Commit Graph

9 Commits

Author SHA1 Message Date
Thorsten Blum
e1fb7430fc lib/bch.c: use swap() to improve code
Use the swap() macro to simplify the functions solve_linear_system() and
gf_poly_gcd() and improve their readability.  Remove the local variable
tmp.

Fixes the following three Coccinelle/coccicheck warnings reported by
swap.cocci:

  WARNING opportunity for swap()
  WARNING opportunity for swap()
  WARNING opportunity for swap()

Link: https://lkml.kernel.org/r/20240708224023.9312-2-thorsten.blum@toblux.com
Signed-off-by: Thorsten Blum <thorsten.blum@toblux.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
2024-07-12 16:39:53 -07:00
John Sanpe
02d7f74a04 lib/bch.c: use bitrev instead of internal logic
Replace internal logic with separate bitrev library.

Link: https://lkml.kernel.org/r/20230730081717.1498217-1-sanpeqf@gmail.com
Signed-off-by: John Sanpe <sanpeqf@gmail.com>
Cc: Bhaskar Chowdhury <unixbhaskar@gmail.com>
Cc: Randy Dunlap <rdunlap@infradead.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
2023-08-18 10:18:58 -07:00
Bhaskar Chowdhury
0523c6922e lib/bch.c: fix a typo in the file bch.c
s/buid/build/

Link: https://lkml.kernel.org/r/20210301123129.18754-1-unixbhaskar@gmail.com
Signed-off-by: Bhaskar Chowdhury <unixbhaskar@gmail.com>
Acked-by: Randy Dunlap <rdunlap@infradead.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2021-05-06 19:24:12 -07:00
Miquel Raynal
1759279ad1 lib/bch: Allow easy bit swapping
It seems that several hardware ECC engine use a swapped representation
of bytes compared to software. This might having to do with how the
ECC engine is wired to the NAND controller or the order the bits are
passed to the hardware BCH logic.

This means that when the software BCH engine is working in conjunction
with data generated with hardware, sometimes we might need to swap the
bits inside bytes, eg:

    0x0A = b0000_1010 -> b0101_0000 = 0x50

Make it possible by adding a boolean to the BCH initialization routine.

Regarding the implementation itself, this is a rather simple approach
that can probably be enhanced in the future by preparing the
->a_{mod,pow}_tab tables with the swapping in mind.

Suggested-by: Boris Brezillon <boris.brezillon@collabora.com>
Signed-off-by: Miquel Raynal <miquel.raynal@bootlin.com>
Reviewed-by: Boris Brezillon <boris.brezillon@collabora.com>
Link: https://lore.kernel.org/linux-mtd/20200519074549.23673-3-miquel.raynal@bootlin.com
2020-05-24 20:48:11 +02:00
Miquel Raynal
c8ae3f744d lib/bch: Rework a little bit the exported function names
There are four exported functions, all suffixed by _bch, which is
clearly not the norm. Let's rename them by prefixing them with bch_
instead.

This is a mechanical change:
    init_bch -> bch_init
    free_bch -> bch_free
    encode_bch -> bch_encode
    decode_bch -> bch_decode

Signed-off-by: Miquel Raynal <miquel.raynal@bootlin.com>
Reviewed-by: Boris Brezillon <boris.brezillon@collabora.com>
Link: https://lore.kernel.org/linux-mtd/20200519074549.23673-2-miquel.raynal@bootlin.com
2020-05-24 20:48:11 +02:00
Gustavo A. R. Silva
a44ce51370 lib/bch.c: replace zero-length array with flexible-array member
The current codebase makes use of the zero-length array language extension
to the C90 standard, but the preferred mechanism to declare
variable-length types such as these ones is a flexible array member[1][2],
introduced in C99:

struct foo {
        int stuff;
        struct boo array[];
};

By making use of the mechanism above, we will get a compiler warning in
case the flexible array does not occur last in the structure, which will
help us prevent some kind of undefined behavior bugs from being
inadvertenly introduced[3] to the codebase from now on.

This issue was found with the help of Coccinelle.

[1] https://gcc.gnu.org/onlinedocs/gcc/Zero-Length.html
[2] https://github.com/KSPP/linux/issues/21
[3] commit 7649773293 ("cxgb3/l2t: Fix undefined behaviour")

Signed-off-by: Gustavo A. R. Silva <gustavo@embeddedor.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Link: http://lkml.kernel.org/r/20200211205119.GA21234@embeddedor
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2020-04-07 10:43:42 -07:00
Arnd Bergmann
f0fe77f601 lib/bch: fix possible stack overrun
The previous patch introduced very large kernel stack usage and a Makefile
change to hide the warning about it.

From what I can tell, a number of things went wrong here:

- The BCH_MAX_T constant was set to the maximum value for 'n',
  not the maximum for 't', which is much smaller.

- The stack usage is actually larger than the entire kernel stack
  on some architectures that can use 4KB stacks (m68k, sh, c6x), which
  leads to an immediate overrun.

- The justification in the patch description claimed that nothing
  changed, however that is not the case even without the two points above:
  the configuration is machine specific, and most boards  never use the
  maximum BCH_ECC_WORDS() length but instead have something much smaller.
  That maximum would only apply to machines that use both the maximum
  block size and the maximum ECC strength.

