Re: [Spice-devel] give me some advise how to improve display effect and decrease client cpu usage
Hi, A recent information about firefox and video pass-through : Firefox 24 (22 was released this week) will have better html5/gstreamer support. As I understood it, gstreamer will be used for H264, MP4 and AAC decoding. I remember that the plans about video pass-through talked about intercepting (with the agent?) gstreamer streams : that would be a very interesting improvement regarding this new Firefox feature! However, I'm not sure how gstreamer is integrated with FF : does it send the stream to the screen/speaker, or is it sent back to Firefox, which is then responsible for the final rendering? In this last option, using gstreamer to send the stream to the client would be quite complicated... I hope this information will help! Frederic. - Mail original - > Hi, > On 06/18/2013 09:25 PM, bigclouds wrote: > > > > so much thanks Yonit: > > 1. > > i basicly understand some of your ideas, please give me more details > > about video pass throught, and hardware accelerate, > > i think it is hardware related, my hardware is 1G cpu, 2 core, AMD, does > > not has special hardware. > > if on android platform, i am not sure if which has such hardware. > video pass-through: with spice current implementation, videos are > decoded inside the guest, the rendered frames reach the server, and > then, if the server manages to classify them as video, the server > re-encodes them (MJPEG), and the client decodes them. > video pass-through means to stream the *original* encoded video (i.e., > you need to catch the video in driver level, or maybe using a > browser/player plugin), and then you decode the original video on the > client side. > This feature might raise some licensing issues, unless decoding of > proprietary codecs is done in the graphics driver of the client. > By hardware acceleration, I meant encoding/decoding using the GPU. Maybe > it will be interesting to compare gstreamer-vaapi jpeg decoding to > libjpeg-turbo decoding. But I'm not familiar with libva, and how common > are the graphics driver it supports. Maybe someone else from the team > have more information about this. > > 2. > > improved caching , if you mean 'display tree' or send queue? this > > part seems hard for me, spice use pixman directly, but there is no > > document about pixman. if you can ,please explain display architeture > > to me. > Caching: the server and client have a synchronized cache of bitmaps > (server hold bitmaps hash keys; client holds the the actual bitmaps). > When a bitmap is sent to the client, if it is already in the cache, only > its id is sent, instead of the whole bitmap. > There is --spice-cache-size option for remote-viewer. You can change the > size of the cache and see if it improves the performance. Then we can > consider implementing caching on disk, and not only memory. > > Cheers, > Yonit. > > thanks. > > > > > > > > > > At 2013-06-19 01:44:29,"Yonit Halperin" wrote: > >>Hi, > >>On 06/18/2013 10:33 AM, bigclouds wrote: > >> > >>> hi,all > >>> i have used spice for a long time,it is good, but in pratice it needs > >>> improvement, like resource usage, network bandwidth,performance. > >>> 1.it eats up much cpu on client side, up to 90% when play video. i am > >>> sure whether decompression or something else uses so much cpu. if > >>> client (remote-viewer) is corss-compiled to arm platform, what will > >>> happen? iordan has succellful compiled it to android platform. > >>> how to decrease cpu usage on client side?(if Mplayer which Fedor has > >>> mention will help ) > >>For improving video streaming cpu & network usage, it would be great to > >>have video pass-through and hardware accelerated decoding. > >>> 2.how to improve display effect and decrease network usage. > >>> now glz algorithm seems the best, but display stuck, and mosaic (video > >>> screen not sync) exist. > >>By default we use auto_glz. Quic is applied on images that are > >>classified as "real life" bitmaps. And GLZ is applied on images that are > >>artifical (e.g., mainly contain text). > >>> better network leads to larger traffic, some time it goes up to 7M/s. > >>> in future spice client will be often used in WAN , in that case the > >>> less trafic the better. > >>> guys , do you have some ideas? > >>Spice bandwidth usage is not deterministic, i.e., if you observe network > >>usage of 7M/s on LAN, it doesn't mean that spice-server will try to push > >>the same amount of data over WAN. Some of the bitmaps may be dropped if > >>possible. > >>For decreasing the bandwidth usage, you can think of features like > >>- improved caching (e.g., try to enlarge the cache size, and if it > >>helps, consider disk based caching). > >>- improved lossless compression (e.g., finding identical sub-regions > >>between bitmaps that are not necessarily "artificial" and compressed by > >>glz today; something like a combination of a dictionary-based encoding > >>and predictive encoding). > >>- Instead of setting TCP_NODELAY=0 ove
Re: [Spice-devel] usbredir and rights management
> The suid helper is a short-lived process, which gets invoked > after a new device has been plugged in, so it cannot differentiate > between newly plugged in and already present devices. Besides that > plugging in devices requires physical access, what is to stop a user > from unplugging and re-plugging a device he wants to get access to, > thereby making it a newly plugged in device? That's a good point :-) In my mind, "already present devices" represent internal USB devices, like internal hubs, fingerprint reader, internal webcam, or, in my laptop, the integrated "trusted platform processor". I'm not an expert, but I can't see any disadvantage in disabling access to all these devices : they are not necessary, and the more you restrict, the less harm the user will be able to do. In the last case (Trusted Platform Processor), maybe giving all access to the device would allow some form of attack, which would be harder with the above-mentionned option. Maybe the filter can cover this option, but the "block by default, authorize only new devices" solution look more secure/flexible to me. Frederic. > > Regards, > > Hans > ___ Spice-devel mailing list Spice-devel@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/spice-devel
Re: [Spice-devel] usbredir and rights management
> > As mentioned in my original mail, the helper uses PolicyKit to ask > for > permission to redirect the device, it is PolicyKit which asks for the > root password, not the helper. In the blog post I linked to are > instructions to change the policy so that local (so behind the > keyboard of the actual machine) users don't need to enter any > password at all. > > Making these kind of (security) policy decisions configurable is > exactly what PolicyKit is intended for. The root password asking > is caused by spice-gtk shipping with what I consider is a sane > default policy. Changing this is easy. > Sorry, I didn't see the link. It explains a lot. Still, I don't know how PolicyKit works (based on policy I imagine?), but it would be a good idea to add a policy allowing newly plugged USB devices (as opposed to devices already present at spice client startup) to be used in a different manner (and the admin can set it to "no password" if he wants to). This would mitigate the issue that you pointed out where "this will give any local users of your machine FULL access to any USB devices plugged in!" Frederic. > Regards, > > Hans > ___ Spice-devel mailing list Spice-devel@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/spice-devel
