On 02/08/2018 10:13 PM, Thomas Huth wrote: > On 30.01.2018 20:49, Paolo Bonzini wrote: >> On 25/01/2018 08:51, Klim Kireev wrote: >>> The following behavior was observed for QEMU configured by libvirt >>> to use guest agent as usual for the guests without virtio-serial >>> driver (Windows or the guest remaining in BIOS stage). >>> >>> In QEMU on first connect to listen character device socket >>> the listen socket is removed from poll just after the accept(). >>> virtio_serial_guest_ready() returns 0 and the descriptor >>> of the connected Unix socket is removed from poll and it will >>> not be present in poll() until the guest will initialize the driver >>> and change the state of the serial to "guest connected". >>> >>> In libvirt connect() to guest agent is performed on restart and >>> is run under VM state lock. Connect() is blocking and can >>> wait forever. >>> In this case libvirt can not perform ANY operation on that VM. >>> >>> The bug can be easily reproduced this way: >>> >>> Terminal 1: >>> qemu-system-x86_64 -m 512 -device pci-serial,chardev=serial1 -chardev >>> socket,id=serial1,path=/tmp/console.sock,server,nowait >>> (virtio-serial and isa-serial also fit) >>> >>> Terminal 2: >>> minicom -D unix\#/tmp/console.sock >>> (type something and press enter) >>> C-a x (to exit) >>> >>> Do 3 times: >>> minicom -D unix\#/tmp/console.sock >>> C-a x >>> >>> It needs 4 connections, because the first one is accepted by QEMU, then two >>> are queued by >>> the kernel, and the 4th blocks. >>> >>> The problem is that QEMU doesn't add a read watcher after succesful read >>> until the guest device wants to acquire recieved data, so >>> I propose to install a separate pullhup watcher regardless of >>> whether the device waits for data or not. >>> >>> Signed-off-by: Klim Kireev <klim.kir...@virtuozzo.com> >>> --- >>> Changelog: >>> v2: Remove timer as a redundant feature >>> >>> v3: Remove read call and return G_SOURCE_REMOVE >>> >>> v4: Move to GSource API >>> >>> v5: Fix git typos >>> >>> chardev/char-socket.c | 22 ++++++++++++++++++++++ >>> 1 file changed, 22 insertions(+) >>> >>> diff --git a/chardev/char-socket.c b/chardev/char-socket.c >>> index 77cdf487eb..a340af6cd3 100644 >>> --- a/chardev/char-socket.c >>> +++ b/chardev/char-socket.c >>> @@ -42,6 +42,7 @@ typedef struct { >>> QIOChannel *ioc; /* Client I/O channel */ >>> QIOChannelSocket *sioc; /* Client master channel */ >>> QIONetListener *listener; >>> + GSource *hup_source; >>> QCryptoTLSCreds *tls_creds; >>> int connected; >>> int max_size; >>> @@ -352,6 +353,12 @@ static void tcp_chr_free_connection(Chardev *chr) >>> s->read_msgfds_num = 0; >>> } >>> >>> + if (s->hup_source != NULL) { >>> + g_source_destroy(s->hup_source); >>> + g_source_unref(s->hup_source); >>> + s->hup_source = NULL; >>> + } >>> + >>> tcp_set_msgfds(chr, NULL, 0); >>> remove_fd_in_watch(chr); >>> object_unref(OBJECT(s->sioc)); >>> @@ -455,6 +462,15 @@ static gboolean tcp_chr_read(QIOChannel *chan, >>> GIOCondition cond, void *opaque) >>> return TRUE; >>> } >>> >>> +static gboolean tcp_chr_hup(QIOChannel *channel, >>> + GIOCondition cond, >>> + void *opaque) >>> +{ >>> + Chardev *chr = CHARDEV(opaque); >>> + tcp_chr_disconnect(chr); >>> + return G_SOURCE_REMOVE; >>> +} >>> + >>> static int tcp_chr_sync_read(Chardev *chr, const uint8_t *buf, int len) >>> { >>> SocketChardev *s = SOCKET_CHARDEV(chr); >>> @@ -528,6 +544,12 @@ static void tcp_chr_connect(void *opaque) >>> tcp_chr_read, >>> chr, chr->gcontext); >>> } >>> + >>> + s->hup_source = qio_channel_create_watch(s->ioc, G_IO_HUP); >>> + g_source_set_callback(s->hup_source, (GSourceFunc)tcp_chr_hup, >>> + chr, NULL); >>> + g_source_attach(s->hup_source, chr->gcontext); >>> + >>> qemu_chr_be_event(chr, CHR_EVENT_OPENED); >>> } >>> >>> >> Queued, thanks. > Hi! > > I'm currently facing some issues with "make check -j4" (i.e. running the > tests in parallel). Git bisect blames this commit - though I'm not sure > whether this is really the right one ... Starting with this commit, I > saw hangs in test-filter-redirector,
Commit 8f6d701044bc87197aac8aa2a627d70ce8471726 fixes it Have you tried to apply it? > but git master rather seems to hang > with vhost-user-test instead. > Stefan also had issues with "make check -j4" today, so it's apparently > not only me ... can somebody else reproduce the problem with the git > master branch? > > Thomas