[oe] [meta-oe][meta-initramfs][PATCH v2] ubi-utils-klibc: catch up with struct renaming in mtd/ubi-user.h

[Andrea Adami]

On Wed, Jun 20, 2012 at 1:02 AM, Andrea Adami <andrea.adami at gmail.com> wrote:
> On Tue, Jun 19, 2012 at 9:47 AM, Andrea Adami <andrea.adami at gmail.com> wrote:
>> * v2 replaces malformed v1 patch
>> * fix build, broken by header upgrade
>> * bump PR
>>
>> Signed-off-by: Andrea Adami <andrea.adami at gmail.com>
>> ---
>>  .../mtd/ubi-utils-klibc/libubi.c.patch             |   17 +++++++++++++++++
>>  1 files changed, 17 insertions(+), 0 deletions(-)
>>
>> diff --git a/meta-initramfs/recipes-devtools/mtd/ubi-utils-klibc/libubi.c.patch b/meta-initramfs/recipes-devtools/mtd/ubi-utils-klibc/libubi.c.patch
>> index db95904..048e39b 100644
>> --- a/meta-initramfs/recipes-devtools/mtd/ubi-utils-klibc/libubi.c.patch
>> +++ b/meta-initramfs/recipes-devtools/mtd/ubi-utils-klibc/libubi.c.patch
>> @@ -9,3 +9,20 @@
>>  {
>>        int fd, ret;
>>
>> +@@ -1352,13 +1352,13 @@
>> +
>> + int ubi_set_property(int fd, uint8_t property, uint64_t value)
>> + {
>> +-      struct ubi_set_prop_req r;
>> ++      struct ubi_set_vol_prop_req r;
>> +
>> +-      memset(&r, 0, sizeof(struct ubi_set_prop_req));
>> ++      memset(&r, 0, sizeof(struct ubi_set_vol_prop_req));
>> +       r.property = property;
>> +       r.value = value;
>> +
>> +-      return ioctl(fd, UBI_IOCSETPROP, &r);
>> ++      return ioctl(fd, UBI_IOCSETVOLPROP, &r);
>> + }
>> +
>> + int ubi_leb_unmap(int fd, int lnum)
>> --
>> 1.7.3.4
>>
>
> This strange renaming issue happened today, with stock linux-libc-headers.
> I'm trying to understand why right now I get still pre-3.0 headers...
> this is mtd/ubi-user.h
>
> Anyhow, this patch is superseded by the next one, committing ver. 1.5.0
>
> Andrea

I've found out oe-core's parent recipe  mtd-utils_1.5.0 is staging its
own headers in sysroot...
We happen to install mtd-utils in our images so the header update has
always been masked.

FWIW ubi-utils-klibc has own do_install and does not touch the sysroot :)

Thinking now about the best solution...

Andrea