Re: [Spice-devel] usbredir and rights management
Hi Hans, Would it be possible to restrict this root helper to make it executable by any users of a specific group without forcing them to give the root password? I would think of something like helper (suid root) -> check user group -> opens up the device give and filter command to ensure that they are compliant with a "normal" access by spice client Anyway, I think that if the user plugs something into an USB port, it is quite logical to grant him "everything" on that device : after all, if he has admin rights inside the guest, spice cannot prevent anything... so the above would be : helper (suid root) -> if a new device is plugged, open it, give every access to the "current" user -> give it to spice client (or transmit commands) How does the desktop handle USB devices by the way : if I plug a USB key, what rights do I have on it? Is there some kind of filter about the commands I can issue? Anyway, I think it is sad to require a root password for USB pass-through since it will block lots of use cases : think about any "managed" environment where the user have minimal rights on their PCs, but can still plug USB keys, webcams, smartphones, or even some more exotic devices (USB-serial converters, authentication dongle for a professional software...) At least, if I get it right, the USB storage case should be dealt with a special layer working at filesystem level. This is a very good option for this case (probably very faster, and which provides good access control), but won't solve the problem of smartphones and dongles (I can't think of more important device that the "normal" user could plug, but those two look good...). there must be a solution to this... Frederic. PS : By the way, GSoC is open, maybe this would make a good subject? What do you think? You could even propose (independently) USB right management and USB storage pass-through for example? - Mail original - > Hi, > > On 02/07/2012 09:40 PM, Dominique Rodrigues wrote: > > Hi, > > > > I have compiled spice-gtk (version 0.9) to support usbredir > > (version 3.3 compiled and installed) on 2 type of OS : Mageia and > > Debian (wheezy). > > > > I appears that usb redirection involves root in different ways for > > these OS. For mageaia, you have to give root password to access to > > USB drives in a virtual machine, > > Yes, this is expected behavior redirecting usb devices requires > direct/raw access to the usb device. If we were > to open up the usb device nodes far enough that this would work > without requiring root rights any user > could do *anything* to *any* usb device, which seems a very poor > default. > > So we have a suid root helper which opens up the usb devices for > spice-gtk based clients, after it has > gotten permission to do so from policykit, you can change the policy > so that a root password is no > longer required, see: http://hansdegoede.livejournal.com/11936.html > Note that if you change the policy away from needing admin rights, > that once more any user can do > *anything* to *any* usb device! > > > while in debian root does not give any possibility (access is not > > allowed). > > Then your spice-gtk is likely compiled without policykit support, > re-run ./configure > and checks its outpu, you are likely missing some -dev packages > needed for the acl helper. > > Regards, > > Hans > ___ > Spice-devel mailing list > Spice-devel@lists.freedesktop.org > http://lists.freedesktop.org/mailman/listinfo/spice-devel > ___ Spice-devel mailing list Spice-devel@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/spice-devel
Re: [Spice-devel] Performance over VPN
> This sounds like you have mouse in server mode. Install spice-vdagent > in > the guest, make sure it is running and pleaser report back if it > solved > things for you You can also (in addition to that) force image compression by adding the appropriate parameter in your libvirt configuration (provided you use libvirt). Check out libvirt XML format for more information. By default, spice should detect your bandwith and set up those parameters appropriately, but if you're in a VPN, these detection can be biased by the VPN compression. Forcing them solves the problem. Frederic. ___ Spice-devel mailing list Spice-devel@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/spice-devel
Re: [Spice-devel] [PATCH spice-gtk 7/8] Check if msg are permitted to be sent in read-only
Is it expected that this commit breaks build?
CC spice-channel.lo
spice-channel.c: In function 'msg_check_read_only':
spice-channel.c:481:14: error: 'SPICE_MSGC_MAIN_MIGRATE_END' undeclared (first
use in this function)
I couldn't find any declaration of this SPICE_MSGC_MAIN_MIGRATE_END in any
other commit...
Thanks for your help,
Frederic.
- Mail original -
> ---
> gtk/spice-channel-priv.h |1 +
> gtk/spice-channel.c | 35 +--
> 2 files changed, 34 insertions(+), 2 deletions(-)
>
> diff --git a/gtk/spice-channel-priv.h b/gtk/spice-channel-priv.h
> index 86b22e1..df661f8 100644
> --- a/gtk/spice-channel-priv.h
> +++ b/gtk/spice-channel-priv.h
> @@ -46,6 +46,7 @@ struct _SpiceMsgOut {
> SpiceMessageMarshallers *marshallers;
> SpiceMarshaller *marshaller;
> SpiceDataHeader *header;
> +gboolean ro_check;
> };
>
> struct _SpiceMsgIn {
> diff --git a/gtk/spice-channel.c b/gtk/spice-channel.c
> index 99af206..c2133ab 100644
> --- a/gtk/spice-channel.c
> +++ b/gtk/spice-channel.c
> @@ -465,6 +465,30 @@ void spice_msg_out_hexdump(SpiceMsgOut *out,
> unsigned char *data, int len)
> hexdump(">> msg", data, len);
> }
>
> +static gboolean msg_check_read_only (int channel_type, int msg_type)
> +{
> +if (msg_type < 100) // those are the common messages
> +return FALSE;
> +
> +switch (channel_type) {
> +/* messages allowed to be sent in read-only mode */
> +case SPICE_CHANNEL_MAIN:
> +switch (msg_type) {
> +case SPICE_MSGC_MAIN_CLIENT_INFO:
> +case SPICE_MSGC_MAIN_MIGRATE_CONNECTED:
> +case SPICE_MSGC_MAIN_MIGRATE_CONNECT_ERROR:
> +case SPICE_MSGC_MAIN_ATTACH_CHANNELS:
> +case SPICE_MSGC_MAIN_MIGRATE_END:
> +return FALSE;
> +}
> +break;
> +case SPICE_CHANNEL_DISPLAY:
> +return FALSE;
> +}
> +
> +return TRUE;
> +}
> +
> G_GNUC_INTERNAL
> SpiceMsgOut *spice_msg_out_new(SpiceChannel *channel, int type)
> {
> @@ -476,6 +500,7 @@ SpiceMsgOut *spice_msg_out_new(SpiceChannel
> *channel, int type)
> out = spice_new0(SpiceMsgOut, 1);
> out->refcount = 1;
> out->channel = channel;
> +out->ro_check = msg_check_read_only(c->channel_type, type);
>
> out->marshallers = c->marshallers;
> out->marshaller = spice_marshaller_new();
> @@ -1547,6 +1572,12 @@ static void
> spice_channel_send_msg(SpiceChannel *channel, SpiceMsgOut *out, gboo
> g_return_if_fail(channel != NULL);
> g_return_if_fail(out != NULL);
>
> +if (out->ro_check &&
> +spice_session_get_read_only(channel->priv->session)) {
> +g_warning("Try to send message while read-only. Please
> report a bug.");
> +return;
> +}
> +
> data = spice_marshaller_linearize(out->marshaller, 0,
>&len, &free_data);
> /* spice_msg_out_hexdump(out, data, len); */
> @@ -1554,9 +1585,9 @@ static void spice_channel_send_msg(SpiceChannel
> *channel, SpiceMsgOut *out, gboo
> spice_channel_buffered_write(channel, data, len);