The largest value for 't' that I could find is '32', which in turn leads
to a 60 byte array instead of 2048 bytes. Making it '64' for future
extension seems also worthwhile, with 120 bytes for the array. Anything
larger won't fit into the OOB area on NAND flash.

With that changed, the warning can be enabled again.

Only linux-4.19+ contains the breakage, so this is only needed
as a stable backport if it does not make it into the release.

Fixes: 02361bc778 ("lib/bch: Remove VLA usage")
Reported-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Cc: stable@vger.kernel.org
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Boris Brezillon <boris.brezillon@bootlin.com>
2018-10-12 09:17:46 +02:00
Kees Cook
02361bc778 lib/bch: Remove VLA usage
In the quest to remove all stack VLA usage from the kernel[1], this
allocates a fixed size stack array to cover the range needed for
bch. This was done instead of a preallocation on the SLAB due to
performance reasons, shown by Ivan Djelic:

 little-endian, type sizes: int=4 long=8 longlong=8
 cpu: Intel(R) Core(TM) i5 CPU         650  @ 3.20GHz
 calibration: iter=4.9143µs niter=2034 nsamples=200 m=13 t=4

   Buffer allocation |  Encoding throughput (Mbit/s)
 ---------------------------------------------------
  on-stack, VLA      |   3988
  on-stack, fixed    |   4494
  kmalloc            |   1967

So this change actually improves performance too, it seems.

The resulting stack allocation can get rather large; without
CONFIG_BCH_CONST_PARAMS, it will allocate 4096 bytes, which
trips the stack size checking:

lib/bch.c: In function ‘encode_bch’:
lib/bch.c:261:1: warning: the frame size of 4432 bytes is larger than 2048 bytes [-Wframe-larger-than=]

Even the default case for "allmodconfig" (with CONFIG_BCH_CONST_M=14 and
CONFIG_BCH_CONST_T=4) would have started throwing a warning:

lib/bch.c: In function ‘encode_bch’:
lib/bch.c:261:1: warning: the frame size of 2288 bytes is larger than 2048 bytes [-Wframe-larger-than=]

But this is how large it's always been; it was just hidden from
the checker because it was a VLA. So the Makefile has been adjusted to
silence this warning for anything smaller than 4500 bytes, which should
provide room for normal cases, but still low enough to catch any future
pathological situations.

[1] https://lkml.kernel.org/r/CA+55aFzCG-zNmZwX4A2FQpadafLfEzK6CC=qPXydAacU1RqZWA@mail.gmail.com

Signed-off-by: Kees Cook <keescook@chromium.org>
Reviewed-by: Ivan Djelic <ivan.djelic@parrot.com>
Tested-by: Ivan Djelic <ivan.djelic@parrot.com>
Acked-by: Boris Brezillon <boris.brezillon@bootlin.com>
Signed-off-by: Boris Brezillon <boris.brezillon@bootlin.com>
2018-06-22 00:29:39 +02:00
Ivan Djelic
437aa565e2 lib: add shared BCH ECC library
This is a new software BCH encoding/decoding library, similar to the shared
Reed-Solomon library.

Binary BCH (Bose-Chaudhuri-Hocquenghem) codes are widely used to correct
errors in NAND flash devices requiring more than 1-bit ecc correction; they
are generally better suited for NAND flash than RS codes because NAND bit
errors do not occur in bursts. Latest SLC NAND devices typically require at
least 4-bit ecc protection per 512 bytes block.

This library provides software encoding/decoding, but may also be used with
ASIC/SoC hardware BCH engines to perform error correction. It is being
currently used for this purpose on an OMAP3630 board (4bit/8bit HW BCH). It
has also been used to decode raw dumps of NAND devices with on-die BCH ecc
engines (e.g. Micron 4bit ecc SLC devices).

Latest NAND devices (including SLC) can exhibit high error rates (typically
a dozen or more bitflips per hour during stress tests); in order to
minimize the performance impact of error correction, this library
implements recently developed algorithms for fast polynomial root finding
(see bch.c header for details) instead of the traditional exhaustive Chien
root search; a few performance figures are provided below:

Platform: arm926ejs @ 468 MHz, 32 KiB icache, 16 KiB dcache
BCH ecc : 4-bit per 512 bytes

Encoding average throughput: 250 Mbits/s

Error correction time (compared with Chien search):

        average   worst      average (Chien)  worst (Chien)
----------------------------------------------------------
1 bit    8.5 µs   11 µs         200 µs           383 µs
2 bit    9.7 µs   12.5 µs       477 µs           728 µs
3 bit   18.1 µs   20.6 µs       758 µs          1010 µs
4 bit   19.5 µs   23 µs        1028 µs          1280 µs

In the above figures, "worst" is meant in terms of error pattern, not in
terms of cache miss / page faults effects (not taken into account here).

The library has been extensively tested on the following platforms: x86,
x86_64, arm926ejs, omap3630, qemu-ppc64, qemu-mips.

Signed-off-by: Ivan Djelic <ivan.djelic@parrot.com>
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
2011-03-11 14:25:50 +00:00