On 10/31/2011 08:40 AM, Peter Maydell wrote:
An early ping since I have no idea who counts as the submaintainer
for this patchset and it definitely needs to go in for 1.0...
I'm in the process of testing them and will apply provided nothing breaks.
Regards,
Anthony Liguori
thanks
-- PMM
On 28 October 2011 10:52, Peter Maydell<peter.mayd...@linaro.org> wrote:
These patches add enough of the TLS abstraction layer to allow us
to make cpu_single_env thread-local on Linux systems. This fixes
the regression described in bug 823902 for the 1.0 release; we
can add the Win32 and POSIX implementations later.
I haven't included Paolo's "Prepare Windows port for thread-local
cpu_single_env" patch -- it would be safe to do so but it isn't
necessary until we actually implement TLS for Win32.
Changes v1->v2:
* fix Paolo's email address
* split the darwin-user change out into a separate patch
* drop the 'tls_' prefix from the cpu_single_env tls var name
Changes v2->v3:
* minor rearrangement of copyright notice in comment
* added a missing Signed-off-by
* fixed the name of the multiple-include-guard #define
Paolo Bonzini (2):
darwin-user/main.c: Drop unused cpu_single_env definition
Make cpu_single_env thread-local
Peter Maydell (1):
qemu-tls.h: Add abstraction layer for TLS variables
cpu-all.h | 4 +++-
darwin-user/main.c | 2 --
exec.c | 2 +-
qemu-tls.h | 52 ++++++++++++++++++++++++++++++++++++++++++++++++++++
4 files changed, 56 insertions(+), 4 deletions(-)
create mode 100644 qemu-tls.h