lennyraposo has quit [Read error: Connection reset by peer]
lennyraposo has joined #linux-sunxi
lennyraposo has quit [Quit: Leaving.]
lennyraposo has joined #linux-sunxi
p1u3sch1_ has quit [Ping timeout: 240 seconds]
p1u3sch1 has joined #linux-sunxi
IgorPec has joined #linux-sunxi
TheSeven has quit [Ping timeout: 272 seconds]
TheSeven has joined #linux-sunxi
keh has joined #linux-sunxi
fredy has quit [Excess Flood]
fredy has joined #linux-sunxi
fire219 has quit [Read error: Connection reset by peer]
zuikis has joined #linux-sunxi
JohnDoe_71Rus has joined #linux-sunxi
IgorPec has quit [Ping timeout: 260 seconds]
jernej has quit [Ping timeout: 276 seconds]
Mr__Anderson has joined #linux-sunxi
apritzel has joined #linux-sunxi
kz1 has left #linux-sunxi [#linux-sunxi]
apritzel has quit [Ping timeout: 244 seconds]
<tuxillo>
moin
jstein has joined #linux-sunxi
fredy has quit [Excess Flood]
<lennyraposo>
hey ssvb
fredy has joined #linux-sunxi
<lennyraposo>
are you about for a little insight/input
<ssvb>
lennyraposo: hi
<lennyraposo>
I wanted to ask about that mali release
<lennyraposo>
more specifically the xf86 portion
<lennyraposo>
I have been trying to get it to work but just every configure option I throw at (./configure --with-drmmode=sunxi)
<lennyraposo>
and for some reason it is not complying
reinforce has joined #linux-sunxi
<lennyraposo>
got any tips?
<lennyraposo>
kernel modules portion is completed
kaspter has joined #linux-sunxi
<lennyraposo>
perhaps I am just not noticing what it is
<mripard>
lennyraposo: which version is it ?
<mripard>
1.4?
<lennyraposo>
let me see
<lennyraposo>
1.4
<MoeIcenowy>
mripard: your "cursor only" issue is found with xf86-video-armsoc?
<mripard>
MoeIcenowy: I haven't tested modesetting for quite some time
<mripard>
but yes
<mripard>
lennyraposo: you don't need that option anymore
<mripard>
it compiles the support for all the SoCs now
<MoeIcenowy>
mripard: this issue occur only on a33, but not on tradition a10/10s/13/20 + r8?
Mr__Anderson has quit [Remote host closed the connection]
<mripard>
MoeIcenowy: yep
Nyuutwo has quit [Read error: Connection reset by peer]
<MoeIcenowy>
mripard: oh terrible
<lennyraposo>
just add prefix for /usr and that should do it then
<MoeIcenowy>
so mysterious allwinner IPs are
Nyuutwo has joined #linux-sunxi
<mripard>
lennyraposo: yep
<lennyraposo>
going well now
<montjoie>
apritzel and for better parano, never compile as your usual user
<wens>
mripard: do you have your i2s board test dts around?
fdcx has joined #linux-sunxi
apritzel has joined #linux-sunxi
<lennyraposo>
thanks for the insight
<mripard>
wens: yes
<wens>
hmm, i got it working
<mripard>
branch sunxi/pen/i2s on my github
<wens>
though i'm getting pops during playback
IgorPec has joined #linux-sunxi
iamfrankenstein1 has joined #linux-sunxi
iamfrankenstein has quit [Ping timeout: 244 seconds]
iamfrankenstein1 is now known as iamfrankenstein
<wens>
codec over i2s sounds better than the internal codec on the bpi-m1+, even with the pops
<wens>
the noise floor for the internal codec is audible :(
apritzel has quit [Ping timeout: 244 seconds]
apritzel has joined #linux-sunxi
lemonzest has joined #linux-sunxi
matthias_bgg has joined #linux-sunxi
matthias_bgg has quit [Remote host closed the connection]
fdcx has quit [Remote host closed the connection]
enrico_ has joined #linux-sunxi
rtp_ is now known as rtp
afaerber has quit [Quit: Ex-Chat]
<KotCzarny>
megi: i love you (nohomo)
<KotCzarny>
for those onrage patches
<KotCzarny>
:)
<NiteHawk>
:D
<KotCzarny>
drat, he isnt there
<KotCzarny>
still, that would finally mean my mini opi chroot farm will start working as expected
reinforce has quit [Quit: Leaving.]
