Patchwork [U-Boot,v2] arm: fix a comment

login
register
mail settings
Submitter Masahiro Yamada
Date May 10, 2013, 4:24 a.m.
Message ID <1368159848-3964-1-git-send-email-yamada.m@jp.panasonic.com>
Download mbox | patch
Permalink /patch/242904/
State Superseded
Delegated to: Albert ARIBAUD
Headers show

Comments

Masahiro Yamada - May 10, 2013, 4:24 a.m.
Signed-off-by: Masahiro Yamada <yamada.m@jp.panasonic.com>
---

Changed for v2:
   - Fix one more comment

 arch/arm/lib/crt0.S |    4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)
Masahiro Yamada - May 14, 2013, 1:43 a.m.
> Changed for v2:
>    - Fix one more comment

I added one more comment line fixed at version 2.

So, I am thinking the commit log should become
"arm: fix comments"
rather than
"arm: fix a comment"

Is it better to re-post my patch as version 3
with only commit log changed?
Am I too worried about it??

I'm new here, and I'm not sure this kind of update is mandatory.
I will appreciate any advice.


Best Regards
Masahiro Yamada
Albert ARIBAUD - May 14, 2013, 6:20 a.m.
Hi Masahiro,

On Tue, 14 May 2013 10:43:07 +0900, Masahiro Yamada
<yamada.m@jp.panasonic.com> wrote:

> 
> > Changed for v2:
> >    - Fix one more comment
> 
> I added one more comment line fixed at version 2.
> 
> So, I am thinking the commit log should become
> "arm: fix comments"
> rather than
> "arm: fix a comment"
> 
> Is it better to re-post my patch as version 3
> with only commit log changed?
> Am I too worried about it??

You should try to read your commit message summary to yourself as if
you did not know what the change is about, because that's what it is
useful for: giving people a good idea of the change. If the summary
tells everything except the actual source code change, that's good.

A summary of "arm: fix comments" indicates two things about the
change: explicitly, that it is about ARM; and implicitly, that it is
cosmetic. But it does not say which part of U-Boot it applies to, even
though this single file can be mentioned in a summary line. Readers
will have to waste some time looking at the change content to find out.

Plus, if it is cosmetic, then the fact that it is about ARM does not
matter that much, does it? e.g., someone looking for ARM changes that
might explain a weird behavior will most certainly not be interested in
this change even though it is ARM. For instance, a cosmetic fix on ARM
code could easily go in the mainline tree without going through the ARM
tree first. So, I'd put "cosmetic:" first, but admittedly, that's a
debatable point.

So... how about a V3 with a commit summary of "cosmetic: arm: fix
comments in arch/arm/lib/crt0.S" ? People reading the summary for no
particular reason will know all there is to know except the detailed
comment fixes; people looking for crt0 changes will see it, and so will
people looking for ARM changes, but "cosmetic:" will help them move it
aside in their search; "cosmetic:" helps deciding whether to apply the
patch and to which tree; etc.

> I'm new here, and I'm not sure this kind of update is mandatory.
> I will appreciate any advice.

No problem.

> Best Regards
> Masahiro Yamada

Amicalement,
Masahiro Yamada - May 15, 2013, 8:46 a.m.
Hello Albert,

Thanks for your advice.

> So... how about a V3 with a commit summary of "cosmetic: arm: fix
> comments in arch/arm/lib/crt0.S" ? People reading the summary for no

I posted a v3 patch as suggested.

Best Regard,
Masahiro Yamada

Patch

diff --git a/arch/arm/lib/crt0.S b/arch/arm/lib/crt0.S
index a9657d1..a5bffb8 100644
--- a/arch/arm/lib/crt0.S
+++ b/arch/arm/lib/crt0.S
@@ -97,13 +97,13 @@  ENTRY(_main)
  * 'here' but relocated.
  */
 
-	ldr	sp, [r8, #GD_START_ADDR_SP]	/* r8 = gd->start_addr_sp */
+	ldr	sp, [r8, #GD_START_ADDR_SP]	/* sp = gd->start_addr_sp */
 	bic	sp, sp, #7	/* 8-byte alignment for ABI compliance */
 	ldr	r8, [r8, #GD_BD]		/* r8 = gd->bd */
 	sub	r8, r8, #GD_SIZE		/* new GD is below bd */
 
 	adr	lr, here
-	ldr	r0, [r8, #GD_RELOC_OFF]		/* lr = gd->start_addr_sp */
+	ldr	r0, [r8, #GD_RELOC_OFF]		/* r0 = gd->reloc_off */
 	add	lr, lr, r0
 	ldr	r0, [r8, #GD_RELOCADDR]		/* r0 = gd->relocaddr */
 	b	relocate_code