @@ -696,8 +704,42 @@ static int disable_devdax_device(struct daxctl_dev *dev)
> int rc;
>
> if (mem) {
> - fprintf(stderr, "%s was already in system-ram mode\n",
> - devname);
> + fprintf(stderr, "%s is in system-ram mode\n", devname);
> + return 1;
> + }
> + if (daxctl_dev_is_famfs_mode(dev)) {
> + fprintf(stderr, "%s is in famfs mode\n", devname);
> + return 1;
> + }
> + if (!daxctl_dev_is_devdax_mode(dev)) {
> + fprintf(stderr, "%s is not in devdax mode\n", devname);
> + return 1;
> + }
> + rc = daxctl_dev_disable(dev);
> + if (rc) {
> + fprintf(stderr, "%s: disable failed: %s\n",
> + daxctl_dev_get_devname(dev), strerror(-rc));
> + return rc;
> + }
> + return 0;
> +}
> +
> +static int disable_famfs_device(struct daxctl_dev *dev)
> +{
> + struct daxctl_memory *mem = daxctl_dev_get_memory(dev);
> + const char *devname = daxctl_dev_get_devname(dev);
> + int rc;
> +
> + if (mem) {
> + fprintf(stderr, "%s is in system-ram mode\n", devname);
> + return 1;
> + }
> + if (daxctl_dev_is_devdax_mode(dev)) {
> + fprintf(stderr, "%s is in devdax mode\n", devname);
> + return 1;
> + }
> + if (!daxctl_dev_is_famfs_mode(dev)) {
> + fprintf(stderr, "%s is not in famfs mode\n", devname);
> return 1;
> }
> rc = daxctl_dev_disable(dev);
disable_devdax_device() and disable_famfs_device() differ only in
which mode they accept vs. reject. With reconfig_mode_* switching on
daxctl_dev_get_mode() first, the caller already knows the mode, so
the internal mode-sanity checks are redundant.
What's left in each function is daxctl_dev_disable() and an error
fprintf, the same code in both, so collapse to a single helper:
static int disable_mode_device(struct daxctl_dev *dev);
No mode parameter: the precondition (caller has matched the mode)
means nothing mode-specific remains.
> @@ -711,6 +753,7 @@ static int disable_devdax_device(struct daxctl_dev *dev)
>
> static int reconfig_mode_system_ram(struct daxctl_dev *dev)
> {
> + struct daxctl_memory *mem = daxctl_dev_get_memory(dev);
> const char *devname = daxctl_dev_get_devname(dev);
> int rc, skip_enable = 0;
>
> @@ -724,11 +767,21 @@ static int reconfig_mode_system_ram(struct daxctl_dev *dev)
> }
>
> if (daxctl_dev_is_enabled(dev)) {
> - rc = disable_devdax_device(dev);
> - if (rc < 0)
> - return rc;
> - if (rc > 0)
> + if (mem) {
> + /* already in system-ram mode */
> skip_enable = 1;
> + } else if (daxctl_dev_is_famfs_mode(dev)) {
> + rc = disable_famfs_device(dev);
> + if (rc)
> + return rc;
> + } else if (daxctl_dev_is_devdax_mode(dev)) {
> + rc = disable_devdax_device(dev);
> + if (rc)
> + return rc;
> + } else {
> + fprintf(stderr, "%s: unknown mode\n", devname);
> + return -EINVAL;
> + }
> }
This if-else chain is repeated in all three reconfig_mode_*() functions below.
Please add a private helper and switch on the result. Something like:
static enum daxctl_dev_mode daxctl_dev_get_mode(struct daxctl_dev *dev)
{
if (daxctl_dev_get_memory(dev))
return DAXCTL_DEV_MODE_RAM;
if (daxctl_dev_is_famfs_mode(dev))
return DAXCTL_DEV_MODE_FAMFS;
if (daxctl_dev_is_devdax_mode(dev))
return DAXCTL_DEV_MODE_DEVDAX;
return DAXCTL_DEV_MODE_UNKNOWN;
}
Then each reconfig_mode_* becomes a switch on the current mode. The
precedence (system-ram first because it's detected via `mem` rather
than driver binding) lives in one place instead of three.
