jeff___m has joined #openwrt-devel
jeff___m has quit [Ping timeout: 480 seconds]
jeff___m has joined #openwrt-devel
hanetzer has quit [Quit: WeeChat 4.1.2]
fakuivan has quit [Remote host closed the connection]
fakuivan has joined #openwrt-devel
jeff___m has quit [Ping timeout: 480 seconds]
goliath has quit [Quit: SIGSEGV]
jeff___m has joined #openwrt-devel
jeff___m has quit [Ping timeout: 480 seconds]
jeff___m has joined #openwrt-devel
tSYS has quit [Quit: *squeak*]
tSYS has joined #openwrt-devel
gch981213 has joined #openwrt-devel
Mangix has joined #openwrt-devel
jeff___m has quit [Remote host closed the connection]
robimarko has joined #openwrt-devel
cmonroe has quit [Ping timeout: 480 seconds]
jeff___m has joined #openwrt-devel
rua has quit [Remote host closed the connection]
rua has joined #openwrt-devel
jeff___m has quit [Remote host closed the connection]
rua has quit [Remote host closed the connection]
rua has joined #openwrt-devel
jeff___m has joined #openwrt-devel
jeff___m has quit [Ping timeout: 480 seconds]
parazyd has joined #openwrt-devel
jeff___m has joined #openwrt-devel
vincejv has quit [Quit: Bye bye! Leaving for now...]
jeff___m has quit [Ping timeout: 480 seconds]
rmilecki has joined #openwrt-devel
<rmilecki> nbd: could you publish mt7996_rom_patch.bin / mt7992_rom_patch.bin (required by mt7996(e) driver) somewhere? maybe https://github.com/openwrt/mt76/tree/master/firmware to start with?
<nbd> need to check with mtk to figure out what version should be published
<rmilecki> nbd: thanks!
<nbd> mtk has a custom sdk tree, but i think they mainly push development versions there, not sure if those are appropriate to put in mt76.git
* f00b4r0 discovers that gitweb has made it impossible to include html README file as used to be the case, *sighs*
jeff___m has joined #openwrt-devel
<rmilecki> Ansuel: i've problem with Aquantia firmware upload support
<rmilecki> my SoC requires those with Mediatek's SDK:
<rmilecki> mdi-reversal = /bits/ 16 <0>;
<Ansuel> ?
<rmilecki> Ansuel: firmware upload has to be followed by that:
<rmilecki> phy_modify_mmd(phydev, MDIO_MMD_PMAPMD, PMAPMD_RSVD_VEND_PROV, PMAPMD_RSVD_VEND_PROV_MDI_CONF, val)
<rmilecki> without it PHY doesn't seem to work
<Ansuel> do you have uboot source?
<rmilecki> no, I've Mediatek's source
<robimarko> You can add that fixup to the driver
<Ansuel> yep should not hard but i guess we need a property for it? I assume it's not something standard that should be applied to any device?
<rmilecki> sure, I can, I plan to
<rmilecki> I wanted to verify that with Ansuel
<robimarko> Yes, its a custom fixup, I dont there is a generic DT binding for it
<Ansuel> aquantia,mdi-reversal i guess?
<robimarko> rmilecki: What is the MMD and register addr?
<robimarko> I am interested in the PHY defaults
jeff___m has quit [Ping timeout: 480 seconds]
<Ansuel> (sad that they didn't include this in the FW directly...)
<rmilecki> include/uapi/linux/mdio.h:#define MDIO_MMD_PMAPMD 1 /* Physical Medium Attachment/
<rmilecki> #define PMAPMD_RSVD_VEND_PROV 0xe400
<rmilecki> #define PMAPMD_RSVD_VEND_PROV_MDI_CONF BIT(0)
<robimarko> Well, 0 is the default value for that bit
<Ansuel> so i guess fw is setting that to 1???
<robimarko> It seems autoneg updates the bit if MDI force configuration is not set as well
<robimarko> So, it would be great to dump that register when it doesnt work
<rmilecki> will do
<robimarko> And see whats going on
<robimarko> BTW, what is the AQR model?
<robimarko> I am looking at the 4th gen datasheet
<rmilecki> Aquantia AQR113C
<rmilecki> i'm not sure about revision
<rmilecki> if there is one
<robimarko> Ok, I am looking at the right datasheet
<robimarko> There are 2 revisions of 113C
<Ansuel> just to confirm mtk sets the bit to 1 and restart autoneg?
