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
Guest4 has joined #linux-amlogic
naoki has joined #linux-amlogic
naoki has quit [Client Quit]
elastic_dog has quit [Quit: elastic_dog]
elastic_dog has joined #linux-amlogic
BlueMatt has quit [Ping timeout: 240 seconds]
BlueMatt has joined #linux-amlogic
Guest4 has quit [Ping timeout: 245 seconds]
JohnnyonFlame has quit [Ping timeout: 240 seconds]
f_ has joined #linux-amlogic
<f_> \o hi all.
<f_> I'm thinking now......is the @AML header actually required only inside BL2?
<f_> By having a libre BL2, we actually (possibly) will have a lot of flexibility on FIP packaging.
<f_> For example: I was able to load BL2 without aligning the FIP to 0x4000.
Terry137322934 has quit [Quit: Bye Bye]
Terry137322934 has joined #linux-amlogic
gabes has quit [Quit: The Lounge - https://thelounge.chat]
gabes has joined #linux-amlogic
<narmstrong> f_: yes by using the standard tf-a we can have more flexibility and align like other vendors and be use standard tools
<minute> working a bit on dsi->edp display again today, trying different settings to work around the desync (?) glitch that happens ~once per minute
<f_> narmstrong: So the @AML header is actually only required by BL2, right?
<minute> also, one oddity: the last 3 lines of the display are missing. instead, the display is shifted down by 3 lines and the first 3 lines take the color of the first pixel in the last line
<minute> something wrong with vertical sync or number of lines perhaps
<narmstrong> f_: probably yes
<f_> So theoretically I can just get rid of aml_encrypt_gxb/amlbootsig, right?
<narmstrong> minute: interesting yes, the hsync/vsync may require some tweaks
<minute> narmstrong: any hunches regarding which parts to tweak?
<narmstrong> f_: you’ll still need to add the right headers to bl2
<f_> The @AML headers?
<narmstrong> Yep
<f_> But..if the bootrom doesn't require those, why even include them (in my opinion)
<minute> increasing the vsync_start in the mode definition makes the downshift / size of that bar bigger
<narmstrong> minute: it could be a vsync signals length issue, amlogic weirdly sets the vsync signal sent to the dsi transceiver
<minute> narmstrong: very interestng
<narmstrong> f_: if the boot rom ignores it then don’t add them!
<f_> Let's try!
<f_> Let's see if the boot ROM requires them.
<minute> looks like the vsync is off by 3 lines... if i set the vsync_start (and also move the end) to 1080+1 instead of 1080+4, it looks correct
<f_> Hah
<f_> The boot ROM does seem to require them.
<minute> looks... like the glitching is gone (don't want to jinx it)
<minute> ah no, now it happened again. it's just very rare now
<minute> maybe still related to nvme?
<minute> no, happens also with system on sd card and nvme not mounted.
<minute> narmstrong: is there anything else that could occassionaly starve the dsi pipeline? can we adjust fifo or axi hurry (is there a noc)?
<minute> i noticed that it might have to do with GPU load, as i get constant horizontal glichting in certain positions in minetest
<narmstrong> minute: I’m afk until Monday but you can try to reduce from 31 to 4 here https://github.com/torvalds/linux/blob/2d1bcbc6cd703e64caf8df314e3669b4786e008a/drivers/gpu/drm/meson/meson_viu.c#L444
<narmstrong> 31 is needed for hdmi with AFBC
<minute> narmstrong: thank you!!
<minute> narmstrong: have a nice weekend
<minute> narmstrong: omg this fixed the problem, you just saved me many hours of hair pulling :D
<f_> Ok so.
<f_> What we currently have is:
<f_> In the FIP header (not @AML) the offsets are all wrong.
<f_> This isn't a problem for Amlogic's BL2, which doesn't care about the FIP header and only cares about the order in which the images are.
<f_> But, for TF-A, which does the right thing by looking at the header and looking at specified offsets and all, it fails....because the offsets are all wrong!
<f_> So basically...Amlogic's BL2 and FIP packaging are flawed.
<f_> Very much so.
JohnnyonFlame has joined #linux-amlogic
<f_> So I'll (probably) modify amlbootsig to update offsets.
<f_> Note: still doesn't boot.
camus has quit [Ping timeout: 240 seconds]
<f_> Made a patch for amlbootsig to fixup ToC offsets.
buzzmarshall has joined #linux-amlogic
<f_> I documented steps to build a BL2.
<f_> It doesn't boot but at least it finds BL31 in the right offset.
<f_> So yes, amlbootsig requires a patch to correctly update ToC offsets.
<f_> aml_encrypt_gxb also requires a fix to that, but meh.
<f_> If anyone wants to fix this proprietary utility...(I certainly don't want to)
<f_> I guess.....chime in?
JohnnyonFlame has quit [Read error: Connection reset by peer]
JohnnyonFlame has joined #linux-amlogic
elastic_dog has quit [Read error: Connection reset by peer]
elastic_dog has joined #linux-amlogic
Danct12 has quit [Remote host closed the connection]
Danct12 has joined #linux-amlogic
f_ has quit [Quit: Lost terminal]
jacobk has joined #linux-amlogic
camus has joined #linux-amlogic
camus has quit [Ping timeout: 240 seconds]
jacobk has quit [Ping timeout: 250 seconds]
jacobk has joined #linux-amlogic
vagrantc has joined #linux-amlogic
Danct12 has quit [Remote host closed the connection]
Danct12 has joined #linux-amlogic
jacobk has quit [Ping timeout: 248 seconds]