| Message ID | 1261129507-13244-1-git-send-email-andre.przywara@amd.com |
|---|---|
| State | New |
| Headers | show |
On 12/18/2009 10:45 AM, Andre Przywara wrote: > If QEMU finds newer kernel header files on compilation time, it will use > advertised features like pipe2 or SOCK_CLOEXEC by just doing a compile test. > If later the executables are executed on an older kernel (<2.6.27, > like Xen Dom0 2.6.18), then QEMU will fail on opening sockets and creating > pipes and returns the rather unspecific "qemu_init_main_loop failed". > This patch fixes this by checking the return values of these calls > for EINVAL and ENOSYS and falling back to the older versions automatically. Agreed, most other apps that use *_CLOEXEC do this (including glibc itself and gnulib; the notable exception is APR and there were quite some flames about this). Paolo
Am 18.12.2009 10:45, schrieb Andre Przywara: > If QEMU finds newer kernel header files on compilation time, it will use > advertised features like pipe2 or SOCK_CLOEXEC by just doing a compile test. > If later the executables are executed on an older kernel (<2.6.27, > like Xen Dom0 2.6.18), then QEMU will fail on opening sockets and creating > pipes and returns the rather unspecific "qemu_init_main_loop failed". > This patch fixes this by checking the return values of these calls > for EINVAL and ENOSYS and falling back to the older versions automatically. > > Signed-off-by: Andre Przywara <andre.przywara@amd.com> > --- > osdep.c | 18 ++++++++++++------ > 1 files changed, 12 insertions(+), 6 deletions(-) > > diff --git a/osdep.c b/osdep.c > index 7509c5b..9949606 100644 > --- a/osdep.c > +++ b/osdep.c > @@ -262,13 +262,15 @@ int qemu_pipe(int pipefd[2]) > > #ifdef CONFIG_PIPE2 > ret = pipe2(pipefd, O_CLOEXEC); > -#else > + if (ret != -1 || errno != ENOSYS) { > + return ret; > + } > +#endif > ret = pipe(pipefd); > if (ret == 0) { > qemu_set_cloexec(pipefd[0]); > qemu_set_cloexec(pipefd[1]); > } > -#endif > > return ret; > } > @@ -283,12 +285,14 @@ int qemu_socket(int domain, int type, int protocol) > > #ifdef SOCK_CLOEXEC > ret = socket(domain, type | SOCK_CLOEXEC, protocol); > -#else > + if (ret != -1 || errno != EINVAL) { > + return ret; > + } > +#endif > ret = socket(domain, type, protocol); > if (ret >= 0) { > qemu_set_cloexec(ret); > } > -#endif > > return ret; > } > @@ -302,12 +306,14 @@ int qemu_accept(int s, struct sockaddr *addr, socklen_t *addrlen) > > #ifdef CONFIG_ACCEPT4 > ret = accept4(s, addr, addrlen, SOCK_CLOEXEC); > -#else > + if (ret != -1 || errno != EINVAL) { Shouldn't this be an ENOSYS? Kevin
Kevin Wolf wrote: >>.... >> @@ -302,12 +306,14 @@ int qemu_accept(int s, struct sockaddr *addr, socklen_t *addrlen) >> >> #ifdef CONFIG_ACCEPT4 >> ret = accept4(s, addr, addrlen, SOCK_CLOEXEC); >> -#else >> + if (ret != -1 || errno != EINVAL) { > > Shouldn't this be an ENOSYS? Oh, right you are. This was the untested part, because my glibc didn't have accept4. I only saw the SOCK_CLOEXEC name on this... Fixed patch follows... Regards, Andre.
diff --git a/osdep.c b/osdep.c index 7509c5b..9949606 100644 --- a/osdep.c +++ b/osdep.c @@ -262,13 +262,15 @@ int qemu_pipe(int pipefd[2]) #ifdef CONFIG_PIPE2 ret = pipe2(pipefd, O_CLOEXEC); -#else + if (ret != -1 || errno != ENOSYS) { + return ret; + } +#endif ret = pipe(pipefd); if (ret == 0) { qemu_set_cloexec(pipefd[0]); qemu_set_cloexec(pipefd[1]); } -#endif return ret; } @@ -283,12 +285,14 @@ int qemu_socket(int domain, int type, int protocol) #ifdef SOCK_CLOEXEC ret = socket(domain, type | SOCK_CLOEXEC, protocol); -#else + if (ret != -1 || errno != EINVAL) { + return ret; + } +#endif ret = socket(domain, type, protocol); if (ret >= 0) { qemu_set_cloexec(ret); } -#endif return ret; } @@ -302,12 +306,14 @@ int qemu_accept(int s, struct sockaddr *addr, socklen_t *addrlen) #ifdef CONFIG_ACCEPT4 ret = accept4(s, addr, addrlen, SOCK_CLOEXEC); -#else + if (ret != -1 || errno != EINVAL) { + return ret; + } +#endif ret = accept(s, addr, addrlen); if (ret >= 0) { qemu_set_cloexec(ret); } -#endif return ret; }
If QEMU finds newer kernel header files on compilation time, it will use advertised features like pipe2 or SOCK_CLOEXEC by just doing a compile test. If later the executables are executed on an older kernel (<2.6.27, like Xen Dom0 2.6.18), then QEMU will fail on opening sockets and creating pipes and returns the rather unspecific "qemu_init_main_loop failed". This patch fixes this by checking the return values of these calls for EINVAL and ENOSYS and falling back to the older versions automatically. Signed-off-by: Andre Przywara <andre.przywara@amd.com> --- osdep.c | 18 ++++++++++++------ 1 files changed, 12 insertions(+), 6 deletions(-)