paulk-aldrin has joined #linux-sunxi
popolon has joined #linux-sunxi
<Wizzup_>
Does the A10/A20 BROM support reading the SD card in 1-bit SPI mode?
keh has quit [Ping timeout: 276 seconds]
zuikis has quit [Ping timeout: 260 seconds]
Uninstall_ has quit [Remote host closed the connection]
<ssvb>
Wizzup_: this is unlikely
<ssvb>
why are you asking?
cnxsoft has quit [Read error: Connection reset by peer]
cnxsoft has joined #linux-sunxi
<Wizzup_>
ssvb: I'm thinking of ways to have my uboot (plus optionally) kernel available on allwinner devices in a way that the code is not easy to change without physical access (and without physical access, very easy). Basically a simple trusted boot like environment
<Wizzup_>
With the spi nor flash this is doable, given that no sd card or nand/emmc is attached
<Wizzup_>
With sd cards, it is possible to force them in permanent read-only more, but an advanced attacker can still flash the firmware of such a microsd card
<Wizzup_>
So I was wondering if I could make the BROM use the spi-mode of microsd cards, and put some simple MCU between the real SD card and the BROM
<Wizzup_>
I'm going to upload the pictures + text of the spi nor flash after I finish work today, btw. :)
<Wizzup_>
ssvb: I am planning on documenting my thoughts on some sort of 'trusted boot' in more detail later, perhaps in a blog post or something, irc is not the best way to explain more complicated ideas
<Wizzup_>
The basic plan is to provide something that makes it almost impossible for attackers with local root, but no phsyical access, to persistently compromise a device
staplr has joined #linux-sunxi
ricardocrudo has joined #linux-sunxi
Mr__Anderson has joined #linux-sunxi
<kronicd>
trusted boot owns
<kronicd>
make it work please
cnxsoft has quit [Quit: cnxsoft]
kaspter has quit [Ping timeout: 260 seconds]
<apritzel>
Wizzup_: so why 1-bit SPI mode then? Because of "simple MCU"?
<apritzel>
Wizzup_: and that would force Linux accesses to use 1-bit mode as well?
<apritzel>
rendering it even slower than it is already?
<Wizzup_>
apritzel: we would only use the SD card for booting
<ssvb>
Wizzup_: I had some ideas about using the MMC1 controller for the removable SD card and using SPI flash as the boot media
<Wizzup_>
ssvb: The SPI flash as boot media is possible, but requires (quite some) hardware changes, as you've seen
<Wizzup_>
I will try to expand on my ideas later - got to go atm
p1u3sch1 has quit [Ping timeout: 244 seconds]
p1u3sch1 has joined #linux-sunxi
staplr has quit [Remote host closed the connection]
Uninstall has joined #linux-sunxi
<ricardocrudo>
I'm getting bug message from kernel (sunxi 3.4, A20) when connecting g_ether to a mac os. It's working perfectly when plugged to linux.
<apritzel>
Wizzup_: check the documentation about the "Secure Peripherals Controller" in the manual (for A64 that's chapter 3.17)
<Wizzup_>
apritzel: ok, I'm using the A10/A20, but I'll check it
<KotCzarny>
wizzup but: Every card has two programmable flags, one for temporary write protection and
<KotCzarny>
one to lock the card forever
<ssvb>
apritzel: if SMHC0 is switched to secure only, then it can't be used for anything practical from Linux, unless the firmware provides some interface to access it
<KotCzarny>
from the page you've pasted
<Wizzup_>
KotCzarny: not every card
<Wizzup_>
some implement it
<ssvb>
apritzel: at least this is how accessing the hard drive worked on Playstation3
<Wizzup_>
and as I said, people have flashed the firmware of sd cards, so that would go around the write protect completely
<KotCzarny>
but is a nice way to prepare such WORM firmwares
<apritzel>
ssvb: yes, and this is what Wizzup_ wants, right?
<Wizzup_>
just like people have run linux on harddisk controllers
<Wizzup_>
(or put persistent backdoors in there)
<ssvb>
apritzel: if I understand it correctly, Wizzup_ wants to securely boot the system, preferably without sacrificing any features
<KotCzarny>
wizzup, did that include dremelling sd cards case?
<Wizzup_>
Yeah, sorry - I will clarify my intentions more soon, but I'm doing irc in between meetings
<apritzel>
ssvb: but how does the 1-bit SPI mode over a MCU solve this?
