[PATCH] dm-raid: do not include dm-core.h

[PATCH] dm-raid: do not include dm-core.h

[Pavel Tikhomirov]

In commit 4cc96131afce ("dm: move request-based code out to dm-rq.[hc]")
we have a note: "DM targets should _never_ include dm-core.h!". And it
is not used in any DM targets except dm-raid now, so let's remove it
from dm-raid for consistency, also use special helpers instead of
accessing dm_table and mapper_device fields directly. This change is
merely a cleanup and should not affect functionality.

Fixes: 7168be3c8a6b ("md: record dm-raid gendisk in mddev")
Signed-off-by: Pavel Tikhomirov <ptikhomirov@virtuozzo.com>
---
 drivers/md/dm-raid.c | 3 +--
 1 file changed, 1 insertion(+), 2 deletions(-)

diff --git a/drivers/md/dm-raid.c b/drivers/md/dm-raid.c
index e8c0a8c6fb51..4fb5ddf50560 100644
--- a/drivers/md/dm-raid.c
+++ b/drivers/md/dm-raid.c
@@ -14,7 +14,6 @@
 #include "raid5.h"
 #include "raid10.h"
 #include "md-bitmap.h"
-#include "dm-core.h"
 
 #include <linux/device-mapper.h>
 
@@ -3305,7 +3304,7 @@ static int raid_ctr(struct dm_target *ti, unsigned int argc, char **argv)
 
 	/* Disable/enable discard support on raid set. */
 	configure_discard_support(rs);
-	rs->md.dm_gendisk = ti->table->md->disk;
+	rs->md.dm_gendisk = dm_disk(dm_table_get_md(ti->table));
 
 	mddev_unlock(&rs->md);
 	return 0;
-- 
2.50.0

[PATCH v2] arch: fix resource leak in jbusmc_probe()

[jackysliu]

From: Siyang Liu <1972843537@qq.com>

In the jbusmc_probe function, the device node mem_node fetched
via of_find_node_by_path("/memory") is not properly freed
on all code paths.
This can lead to leakage of device node reference counts,
which may result in kernel resources not being released.

This issue was detected by rule based static tools
developed by Tencent.

Signed-off-by: Siyang Liu <1972843537@qq.com>
---
 arch/sparc/kernel/chmc.c | 10 +++++++---
 1 file changed, 7 insertions(+), 3 deletions(-)

diff --git a/arch/sparc/kernel/chmc.c b/arch/sparc/kernel/chmc.c
index d4c74d6b2e1b..fd20e4ee0971 100644
--- a/arch/sparc/kernel/chmc.c
+++ b/arch/sparc/kernel/chmc.c
@@ -412,7 +412,7 @@ static int jbusmc_probe(struct platform_device *op)
 	mem_regs = of_get_property(mem_node, "reg", &len);
 	if (!mem_regs) {
 		printk(KERN_ERR PFX "Cannot get reg property of /memory node.\n");
-		goto out;
+		goto out_put;
 	}
 	num_mem_regs = len / sizeof(*mem_regs);
 
@@ -420,7 +420,7 @@ static int jbusmc_probe(struct platform_device *op)
 	p = kzalloc(sizeof(*p), GFP_KERNEL);
 	if (!p) {
 		printk(KERN_ERR PFX "Cannot allocate struct jbusmc.\n");
-		goto out;
+		goto out_put;
 	}
 
 	INIT_LIST_HEAD(&p->list);
@@ -473,6 +473,10 @@ static int jbusmc_probe(struct platform_device *op)
 
 	err = 0;
 
+out_put:
+	of_node_put(mem_node);
+	goto out;
+
 out:
 	return err;
 
@@ -481,7 +485,7 @@ static int jbusmc_probe(struct platform_device *op)
 
 out_free:
 	kfree(p);
-	goto out;
+	goto out_put;
 }
 
 /* Does BANK decode PHYS_ADDR? */
-- 
2.43.5

Re: [PATCH] dm-raid: do not include dm-core.h

[Yu Kuai]

在 2025/07/21 11:49, Pavel Tikhomirov 写道:
> In commit 4cc96131afce ("dm: move request-based code out to dm-rq.[hc]")
> we have a note: "DM targets should _never_ include dm-core.h!". And it
> is not used in any DM targets except dm-raid now, so let's remove it
> from dm-raid for consistency, also use special helpers instead of
> accessing dm_table and mapper_device fields directly. This change is
> merely a cleanup and should not affect functionality.
> 
> Fixes: 7168be3c8a6b ("md: record dm-raid gendisk in mddev")
> Signed-off-by: Pavel Tikhomirov <ptikhomirov@virtuozzo.com>
> ---
>   drivers/md/dm-raid.c | 3 +--
>   1 file changed, 1 insertion(+), 2 deletions(-)
> 
Reviewed-by: Yu Kuai <yukuai3@huawei.com>
Thanks

