osmo_apdu_segment_in() doesn't have the context of the previous APDU,
for nearly all calls this is fine, except for GP GET DATA, which can be
a case 2 or case 4 APDU.
GP GET DATA defines the Le field as 0x00 which is used by osmo_apdu_segment_in
to detect if it is a case 2 or case 4.
However a card can respond with the SW 6cXX, which means the previous
command must be resend with a Le field XX, which got misinterpreted by
the osmo_apdu_segment_in() as a case 4 resulting in a timeout waiting for
more data.
Modem <-> Card
81cadf2000 ->
<- 6c0b
81cadf200b ->
-- Simtrace waits for more data because it wrongly decodes it as APDU case 4
instead of case 2.
Related: SYS#8147
Change-Id: Ie238662e7e6a10dd4283ede0c8d8d73bc375c247
Use .tarball-version from the topdir, because it only gets written there
when generating the OBS package. Remove the duplicate git-version-gen
script and use the one from the topdir to generate a version string if
building from the git tree.
Related: OS#4413
Change-Id: I4b197a218ab44632ff182ffbd72e15c2b20db341
Make building the debian packages work again. I've verified that it
works in my own OBS namespace.
This patch also adds missing pkgconf variables in host/Makefile.am, so
libosmo-simtrace2.pc installs properly.
Related: OS#4283
Fixes: 964cda309d ("host: use autotools and split shared code to libosmo-simtrace2")
Change-Id: I2377de1e8b149520922217a1ab16f6e22fe6462a
Alexander Couzens