Le jeudi 13 août 2015 à 08:13 -0400, Tom Rini a écrit :
On Fri, Aug 07, 2015 at 11:44:46AM +0200, Paul Kocialkowski wrote:
Le mardi 04 août 2015 à 14:27 -0400, Tom Rini a écrit :
On Tue, Aug 04, 2015 at 08:22:39PM +0200, Paul Kocialkowski wrote:
Le mardi 04 août 2015 à 14:16 -0400, Tom Rini a écrit :
On Tue, Aug 04, 2015 at 08:02:40PM +0200, Paul Kocialkowski wrote:
Le lundi 03 août 2015 à 22:08 -0400, Tom Rini a écrit : > On Mon, Jul 20, 2015 at 03:17:13PM +0200, Paul > Kocialkowski wrote: > > > The OMAP3 has some die-specific ID bits that we can use > > to give the device a > > (more or less) unique serial number. This is > > particularly useful for e.g. USB. > > > > Signed-off-by: Paul Kocialkowski contact@paulk.fr > > --- > > board/lge/sniper/sniper.c | 13 +++++++++++++ > > 1 file changed, 13 insertions(+) > > > > diff --git a/board/lge/sniper/sniper.c > > b/board/lge/sniper/sniper.c > > index 44d422d..f26855d 100644 > > --- a/board/lge/sniper/sniper.c > > +++ b/board/lge/sniper/sniper.c > > @@ -70,7 +70,9 @@ int board_init(void) > > > > int misc_init_r(void) > > { > > + char serial_string[17] = { 0 }; > > char reboot_mode[2] = { 0 }; > > + u32 dieid[4] = { 0 }; > > > > /* Reboot mode */ > > > > @@ -82,6 +84,17 @@ int misc_init_r(void) > > omap_reboot_mode_clear(); > > } > > > > + /* Serial number */ > > + > > + get_dieid((u32 *)&dieid); > > + > > + if (!getenv("serial#")) { > > + snprintf(serial_string, > > sizeof(serial_string), > > + "%08x%08x", dieid[0], > > dieid[3]); > > + > > + setenv("serial#", serial_string); > > + } > > + > > return 0; > > } > > Shouldn't this be in more generic code so everyone gets > this set now? > Thanks!
Well, we had a similar discussion for sunxi, and the outcome was that serial number could be obtained from other places on other devices (e.g. EEPROM) or be calculated from the dieid bits in a different way, so I prefer to keep this board-specific instead of omap3-generic for now.
This merely matches what is done on Android OMAP devices, but one could do it another way, too.
What do you think?
I think, ug, arch/arm/cpu/armv7/omap -common/utils.c::usb_set_serial_num_from_die_id() should be called set_serial_num_from_die_id() and we can use that for this board too even if it's not ideal.
Oh okay then, I don't have any problem with making this code common, especially if it's not called by every omap3 board then.
I agree with your proposal. Should I submit a v2 with a patch in that direction?
Sounds good, thanks!
Taking a closer look at things, it appears that various (non-omap3) boards are grabbing the Die ID bits on their own and then calling those functions (usb_fake_mac_from_die_id, usb_set_serial_num_from_die_id).
IMHO, we should have a common naming scheme for the function to get the dieid (omap_dieid), define that for each omap platform and have it called in omap-common code (with one function for the serial number and one for the fake mac), just like what I did with omap_sys_boot_device.
Then, each board would simply call those functions directly, without having to care about how to obtain the die id bits.
This seems like a series that would deserve to live on its own, so I suggest that you merge Optimus Black support as-is for now and I'll submit another series to implement that behaviour on top.
What do you think?
Yes, OK, follow-up series to clean-up that for everyone.
Good, thanks for merging this, I'll get around producing that series sometime next week, I'm at CCCamp for now!