i2c: improve s3c24x0 with High-speed and new SYS_I2C framework support
Get series with b4
b4 am [email protected]Versions newest first
- v1 Latest0 of 3 patches 0 repliesNew
Patches in v1 No applies recorded yet
None of the 3 patches the cover letter announces are in this archive.
Reviews trailers per patch
No Reviewed-by, Acked-by or Tested-by trailers for v1 yet.
Cover letter [U-Boot] [PATCH 0/3] i2c: improve s3c24x0 with High-speed and new SYS_I2C framework support
Hello Naveen,
Am 01.10.2013 08:19, schrieb Naveen Krishna Ch:
Quoted 49 lines
Hello Heiko,
Quoted 46 lines · Heiko Schocher
Hello Naveen,
Am 30.09.2013 12:03, schrieb Naveen Krishna Ch:
Quoted 42 lines
Helo Heiko,
Thanks for timely reply.
Quoted 37 lines · Heiko Schocher
Hello Naveen,
Am 30.09.2013 08:58, schrieb Naveen Krishna Chatradhi:
Quoted 32 lines
This patchset fixes few bugs in the existing
s3c24x0.ccode (standard i2c)
and add support for High-speed i2c bus controller available on Exynos5 SoCs
from Samsung.Exynos5250 channels [0 ~ 3] can configured for I2C contoller or
new HSI2C controller
channels [3 ~ 7] are standard I2C controller channels
Exynos5420 channels [0 ~ 3] are standard I2C controller channels and
channels [4 ~ 10] are High-speed controller channelsPatchset:
1. exynos: i2c: Fix i2c driver to handle NACKs properly Improvements and fixes from Vadim Bendebury for standard i2c calls2. exynos: i2c: Change FDT bus setup code to enumerate ports correctly FDT bus setup code from Simon Glass
3. PATCH v5: i2c: s3c24xx: add hsi2c controller support High-speed controller register description and defines new i2c
read/write,
probe andset_buscalls.This is been reviewed earlier at
http://lists.denx.de/pipermail/u-boot/2013-May/153245.html
Thanks for review and improvements from Vadim Bendebury.Question:
4. RFC: samsung: i2c: Enable newCONFIG_SYS_I2Cframework I've tried to implement the new I2C multi bus framework in u-boot tree, takingtegra-i2c.cas reference.
[...]
Quoted 17 lines
Quoted 9 lines
Quoted 7 lines
channel:
--0----1----2----3------4--------5---------6-------7--------8-------9-------10
controller: i2c, i2c, i2c, i2c, hsi2c, hsi2c, hsi2c, hsi2c, hsi2c, hsi2c, hsi2c.
But the "i2c bus" command is showing the bus number according to the "name"
string comparison. Which seems wrong. Isn't it ??Quoted 8 lines
Hmm.. you are right, seems that the compiler sorts them alphabetical ... So, two ways:
- use another name, saying first a two digit for the channel ?
- or, useCONFIG_SYS_NUM_I2C_BUSES,CONFIG_SYS_I2C_DIRECT_BUSandCONFIG_SYS_I2C_BUSESas described in the README (grep forCONFIG_SYS_I2C_BUSESininclude/configsand you will find some examples ...) and you can define, which adapter is on whichi2c_busnumber ...
Quoted 1 line
Will try and implement it this way.
Ok.
Quoted 62 lines
Quoted 32 lines
Quoted 22 lines
Quoted 11 lines
Quoted 6 lines
c). What's the alternative for the
board_i2c_init(),i2c_get_bus_num_fdt(),i2c_reset_port_fdt().
"Functions to get the I2C bus number and reset I2C bus using FDT
node"I think, these functions are still needed.
Hmm.. that needs a general discussion, how to use fdt and i2c I think.
Quoted 7 lines
I would prefer a way (not really nowing, if it is possible for all configurations) where, if using fdt, the DT gets parsed and the availiable i2c buses gets created ... After that, "normal" i2c access with
i2c_set_bus_num(), i2c_read/write should be possible ... this is currently not possible with the i2c framework, but should not be so hard to do. Except the restriction, that it would not work in SPL case, or before relocation for i2c buses announced through dtonce
i2c_init_board()is doneboard_i2c_init()is not quite needed usingi2c_init_boardwe can avoid the relocation problem aswell.Quoted 8 lines
by the definition of
i2c_get_bus_num()indrivers/i2c/i2c_core.cunsigned int
i2c_get_bus_num(void){ return gd->cur_i2c_bus; }we don't need a special function
i2c_get_bus_num_fdt()Quoted 7 lines
Ah, ok!
Quoted 1 line
IMHO,
i2c_reset_port_fdt()is the only function to be discussedAnd why is
i2c_init()not good? Why must we have here a new function?That's right, even if there is a need for
i2c_reset_port_fdt(). it must be a i2c-core function instead of being in a driver.Quoted 21 lines
Quoted 13 lines
Quoted 12 lines
Define
CONFIG_SYS_I2C_MAX_HOPS->CONFIG_SYS_I2C_DIRECT_BUSis not
defined
so i2c_bus[] is used indrivers/i2c/i2c_core.c. Define the fix i2c buses in
the board config file withCONFIG_SYS_I2C_BUSES(if you have no fix buses,
let this empty) and add a function indrivers/i2c/i2c_core.c, which adds new i2c buses to i2c_bus[] after the fix buses and call this new function,
from where you interpret the fdt ...Quoted 2 lines
I din't quite understood this.
Can you point me to some readme or Doc or discussion Please.just the U-Boot README ... The above was just a fast idea, how it is possible to add i2c buses from the info in the fdt ...
Here some theoretical code, how it could look like:
[...]
Quoted 1 line
Thanks for this explanation.
! It is just theoretical ... you must try it ;-)
Quoted 6 lines
Quoted 6 lines
And maybe here and there some adaptions for getting this running... Thinking of
i2c_set_bus_num(), there must be now a check fori2c_fix_busI think ...Adapt the README ...
Quoted 8 lines
Quoted 4 lines
And then, from the place where you interpret the fdt, call if you have the information for one i2c bus,
i2c_add_one_bus()...Not sure, if I overlooked here something ...
Will try and do this.
Mean while can we get the other 3 patches reviewed. Patch 1: http://patchwork.ozlabs.org/patch/278933/
This patch seems good to me, added to u-boot-i2c.git as it seems it is a bugfix.
Quoted 2 lines
Patch 2: http://patchwork.ozlabs.org/patch/278931/
Patch 3: http://patchwork.ozlabs.org/patch/278930/
Think, they are OK, but they go after the next release to mainline, as we are close to the next release ...
Quoted 1 line
They were tested without the change to the new i2c framework.
Thanks for your work!
bye,
Heiko
Signature · 2 lines
DENX Software Engineering GmbH, MD: Wolfgang Denk & Detlev Zundel HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany