clk: tegra: (clk-tegra30) Add missing of_node_put()
Subsystems (3)
PATCH v2 1/1 clk: tegra: (clk-tegra30) Add missing of_node_put()
In tegra30_clock_init, of_find_matching_node() will return a node pointer with refcount incremented. We should use of_node_put() when the node pointer is not used anymore.
---
changelog:
v2: use real name for Sob
v1: fix missing bug
drivers/clk/tegra/clk-tegra30.c | 1 + 1 file changed, 1 insertion(+)
drivers/clk/tegra/clk-tegra30.c +1 -0
@@ -1320,6 +1320,7 @@ static void __init tegra30_clock_init(struct device_node *np) } pmc_base =of_iomap (node,0 ); +of_node_put (node);if (!pmc_base) {pr_err ("Can't map pmc registers\n" );BUG ();
Signature · 1 line
2.25.1
Stephen Boyd
maintainerThe subject should match historical subjects
$ git log --oneline -3 -- drivers/clk/tegra/clk-tegra30.c
shows mostly "clk: tegra: ". Can you also combine this with the other tegra patch? Don't think we need two patches for essentially the same thing.
Quoting Liang He (2022-06-15 20:36:22)
Quoted 9 lines
In
tegra30_clock_init,of_find_matching_node()will return a node pointer with refcount incremented. We should useof_node_put()when the node pointer is not used anymore.Signed-off-by Liang He <[email protected]>---
changelog:v2: use real name for Sob
Thanks!
Quoted 1 line
v1: fix missing bug
Quoted 13 lines · Stephen Boyd
The subject should match historical subjects
$ git log --oneline -3 -- drivers/clk/tegra/clk-tegra30.cshows mostly "clk: tegra: ". Can you also combine this with the other tegra patch? Don't think we need two patches for essentially the same thing.
Quoting Liang He (2022-06-15 20:36:22)
Quoted 4 lines
In
tegra30_clock_init,of_find_matching_node()will return a node pointer with refcount incremented. We should useof_node_put()when the node pointer is not used anymore.
Quoted 9 lines
Quoted 5 lines
Signed-off-by Liang He <[email protected]>---
changelog:v2: use real name for Sob
Thanks!
Quoted 1 line
v1: fix missing bug
Sorry for my fault. I have been advised to use real name and I resend a PATCH v2 with my real name, but really with the same patch code.
So how can I withdraw the first patch or resend other thing? I am confused.
Can you help me, Conor?
Quoted 22 lines · Liang He
Quoted 21 lines · Stephen Boyd
The subject should match historical subjects
$ git log --oneline -3 -- drivers/clk/tegra/clk-tegra30.cshows mostly "clk: tegra: ". Can you also combine this with the other tegra patch? Don't think we need two patches for essentially the same thing.
Quoting Liang He (2022-06-15 20:36:22)
Quoted 9 lines
In
tegra30_clock_init,of_find_matching_node()will return a node pointer with refcount incremented. We should useof_node_put()when the node pointer is not used anymore.Signed-off-by Liang He <[email protected]>---
changelog:v2: use real name for Sob
Thanks!
Quoted 7 lines
Quoted 1 line
Quoted 1 line
v1: fix missing bug
Sorry for my fault. I have been advised to use real name and I resend a PATCH v2 with my real name, but really with the same patch code.
So how can I withdraw the first patch or resend other thing? I am confused.
Can you help me, Conor?
Yeah, sure. I think you're just getting a bit confused by conflicting responses from different people. Some of the things I said on whatever the original patch I replied to apply everywhere - like using your real name or adding changelogs.
However, different subsystem maintainers have a different opinions about how patches for their subsystem should look. I would imagine that it was Guenter Roeck that asked you to use the "subsystem: (driver) action" subject, which is how hwmon patches are done - but not other subsystems.
What Stephen is asking, is that you run
$ git log --oneline -3 -- drivers/clk/tegra/clk-tegra30.c to figure out what the subject should be, based on previous subjects. That's good advice to follow for any patch you send :)
This other comment was:
Quoted 1 line
Don't think we need two patches for essentially the same thing.
I assume you sent two different patches for the same driver, or the same directory of drivers?
He's just asking that you squash the two commits together into one commit.
From your other email:
Quoted 2 lines
Sorry, what do you mean same subject line comment? I add a [PATCH v2] tag when I use 'git format-patch -v 2'.
I assume Stephen meant: the same comments about "(clk-tegra30)" from this patch apply to that patch too. There's nothing wrong with having "[PATCH v2]".
Quoted 2 lines
You mean the two commit has same subject content?
Sorry, I am a beginner to submit patch commit.
Ohh don't worry, we have all been there (and in my case not too long ago either...)
Hope that helps!
Thanks,
Conor.
Quoted 47 lines · At 2022-06-16 16:42:26, [email protected] wrote:
Quoted 22 lines · Liang He
Quoted 21 lines · Stephen Boyd
The subject should match historical subjects
$ git log --oneline -3 -- drivers/clk/tegra/clk-tegra30.cshows mostly "clk: tegra: ". Can you also combine this with the other tegra patch? Don't think we need two patches for essentially the same thing.
Quoting Liang He (2022-06-15 20:36:22)
Quoted 9 lines
In
tegra30_clock_init,of_find_matching_node()will return a node pointer with refcount incremented. We should useof_node_put()when the node pointer is not used anymore.Signed-off-by Liang He <[email protected]>---
changelog:v2: use real name for Sob
Thanks!
Quoted 7 lines
Quoted 1 line
Quoted 1 line
v1: fix missing bug
Sorry for my fault. I have been advised to use real name and I resend a PATCH v2 with my real name, but really with the same patch code.
So how can I withdraw the first patch or resend other thing? I am confused.
Can you help me, Conor?
Yeah, sure. I think you're just getting a bit confused by conflicting responses from different people. Some of the things I said on whatever the original patch I replied to apply everywhere - like using your real name or adding changelogs.
However, different subsystem maintainers have a different opinions about how patches for their subsystem should look. I would imagine that it was Guenter Roeck that asked you to use the "subsystem: (driver) action" subject, which is how hwmon patches are done - but not other subsystems.What Stephen is asking, is that you run
$ git log --oneline -3 -- drivers/clk/tegra/clk-tegra30.c to figure out what the subject should be, based on previous subjects. That's good advice to follow for any patch you send :)
Quoted 8 lines
This other comment was:
Quoted 1 line
Don't think we need two patches for essentially the same thing.
I assume you sent two different patches for the same driver, or the same directory of drivers?
He's just asking that you squash the two commits together into one commit.
Ok, thanks, Conor, I got it and I will prepare my next merged patch for these files which are indeed in the same driver directory.
Quoted 11 lines
From your other email:
Quoted 2 lines
Sorry, what do you mean same subject line comment? I add a [PATCH v2] tag when I use 'git format-patch -v 2'.
I assume Stephen meant: the same comments about "(clk-tegra30)" from this patch apply to that patch too. There's nothing wrong with having "[PATCH v2]".
Quoted 2 lines
You mean the two commit has same subject content?
Sorry, I am a beginner to submit patch commit.
Quoted 8 lines
Ohh don't worry, we have all been there (and in my case not too long ago either...)
Hope that helps!
Thanks,
Conor.
Thanks again, Conor.
Liang