> @@ -750,7 +803,7 @@ static int disable_system_ram_device(struct daxctl_dev *dev)
> int rc;
>
> if (!mem) {
> - fprintf(stderr, "%s was already in devdax mode\n", devname);
> + fprintf(stderr, "%s is not in system-ram mode\n", devname);
> return 1;
> }
>
> @@ -786,12 +839,31 @@ static int disable_system_ram_device(struct daxctl_dev *dev)
>
> static int reconfig_mode_devdax(struct daxctl_dev *dev)
> {
> + struct daxctl_memory *mem = daxctl_dev_get_memory(dev);
> + const char *devname = daxctl_dev_get_devname(dev);
> int rc;
>
> if (daxctl_dev_is_enabled(dev)) {
> - rc = disable_system_ram_device(dev);
> - if (rc)
> - return rc;
> + if (mem) {
> + rc = disable_system_ram_device(dev);
> + if (rc)
> + return rc;
> + } else if (daxctl_dev_is_famfs_mode(dev)) {
> + rc = disable_famfs_device(dev);
> + if (rc)
> + return rc;
> + } else if (daxctl_dev_is_devdax_mode(dev)) {
> + /* already in devdax mode, just re-enable */
> + rc = daxctl_dev_disable(dev);
> + if (rc) {
> + fprintf(stderr, "%s: disable failed: %s\n",
> + devname, strerror(-rc));
> + return rc;
> + }
> + } else {
> + fprintf(stderr, "%s: unknown mode\n", devname);
> + return -EINVAL;
> + }
> }
Same if-else chain, second copy. See the helper suggestion above.
> rc = daxctl_dev_enable_devdax(dev);
> @@ -801,6 +873,43 @@ static int reconfig_mode_devdax(struct daxctl_dev *dev)
> return 0;
> }
>
> +static int reconfig_mode_famfs(struct daxctl_dev *dev)
> +{
> + struct daxctl_memory *mem = daxctl_dev_get_memory(dev);
> + const char *devname = daxctl_dev_get_devname(dev);
> + int rc;
> +
> + if (daxctl_dev_is_enabled(dev)) {
> + if (mem) {
> + fprintf(stderr,
> + "%s is in system-ram mode; must be in devdax mode to convert to famfs\n",
> + devname);
> + return -EINVAL;
> + } else if (daxctl_dev_is_famfs_mode(dev)) {
> + /* already in famfs mode, just re-enable */
> + rc = daxctl_dev_disable(dev);
> + if (rc) {
> + fprintf(stderr, "%s: disable failed: %s\n",
> + devname, strerror(-rc));
> + return rc;
> + }
> + } else if (daxctl_dev_is_devdax_mode(dev)) {
> + rc = disable_devdax_device(dev);
> + if (rc)
> + return rc;
> + } else {
> + fprintf(stderr, "%s: unknown mode\n", devname);
> + return -EINVAL;
> + }
> + }
Same if-else chain, third copy.
> +
> + rc = daxctl_dev_enable_famfs(dev);
> + if (rc)
> + return rc;
> +
> + return 0;
> +}
> +
> static int do_create(struct daxctl_region *region, long long val,
> struct json_object **jdevs)
> {
> @@ -887,6 +996,9 @@ static int do_reconfig(struct daxctl_dev *dev, enum dev_mode mode,
> case DAXCTL_DEV_MODE_DEVDAX:
> rc = reconfig_mode_devdax(dev);
> break;
> + case DAXCTL_DEV_MODE_FAMFS:
> + rc = reconfig_mode_famfs(dev);
> + break;
> default:
> fprintf(stderr, "%s: unknown mode requested: %d\n",
> devname, mode);
> diff --git a/daxctl/json.c b/daxctl/json.c
> index 3cbce9dcd651..2a4b12c2f925 100644
> --- a/daxctl/json.c
> +++ b/daxctl/json.c
> @@ -48,8 +48,19 @@ struct json_object *util_daxctl_dev_to_json(struct daxctl_dev *dev,
>
> if (mem)
> jobj = json_object_new_string("system-ram");
> - else
> + else if (daxctl_dev_is_famfs_mode(dev))
> + jobj = json_object_new_string("famfs");
> + else if (daxctl_dev_is_devdax_mode(dev))
> jobj = json_object_new_string("devdax");
The 'else if' above, and the 'else' below, both assign "devdax",
so the 'else if' is redundant. And once daxctl_dev_get_mode()
exists, this becomes a switch on the result with "devdax"
as the UNKNOWN fallback.
