EDIT: as it turns out, 21061 had an older kernel version than 33375, and therefore the webflash upgrade function should work properly, as I found after looking through all the key commits. _________________ "Life is but a fleeting moment, a vapor that vanishes quickly; All is vanity"
Contribute To DD-WRT Pogo - A minimal level of ability is expected and needed... DD-WRT Releases 2023 (PolitePol)
DD-WRT Releases 2023 (RSS Everything)
----------------------
Linux User #377467 counter.li.org / linuxcounter.net
Last edited by kernel-panic69 on Thu Jun 24, 2021 6:24; edited 1 time in total
(Not the slightest idea at the moment whether this router uses ar7100_flash.c or ar7240_flash.c or whatever, looks like the name of a C source file to me.)
I used DD-WRT web interface, Administration | Firmware Upgrade, I left "Don't reset" selected as you recommended, which spared me to configure again most settings (it only changed a few, for instance the wireless settings and the time zone, easily reconfigured, and a few OpenVPN cipher settings, NordVPN server is complaining about some deprecated options but OpenVPN is now working, I'll check about those settings for possible finer tuning).
Thank you. (Going to check DD-WRT "donate" modalities".)
P.S.: I don't see how to mark this thread as SOLVED.
Joined: 08 May 2018 Posts: 14244 Location: Texas, USA
Posted: Thu Jun 24, 2021 5:38 Post subject:
blkt wrote:
r21061 has old dd-wrt header, before a bug introduced with r33375, so it properly detected bootloader size.
BrainSlayer wrote:
in older firmwares the string isnt tp-link but dd-wrt. i changed that string to tp-link in that firmware and this all caused the bug since my mtd partition code was looking for the wrong string to detect the bootloader size
Now, read that all again. The bug was the old header that existed *before* 33375 being looked for by the webupgrade / mtd code *in* 33375. So, it shouldn't have flashed the upgrade, *unless* the checks for the header (or kernel version) were different, etc. in 21061 (which was actually the case, different kernel version with good code in it). I didn't just pull the 33375 thread or the other threads out of my orifice for no reason. See the image...
The bug was the old header that existed *before* 33375 being looked for by the webupgrade / mtd code *in* 33375. So, it shouldn't have flashed the upgrade, *unless* the checks for the header (or kernel version) were different, etc. in 21061 (which was actually the case, different kernel version with good code in it).
I can back-trace through all the relevant commits, too, if you really want.
Do want, for posterity of course. Would be nice to know the first build requiring fullflash after webupgrade.
Joined: 08 May 2018 Posts: 14244 Location: Texas, USA
Posted: Thu Jun 24, 2021 15:47 Post subject:
The answer is already in this thread for when a fullflash was required (33375 thru 42925, fixed with commit 42928 for build 42954). The other relevant commits (kernel updates, changing from older kernel (pre-3.10, 3.18, etc.) start here:
I would have to look further into when 3.10 and 3.18 were actually ported to TP-Links specifically, but both were added after 21061. I'm not going to let my OCD chase that rabbit farther than I already have, but feel free to let your clicker do the clicking.
Anyhow, this was a case of my presuming that there was a likelihood of the CLI magic voodoo being required, where it turned out that it wasn't. No harm, no foul. _________________ "Life is but a fleeting moment, a vapor that vanishes quickly; All is vanity"
Contribute To DD-WRT Pogo - A minimal level of ability is expected and needed... DD-WRT Releases 2023 (PolitePol)
DD-WRT Releases 2023 (RSS Everything)
----------------------
Linux User #377467 counter.li.org / linuxcounter.net
That build must've been old enough that it didn't check the header on upgrading via webUI; but you are absolutely positive that it is upgraded?
Absolutely positive, upgrade confirmed, and it's working flawlessly so far.
(I've also unblocked the super channel feature, although for now I'm leaving it deactivated in the Wireless advanced settings, basically because I don't need it and I fear to submit the final output stage of the radio to higher SWR ratio.)
Last edited by overw on Thu Jun 24, 2021 17:11; edited 1 time in total
Would be nice to know the first build requiring fullflash after webupgrade.