> else
> spice_channel_write(channel, data, len);
> -if (free_data) {
> +
> +if (free_data)
> free(data);
> -}
> }
>
> /* coroutine context */
> --
> 1.7.7
>
> ___
> Spice-devel mailing list
> Spice-devel@lists.freedesktop.org
> http://lists.freedesktop.org/mailman/listinfo/spice-devel
>
___
Spice-devel mailing list
Spice-devel@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/spice-devel
Re: [Spice-devel] [master PATCH 2/2] server: netstat: send random data not zeros
> > That may be overkill, but you could also "scramble" it by xoring it > > with a randomly generated sequence (and send the seed to the > > client to allow him to reproduce the sequence). The only overhead > > would thus be the seed. > > By the way, that could be applied to any message, not necessarily > > the first image. > > > > Frederic. > > Why should we do that? If we don't compress images well enough, and a > WAN accelerator (or even if the channel is within SSL with > compression > enabled), let them reduce the bandwidth for us. > Y. Of course, but I proposed that only for the "bandwith estimation" phase, to avoid data compression at that time. Still, if images are compressed by the WAN accelerator afterward, that may be a good thing to estimate bandwith with the compression. Frederic. ___ Spice-devel mailing list Spice-devel@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/spice-devel
Re: [Spice-devel] [master PATCH 2/2] server: netstat: send random data not zeros
> That seems like a good idea. > Note that if the first image is not compressed, we may have a similar > problem > with WAN accelerators. Also maybe it is better to always compress it. > > This random buffer was intended as a quick fix. > In the long run, we would like to measure network statistics > dynamically, e.g. > when sending large images (as you suggested for the first one). > That may be overkill, but you could also "scramble" it by xoring it with a randomly generated sequence (and send the seed to the client to allow him to reproduce the sequence). The only overhead would thus be the seed. By the way, that could be applied to any message, not necessarily the first image. Frederic. ___ Spice-devel mailing list Spice-devel@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/spice-devel
Re: [Spice-devel] spicec bug - only works on X display :0.0 (no multi-display support)
> maybe you have an alternative? > vncserver + vncclient on localhost ? ___ Spice-devel mailing list Spice-devel@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/spice-devel
Re: [Spice-devel] Using systemd/udev acl management to open up additional /dev nodes on request
hi Hans, Instead of using a privileged helper, wouldn't it be better to do it just like another virtualization solution (from the society that also licences a very-well known object-oriented programming language) does : create a special user group, add an udev rule that associates that group to the devices in /dev/bus/usb as they get plugged, and thus allow people of that group to use usb devices transparently? Of course, the main problem that it raises is that it will break that other virtualization solution's, since devices cannot be assigned to 2 different user groups... Frederic. - Mail original - > Hi, > > Currently when people want to use usbredirection to a virtual machine > from > spice-client, they must launch the spice-client as root so that it > can > access device nodes under /dev/bus/usb. > > Since the purpose is for usbredirection to just work plug and play > for > virtual machines, this needs to change. > > My plan is to write a (privileged) helper program which will: > 1) Check if it is invoked from a console session (using ConsoleKit > or the new ConsoleKit equivalent functionality in systemd in > F-16) > 2) Poke PolicyKit asking it if it is ok for the user to get access > to raw usb devices > 3) Do something to actually open up the device to the spice-client, > there are 2 options: > a) relax permissions (set an acl) > b) open the device node and hand over a fd, but since I'm using > libusb > to access the device nodes this is not really an option, leaving > only a. > > 3) Is a part where I've some systemd/udev questions about. Currently > udev already does similar opening up of acl's for the active console > user for things like soundcards, etc. I wonder if somehow I could > hook > into udev to make use of this for the usb device nodes (after having > done the policykit tests? > > Thanks & Regards, > > Hans > ___ > Spice-devel mailing list > Spice-devel@lists.freedesktop.org > http://lists.freedesktop.org/mailman/listinfo/spice-devel > ___ Spice-devel mailing list Spice-devel@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/spice-devel
Re: [Spice-devel] Missing key and mouse release events
I didn't fill a bug at that time, but you'll find every details here : (The last message has an attachment with the corresponding wireswhark dump) http://comments.gmane.org/gmane.comp.emulators.spice.devel/2948 - Mail original - > On Sat, Jul 23, 2011 at 03:47:30AM -0400, John A. Sullivan III wrote: > > Hello, all. This is happening enough now that I don't think it is > > my > > imagination. It seems that SPICE sometimes either does not detect > > that > > I have released a key or is slow in doing so. > > > > On quite a number of occasions, my input started behaving > > strangely. > > Menu options would not work, it appeared my CAPS lock was on but > > not > > quite. Then I realized that the SHIFT key was stuck. I pressed > > and > > released it again and normal functionality returned. Again, this > > was > > not a one-off event but happened several times. > > > > Just now, I was testing a strange but potentially common scenario > > in our > > environment where my SPICE client was running an X2Go session on > > the > > SPICE guest to a third system. Sometimes the scroll bars did not > > seem > > to realize I had released the mouse button. On occasion, I would > > press > > delete and an entire block of text would disappear or I'd press > > space > > and a row of spaces would appear. > > > > Has anyone else seen this behavior? I thought I would ask before > > opening > > it as a bug. Thanks - John > > > > Please open a bug. I remember seeing missed keys but not recently and > I > don't know how to reproduce, would be good if you provided the > details. > > > ___ > > Spice-devel mailing list > > Spice-devel@lists.freedesktop.org > > http://lists.freedesktop.org/mailman/listinfo/spice-devel > ___ > Spice-devel mailing list > Spice-devel@lists.freedesktop.org > http://lists.freedesktop.org/mailman/listinfo/spice-devel > ___ Spice-devel mailing list Spice-devel@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/spice-devel
Re: [Spice-devel] Out of surface in Xorg.0.log