<Wizzup_>
KotCzarny: not afaik
<Wizzup_>
KotCzarny: I think without physical access
<KotCzarny>
you could try write protecting one card and trying to hack it
<KotCzarny>
i mean, permanenlty wp
<Wizzup_>
apritzel: you can put a device that you fully control between the sd card and check the commands / writes
<Wizzup_>
KotCzarny: I am not as powerful as the nsa, but it's still a black box
<Wizzup_>
I will make my case more clearly later :D
<KotCzarny>
i think xD cards were exposing nand directly
<Wizzup_>
no, there's a lot of firmware
<Wizzup_>
they also do wear leveling, etc
Gerwin_J has quit [Quit: Gerwin_J]
<KotCzarny>
The fastest xD card offers less than 10% of the speed of current (2009) Compact Flash cards
<KotCzarny>
heh
<KotCzarny>
Directly derived from the Smart Media card. Thus, has no wear leveling controller
<KotCzarny>
unless wikipedia is outdated
<KotCzarny>
Comparing the pinout of an xD card[17][18] to the pinout of a NAND flash chip in a standard TSOP package,[17] one finds a nearly one-to-one correspondence between the active pins of the two devices
<KotCzarny>
so yeah, firmware is only in readers
<apritzel>
Wizzup_: I think I got this, I was just wondering how poor Linux' performance would be if the SD card is limited to 1-bit mode
<apritzel>
ssvb: ^^^ which goes against the idea of not sacrificing anything
* apritzel
wonders if EFI could provide a driver to the SD card, preventing any access to the firmware bits
<apritzel>
and Linux just uses the EFI driver (direct access to SMHC0 is forbidden because secure only)
<ssvb>
apritzel: it makes more sense for SMHC2 (eMMC), but I wonder how it would affect performance
<apritzel>
I guess the bottleneck here is the SD interface, not passing the request through firmware
<ssvb>
hdparm -t /dev/ps3da
<ssvb>
Timing buffered disk reads: 124 MB in 3.02 seconds = 41.09 MB/sec
<ssvb>
hmm, I wonder if it is good or bad, but the playstation3 firmware also supposedly does encryption on the fly
fire219 has joined #linux-sunxi
<KotCzarny>
you dont control the keys
<KotCzarny>
which is bad
<KotCzarny>
and probably done to lock in the hardware
<ssvb>
well, the point is that the access to actual hardware is not done directly, but via some interface provided by the firmware
<ssvb>
it can't speed up things for sure, but I wonder whether the performance loss can be reduced to the very minimum
<Wizzup_>
apritzel: I would have external storage next to the sd, and the sd would only be used to load u-boot plus possibly a kernel, so performance is not an issue
diego_r has joined #linux-sunxi
<Wizzup_>
n.b. the plan is not to (ever) lock users out, but rather provide trusted boot to users, where replacing the keys is as simple as pulling the WP pin on the flash & writing new keys, etc
<Wizzup_>
(requiring physical access)
<apritzel>
Wizzup_: so why not easily _not_ provide any storage on SMHC0 & 2?
<apritzel>
and then use SPI flash, either making this secure only as well or using some write protect pins
<apritzel>
Wizzup_: so use something h/w protected on SMHC2 then?
<KotCzarny>
wizzup, emulating sd by some other device?
<KotCzarny>
'root over spi' ?
<KotCzarny>
;)
<Wizzup_>
apritzel: spi flash has write protects pins for free
<Wizzup_>
apritzel: I am not sure if SMHC0 is avail on A10/A20
<apritzel>
that's the normal SD card interface, I think
<apritzel>
but I don't think A20 can't switch peripherals into secure
<apritzel>
*can*
<Wizzup_>
And you'd have to trust the firmware in the peripherals, no? I guess I'm ignorant of this implementation
<apritzel>
Wizzup_: I am talking entirely about SoC peripherals here
<apritzel>
If you control SMHC (which is the Allwinner manual name for the MMC controller), you can't bypass this
<Wizzup_>
Granted that it doesn't do anything silly, but I agree
<Wizzup_>
The nice thing about SPI flash is that it'd be completely separate and that there's a physical write protect pin
<apritzel>
sure, but it's lower priority than MMC, so doesn't help
<apritzel>
so either you just don't connect anything to MMC2 (usually eMMC) and MMC0 (usually SD)
<apritzel>
or you make one of those secure
<Wizzup_>
What do you mean lower priority?
