search for: nbd_local_port

As the subject says, support socket activation in virt-p2v. I have added upstream support for socket activation to nbdkit already: https://github.com/libguestfs/nbdkit/commit/7ff39d028c6359f5c0925ed2cf4a2c4c751af2e4 I posted a patch for qemu-nbd, still waiting on more reviews for that one: https://www.mail-archive.com/qemu-devel@nongnu.org/msg427246.html I tested this against old and new qemu

qemu-nbd is a fine NBD server. However it's not easy to compile a reliably working version on RHEL 5. This patch series allows nbdkit to be used as an alternative (qemu-nbd is still the default). Rich.

Changes since v1: - Set probing source port to be nbd_local_port+1 instead of always using 50124 to deal with multi-disk scenario. - Set SO_REUSEADDR on client socket to avoid issues with old connections in TIME_WAIT. I've been running this for a few hours now using the updated multi-disk test and haven't seen any problems.

--- p2v/conversion.c | 43 +++++++++++++++++++++++++++++++++++++------ 1 file changed, 37 insertions(+), 6 deletions(-) diff --git a/p2v/conversion.c b/p2v/conversion.c index 4ff7ecc..14e7b3b 100644 --- a/p2v/conversion.c +++ b/p2v/conversion.c @@ -398,7 +398,8 @@ wait_qemu_nbd (int nbd_local_port, int timeout_seconds) { int sockfd; int result = -1; - struct sockaddr_in addr; + int reuseaddr = 1; + struct sockaddr_in src_addr, dst_addr; time_t start_t, now_t; struct timeval timeout = { .tv_usec = 0 }; char magic[8]; /* NBDMAGIC */ @@ -413,10 +414,40 @@ wait_qemu_nbd (int n...

...-git a/p2v/conversion.c b/p2v/conversion.c index 4ff7ecc..5c0c78e 100644 --- a/p2v/conversion.c +++ b/p2v/conversion.c @@ -44,6 +44,16 @@ /* How long to wait for qemu-nbd to start (seconds). */ #define WAIT_QEMU_NBD_TIMEOUT 10 +/* Source port for probing qemu-nbd. Should be one greater than + * nbd_local_port in ssh.c (50123). It's not guaranteed to always + * bind to 50124, but it will hint the kernel to start there and try + * incrementally higher ports if needed. This avoids the case where + * the kernel selects 50123 as our source port, and we immediately + * connect to ourself. + * See: https...

...#include <libintl.h> #include <sys/types.h> #include <sys/wait.h> +#include <arpa/inet.h> +#include <netinet/in.h> #include <glib.h> @@ -49,6 +51,7 @@ struct data_conn { static int send_quoted (mexp_h *, const char *s); static pid_t start_qemu_nbd (int nbd_local_port, const char *device); +static int wait_qemu_nbd (int nbd_local_port, int timeout_seconds); static void cleanup_data_conns (struct data_conn *data_conns, size_t nr); static char *generate_libvirt_xml (struct config *, struct data_conn *); static const char *map_interface_to_network (struct config...

(I don't have a BZ# for this yet, but I'm expecting it to be filed as an RFE) Currently if the user is in the virt-p2v GUI and cancels the conversion, all that happens is we abruptly close the ssh session to virt-v2v. That is bad .. possibly (or maybe not). But in any case there is an alternative: we can send a ^C key to the virt-v2v process, which it could catch and handle gracefully,

In the run dialog, I have implemented an ANSI colour escape sequence interpreter, so that colours displayed by the remote virt-v2v are now shown to the user. (https://bugzilla.redhat.com/show_bug.cgi?id=1314244) This requires virt-v2v to send colours. It wasn't doing that because the output was a pipe (as we capture the output into the log file). So I added a global --colours option to make

In an effort to reduce the code, start to use few bits of GLib: - replace the gnulib c-type module - replace the gnulib getprogname module - use g_spawn_sync to launch curl, and drop a file reading function Pino Toscano (6): Include glib.h in p2v.h Use g_ascii_isspace instead of c_isspace from gnulib Use g_get_prgname instead of getprogname from gnulib build: remove no more used gnulib

Now that virt-p2v has its own repository [1] and releases [2], it is time to remove it from libguestfs. [1] https://github.com/libguestfs/virt-p2v [2] http://download.libguestfs.org/virt-p2v/ Pino Toscano (2): Remove virt-p2v Remove remaining virt-p2v bits .gitignore | 4 - Makefile.am | 7 +- bash/Makefile.am

Hi, this series update libguestfs to a recent gnulib version, so that we can use its new getprogname module, and solve altogether one of the porting issues (the need for 'program_name' by the error module of gnulib), and have a single way to get the name of the current program. A number of changes in tools mostly, although mechanical. Thanks, Pino Toscano (3): Update gnulib to latest

...= caml__frame; \ + return (caml__temp_result); \ + }while(0) #endif /* These prototypes are solely to quiet gcc warning. */ diff --git a/p2v/conversion.c b/p2v/conversion.c index b19bd49..e2f37cd 100644 --- a/p2v/conversion.c +++ b/p2v/conversion.c @@ -584,7 +584,7 @@ wait_qemu_nbd (int nbd_local_port, int timeout_seconds) } result = 0; -cleanup: + cleanup: close (sockfd); return result; @@ -615,25 +615,25 @@ cleanup_data_conns (struct data_conn *data_conns, size_t nr) /* Macros "inspired" by src/launch-libvirt.c */ /* <element */ -#define start_element(element)...

Result of earlier copy and paste. --- align/scan.c | 35 ++++++++++--------- cat/cat.c | 39 +++++++++++---------- cat/filesystems.c | 69 +++++++++++++++++++------------------- cat/log.c | 35 ++++++++++--------- cat/ls.c | 61 +++++++++++++++++---------------- df/main.c | 43 ++++++++++++------------ diff/diff.c | 67

libguestfs has used double and triple underscores in identifiers. These aren't valid for global names in C++. (http://stackoverflow.com/a/228797) These large but completely mechanical patches change the illegal identifiers to legal ones. Rich.