Hi spice developpers, The problem came back, and won't leave It happens quickly when I start "digikam" (version 2.0.0-rc, built from source, works perfectly on an other physical PC). The guest is Fedora 15, everything up-to-date. I even use seabios version that Gerd suggested. As soon as digikam is launched, I see "bad bpp" messages (I think I launched it at time "100" in the following log) and, when the program is initialized, "Out of surfaces" messages flood the log. I don't know what digikam does, but I can say it's certainly not too much : it is a fresh install, without any collection yet! The flood starts at the "tip of the day" screen. I think libvirt log may be useful to, so I pasted it afterward. See Xorg.0.log : [52.550] primary is 0x1f62e40 [54.060]Bad bpp: 1 (1) [54.101]Bad bpp: 1 (1) [96.725]Zero width or height [ 100.723]Bad bpp: 1 (1) [ 100.723]Bad bpp: 1 (1) [ 101.551]Bad bpp: 1 (1) [ 101.551]Bad bpp: 1 (1) [ 101.555]Bad bpp: 1 (1) [ 101.555]Bad bpp: 1 (1) [ 101.560]Bad bpp: 1 (1) [ 101.560]Bad bpp: 1 (1) [ 101.564]Bad bpp: 1 (1) [ 101.564]Bad bpp: 1 (1) [ 101.568]Bad bpp: 1 (1) [ 101.568]Bad bpp: 1 (1) [ 101.572]Bad bpp: 1 (1) [ 101.572]Bad bpp: 1 (1) [ 101.575]Bad bpp: 1 (1) [ 101.576]Bad bpp: 1 (1) [ 101.582]Bad bpp: 1 (1) [ 101.582]Bad bpp: 1 (1) [ 101.585]Bad bpp: 1 (1) [ 101.585]Bad bpp: 1 (1) [ 101.588]Bad bpp: 1 (1) [ 101.588]Bad bpp: 1 (1) [ 101.592]Bad bpp: 1 (1) [ 101.592]Bad bpp: 1 (1) [ 101.710]Bad bpp: 1 (1) [ 101.710]Bad bpp: 1 (1) [ 101.937]Bad bpp: 1 (1) [ 101.943]Bad bpp: 1 (1) [ 104.492]Bad bpp: 1 (1) [ 104.492]Bad bpp: 1 (1) [ 110.228] Out of surfaces [ 110.979] Out of surfaces [ 111.039] Out of surfaces [ 111.164] Out of surfaces [ 111.187] Out of surfaces [ 111.212] Out of surfaces [ 111.236] Out of surfaces [ 111.260] Out of surfaces [ 111.284] Out of surfaces [ 111.307] Out of surfaces (doesn't really stop) libvirt qemu log : Note : "oom" number escalation happens exactly when the "Out of surfaces" flood starts. red_dispatcher_set_cursor_peer: handle_dev_input: cursor connect spice_server_add_interface: SPICE_INTERFACE_TABLET handle_dev_input: mouse mode 2 handle_dev_destroy_surfaces: handle_dev_destroy_surfaces: handle_dev_input: oom current 7 pipe 0 handle_dev_destroy_surfaces: handle_dev_input: oom current 34 pipe 0 handle_dev_input: oom current 25 pipe 0 handle_dev_input: oom current 830 pipe 0 handle_dev_input: oom current 1000 pipe 0 handle_dev_input: oom current 1000 pipe 0 handle_dev_input: oom current 1000 pipe 0 handle_dev_input: oom current 1000 pipe 0 (snip ... this line repeats roughly 60 times) handle_dev_input: oom current 1000 pipe 0 handle_dev_input: oom current 1000 pipe 0 handle_dev_input: oom current 1000 pipe 0 handle_dev_input: oom current 1000 pipe 0 handle_dev_input: oom current 1000 pipe 0 handle_dev_input: oom current 950 pipe 0 handle_dev_input: oom current 894 pipe 0 handle_dev_input: oom current 830 pipe 0 handle_dev_input: oom current 784 pipe 0 handle_dev_input: oom current 727 pipe 0 handle_dev_input: oom current 675 pipe 0 handle_dev_input: oom current 612 pipe 0 handle_dev_input: oom current 548 pipe 0 handle_dev_input: oom current 484 pipe 0 handle_dev_input: oom current 420 pipe 0 handle_dev_input: oom current 356 pipe 0 handle_dev_input: oom current 292 pipe 0 handle_dev_input: oom current 228 pipe 0 handle_dev_input: oom current 164 pipe 0 handle_dev_input: oom current 100 pipe 0 handle_dev_input: oom current 36 pipe 0 handle_dev_input: oom current 0 pipe 0 handle_dev_input: oom current 1 pipe 0 handle_dev_input: oom current 0 pipe 0 handle_dev_input: oom current 1 pipe 0 (then, there are lots of similar messages, with alternatively "current 1" and "current 0"). I tried to increase video size to 256Mb, but X doesn't start : Xorg.0.log says : [22.197] (II) qxl(0): Creating default Display subsection in Screen section "Default Screen Section" for depth/fbbpp 24/32 [22.197] (==) qxl(0): Depth 24, (--) framebuffer bpp 32 [22.197] (==) qxl(0): RGB weight 888 [22.197] (==) qxl(0): Default visual is TrueColor [22.197] (==) qxl(0): Using gamma correction (1.0, 1.0, 1.0) [22.236] (II) UnloadModule: "qxl" [22.236] (II) Unloading qxl [22.236] (EE) Screen(s) found, but none have a usable configuration. [22.236] Fatal server error: [22.236] no screens found [22.236] I hope there is a solution to this, so tell me if there is anything I can do to help! Frederic. - Mail original - > > > How does /proc/iomem look like with 128Mb? > > I couldn't log in to the guest, since the host was overloaded with > the MSI-X message. I should try with logging deactivated. > > Anyway, the problem occured again (still with 64Mb
Re: [Spice-devel] Out of surface in Xorg.0.log
> How does /proc/iomem look like with 128Mb? I couldn't log in to the guest, since the host was overloaded with the MSI-X message. I should try with logging deactivated. Anyway, the problem occured again (still with 64Mb), and here is iomem : - : reserved 0001-0009f3ff : System RAM 0009f400-0009 : reserved 000f-000f : reserved 0010-dfffcfff : System RAM 0100-0147e0ed : Kernel code 0147e0ee-01b4137f : Kernel data 01c34000-01da : Kernel bss dfffd000-dfff : reserved f000-f3ff : :00:02.0 f400-f7ff : :00:02.0 f800-f8001fff : :00:02.0 f801-f801 : :00:02.0 f802-f8020fff : :00:03.0 f802-f8020fff : virtio-pci f803-f803 : :00:03.0 f804-f8043fff : :00:04.0 f804-f8043fff : ICH HD audio f8044000-f8044fff : :00:05.0 f8044000-f8044fff : virtio-pci f8045000-f8045fff : :00:06.0 f8045000-f8045fff : virtio-pci fec0-fec003ff : IOAPIC 0 fed0-fed003ff : HPET 0 fee0-fee00fff : Local APIC fffbc000- : reserved 1-11fff : System RAM At the same time, libvirt log had about 10 "handle_dev_input: oom current X pipe 0" message per second. > Using the bios from http://spice-space.org/download/seabios/ might > help. I just downloaded it, and actually it helps : the vm starts without problem with 128Mb. It seems to be working fine after few minutes, including when I'm watching a video in youtube (this is my usual test...). By the way, I noticed that upscaling seems to work in windows guests, but not linuxes : is it expected? (actually, the video seen in linux is very slow in fullscreen and very smooth in normal mode, while it is always smooth with windows guests). Thanks for your help, Frederic. > > cheers, >Gerd > ___ Spice-devel mailing list Spice-devel@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/spice-devel
Re: [Spice-devel] Out of surface in Xorg.0.log
> [ 664.762] Out of surfaces > [ 664.802] Out of surfaces > I suspected that this was somewhat related to the resolution I did set in X : 1920x1080. I then set the VRAM size to 128Mb instead of 64Mb : unfortunately, the host was overloaded with the following message in libvirt log : "MSI-X: only dword read is allowed!" I tried 96Mb instead of 128, but I got the same problem. The only solution is to use 64Mb, but it runs out of "surfaces" then... Any help will be greatly appreciated. Frederic. ___ Spice-devel mailing list Spice-devel@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/spice-devel
[Spice-devel] Out of surface in Xorg.0.log