<Wizzup_>
Yes, that is what I would do - do not connect to those busses
<Wizzup_>
Because you cannot have anything in there, or you'd be able to control the boot order
<Wizzup_>
In more recent SoCs I think you can also change the brom order
<apritzel>
lower boot priority, the SPI flash will not be used if there is something on MMC0 or MMC2
<Wizzup_>
I know
<Wizzup_>
so I'd for example use SPI flash + sata
<Wizzup_>
or usb
<apritzel>
yes
<apritzel>
but why were you asking about 1-bit SPI mode, then?
<Wizzup_>
apritzel: because SPI flash may require (a lot of) hardware mods
<Wizzup_>
whereas using the mmc0 interface but making it trustworthy may be easier
<Wizzup_>
trustworthy to some degree, of course
<apritzel>
so you are looking for a solution for existing A20 based boards, then?
<apritzel>
which don't expose SPI flash PCx on headers
<Wizzup_>
Olimex has open board designs, we could modify that
<Wizzup_>
especially if we can interest olimex for that
<apritzel>
now we just need a modern SoC with SATA ;-)
<Wizzup_>
is a20 not modern?
<apritzel>
no
<apritzel>
it's ancient: just two A7 cores, limited to at most 1GHz, memory to 400(?) MHz, 32-bit, ....
<Wizzup_>
It's good enough for me :_
<Wizzup_>
:)
<wens>
"modern" socs might have less support :p
<apritzel>
sure, it fits many purposes, but still
JohnDoe_71Rus has quit [Quit: Miranda IM! Smaller, Faster, Easier. http://miranda-im.org]
<apritzel>
wens: that's why we are here, right?
<fire219>
it's comparable to the SoC in 2011 flagship phones. i'd call that ancient for sure
xinj has joined #linux-sunxi
<maz>
fire219: for a small server, that's just great. and I haven't seen a phone with SATA and Gb Ethernet...
<fire219>
i'm sure it could be done, it just wouldn't sell well to anyone except nerds like us ;)
<maz>
fire219: let me rephrase that: I haven't seen an SoC fitting in a phone with...
<fire219>
if the A20 is enough for you, then good. but its a dinosaur of a chip and far too slow for a lot of tasks...
paulk-aldrin has quit [Quit: Leaving]
<KotCzarny>
apritzel, why sata? let it be pci-e
<KotCzarny>
more universal
<fire219>
a cheap SBC with an M.2 NVMe (PCIe) slot on it... i'd like to see that
<apritzel>
KotCzarny: whatever, but given that AW actually even _dropped_ SATA, I don't dare to dream of anything cheap with PCI
<maz>
fire219: between a dinosaur that provides a service and a useless SoC (or even a pipe dream, by the look of it), I've made my choice! ;-)
<fire219>
isn't the point of this community to make "useless" SoCs useful? ;)
<KotCzarny>
yes, but this community dreams!
<KotCzarny>
of more noble things
<maz>
fire219: definitely. but most people just talk, and only a few are actually doing any work, unfortunately.
paulk-aldrin has joined #linux-sunxi
<KotCzarny>
maz, userbase is also important
cnxsoft has joined #linux-sunxi
<Amit_T>
Could anyone please tell me how can I set IPv4 address from EFI shell ?
nove has joined #linux-sunxi
petr has quit [Remote host closed the connection]
petr has joined #linux-sunxi
lennyraposo has quit [Quit: Leaving.]
Keziolio_ has quit [Ping timeout: 260 seconds]
afaerber has joined #linux-sunxi
Keziolio has joined #linux-sunxi
cnxsoft has quit [Quit: cnxsoft]
IgorPec has quit [Ping timeout: 244 seconds]
lennyraposo has joined #linux-sunxi
iaglium has joined #linux-sunxi
IgorPec has joined #linux-sunxi
BenG83 has joined #linux-sunxi
ricardocrudo has quit [Remote host closed the connection]
gzamboni has quit [Ping timeout: 276 seconds]
xinj has quit [Ping timeout: 244 seconds]
gzamboni has joined #linux-sunxi
<MoeIcenowy>
wens: is a23/33 voltage/frequency adjusting now available in mainline kernel?
Nacho__ has quit [Ping timeout: 244 seconds]
phipli has joined #linux-sunxi
<phipli>
anyone had much luck with disputes on aliexpress?
<phipli>
seller is being a pain in the backside - I'm wondering what happens when it goes to arbitration tomorrow.
