narmstrong changed the topic of #linux-amlogic to: Amlogic mainline kernel development discussion - our wiki http://linux-meson.com/ - ml linux-amlogic@lists.infradead.org - official channel moved from Freenode - publicly logged on https://libera.irclog.whitequark.org/linux-amlogic
vagrantc has quit [Quit: leaving]
vagrantc has joined #linux-amlogic
chewitt_ has quit [Read error: Connection reset by peer]
chewitt has joined #linux-amlogic
chewitt has quit [Read error: Connection reset by peer]
chewitt_ has joined #linux-amlogic
chewitt_ is now known as chewitt
GNUtoo has quit [Remote host closed the connection]
GNUtoo has joined #linux-amlogic
buzzmarshall has quit [Quit: Konversation terminated!]
vagrantc has quit [Ping timeout: 240 seconds]
chewitt has quit [Quit: Zzz..]
rektide has quit [Ping timeout: 250 seconds]
rektide has joined #linux-amlogic
<narmstrong> gbisson: indeed we revert this patch in our yukawa u-boot https://gitlab.com/baylibre/amlogic/atv/u-boot/-/commits/mkorpershoek/u-boot/v2021.10/integ/
<gbisson> narmstrong: arf :( thanks for the pointer to the yukawa repo
<gbisson> narmstrong: something is fishy in dwc2_udc_otg.c then, I know NXP made this change along with the previous one: 723fd5668ff usb: gaget: ci: set ep's desc when enable ep
<gbisson> but this seems to be taken care of in dwc2 driver already, so either the data we get is invalid or we truly need to revert the patch upstream
<gbisson> and that explains why my fastboot test worked on the other platform (I tested on NXP i.MX)
<narmstrong> It’s weird we are the only affected platform since dwc2 is used on like rpi
<gbisson> Oh I actually might have an rpi to try on, need to find it
<gbisson> TBH I don't believe many upstream user use fastboot
<gbisson> is fastboot even enabled on rpi config?
<narmstrong> gbisson: probably not indeed
<narmstrong> gbisson: does this fix ? https://termbin.com/x5yd
<narmstrong> it may explain why you get an abort on libretech-cc since the first 16MiB of physical memory is reserved, unlike new SoCs
<narmstrong> so if desc NULL pointer is used, it would cause an abort
<narmstrong> why desc is null ? maybe becaus ep_enable is not called
<gbisson> ok, I'll try that after lunch if that's ok
<gbisson> but I'd find it weird that ep_enable is not called tho
<gbisson> because it sets many things, not just the descriptor
<gbisson> + in my case I was seeing the maxp size as 34832 just like you, wouldn't it be a random number if we were accessing random memory?
<gbisson> huh I take it back, you're correct it's null
<gbisson> so you're patch works
<gbisson> I'll check on lepotato
<gbisson> ok so it doesn't fix lepotato, it doesn't even reach that code, it aborts before that
<gbisson> nvm it also fixes it for lepotato ;)
<gbisson> I had a printf that was still printing the usb_endpoint_maxp() output, hence the abort again
<gbisson> not sure if you plan on pushing that patch or if we'd rather look at why enable_ep isn't called
<gbisson> but if you commit it, cc me I'll add a Tested-by ;)
<gbisson> + let me know if you want me to add support for Fastboot for lepotato (but seems like no one needs it)
<gbisson> ok so it looks like ep_enable is called, so it might be on the wrong ep
<gbisson> I need to go back to my other task
camus1 has joined #linux-amlogic
camus has quit [Ping timeout: 256 seconds]
camus1 is now known as camus
<gbisson> so final thing on that for today, it is called on the proper ep, but *someone* resets it
<gbisson> (didn't know termbin, it's great)
<gbisson> ok found it, I'll submit a patch upstream in a few
<gbisson> we actually need the same patch as the chipidea one, I thought it was done but no (ep vs. _ep)
<gbisson> narmstrong: patch sent, I cc'd you, thanks for the support ;)
camus has quit [Ping timeout: 240 seconds]
camus has joined #linux-amlogic
<narmstrong> gbisson: interesting, thanks for debugging !!
<gbisson> narmstrong: regarding your off-tree changes, I believe you can get rid of the MMC_BOOT_OFFSET one
<gbisson> narmstrong: I'm using this variable "fastboot_raw_partition_bootloader=0x1 0xfff mmcpart 1" which allows to use "bootloader" as a standard part name and it flashes boot1 at offset 1
<narmstrong> gbisson: oh interesting, I'll transfer this to the guys in charge
Guest43 has joined #linux-amlogic
Guest43 has quit [Client Quit]
glaroque has joined #linux-amlogic
mkorpershoek has joined #linux-amlogic
jernej has joined #linux-amlogic
chewitt has joined #linux-amlogic
mkorpershoek has quit [Quit: Client closed]
Stricted- has quit [Ping timeout: 250 seconds]
chewitt has quit [Read error: Connection reset by peer]
chewitt_ has joined #linux-amlogic
Stricted has joined #linux-amlogic
chewitt_ is now known as chewitt
chewitt has quit [Quit: Zzz..]
chewitt has joined #linux-amlogic
jernej has quit [Ping timeout: 256 seconds]
vagrantc has joined #linux-amlogic