Hi all, I've been using spice for quite a long time, but I hit a problem quite recently. My setup is as follow : host : Fedora15+virt-preview guest : Fedora15+virt-preview, qxl drivers installed with standard yum, Gnome3 "in standard fallback mode" client : Fedora15+virt-preview, with vinagre everything up-to-date qemu command line : /usr/bin/qemu-kvm -S -M pc-0.14 -enable-kvm -m 4096 -smp 2,sockets=2,cores=1,threads=1 -name spice-f15 -uuid /**/ -nodefconfig -nodefaults -chardev socket,id=charmonitor,path=/var/lib/libvirt/qemu/spice-f15.monitor,server,nowait -mon chardev=charmonitor,id=monitor,mode=control -rtc base=utc -boot order=c,menu=off -device virtio-serial-pci,id=virtio-serial0,bus=pci.0,addr=0x5 -drive file=/dev/raid/vm_spice_f15,if=none,id=drive-virtio-disk0,boot=on,format=raw,cache=writeback -device virtio-blk-pci,bus=pci.0,addr=0x6,drive=drive-virtio-disk0,id=virtio-disk0 -netdev tap,fd=23,id=hostnet0 -device virtio-net-pci,netdev=hostnet0,id=net0,mac=/***/,bus=pci.0,addr=0x3 -chardev pty,id=charserial0 -device isa-serial,chardev=charserial0,id=serial0 -chardev spicevmc,id=charchannel0,name=vdagent -device virtserialport,bus=virtio-serial0.0,nr=1,chardev=charchannel0,id=channel0,name=com.redhat.spice.0 -usb -device usb-tablet,id=input0 -spice port=5935,addr=0.0.0.0,disable-ticketing -vga qxl - global qxl-vga.vram_size=67108864 -device intel-hda,id=sound0,bus=pci.0,addr=0x4 -device hda-duplex,id=sound0-codec0,bus=sound0.0,cad=0 -device virtio-balloon-pci,id=balloon0,bus=pci.0,addr=0x7 I noticed that after some use time, UI becomes suddently very slow. I was using Firefox in the guest, but closing it didn't change anything. I finally found that the Xorg.0.log shows lots "Out of surface" (in the worst case, more than 10/seconds). A long time ago in the log, I found this (by searching for the first occurence of "Out of Surface") : [ 658.897] - 1th attempt [ 658.897] - OOM at 1846 46 24 [ 658.897] Cache contents: null null null null null null null null null null null null null null null null null null null null null null null null null null null null null null null null null null null null null null null null null null null null null null 836 828 864 874 847 895 1016 890 857 893 878 896 858 866 835 1014 921 843 total: 18 [ 658.907] - 2th attempt [ 658.907] - OOM at 1846 46 24 [ 658.907] Cache contents: null null null null null null null null null null null null null null null null null null null null null null null null null null null null null null null null null null null null null null null null null null null null null null 836 828 864 874 847 895 1016 890 857 893 878 896 858 866 835 1014 921 843 total: 18 [ 658.918] - 3th attempt [ 658.918] - OOM at 1846 46 24 [ 658.918] Cache contents: null null null null null null null null null null null null null null null null null null null null null null null null null null null null null null null null null null null null null null null null null null null null null null 836 828 864 874 847 895 1016 890 857 893 878 896 858 866 835 1014 921 843 total: 18 [ 658.930] Out of video memory: Could not allocate 347048 bytes [ 662.915] - 0th attempt [ 662.915] - OOM at 1846 46 24 [ 662.915] Cache contents: null null null null null null null null null null null null null null null null null null null null null null null null null null null null null null null null null null null null null null null null null null null null null null 836 828 864 874 847 895 1016 890 857 893 878 896 858 866 835 1014 921 843 total: 18 [ 662.926] - 1th attempt [ 662.926] - OOM at 1846 46 24 [ 662.926] Cache contents: null null null null null null null null null null null null null null null null null null null null null null null null null null null null null null null null null null null null null null null null null null null null null null 836 828 864 874 847 895 1016 890 857 893 878 896 858 866 835 1014 921 843 total: 18 [ 662.936] - 2th attempt [ 662.936] - OOM at 1846 46 24 [ 662.936] Cache contents: null null null null null null null null null null null null null null null null null null null null null null null null null null null null null null null null null null null null null null null null null null null null null null 836 828 864 874 847 895 1016 890 857 893 878 896 858 866 835 1014 921 843 total: 18 [ 662.948] - 3th attempt [ 662.948] - OOM at 1846 46 24 [ 662.949] Cache contents: null null null null null null null null null null null null null null null null null null null null null null null null null null null null null null null null null null null null null null null null null null null null null null 836 828 864 874 847 895 1016 890 857 893 878 896 858 866 835 1014 921 843 total: 18 [ 662.961] - 4th attempt [ 662.9
Re: [Spice-devel] spice-gtk : magic key options
At last, I updated my client to Fedora 15, which lets me go further than Ubuntu
regarding spice client usage...
As you suggested, I tried vinagre instead of spic{y,ec} : it is quite better,
especially regarding usability thanks to the bookmarks, the popup you talked
about, etc... I haven't got any problem with dual-screen yet, since my
secondary monitor just died yesterday! It'll be for another time...
But (and there is a but...), the problem with "Alt" is still here!
In particular alt+click (does nothing) and alt+space+click (opens vinagre's
window menu).
Since I thought it was mostly related to the "drag window" functionality of the
window manager (in this case, mutter, under Ubuntu it was metacity), I found
this message :
https://bbs.archlinux.org/viewtopic.php?id=118585
Quote :
>The controls for changing the keyboard layout is a little hidden in Gnome 3.
>Pull up the Activities Menu, search for Region and Languages.
>Next select Layout tab, and then select Options.
>There, you can change Alt/Win Key behavior.
The only satisfying option is to choose "Left Alt key is switched with left
logo key". At last, Alt works as expected inside vinagre, but the main problem
is that now, I have to use "Logo+Tab" when I want to switch windows in Gnome,
and I am really not used to that...
Maybe an option would be to globally disable Alt key behaviour whenever the
focus is inside the guest, and re-enable it afterward, but this needs to be
robust... (otherwise, if vinagre crashes the user looses control over its
windows...)
I'm adding David King (maintainer of Vinagre) in this discussion since he may
be interested with it, and he may have an advice on the question.
I'll keep trying other solutions but at least I can zoom out in PS now!
I will add a bug about it in bugzilla : do you think it is better in vinagre or
spice-gtk ?
Frederic.
- Mail original -
> Hi Frédéric,
>
> Although it looks like we have more problems than before, they were
> mostly expected. The primary goal of spice-gtk is not to be a
> client, but rather a client widget. spicy is only our experimental
> client that help us to quickly evaluate a change API, feature or
> behaviour. There is not yet a fully featured (and solid) spice-gtk
> based client. The best is still spicy, for now.. Vinagre+Spice isn't
> bad either, but it lacks multi-head (look, it has the pop-up menu
> you talked about!), and we have plan to make its UI better
> (https://live.gnome.org/Design/Apps/RemoteAccess).
>
> > -I had to go past the warnings about new options (I'm still on
> > Ubuntu,
> > so --with-gtk2.0 helps, and --disable-werror since there is no
> > libjpeg-turbo ppa available yet), but that's done.
>
> Regarding the warning during compilation, that's what we wanted, make
> 3.0 default, and jpeg-turbo *strongly* recommended (I think it
> reduced by x2 the cpu of mjpeg video decoding)
>
> Regarding key handling, the widget is embedded inside a widget tree.
> As such it cannot fully control the behavior of application key
> handling. So it is left to the application (spicy or virt-manager
> for example), to deal with key bindings, configuration, grabbing
> cooperation etc.. There is room for improvements.
>
> > -now that it works : Alt+space still pops up the menu (not the file
> > menu, but that one with the reduce, restore, move window options)
>
> That shouldn't happen if the mouse is over the widget. I think it
> shouldn't ungrab & reactivate accels&mnemonics on mouse leave. Would
> that work in practise? Somebody have an opinion?