> + else {
> + /* Legacy condition; if a daxdev is not in any "mode", that
> + * means no driver is bound. We report that as a disabled
> + * device in devdax mode. (the disabled modifier is added later
> + * in this function if applicable)
> + */
> + jobj = json_object_new_string("devdax");
> + }
> +
> if (jobj)
> json_object_object_add(jdev, "mode", jobj);
>
> diff --git a/daxctl/lib/libdaxctl-private.h b/daxctl/lib/libdaxctl-private.h
> index ae45311e5d57..0bb73e8c04bf 100644
> --- a/daxctl/lib/libdaxctl-private.h
> +++ b/daxctl/lib/libdaxctl-private.h
> @@ -21,12 +21,14 @@ static const char *dax_subsystems[] = {
> enum daxctl_dev_mode {
> DAXCTL_DEV_MODE_DEVDAX = 0,
> DAXCTL_DEV_MODE_RAM,
> + DAXCTL_DEV_MODE_FAMFS,
> DAXCTL_DEV_MODE_END,
> };
>
> static const char *dax_modules[] = {
> [DAXCTL_DEV_MODE_DEVDAX] = "device_dax",
> [DAXCTL_DEV_MODE_RAM] = "kmem",
> + [DAXCTL_DEV_MODE_FAMFS] = "fsdev_dax",
> };
Add a DAXCTL_DEV_MODE_UNKNOWN sentinel here to support the
daxctl_dev_get_mode() helper cleanly. Doesn't have to be assigned a
value, can reuse END or add it alongside, your call.
> enum memory_op {
> diff --git a/daxctl/lib/libdaxctl.c b/daxctl/lib/libdaxctl.c
> index 02ae7e50b123..33121dcb1d1b 100644
> --- a/daxctl/lib/libdaxctl.c
> +++ b/daxctl/lib/libdaxctl.c
> @@ -385,13 +385,13 @@ static bool device_model_is_dax_bus(struct daxctl_dev *dev)
> return false;
> }
>
> -DAXCTL_EXPORT int daxctl_dev_is_system_ram_capable(struct daxctl_dev *dev)
> +static int daxctl_dev_bound_to_module(struct daxctl_dev *dev, const char *mod_name)
> {
> const char *devname = daxctl_dev_get_devname(dev);
> struct daxctl_ctx *ctx = daxctl_dev_get_ctx(dev);
> const char *mod_base;
> char *mod_path;
> - char path[200];
> + char path[PATH_MAX];
> const int len = sizeof(path);
>
> if (!device_model_is_dax_bus(dev))
Nice!
> @@ -406,11 +406,13 @@ DAXCTL_EXPORT int daxctl_dev_is_system_ram_capable(struct daxctl_dev *dev)
> }
>
> mod_path = realpath(path, NULL);
> - if (!mod_path)
> + if (!mod_path) {
> + dbg(ctx, "%s: realpath failed for driver link\n", devname);
> return false;
> + }
>
> mod_base = path_basename(mod_path);
> - if (strcmp(mod_base, dax_modules[DAXCTL_DEV_MODE_RAM]) == 0) {
> + if (strcmp(mod_base, mod_name) == 0) {
> free(mod_path);
> return true;
> }
> @@ -419,6 +421,30 @@ DAXCTL_EXPORT int daxctl_dev_is_system_ram_capable(struct daxctl_dev *dev)
> return false;
> }
>
> +DAXCTL_EXPORT int daxctl_dev_is_system_ram_capable(struct daxctl_dev *dev)
> +{
> + return daxctl_dev_bound_to_module(dev, dax_modules[DAXCTL_DEV_MODE_RAM]);
> +}
With daxctl_dev_is_famfs_mode() and daxctl_dev_is_devdax_mode() now
alongside it, the _capable vs _mode naming split looks inconsistent.
Post-refactor all three delegate to daxctl_dev_bound_to_module(), which
checks current driver binding.
Either document the semantic distinction more explicitly or consider
adding daxctl_dev_is_system_ram_mode() as the preferred interface.
> +
> +/*
> + * Check if device is currently in famfs mode (bound to fsdev_dax driver).
> + * Returns false for disabled devices: the DAX bus does not retain the previous
> + * driver binding after unbind, so mode cannot be determined without a driver.
> + */
> +DAXCTL_EXPORT int daxctl_dev_is_famfs_mode(struct daxctl_dev *dev)
> +{
> + return daxctl_dev_bound_to_module(dev, dax_modules[DAXCTL_DEV_MODE_FAMFS]);
> +}
> +
> +/*
> + * Check if device is currently in devdax mode (bound to device_dax driver).
> + * Returns false for disabled devices; see daxctl_dev_is_famfs_mode().
> + */
> +DAXCTL_EXPORT int daxctl_dev_is_devdax_mode(struct daxctl_dev *dev)
> +{
> + return daxctl_dev_bound_to_module(dev, dax_modules[DAXCTL_DEV_MODE_DEVDAX]);
> +}
These comments claim "Returns false for disabled devices" as if that
were unique to these two helpers, but it isn't. daxctl_dev_is_system_ram_capable
returns false for disabled devices too. Either drop the comment or move it to
to daxctl_dev_bound_to_module().
> +
> /*
> * This checks for the device to be in system-ram mode, so calling
> * daxctl_dev_get_memory() on a devdax mode device will always return NULL.
