diff options
author | kfraser@localhost.localdomain <kfraser@localhost.localdomain> | 2007-03-07 10:53:34 +0000 |
---|---|---|
committer | kfraser@localhost.localdomain <kfraser@localhost.localdomain> | 2007-03-07 10:53:34 +0000 |
commit | 0cb97f4d8bff4b430318d74d7235ec3568ec5771 (patch) | |
tree | 04051fe5650389718baf1bda3ef8c1992c7b4c1c /docs/misc/dump-core-format.txt | |
parent | 9de4f0c67def5ab75c4f5a9a4428b719e149e8b2 (diff) | |
download | xen-0cb97f4d8bff4b430318d74d7235ec3568ec5771.tar.gz xen-0cb97f4d8bff4b430318d74d7235ec3568ec5771.tar.bz2 xen-0cb97f4d8bff4b430318d74d7235ec3568ec5771.zip |
dump-core: always use elfclass64 format to support x86_32pae.
Without this patch xen dump-core format can't handle over 4GB case.
Signed-off-by: Isaku Yamahata <yamahata@valinux.co.jp>
Diffstat (limited to 'docs/misc/dump-core-format.txt')
-rw-r--r-- | docs/misc/dump-core-format.txt | 14 |
1 files changed, 10 insertions, 4 deletions
diff --git a/docs/misc/dump-core-format.txt b/docs/misc/dump-core-format.txt index ce759f39a0..99384ca94a 100644 --- a/docs/misc/dump-core-format.txt +++ b/docs/misc/dump-core-format.txt @@ -26,11 +26,12 @@ For xen related structure, please see the xen header files. Elf header ---------- The elf header members are set as follows + e_ident[EI_CLASS] = ELFCLASS64 = 2 e_ident[EI_OSABI] = ELFOSABI_SYSV = 0 e_type = ET_CORE = 4 -e_ident[EI_CLASS], e_ident[EI_DATA] and e_flags are set according -to an architecture which a file is created. Other members are set as usual. - +ELFCLASS64 is always used independent of architecture. +e_ident[EI_DATA] and e_flags are set according to the dumping system's +architecture. Other members are set as usual. Sections -------- @@ -221,5 +222,10 @@ format_version descriptor Format version history ---------------------- -The currently only (major, minor) = (0, 1) is used. +Currently only (major, minor) = (0, 1) is used. [When the format is changed, it would be described here.] + +(0, 1) update +- EI_CLASS member of elf header was changed to ELFCLASS64 independent of + architecture. This is mainly for x86_32pae. + The format version isn't bumped because analysis tools can distinguish it. |