>
> > -Alt alone does the same as before : if I'm in panning mode in PS,
> > the
> > cursor changes to "zoom out" as expected, but then nothing happens
> > if
> > I click or click+drag
>
> It seems the window manager still intercepts the mouse, although
> spice-gtk should have the grab. It fails at least for
> alt+right-click (on f15/gnome-shell, with spic{y,ec}). Can you try
> also with spicec? I don't have PS.
>
> > -once in fullscreen, nothing goes out of the client! That means
> > that
> > Shift+F11 does not go back to window mode... Fortunately I am in
> > dual
> > screen (guest has single screen), so I moved the mouse away to the
> > second monitor, hit alt+tab to activate the client again, and only
> > then, Alt+F4 or shift+F11 work.
>
> Actually Shift+F12 should leave fullscreen: spicy notice the ungrab
> combination has been pressed and leave fullscreen. This is perhaps
> not ideal, and we have to decide if spice-gtk widget should handle
> the various key bindings, that is normally handled also by the
> application... With your idea of magic key, we could make it easy
> for an application to create custom bindings without being too
> intrusive.
>
> > -Actually, I noticed that when the mouse is grabbed (even in window
> > mode), shift+F11 does not work... It only work when the mouse is
> > outside the guest area
>
> Correct, same reason, we drop all accels & mnemonics from the app. So
> we need what I
Re: [Spice-devel] spice-gtk : magic key options
Thanks Marc-André for your answers. With your advice, I just tested latest git version, and had some problems : -I had to go past the warnings about new options (I'm still on Ubuntu, so --with-gtk2.0 helps, and --disable-werror since there is no libjpeg-turbo ppa available yet), but that's done. -now that it works : Alt+space still pops up the menu (not the file menu, but that one with the reduce, restore, move window options) -Alt alone does the same as before : if I'm in panning mode in PS, the cursor changes to "zoom out" as expected, but then nothing happens if I click or click+drag -once in fullscreen, nothing goes out of the client! That means that Shift+F11 does not go back to window mode... Fortunately I am in dual screen (guest has single screen), so I moved the mouse away to the second monitor, hit alt+tab to activate the client again, and only then, Alt+F4 or shift+F11 work. -Actually, I noticed that when the mouse is grabbed (even in window mode), shift+F11 does not work... It only work when the mouse is outside the guest area -I'm not sure if that was the case before, but now, the window has always the same size at opening (and not very big... actually fits exactly with 640x480 in the guest). If I was previously in fullscreen for example, the window opens small, and the agent fits the guest's resolution to 640x480 : not very handy... At client opening, it would be a good idea to try to match the window to the guest resolution, eventually by directly going to fullscreen. -I found the configuration file, but not the option to put into it... -Using a magic key as suggested, I hope you'll put shortcut for all the other options. In particular, it would be handy to be able to disable the "match guest resolution to window size" once in fullscreen : that way, the guest's resolution matches the client fullscreen resolution when it goes fullscreen, and then sticks to it... FYI, in VB there is a popup menu (at the top or bottom of the screen, configurable) that opens when the mouse reaches the border. It contains all options related to the client : de/activate this kind of option, plug/unplug devices, etc... That's a lot of suggestion for tonight, so I'll come back later to you if necessary! Anyway, I'll keep testing and feedbacking you! Frederic. - Mail original - > Hi Frédéric, > > - Original Message - > > works very well over 100Mbit network, 1440*1280 fullscreen), some > > key > > combinations don't work well. In particular, alt works alone > > (client's > > cursor gets changed), but you can't press another key together with > > it > > nor click (zoom out in photoshop is with alt+space+click, or > > alt+click > > when already in pan mode, which doesn't work either). When I try > > that, > > the gtk client's menu appears, and that is very frustrating... > > > That's my first point : at the moment, only shift+F11/F12 are > > supposed > > to interact with the outside, but combinations involving Alt > > interfere. > > Hopefully, this is fixed in git. We now disable all Gtk mnemonics & > accels. If you have a chance, please give it a try and tell us if > the problem persist. > > > My second point would be about a configuration option : I used to > > use > > Virtualbox, and there is a possibility to choose the key that gets > > intercepted. Using specifically "Right shift" (which is, otherwise, > > almost never used), I solved all my problems : I'm not sure if > > there > > is already a configuration framework for storing user-defined > > options > > in spicy, but maybe that would be a good thing... > > spice-gtk have support for custom grab key combination. It is used by > virt-manager "Preferences" for instance. Spicy however, doesn't have > configuration UI for it. Adding it to spicy.ini configuration file > should be easy, if you don't mind editing ini files. I'll keep that > in mind before the next release. Feel free to open a bug in > bugzilla.freedesktop.org. > > > I talked only about a single "magic key", regarding how VB works : > > the > > base key (or combination, it can be "Alt+Ctrl" for example) is > > defined, and gives access to all shortcuts : key+F for fullscreen, > > "magic key"+other key for the other functions > > That seems neat, indeed. I very much like it, and I might well try to > implement that for spicy. > > > Sorry for not helping much in this project that I really believe > > in, > > but I hope that at least my user feedback will help you! > > Your feedback is great, thanks a lot! > > -- > Marc-André Lureau > > ___ Spice-devel mailing list Spice-devel@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/spice-devel
[Spice-devel] spice-gtk : magic key options
Hi All, I'm using spicy more and more, and all I can say is that combined with spice agent, it's getting more and more pleasant! I would like to make a suggestion regarding the key combinations grabbed by the client : when using photoshop for example (Yes, it works very well over 100Mbit network, 1440*1280 fullscreen), some key combinations don't work well. In particular, alt works alone (client's cursor gets changed), but you can't press another key together with it nor click (zoom out in photoshop is with alt+space+click, or alt+click when already in pan mode, which doesn't work either). When I try that, the gtk client's menu appears, and that is very frustrating... That's my first point : at the moment, only shift+F11/F12 are supposed to interact with the outside, but combinations involving Alt interfere. My second point would be about a configuration option : I used to use Virtualbox, and there is a possibility to choose the key that gets intercepted. Using specifically "Right shift" (which is, otherwise, almost never used), I solved all my problems : I'm not sure if there is already a configuration framework for storing user-defined options in spicy, but maybe that would be a good thing... I found that "right shift" is a very good option, but I know it can't be universal, since some keyboard may not make the difference between the two, and some may even have no right shift key... I talked only about a single "magic key", regarding how VB works : the base key (or combination, it can be "Alt+Ctrl" for example) is defined, and gives access to all shortcuts : key+F for fullscreen, "magic key"+other key for the other functions Sorry for not helping much in this project that I really believe in, but I hope that at least my user feedback will help you! Frederic. ___ Spice-devel mailing list Spice-devel@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/spice-devel
Re: [Spice-devel] Multi-monitor
The last time this question was asked on this mailing list, the answer was that QXL driver for xorg does not (yet...) allow dual monitor setup. I doubt it has changed since. Frederic. - Mail original - > i'm having the same issue for a while and i haven't found anything > on google or spice documents. anyone here has found the solution? > > cheers, > al > > > > ___ > Spice-devel mailing list > Spice-devel@lists.freedesktop.org > http://lists.freedesktop.org/mailman/listinfo/spice-devel > ___ Spice-devel mailing list Spice-devel@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/spice-devel
[Spice-devel] spice-gtk 0.6 : Ctrl+W closes the client
Hi all, I finally decided to give a try to the gtk version of spice client. First, I have to say that I'm quite impressed (and I understand why it tends to become the baseline client). Still, there is one bug that I would consider blocking (as far as I'm concerned) : I'm pretty used to the "Ctrl+W" combination to close windows, firefox or eclipse tabs, etc... However, is it expected that it is grabbed by the client itself (i.e. not the guest) and that pressing Ctrl+W closes the whole client window? I would expect the following : if focus is on toolbar, menu, etc, Ctrl+W closes the client, while if focus is on the guest (i.e. mouse over the canvas), the Ctrl+W is sent along to the guest, and spicy client doesn't close. Thanks for taking that into account if possible! Frederic. ___ Spice-devel mailing list Spice-devel@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/spice-devel
Re: [Spice-devel] Bug : KeyEvent up not sent?