> @@ -983,6 +1009,11 @@ DAXCTL_EXPORT int daxctl_dev_enable_ram(struct daxctl_dev *dev)
> return daxctl_dev_enable(dev, DAXCTL_DEV_MODE_RAM);
> }
>
> +DAXCTL_EXPORT int daxctl_dev_enable_famfs(struct daxctl_dev *dev)
> +{
> + return daxctl_dev_enable(dev, DAXCTL_DEV_MODE_FAMFS);
> +}
OK.
> +
> DAXCTL_EXPORT int daxctl_dev_disable(struct daxctl_dev *dev)
> {
> const char *devname = daxctl_dev_get_devname(dev);
> diff --git a/daxctl/lib/libdaxctl.sym b/daxctl/lib/libdaxctl.sym
> index 309881196c86..2a812c6ad918 100644
> --- a/daxctl/lib/libdaxctl.sym
> +++ b/daxctl/lib/libdaxctl.sym
> @@ -104,3 +104,10 @@ LIBDAXCTL_10 {
> global:
> daxctl_dev_is_system_ram_capable;
> } LIBDAXCTL_9;
> +
> +LIBDAXCTL_11 {
> +global:
> + daxctl_dev_enable_famfs;
> + daxctl_dev_is_famfs_mode;
> + daxctl_dev_is_devdax_mode;
> +} LIBDAXCTL_10;
If you resolve the _capable/_mode naming inconsistency by adding
daxctl_dev_is_system_ram_mode(), please export it here as well.
> diff --git a/daxctl/libdaxctl.h b/daxctl/libdaxctl.h
> index 53c6bbdae5c3..84fcdb40c7a9 100644
> --- a/daxctl/libdaxctl.h
> +++ b/daxctl/libdaxctl.h
> @@ -72,12 +72,15 @@ int daxctl_dev_is_enabled(struct daxctl_dev *dev);
> int daxctl_dev_disable(struct daxctl_dev *dev);
> int daxctl_dev_enable_devdax(struct daxctl_dev *dev);
> int daxctl_dev_enable_ram(struct daxctl_dev *dev);
> +int daxctl_dev_enable_famfs(struct daxctl_dev *dev);
> int daxctl_dev_get_target_node(struct daxctl_dev *dev);
> int daxctl_dev_will_auto_online_memory(struct daxctl_dev *dev);
> int daxctl_dev_has_online_memory(struct daxctl_dev *dev);
>
> struct daxctl_memory;
> int daxctl_dev_is_system_ram_capable(struct daxctl_dev *dev);
> +int daxctl_dev_is_famfs_mode(struct daxctl_dev *dev);
> +int daxctl_dev_is_devdax_mode(struct daxctl_dev *dev);
Mirror whatever you do about _capable vs _mode here. If you add
daxctl_dev_get_mode() as a public library function, declare it here
too.
> struct daxctl_memory *daxctl_dev_get_memory(struct daxctl_dev *dev);
> struct daxctl_dev *daxctl_memory_get_dev(struct daxctl_memory *mem);
> const char *daxctl_memory_get_node_path(struct daxctl_memory *mem);
next prev parent reply other threads:[~2026-05-13 5:32 UTC|newest]
Thread overview: 8+ messages / expand[flat|nested] mbox.gz Atom feed top
[not found] <20260430153331.84139-1-john@jagalactic.com>
2026-04-30 15:33 ` [PATCH V5 0/2] daxctl: Add support for famfs mode John Groves
2026-04-30 15:34 ` [PATCH V5 1/2] " John Groves
2026-04-30 16:20 ` Dave Jiang
2026-05-13 5:32 ` Alison Schofield [this message]
2026-04-30 15:34 ` [PATCH V5 2/2] Add test/daxctl-famfs.sh to test famfs mode transitions: John Groves
2026-04-30 16:27 ` Dave Jiang
2026-04-30 16:50 ` John Groves
2026-05-13 5:34 ` Alison Schofield
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=agQM3KXNMvtbpb_0@aschofie-mobl2.lan \
--to=alison.schofield@intel.com \
--cc=John@groves.net \
--cc=Jonathan.Cameron@huawei.com \
--cc=ajayjoshi@micron.com \
--cc=arramesh@micron.com \
--cc=dave.jiang@intel.com \
--cc=dev.srinivasulu@gmail.com \
--cc=djbw@kernel.org \
--cc=jgroves@fastmail.com \
--cc=jgroves@micron.com \
--cc=john@jagalactic.com \
--cc=linux-cxl@vger.kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=nvdimm@lists.linux.dev \
--cc=venkataravis@micron.com \
--cc=vishal.l.verma@intel.com \
/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