> diff --git a/drivers/md/dm-raid.c b/drivers/md/dm-raid.c
> index e8c0a8c6fb51..4fb5ddf50560 100644
> --- a/drivers/md/dm-raid.c
> +++ b/drivers/md/dm-raid.c
> @@ -14,7 +14,6 @@
>   #include "raid5.h"
>   #include "raid10.h"
>   #include "md-bitmap.h"
> -#include "dm-core.h"
>   
>   #include <linux/device-mapper.h>
>   
> @@ -3305,7 +3304,7 @@ static int raid_ctr(struct dm_target *ti, unsigned int argc, char **argv)
>   
>   	/* Disable/enable discard support on raid set. */
>   	configure_discard_support(rs);
> -	rs->md.dm_gendisk = ti->table->md->disk;
> +	rs->md.dm_gendisk = dm_disk(dm_table_get_md(ti->table));
>   
>   	mddev_unlock(&rs->md);
>   	return 0;
> 

Re: [PATCH] dm-raid: do not include dm-core.h

[Mikulas Patocka]

On Mon, 21 Jul 2025, Pavel Tikhomirov wrote:

> In commit 4cc96131afce ("dm: move request-based code out to dm-rq.[hc]")
> we have a note: "DM targets should _never_ include dm-core.h!". And it
> is not used in any DM targets except dm-raid now, so let's remove it
> from dm-raid for consistency, also use special helpers instead of
> accessing dm_table and mapper_device fields directly. This change is
> merely a cleanup and should not affect functionality.
> 
> Fixes: 7168be3c8a6b ("md: record dm-raid gendisk in mddev")
> Signed-off-by: Pavel Tikhomirov <ptikhomirov@virtuozzo.com>
> ---
>  drivers/md/dm-raid.c | 3 +--
>  1 file changed, 1 insertion(+), 2 deletions(-)
> 
> diff --git a/drivers/md/dm-raid.c b/drivers/md/dm-raid.c
> index e8c0a8c6fb51..4fb5ddf50560 100644
> --- a/drivers/md/dm-raid.c
> +++ b/drivers/md/dm-raid.c
> @@ -14,7 +14,6 @@
>  #include "raid5.h"
>  #include "raid10.h"
>  #include "md-bitmap.h"
> -#include "dm-core.h"
>  
>  #include <linux/device-mapper.h>
>  
> @@ -3305,7 +3304,7 @@ static int raid_ctr(struct dm_target *ti, unsigned int argc, char **argv)
>  
>  	/* Disable/enable discard support on raid set. */
>  	configure_discard_support(rs);
> -	rs->md.dm_gendisk = ti->table->md->disk;
> +	rs->md.dm_gendisk = dm_disk(dm_table_get_md(ti->table));
>  
>  	mddev_unlock(&rs->md);
>  	return 0;
> -- 
> 2.50.0
> 

Applied, thanks.

(I deleted the Fixes line, because it is not really a bugfix and it 
doesn't have to be backported to the stable kernel branches)

Mikulas

Re: [PATCH] arch: fix resource leak in chmc.c

[Markus Elfring]

> In the jbusmc_probe function, the device node mem_node fetched 
> via of_find_node_by_path("/memory") is not properly freed 
> on all code paths. 
…

You may occasionally put more than 62 characters into text lines
of such a change description.

See also:
https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/Documentation/process/submitting-patches.rst?h=v6.16-rc6#n94


* How do you think about to add any tags (like “Fixes” and “Cc”) accordingly?

* Will an other subsystem specification be more appropriate?

* Would it be more helpful to refer to “jbusmc_probe()” instead of “chmc.c”
  in the summary phrase?


Regards,
Markus

[PATCH v2] arch: fix resource leak in jbusmc_probe()

[jackysliu]

From: Siyang Liu <1972843537@qq.com>

In the jbusmc_probe function, the device node mem_node fetched
via of_find_node_by_path("/memory") is not properly freed
on all code paths.
This can lead to leakage of device node reference counts,
which may result in kernel resources not being released.

This issue was detected by rule based static tools
developed by Tencent.

