xbow.c kmalloc fixes
So what should be done about these SN2 cleanups? I don't want to accept dozens of small "fixes" (which really are mostly cleanups) as that would defeat the idea of the code-fixes-only decree. On the other hand, I really would like to see the SN2 code cleaned up so we actually have a sane base to do bug fixes on top of.
Perhaps it would actually be a better idea to have one big SN2 cleanup patch which gets applied directly by Andrew? It seems like he would be OK with a one-time exception (since the code had been in the works and since the old code is so gross) and as long as the patch doesn't break GENERIC compiles, I'm going to be OK with it anyhow. How about it?
--david
Quoted 5 lines · David Mosberger
So what should be done about these SN2 cleanups? I don't want to accept dozens of small "fixes" (which really are mostly cleanups) as that would defeat the idea of the code-fixes-only decree. On the other hand, I really would like to see the SN2 code cleaned up so we actually have a sane base to do bug fixes on top of.
Unfortunately, many of the cleanups and fixes (of which there are many) are now intertwined.
Quoted 6 lines
Perhaps it would actually be a better idea to have one big SN2 cleanup patch which gets applied directly by Andrew? It seems like he would be OK with a one-time exception (since the code had been in the works and since the old code is so gross) and as long as the patch doesn't break GENERIC compiles, I'm going to be OK with it anyhow. How about it?
Sounds good to me. Pat, do you have one big patch you can send to Andrew that includes Jes' latest checkins? Let me know if you need help testing it.
Thanks,
Jesse
Quoted 5 lines · David Mosberger
So what should be done about these SN2 cleanups? I don't want to accept dozens of small "fixes" (which really are mostly cleanups) as that would defeat the idea of the code-fixes-only decree. On the other hand, I really would like to see the SN2 code cleaned up so we actually have a sane base to do bug fixes on top of.
What about letting the SGI folks creeate a sn-cleanups-2.6 BK tree and let them deal with Linus themselves with your below blanket approval?
Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line · On Fri, 17 Oct 2003 10:37:54 -0700, [email protected] (Jesse Barnes) said:
Jesse> On Fri, Oct 17, 2003 at 10:31:27AM -0700, David Mosberger
Quoted 5 lines · Jesse> wrote:
Quoted 5 lines
So what should be done about these SN2 cleanups? I don't want to accept dozens of small "fixes" (which really are mostly cleanups) as that would defeat the idea of the code-fixes-only decree. On the other hand, I really would like to see the SN2 code cleaned up so we actually have a sane base to do bug fixes on top of.
Jesse> Unfortunately, many of the cleanups and fixes (of which there Jesse> are many) are now intertwined.
Yes, I understand and I don't think it makes sense to seprate them out, since the cleanups are so badly needed, they should count as fixes...
Quoted 6 lines
Quoted 6 lines
Perhaps it would actually be a better idea to have one big SN2 cleanup patch which gets applied directly by Andrew? It seems like he would be OK with a one-time exception (since the code had been in the works and since the old code is so gross) and as long as the patch doesn't break GENERIC compiles, I'm going to be OK with it anyhow. How about it?
Jesse> Sounds good to me. Pat, do you have one big patch you can Jesse> send to Andrew that includes Jes' latest checkins? Let me Jesse> know if you need help testing it.
Great!
--david
Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line · Christoph Hellwig
Christoph> On Fri, Oct 17, 2003 at 10:31:27AM -0700, David Mosberger
Quoted 5 lines · Christoph> wrote:
Quoted 5 lines
So what should be done about these SN2 cleanups? I don't want to accept dozens of small "fixes" (which really are mostly cleanups) as that would defeat the idea of the code-fixes-only decree. On the other hand, I really would like to see the SN2 code cleaned up so we actually have a sane base to do bug fixes on top of.
Christoph> What about letting the SGI folks creeate a Christoph> sn-cleanups-2.6 BK tree and let them deal with Linus Christoph> themselves with your below blanket approval?
That's what I was suggesting (apart from patch vs bk; I don't care about the mechanism of getting the updates into Linus tree; I just don't want to be stuck with a large SN2 update which may or may not be accepted by Linus/Andrew).
--david
Quoted 14 lines · David Mosberger
Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line · Christoph Hellwig
Christoph> On Fri, Oct 17, 2003 at 10:31:27AM -0700, David Mosberger
Quoted 5 lines · Christoph> wrote:
Quoted 5 lines
So what should be done about these SN2 cleanups? I don't want to accept dozens of small "fixes" (which really are mostly cleanups) as that would defeat the idea of the code-fixes-only decree. On the other hand, I really would like to see the SN2 code cleaned up so we actually have a sane base to do bug fixes on top of.