Just tried again : I launched my F14 guest, and could trigger the bug after the 3rd key. I pressed 1, 2, 3, just enough time to have "1112233...", and at 3, it kept writing while I released. I have a wireshark session of it, but don't really know what to do with it (I know there is work on a wireshark dissector, but I don't think it has been pushed upstream yet, since I can't find "spice" in the protocol list). Frederic. -- frede...@grelot.net Tél : 09 54 76 14 26 / 06 82 23 46 17 while(!asleep()) sheep++ - Mail original - > Hi all, > > I've not been able to provide more information about this bug, but it > happened again today, using : > -same server > -wired ethernet (Gb, 1 switch only, never had any problems) > -recent client from git (cloned, say 1 week ago) > -latest qemu/spice from virt-preview repo > -more decent (i.e. powerfull) client hardware > -F14 and F15 alpha (x86_64) guests, even got the bug during F15_alpha > install (anaconda) > > The problem was identical : I press one key, and it stays pressed > although I released it. > I just thought about this : I don't remember using the mouse at the > same "time", but can not say for sure. > > Hope this helps! > > Could you give an idea of the amount of traffic between spice client > and server? the bug does not occur really often, so maybe I should > setup wireshark, but I don't know if it will hold... > > Thanks for your help (and your work!) > > Frederic. > > > > - Mail original - > > On Thu, Feb 24, 2011 at 11:45:15PM +0100, Frédéric Grelot wrote: > > > Hi all, > > > > > > I'm currently using Spice to develop on a remote computer (more > > > powerfull than my laptop, which makes me spare quite a lot of > > > time > > > waiting for the builds!), and I come to a quite annoying bug when > > > I use it : relatively often, some event seem not to be sent. > > > The most noticeable is a keyevent up, since the character keeps > > > printing, but I cannot say that other events are not dropped. I > > > waited for few seconds and the event is definitely lost : the > > > only > > > way to stop the characters from printing (or removing if I > > > pressed > > > backspace!) is to press/release another key. > > > I don't really know how to get more information about this bug : > > > there is nothing in the logs (client, server, and tail -f > > > /var/log/* on the guest shows nothing). > > > My versions are as follow : > > > -server : git, spice.kvm.v28, fedora 14, all updates > > > -client : built from git, 0.7.3, Ubuntu 10.10, all updates > > > -guest : fedora 14, all updates > > > > > > This is quite annoying, since it happens almost once every 20 key > > > pressed I would say, which is quite a lot! > > > It happens whether I am in fullscreen or not. > > > > > > Tell me if you need more information about this, I would be > > > pleased > > > to help! > > > > What kind of network? I mean, can you reproduce this with > > localhost? > > (client == server) Or can you create > > some synthetic environment to trigger it? Is it a specific key (ok, > > I'm reaching)? Does it happen when the guest > > is in X only, or also console? > > > > > > Frederic. > > > ___ > > > Spice-devel mailing list > > > Spice-devel@lists.freedesktop.org > > > http://lists.freedesktop.org/mailman/listinfo/spice-devel > > > ___ > Spice-devel mailing list > Spice-devel@lists.freedesktop.org > http://lists.freedesktop.org/mailman/listinfo/spice-devel > ___ Spice-devel mailing list Spice-devel@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/spice-devel
Re: [Spice-devel] Bug : KeyEvent up not sent?
Hi all, I've not been able to provide more information about this bug, but it happened again today, using : -same server -wired ethernet (Gb, 1 switch only, never had any problems) -recent client from git (cloned, say 1 week ago) -latest qemu/spice from virt-preview repo -more decent (i.e. powerfull) client hardware -F14 and F15 alpha (x86_64) guests, even got the bug during F15_alpha install (anaconda) The problem was identical : I press one key, and it stays pressed although I released it. I just thought about this : I don't remember using the mouse at the same "time", but can not say for sure. Hope this helps! Could you give an idea of the amount of traffic between spice client and server? the bug does not occur really often, so maybe I should setup wireshark, but I don't know if it will hold... Thanks for your help (and your work!) Frederic. - Mail original - > On Thu, Feb 24, 2011 at 11:45:15PM +0100, Frédéric Grelot wrote: > > Hi all, > > > > I'm currently using Spice to develop on a remote computer (more > > powerfull than my laptop, which makes me spare quite a lot of time > > waiting for the builds!), and I come to a quite annoying bug when > > I use it : relatively often, some event seem not to be sent. > > The most noticeable is a keyevent up, since the character keeps > > printing, but I cannot say that other events are not dropped. I > > waited for few seconds and the event is definitely lost : the only > > way to stop the characters from printing (or removing if I pressed > > backspace!) is to press/release another key. > > I don't really know how to get more information about this bug : > > there is nothing in the logs (client, server, and tail -f > > /var/log/* on the guest shows nothing). > > My versions are as follow : > > -server : git, spice.kvm.v28, fedora 14, all updates > > -client : built from git, 0.7.3, Ubuntu 10.10, all updates > > -guest : fedora 14, all updates > > > > This is quite annoying, since it happens almost once every 20 key > > pressed I would say, which is quite a lot! > > It happens whether I am in fullscreen or not. > > > > Tell me if you need more information about this, I would be pleased > > to help! > > What kind of network? I mean, can you reproduce this with localhost? > (client == server) Or can you create > some synthetic environment to trigger it? Is it a specific key (ok, > I'm reaching)? Does it happen when the guest > is in X only, or also console? > > > > Frederic. > > ___ > > Spice-devel mailing list > > Spice-devel@lists.freedesktop.org > > http://lists.freedesktop.org/mailman/listinfo/spice-devel > ___ Spice-devel mailing list Spice-devel@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/spice-devel
[Spice-devel] handle_dev_input: oom current 0 pipe X
