Linux Kernel refresh patches

For Developers
21

103 is stable for me but I still need to test 105 and then do a PR.

1 Like
22

I just finished creating the firmware for my WNDR4300, the patch for .105 can be found here

I will do a runtime test, if that is okay I will do a PR.

23

Runtime test for .105 was succesful I created a PR.

@Borromini I hope the update resolves the issue you experience with .103

24

Thanks, I thinks it's just my power outlet being wonky though. Compiling as we speak, will run-test ramips and x86 and let you know.

1 Like
25

@Borromini I experience an issue applying the following patch which should be fixed:
0072-mtd-backport-v4.7-0day-patches-from-Boris.patch

I also made a new PR for this fix and updated to .106 the patch can be found here

26

Thanks. Did you see Hauke's remark on the mailing list?

27

@Borromini, I saw Hauke's remark. What I did is I try to fix the pathes for Mediatek en Oxnas:
0072-mtd-backport-v4.7-0day-patches-from-Boris.patch

But Kevin Darbyshire-Bryant send me an remark on the new PR about it.

This part of the patch needs an extra line:
2509 @@ -2662,7 +2697,7 @@ err_out:
2510 static int panic_nand_write(struct mtd_info *mtd, loff_t to, size_t len,
2511 size_t *retlen, const uint8_t *buf)
2512 {
2513 - struct nand_chip *chip = mtd->priv;
2514 + struct nand_chip *chip = mtd_to_nand(mtd);
2515 struct mtd_oob_ops ops;
2516 int ret;

Between the following two lines there need to be a line added:
2514 + struct nand_chip *chip = mtd_to_nand(mtd);
2515 struct mtd_oob_ops ops;

This should be added: int chipnr = (int)(to >> chip->chip_shift);

But I don't figure out how to update the patch properly, any idea?

28

@bladeoner I made a patch for that single line; grab it here. This patch will update the existing 17.01 codebase to include that line. You can simply apply (don't use git am, just apply the patch) it and rework your .106 bump on top of that.

I'm not sure what your .106 bump is supposed to do here, since it seems to be removing all that code at present, and the code you show above is how it looks in the 17.01 tree.

Your .106 patch has this, effectively removing that part from the 0072 patch:

-- a/target/linux/mediatek/patches-4.4/0072-mtd-backport-v4.7-0day-patches-from-Boris.patch
+++ b/target/linux/mediatek/patches-4.4/0072-mtd-backport-v4.7-0day-patches-from-Boris.patch
@@ -2506,16 +2506,7 @@ Signed-off-by: John Crispin <blogic@openwrt.org>
  
  	uint8_t *oob = ops->oobbuf;
  	uint8_t *buf = ops->datbuf;
-@@ -2662,7 +2697,7 @@ err_out:
- static int panic_nand_write(struct mtd_info *mtd, loff_t to, size_t len,
- 			    size_t *retlen, const uint8_t *buf)
- {
--	struct nand_chip *chip = mtd->priv;
-+	struct nand_chip *chip = mtd_to_nand(mtd);
- 	struct mtd_oob_ops ops;
- 	int ret;
-

I may be the blind leading the blind here. I checked your Github PR but I'm not seeing any comments by @ldir so I hope he can clarify (he probably knows more about it than the both of us).

P.S. Use the backticks (`) for code or the </> button in your editor for chunks, so they stay readable (and usable).

BTW, I am experiencing weird slowdowns on my DIR-860L. Not sure if that's in any way related to those kernel bumps. The DIR-860L is still on .105 for that matter.

29

@Borromini thank you for the patch.

KevinDB send a remark the same as Hauke with the mailinglist so its not in the PR. Actually saying the same the part you mentioned shouldn’t be removed from the patch. I did a refresh with the script you shared with me and Quilt which caused that.

For now I closed the PR, the patch is still in Patchwork I asked KevinDB to remove it if he can, don’t know how to remove it there.