Skip to contentSkip to conversation

CONFIG_SYS_HZ: Please revert breaking net commit

U-Boot Detlef Vollmann 2 participants 2 messages Feb 24, 2009

lore
Share & tools
lore
#0 lore raw

Hello,

From CHANGELOG:
<quote>
commit 49f3bdbba8071f56d950a9498b6cdb998b35340a

Author: Bartlomiej Sieka <[email protected]>
Date:   Wed Oct 1 15:26:28 2008 +0200

net: express the first argument to NetSetTimeout() in milliseconds

Enforce millisecond semantics of the first argument to
NetSetTimeout() --
the change is transparent for well-behaving boards (CFG_HZ == 1000 and get_timer() countiing in milliseconds).

Rationale for this patch is to enable millisecond granularity for network-related timeouts, which is needed for the upcoming automatic software update feature.

Summary of changes:
- do not scale the first argument to NetSetTimeout() by CFG_HZ
- change timeout values used in the networking code to milliseconds

</quote>

This commit breaks *silently* all boards that don't return milliseconds from get_timer, i.e. have a different value for CONFIG_SYS_HZ than 1000.
I'm open to any discussion about removing CONFIG_SYS_HZ, but if it is removed, it should be removed explicitely, and not just silently ignored.

So I see following reasons to revert the above commit:
- silently breaks existing boards
- silently ignores a still valid CONFIG setting
- bad programming style anyway: even if CONFIG_SYS_HZ would go away, there should be some internal define that should be used in code
- the intended benefit 'millisecond granularity for network-related timeouts' can be achieved while still keeping CONFIG_SYS_HZ

Best Regards,
Detlef

Signature · 2 lines
Detlef Vollmann   vollmann engineering gmbh
Linux and C++ for Embedded Systems    http://www.vollmann.ch/
#1 lore raw

Dear Detlef Vollmann,

Quoted 2 lines · In message <[email protected]> you wrote:

commit 49f3bdbba8071f56d950a9498b6cdb998b35340a

...

Quoted 3 lines

This commit breaks *silently* all boards that don't return milliseconds from get_timer, i.e. have a different value for CONFIG_SYS_HZ than 1000.

No, this is not true. All these boards ARE already broken because they use anincorrect timer setup. The patch just makes the problem visible.

Quoted 3 lines

I'm open to any discussion about removing CONFIG_SYS_HZ, but if it is removed, it should be removed explicitely, and not just silently ignored.

There is no need in the code to use CONFIG_SYS_HZ, because the code just used get_timer() as specified, i. e. it expects correct behaviour. If your implementation is broken, you should fix your implementation.

BTW: how comes it takes you nearly half a year to detect this?

Quoted 2 lines

So I see following reasons to revert the above commit:
- silently breaks existing boards

These boards were already broken before.

Quoted 1 line

- silently ignores a still valid CONFIG setting

It does not ignore it. It does not need it, because it is specified that get_timer() uses milisecond resolution.

Quoted 3 lines

- bad programming style anyway: even if CONFIG_SYS_HZ would go away, there should be some internal define that should be used in code

I don't understand what you mean.

Quoted 2 lines

- the intended benefit 'millisecond granularity for network-related timeouts' can be achieved while still keeping CONFIG_SYS_HZ

Instead of complaining about a patch that was discussed and cehcekd in many months ago you could rather spend your time on fixing your broken boards.

Best regards,

Wolfgang Denk

Signature · 4 lines
DENX Software Engineering GmbH,     MD: Wolfgang Denk & Detlev Zundel
HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany
Phone: (+49)-8142-66989-10 Fax: (+49)-8142-66989-80 Email: wd at denx.de
This is now.  Later is later.