Nacho has joined #linux-sunxi
<phipli>
Got a "Beelink X2" H3 based TV media box. I think the emmc is non-functional. It wont boot the android it came with (straight out the box), wont re-flash (fails at 7%)... but boots from an SD card (not that I'm going to tell them that, because they'll say I broke it)
apritzel1 has joined #linux-sunxi
<MoeIcenowy>
phipli: while buying something for hacl
<MoeIcenowy>
hack
<MoeIcenowy>
it will be a good habit to check the functionalitiy with the stock rom at first
<phipli>
they've had me jumping through pointless fake troubleshooting for over a week "Dear, must plug in the ethernet you will" "Dear, power on while standing on a donkey?"
<phipli>
MoeIcenowy : that is what I did
<MoeIcenowy>
:-)
<MoeIcenowy>
Thus you should directly apply for Aliexpress service
<MoeIcenowy>
(I means apply for service with Aliexpress rather than contacting the seller
<phipli>
I've escalated the complaint
<phipli>
but it seems to have 3 or 4 days before it goes to them
<phipli>
was there another option I missed?
<MoeIcenowy>
I don't know about Aliexpree
<MoeIcenowy>
Aliexperss
<MoeIcenowy>
But for the Chinese variant Taobao
<phipli>
They've now both asked me to send it back (right around the world) to them, and to perform a firmware update, which either requires the device to be working, or you to have a USB A to USB A cable.
<phipli>
Joke's on them. I have one.
Nacho has quit [Quit: No Ping reply in 180 seconds.]
<MoeIcenowy>
phipli: In my memory the A to A cable should be a non-standard one?
<phipli>
yup
<MoeIcenowy>
Maybe you should send it back
<MoeIcenowy>
I think for this kind of problem
Mr__Anderson has quit [Remote host closed the connection]
<phipli>
It will cost a significant amount in registered post
<MoeIcenowy>
you can ask them for the shipping fee
<phipli>
yeah
<phipli>
I was considering that
Nacho has joined #linux-sunxi
<phipli>
If you know what I mean, the thing is honestly not working out of the box - It is a little frustrating that they wont believe me
<MoeIcenowy>
At least on Taobao the seller should pay the shipping fee for quality problem
<phipli>
I've uploaded 3 videos and about 6 photos
<MoeIcenowy>
To be honest you are too unfortunate.
<phipli>
after everyone they say something like "could you please also show that it is plugged into the router - we think you haven't plugged it in correctly"
<phipli>
Yes. First problem I've had
<phipli>
I usually use banggood because you can get refunds through paypal with them
JohnDoe_71Rus has joined #linux-sunxi
<MoeIcenowy>
and in my opinion booting it requires no network...
<MoeIcenowy>
(As I know the service workers on Aliexpress know little about technical problems
reinforce has joined #linux-sunxi
<lennyraposo>
mali no worky thus far on the pine
<MoeIcenowy>
lennyraposo: you cannot expect it to be working
<MoeIcenowy>
except on Android
<lennyraposo>
I have the AllWinner driver in hand and binaries ;)
<MoeIcenowy>
lennyraposo: aarch64 libMali.so for X11/FB ?
<lennyraposo>
yes
<MoeIcenowy>
why can you have it?
<lennyraposo>
xf86-server-armsoc
<lennyraposo>
it's available as of yesterday
<MoeIcenowy>
give me a link
<phipli>
MoeIcenowy : yes - they've been telling me to do stupid things. I've followed the instructions to try and get the refund. They keep making me film all the stupid things. Bet they're having a right joke about it
<longsleep>
lennyraposo: hey, why is it required to build a new kernel for the mail stuff received from allwinner?
<BenG83>
I think they send some fb driver along with it, not sure since I havent seen the files yet...
<longsleep>
BenG83: yeah, i am looking at it now
<longsleep>
lennyraposo: ok got it, i will create a experimental branch adding/replacing the BSP mali driver with the one from the new tarball and also adding the drm driver
<longsleep>
BenG83: they sent the mali blobs and an X11 video driver - but the reason why the Kernel needs to rebuild is that they updated the mali related modules as well
<BenG83>
lennyraposo had one problem that the x driver was not able to find a symbol
<lennyraposo>
you wanna keep the 2 separate for now?
jernej has joined #linux-sunxi
<longsleep>
lennyraposo: until it works yes
<lennyraposo>
ok
<lennyraposo>
will let you knwo of the results
<longsleep>
lennyraposo: once it works i will merge and that branch will be removed
<longsleep>
lennyraposo: well the mali_drm inside the mali driver does not work still, so i guess they want us to use this sunxi_drm thing whatever it is
scream has quit [Remote host closed the connection]
<lennyraposo>
they do
<lennyraposo>
for testing purposes I copied it over