fw_env: auto-detect NAND characteristics

Hello,
Please find attached a patch proposal which modifies fw_env so NAND chip erase size as well as environment sector counts are guessed at runtime if they are both statically configured as zero. This actually extend a previous similar behavior which was NOR-specific, useful when a same RootFS is executed with several flash chips.
Be aware that I did not test this exact patch for real. I extracted it from my older u-boot-flsc 2016 tree on which it works well, and built it using a sandbox configuration, without running it. I only checked my gcc was happy with those lines (no warning added). My feeling is that this patch is small enough to be confident it will work on master as it worked on my old tree.
Opened to any review or request, best rergards, Anthony Loiseau

From: Anthony Loiseau anthony.loiseau@allcircuits.com
As already done for NOR chips, if device ESIZE and ENVSECTORS static configurations are both zero, then autodetect them at runtime.
Cc: Joe Hershberger joe.hershberger@ni.com Cc: Stefan Agner stefan.agner@toradex.com Signed-off-by: Anthony Loiseau anthony.loiseau@allcircuits.com --- tools/env/README | 3 +++ tools/env/fw_env.c | 11 +++++++++-- 2 files changed, 12 insertions(+), 2 deletions(-)
diff --git a/tools/env/README b/tools/env/README index 480a893202..b8c6a7e197 100644 --- a/tools/env/README +++ b/tools/env/README @@ -58,6 +58,9 @@ DEVICEx_ENVSECTORS defines the number of sectors that may be used for this environment instance. On NAND this is used to limit the range within which bad blocks are skipped, on NOR it is not used.
+If DEVICEx_ESIZE and DEVICEx_ENVSECTORS are both zero, then a runtime +detection is attempted for NOR and NAND mtd types. + To prevent losing changes to the environment and to prevent confusing the MTD drivers, a lock file at /run/fw_printenv.lock is used to serialize access to the environment. diff --git a/tools/env/fw_env.c b/tools/env/fw_env.c index c9a8774ace..9f57f19d80 100644 --- a/tools/env/fw_env.c +++ b/tools/env/fw_env.c @@ -1659,8 +1659,15 @@ static int check_device_config(int dev) } DEVTYPE(dev) = mtdinfo.type; if (DEVESIZE(dev) == 0 && ENVSECTORS(dev) == 0 && - mtdinfo.type == MTD_NORFLASH) - DEVESIZE(dev) = mtdinfo.erasesize; + mtdinfo.erasesize > 0) { + if (mtdinfo.type == MTD_NORFLASH) + DEVESIZE(dev) = mtdinfo.erasesize; + else if (mtdinfo.type == MTD_NANDFLASH) { + DEVESIZE(dev) = mtdinfo.erasesize; + ENVSECTORS(dev) = + mtdinfo.size / mtdinfo.erasesize; + } + } if (DEVESIZE(dev) == 0) /* Assume the erase size is the same as the env-size */ DEVESIZE(dev) = ENVSIZE(dev);

Hi Anthony,
On Thu, Dec 21, 2023 at 1:18 PM Anthony Loiseau anthony@loiseau.fr wrote:
Hello,
Please find attached a patch proposal which modifies fw_env so NAND chip
Your patch did not come. Could you please send it via git send-email?
Also, run ./scripts/get_maintainer.pl 0001-your.patch and Cc the maintainers when submitting it.
Thanks
participants (2)
-
Anthony Loiseau
-
Fabio Estevam