<rmilecki> Mediatek is their SDK has:
<rmilecki> mdi-reversal = /bits/ 16 <1>;
<rmilecki> that doesn't work for my board and I had to replace it with
<rmilecki> mdi-reversal = /bits/ 16 <0>;
<robimarko> That is weird if the kernel driver is loading the FW as it should default to 0 anyway
<rmilecki> Mediatek's firmware upload: https://pastebin.com/j7jx6dHK
<Ansuel> rmilecki do you have by chance a link to their driver? I'm curious how they handle it
<rmilecki> sec
<robimarko> Maybe their 1 means 0 in the code
<Ansuel> mhhh nope it seems they really enable that bit based on what it's in the DT
<Ansuel> interesting
<Ansuel> i mean nothing bad it's a cheap and easy addition just strange...
<Ansuel> a quick check with upstream driver it doesn't seems other PHY have something similar so I guess it has to be submitted as a custom property
<rmilecki> Gitiles doesn't have friendly UI, it took me a moment to find it
<Ansuel> wow they add lots of stuff for simple fw download
<rmilecki> I was surprised with your small implementation of fw upload after looking at Mediatek's code ;)
<Ansuel> robimarko that provisioning thing looks interesting wonder if qcom implementation didn't had it
<Ansuel> (btw on a side note been 2 days i'm investigating a dtb corruption with the arm decompressor... it's either a bug in how the dtb is getting compiled or a bug in how it's parsed by the fdt functions... resulting in only one of the memory node getting detected if the dtb is big enough)
<Ansuel> (EXTREME SUFFERING)
<Ansuel> (all of that to permit use of AUTO_ZRELADDR on ipq806x -.-)
cmonroe has joined #openwrt-devel
goliath has joined #openwrt-devel
<rmilecki> Ansuel: robimarko: so before firmware load devad=MDIO_MMD_PMAPMD regnum=PMAPMD_RSVD_VEND_PROV is 0x0000
<rmilecki> after firmware load it's 0x0007
<rmilecki> to make PHY work on my board I need it 0x0006
<Ansuel> are you sure it's the correct fw?
<Ansuel> the fw have the bit set to one in the provision data (the fw can set phy regs)
<Ansuel> so your fw set the mdi reverse but the phy actually isn't... aka a broken fw for the installed phy
<Ansuel> OK WTF NOW? THE APPENDED DTB TO THE ZIMAGE IS GETTING CORRUPTED??????
<robimarko> Ansuel: It seems that MTK copy/pasted half of the standalone AQR API to load the FW
<robimarko> I based the implementation on the upstream U-Boot
<robimarko> rmilecki: That register is provisioned by the FW, it was set to that value by the vendor FW configuration
<Ansuel> I discovered something wonderful.... the sp register makes the appended dtb to the zimage corrupt
<Ansuel> that is actually good since i had lots of fear bootm was corrupting stuff while relocating the loaded image...
Forst has quit [Quit: Quitting]
jeff___m has joined #openwrt-devel
jeff___m has quit [Ping timeout: 480 seconds]
mentalow has quit [Quit: :]]
mentalow has joined #openwrt-devel
Forst has joined #openwrt-devel
goliath has quit [Quit: SIGSEGV]
vincejv has joined #openwrt-devel
jeff___m has joined #openwrt-devel
jeff___m has quit [Ping timeout: 480 seconds]
vincejv has quit [Quit: Bye bye! Leaving for now...]
<robimarko> Ansuel: Is the DTB properly aligned?
<Ansuel> it was a problem with moving the stack pointer and calling C functions (assembly is always fun) i'm making progress on this and I think I'm very near on having a final patch to propose
<Ansuel> The only thing is that it does patch the arm compressor code so it will affect every target that makes use of those flags
<Ansuel> (but my idea is that it was broken from the start and things had to disable AUTO_Z... config to make the thing bootable)
<Ansuel> lovely r7800 uboot atags
<Ansuel> they pass 0x41500000 as starting
<Ansuel> and 0x60000000 as size
<Ansuel> 415 is more or less O.K. problem is the gigantic wrong ram size ahahahah
jeff___m has joined #openwrt-devel
jeff___m has quit [Ping timeout: 480 seconds]
vincejv has joined #openwrt-devel
fakuivan has quit [Remote host closed the connection]
fakuivan has joined #openwrt-devel
jeff___m has joined #openwrt-devel
jeff___m has quit [Ping timeout: 480 seconds]
<Ansuel> sent the patch upstream... really curious what they will think of it...
robimarko has quit [Quit: https://quassel-irc.org - Chat comfortably. Anywhere.]
jeff___m has joined #openwrt-devel
jeff___m has quit [Ping timeout: 480 seconds]
jeff___m has joined #openwrt-devel
cmonroe has quit [Ping timeout: 480 seconds]
goliath has joined #openwrt-devel
cmonroe has joined #openwrt-devel
gch981213 has quit [Quit: https://quassel-irc.org - Chat comfortably. Anywhere.]
jeff___m has quit [Remote host closed the connection]
jeff___m has joined #openwrt-devel