As usual before a release I spent a day trying to clean out all their
warnings. A lot of them were clearly false positives, a lot of them were
somewhat questionable but I could see at least the theoretical point and
added some more checks. And a few real bugs were found.
I ignored this last time
From: Linus Torvalds
Date: Sat, 3 Oct 2015 10:29:40 -0400
Subject: [PATCH] Clean up Divesoft Freedome time parsing
So Anton Lundin says that the 32-bit timestamp for the Divesoft Freedom
is indeed a signed offset from Jan 1, 2000.
This does that, but also
On Oct 3, 2015 11:06, "Anton Lundin" wrote:
>
> I looked at the wrong typedef and its a uint32_t. Sorry for the mess.
Christ. Oh well, fixing up the comment and removing the cast to (signed
char) would just fix that.
Dirk, did you already apply it? Just edit the patch if not.
On Sat, Oct 03, 2015 at 09:38:10AM +0200, Anton Lundin wrote:
> On 02 October, 2015 - Linus Torvalds wrote:
>
> > Commit 31fb2e4c62ab ("Avoid possible sign extension") handled the
> > problem when a "unsigned char" is shifted 24 bits left, and becomes a
> > "signed int". By casting the result to
On Oct 3, 2015 7:03 AM, "Dirk Hohndel" wrote:
>
> Which then means that the old code was actually correct and the fix and
> the fix of the fix were actually wrong?
Well, the final add (to convert from 2000-based numbers to 1970-based ones)
should still be done in timestamp_t.
On Oct 3, 2015 11:38 AM, "Anton Lundin" wrote:
>
> It ain't that easy trying to read code and write email while being a
> human climbing tree and trying to keep the little one from eating all
> every cable.
Well, it's unlikely to be a problem. At least for a while. If it's
On 03 October, 2015 - Anton Lundin wrote:
> On 02 October, 2015 - Linus Torvalds wrote:
>
> > Commit 31fb2e4c62ab ("Avoid possible sign extension") handled the
> > problem when a "unsigned char" is shifted 24 bits left, and becomes a
> > "signed int". By casting the result to uint32_t, that
On Sat, Oct 03, 2015 at 10:36:11AM -0400, Linus Torvalds wrote:
>
> This is a bigger patch than required (it could have just changed the
> "uint32_t" into "int32_t"), but I think the code ends up being easier to
> read, and it should be as efficient (although because we compile with
>
On 03 October, 2015 - Linus Torvalds wrote:
> On Oct 3, 2015 11:06, "Anton Lundin" wrote:
> >
> > I looked at the wrong typedef and its a uint32_t. Sorry for the mess.
>
> Christ.
Don't involve your imaginary friends in this =)
It ain't that easy trying to read code and
On Saturday 03 October 2015 09:38:10 Anton Lundin wrote:
> The root bug was mine. Another one of C's wonderful things that i didn't
> know about.
Yeah...
unsigned short us = 0x;
unsigned u = 0x;
us * us;// this is undefined behaviour
u *
On 03 October, 2015 - Linus Torvalds wrote:
>
> From: Linus Torvalds
> Date: Sat, 3 Oct 2015 10:29:40 -0400
> Subject: [PATCH] Clean up Divesoft Freedome time parsing
>
> So Anton Lundin says that the 32-bit timestamp for the Divesoft Freedom
> is indeed a signed
This reverts commit 3d8e5b638ad4c1fbb43f6dd5f535bf0b33a51f0b.
Calculating the next gradient should be based on the tissue loading at the end
of the previous iteration, so it was wrong to restore the deco state first.
This has a tiny affect on the calculated profile, and makes one of the tests
Reverting commit 3d8e5b638ad4c1fbb43f6dd5f535bf0b33a51f0b makes the CVA
calculation marginally less conservative, and one of the tests fails as a
result. This tiny adjustment to the conservatism fixes that.
Signed-off-by: Rick Walsh
---
deco.c | 4 ++--
1 file changed, 2
On 02 October, 2015 - Linus Torvalds wrote:
> Commit 31fb2e4c62ab ("Avoid possible sign extension") handled the
> problem when a "unsigned char" is shifted 24 bits left, and becomes a
> "signed int". By casting the result to uint32_t, that signed case won't
> happen.
>
The root bug was mine.
14 matches
Mail list logo