Signed-off-by: Siyang Liu <1972843537@qq.com>
---
 arch/sparc/kernel/chmc.c | 10 +++++++---
 1 file changed, 7 insertions(+), 3 deletions(-)

diff --git a/arch/sparc/kernel/chmc.c b/arch/sparc/kernel/chmc.c
index d4c74d6b2e1b..fd20e4ee0971 100644
--- a/arch/sparc/kernel/chmc.c
+++ b/arch/sparc/kernel/chmc.c
@@ -412,7 +412,7 @@ static int jbusmc_probe(struct platform_device *op)
 	mem_regs = of_get_property(mem_node, "reg", &len);
 	if (!mem_regs) {
 		printk(KERN_ERR PFX "Cannot get reg property of /memory node.\n");
-		goto out;
+		goto out_put;
 	}
 	num_mem_regs = len / sizeof(*mem_regs);
 
@@ -420,7 +420,7 @@ static int jbusmc_probe(struct platform_device *op)
 	p = kzalloc(sizeof(*p), GFP_KERNEL);
 	if (!p) {
 		printk(KERN_ERR PFX "Cannot allocate struct jbusmc.\n");
-		goto out;
+		goto out_put;
 	}
 
 	INIT_LIST_HEAD(&p->list);
@@ -473,6 +473,10 @@ static int jbusmc_probe(struct platform_device *op)
 
 	err = 0;
 
+out_put:
+	of_node_put(mem_node);
+	goto out;
+
 out:
 	return err;
 
@@ -481,7 +485,7 @@ static int jbusmc_probe(struct platform_device *op)
 
 out_free:
 	kfree(p);
-	goto out;
+	goto out_put;
 }
 
 /* Does BANK decode PHYS_ADDR? */
-- 
2.43.5

Re: [PATCH v2] arch: fix resource leak in jbusmc_probe()

[Markus Elfring]

…> This can lead to leakage of device node reference counts,
> which may result in kernel resources not being released.
…

See also one more:
https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/Documentation/process/submitting-patches.rst?h=v6.16-rc6#n94


* Will you become more interested in additional tags (like “Fixes” and “Cc”)?

* Would the subsystem specification “sparc” be more appropriate
  in the summary phrase?


…> ---
>  arch/sparc/kernel/chmc.c | 10 +++++++---
…

How do you think about to improve your version management?
https://lore.kernel.org/all/?q=%22This+looks+like+a+new+version+of+a+previously+submitted+patch%22
https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/Documentation/process/submitting-patches.rst?h=v6.16-rc6#n784

Regards,
Markus

Re: [PATCH v2] arch: fix resource leak in jbusmc_probe()

[Krzysztof Kozlowski]

On 21/07/2025 05:57, jackysliu wrote:
> From: Siyang Liu <1972843537@qq.com>
> 
> In the jbusmc_probe function, the device node mem_node fetched
> via of_find_node_by_path("/memory") is not properly freed
> on all code paths.
> This can lead to leakage of device node reference counts,
> which may result in kernel resources not being released.

Please wrap commit message according to Linux coding style / submission
process (neither too early nor over the limit):
https://elixir.bootlin.com/linux/v6.4-rc1/source/Documentation/process/submitting-patches.rst#L597


Do not attach (thread) your patchsets to some other threads (unrelated
or older versions). This buries them deep in the mailbox and might
interfere with applying entire sets.

> 
> This issue was detected by rule based static tools
> developed by Tencent.

What tools? This must be specific, give their name.

Previously Tencent reports were often bogus or low quality.

> 
> Signed-off-by: Siyang Liu <1972843537@qq.com>
> ---
>  arch/sparc/kernel/chmc.c | 10 +++++++---
>  1 file changed, 7 insertions(+), 3 deletions(-)
> 
> diff --git a/arch/sparc/kernel/chmc.c b/arch/sparc/kernel/chmc.c
> index d4c74d6b2e1b..fd20e4ee0971 100644
> --- a/arch/sparc/kernel/chmc.c
> +++ b/arch/sparc/kernel/chmc.c
> @@ -412,7 +412,7 @@ static int jbusmc_probe(struct platform_device *op)
>  	mem_regs = of_get_property(mem_node, "reg", &len);
>  	if (!mem_regs) {
>  		printk(KERN_ERR PFX "Cannot get reg property of /memory node.\n");
> -		goto out;
> +		goto out_put;

You did not implement at all what I asked. I wanted the code to be
simpler, you just made it more complicated. Read again previous feedback.

Best regards,
Krzysztof