Christoph> What about letting the SGI folks creeate a Christoph> sn-cleanups-2.6 BK tree and let them deal with Linus Christoph> themselves with your below blanket approval?
Quoted 6 lines
That's what I was suggesting (apart from patch vs bk; I don't care about the mechanism of getting the updates into Linus tree; I just don't want to be stuck with a large SN2 update which may or may not be accepted by Linus/Andrew).
--david
Hi David,
On top of cleanup, we will also have new IO hardware support for 2.6 that we are doing on 2.4 and shipping with 2.4 kernels. I do like the idea of a single mod for 2.6 that covers:
1. Christoph's various patches.
2. Our very own cleanup effort that has been in progress.
3. New SGI IO Hardware support.It does not matter to me if you want us to wait until this code-fixes-only period is over. We will be doing all our regression on 2.4 and when we merge over to 2.6, we will also do our regression tests there too. None of these changes will have any effect on any other platform except to ensure that Generic builds and boots on other Platforms(including SGI).
We are currently adjusting our development effort to better update and refresh the latest levels of Linux.
Thanks.
colin
Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line
"Colin" = Colin Ngam <[email protected]> writes:
Colin> It does not matter to me if you want us to wait until this Colin> code-fixes-only period is over. We will be doing all our Colin> regression on 2.4 and when we merge over to 2.6, we will also Colin> do our regression tests there too. None of these changes will Colin> have any effect on any other platform except to ensure that Colin> Generic builds and boots on other Platforms(including SGI).
Hi Colin,
I think we need to get moving on the 2.6 push asap, the 2.6 tree is closing soon so we need to get it in as soon as possible ot we risk missing the train.
I don't think the new hardware supports is as big a problem to put in later since it's always legal to push new 'drivers' to a stable kernel.
Cheers,
Jes
Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line
"David" = David Mosberger <[email protected]> writes:
David> So what should be done about these SN2 cleanups? I don't want David> to accept dozens of small "fixes" (which really are mostly David> cleanups) as that would defeat the idea of the code-fixes-only David> decree. On the other hand, I really would like to see the SN2 David> code cleaned up so we actually have a sane base to do bug fixes David> on top of.
Hi David,
I am in two minds over this, I agree that getting the sn2 code cleaned up is a big win and a lot of the changes like those I posted are clearly bug fixes (not checking kmalloc returns should always be fixed). However having a big patch will take time and means we will have to operate with an internal tree until we get the whole thing into shape to push, but the seperate tree is one of the things we have been trying to avoid with 2.6. What came first, the chicken or the soft boiled egg ;-(
Cheers,
Jes
Quoted 3 lines
Quoted 3 lines
Quoted 3 lines
Quoted 3 lines
Quoted 3 lines · Jes Sorensen
"David" = David Mosberger <[email protected]> writes:
David> So what should be done about these SN2 cleanups? I don't David> want to accept dozens of small "fixes" (which really are David> mostly cleanups) as that would defeat the idea of the David> code-fixes-only decree. On the other hand, I really would David> like to see the SN2 code cleaned up so we actually have a David> sane base to do bug fixes on top of.
Jes> I am in two minds over this, I agree that getting the sn2 code Jes> cleaned up is a big win and a lot of the changes like those I Jes> posted are clearly bug fixes (not checking kmalloc returns Jes> should always be fixed). However having a big patch will take Jes> time and means we will have to operate with an internal tree Jes> until we get the whole thing into shape to push, but the Jes> seperate tree is one of the things we have been trying to avoid Jes> with 2.6. What came first, the chicken or the soft boiled egg Jes> ;-(
The issue may be moot already. Now that test8 is out, I'm not sure Andrew/Linus would still accept large at all patches anymore.
--david
Quoted 6 lines · David Mosberger
The issue may be moot already. Now that test8 is out, I'm not sure Andrew/Linus would still accept large at all patches anymore.
--david
Hi David,
When is the next opportunity we can update the code base at sn directory?
Thanks.
colin
Quoted 5 lines
-
To unsubscribe from this list: send the line "unsubscribe linux-ia64" in the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html
Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line · Colin Ngam
Colin> When is the next opportunity we can update the code base at sn Colin> directory?
Good question. I sounds like the rules may be relaxed some, once 2.6.0 is out. I'd recommend to discuss this with Andrew to see what he thinks.
--david