Re: [U-Boot] [PATCH] IXP425: Fixing timer code Part 1/1

2009-01-03 Thread Jean-Christophe PLAGNIOL-VILLARD
On 20:08 Sun 21 Dec , Stefan Althoefer wrote: > This fixes bugs in non-interupt timer code for IXP425 > > Fixed udelay() implementation to be more accurate. > The timing error of udelay() was caused by > too much looping overhead. The actual timing routine > was called repeatedly for each micr

Re: [U-Boot] [PATCH] IXP425: Fixing timer code Part 1/1

2008-12-21 Thread Stefan Althoefer
This fixes bugs in non-interupt timer code for IXP425 Fixed udelay() implementation to be more accurate. The timing error of udelay() was caused by too much looping overhead. The actual timing routine was called repeatedly for each microsecond. When bigger slices are used, it becomes more accurate

Re: [U-Boot] [PATCH] IXP425: Fixing timer code Part 1/1

2008-12-16 Thread Stefan Althoefer
Wolfgang Denk schrieb: > Dear Stefan Althoefer, > > In message <49384728.lmggwd1otzmugoap%stefan.althoe...@web.de> you wrote: >> With this patch, this is nonsense. Instead you should use: >> >> #ifdef CONFIG_USE_IRQ >> /* Interrupt driven timer wants system tick here */ >> #define CONFIG_SYS_HZ

Re: [U-Boot] [PATCH] IXP425: Fixing timer code Part 1/1

2008-12-15 Thread Wolfgang Denk
Dear Stefan Althoefer, In message <49384728.lmggwd1otzmugoap%stefan.althoe...@web.de> you wrote: > > With this patch, this is nonsense. Instead you should use: > > #ifdef CONFIG_USE_IRQ > /* Interrupt driven timer wants system tick here */ > #define CONFIG_SYS_HZ 1000 > #else >

[U-Boot] [PATCH] IXP425: Fixing timer code Part 1/1

2008-12-04 Thread Stefan Althoefer
[PATCH] IXP425: Fixing timer code The non-interrupted timer code is inaccurate. I found that the timing error of udelay() is caused by to much looping overhead. The actual timing routine was called repeatedly for each microsecond. When I used bigger slices it became more accurate. Some IXP425 po