Hi all, I'm using spice with Fedora 14 host, but I'm having some problems since the last time I updated (I think). I'm using virt-preview repository, and recently updated to latest qemu (previously, I was using my own built version, because of the bug regarding an assert in kvm code). When the guest is Fedora 14, as soon as the X server is started, I get "handle_dev_input: oom current 0" errors on the console (lots). If I try to open a session, I get more and more, and eventually the X server crashes (it always ends like that, sometimes after 2 seconds, sometimes 2 minutes... anyway it's not hours). The above mentionned errors are on host side, and I didn't noticed anything wrong on guest logs. It has "decent" qxl drivers (updated drivers from fedora repositories). I'm using libvirt/virt-manager. I tried to increase the video memory size (by editing the xml file), but it doesn't change the qemu command line anyway (libvirt bug?) The bug triggers even without any client connected. It disappears if I use Cirrus card instead of QXL. I haven't got any problem with my Windows XP guest. If you need more details I'll be happy to provide them, and hope you'll be able to help me (or, at least, solve a bug thanks to my report!) By the way, the command line is : /usr/bin/qemu-kvm -S -M pc-0.13 -enable-kvm -m 2048 -smp 1,maxcpus=4,sockets=4,cores=1,threads=1 -name spice-f14-i386 -uuid 21b91c42-9c60-1b95-e0bc-4f938a1e23c6 -nodefconfig -nodefaults -chardev socket,id=charmonitor,path=/var/lib/libvirt/qemu/spice-f14-i386.monitor,server,nowait -mon chardev=charmonitor,id=monitor,mode=control -rtc base=utc -boot c -drive file=/dev/vg_raid10/vm_spice_f14_root,if=none,id=drive-virtio-disk0,boot=on,format=raw,cache=writeback -device virtio-blk-pci,bus=pci.0,addr=0x4,drive=drive-virtio-disk0,id=virtio-disk0 -netdev tap,fd=28,id=hostnet0 -device virtio-net-pci,netdev=hostnet0,id=net0,mac=52:54:00:03:54:16,bus=pci.0,addr=0x3 -usb -device usb-tablet,id=input0 -spice port=5932,addr=0.0.0.0,disable-ticketing -vga qxl -device virtio-balloon-pci,id=balloon0,bus=pci.0,addr=0x5 Thanks for your help. Frederic. ___ Spice-devel mailing list Spice-devel@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/spice-devel
Re: [Spice-devel] smartcard usage
You've given lots of explanation about what you did and now it works! : if you have time, it may be a good idea to put that into a wiki page at http://www.spice-space.org/page/Main_Page since you already made most of the work :-) Frederic. - Mail original - > On 03/01/2011 10:00 AM, william wrote: > > On 03/01/2011 08:13 AM, william wrote: > >> On 03/01/2011 12:23 AM, Robert Relyea wrote: > >>> On 02/28/2011 08:34 AM, william wrote: > On 02/26/2011 08:49 PM, Alon Levy wrote: > > On Fri, Feb 25, 2011 at 12:06:33PM +0100, william wrote: > >> On 02/24/2011 08:10 PM, Alon Levy wrote: > >>> On Thu, Feb 24, 2011 at 05:46:33PM +0100, william wrote: > On 02/24/2011 05:09 PM, Alon Levy wrote: > > On Thu, Feb 24, 2011 at 04:28:13PM +0100, william wrote: > >> On 02/24/2011 12:09 PM, Alon Levy wrote: > >>> On Thu, Feb 24, 2011 at 10:17:21AM +0100, > >>> k...@cobradevil.org > >>> wrote: > Dear list, > > i have tried to get smartcard support running but i'm a > bit > lost :) > probably because it's not finished yet. > > we have smartcards with certificates like us dod and i > would > like to use > those from a client on a remote server for > authentication and > such. > I have followed the build instructions: > http://spice-space.org/page/Building_Instructions on a > ubuntu > system and > have managed to get those compiled. > > But when i try to start a vm with smartcard passthrough > it > asks me to give > a driver name? > > ./x86_64-softmmu/qemu-system-x86_64 -chardev > socket,server,host=0.0.0.0,port=2001,id=ccid,nowait > -device > ccid-card-passthru,chardev=ccid -drive > file=/var/lib/libvirt/images/test.img,if=ide -soundhw > ac97 -L > pc-bios > -nographic -vga qxl -spice port=5930,disable-ticketing > -usbdevice tablet > -enable-kvm -m 512 > > do_spice_init: starting 0.6.3 > spice_server_add_interface: SPICE_INTERFACE_KEYBOARD > spice_server_add_interface: SPICE_INTERFACE_MOUSE > spice_server_add_interface: SPICE_INTERFACE_QXL > red_worker_main: begin > spice_server_add_interface: SPICE_INTERFACE_RECORD > spice_server_add_interface: SPICE_INTERFACE_PLAYBACK > qemu-system-x86_64: -device > ccid-card-passthru,chardev=ccid: > Parameter > 'driver' expects a driver name > Try with argument '?' for a list. > > Am i starting the vm the right way or am i missing > something? > >>> You are doing the right steps with the wrong qemu. To be > >>> explicit: qemu hasn't > >>> accepted the patches for the smartcard devices yet, so I > >>> don't > >>> know where you > >>> got the qemu executable but unless you built it by hand > >>> and > >>> applied the patches > >>> on the list, or easier used the pull url I provide in the > >>> patches I sent (like v20 > >>> git://anongit.freedesktop.org/~alon/qemu usb_ccid.v20) > >>> you > >>> won't have them. > >>> > >>> Alon > >>> > >> Sorry for the priv mail :( > >> i can start the vm now with the usb_ccid.v19 git 20 gives > >> me > >> compile errors > >> > >> ./x86_64-softmmu/qemu-system-x86_64 -chardev > >> socket,server,host=0.0.0.0,port=2001,id=ccid,nowait > >> -device > >> usb-ccid > >> -device ccid-card-passthru,chardev=ccid -drive > >> file=/var/lib/libvirt/images/test.img,if=ide -soundhw > >> ac97 -L > >> pc-bios -nographic -spice port=5930,disable-ticketing > >> -usbdevice > >> tablet -enable-kvm -m 512 -device > >> virtio-net-pci,vlan=0,id=net0,mac=52:54:00:f4:f5:0b -net > >> user > >> do_spice_init: starting 0.7.3 > >> spice_server_add_interface: SPICE_INTERFACE_KEYBOARD > >> spice_server_add_interface: SPICE_INTERFACE_MOUSE > >> spice_server_add_interface: SPICE_INTERFACE_RECORD > >> spice_server_add_interface: SPICE_INTERFACE_PLAYBACK > >> spice_server_add_interface: SPICE_INTERFACE_QXL > >> red_worker_main: begin > >> handle_dev_input: start > >> > >> I also installed spice 0.7.3 > >> > >> When starting the spicec client i can connect but how can > >> i > >> share > >> say a local dev
Re: [Spice-devel] Bug : KeyEvent up not sent?
- Mail original - > > > > What kind of network? I mean, can you reproduce this with > > localhost? > > (client == server) Or can you create > > some synthetic environment to trigger it? Is it a specific key (ok, > > I'm reaching)? Does it happen when the guest > > is in X only, or also console? > > Network is a Wifi, I cannot test client==server (server is > headless...), but I will test with an other client plugged through > ethernet. It is not for a specific key, and I did not test it in the > console. I tried on another client, over ethernet (directly connected to the server through a Gigabit Switch, no expected loss) : the bug still happens. I did not trigger it in the console however (although I did not spend so much time in the console : I'm really not used to vi... But by comparison with my tests over X, the bug should have appeared I think). About you question "is it a specific key" : I can say that I noticed that (at least) for the following : -normal key press -baskspace -return -Ctrl+Z (lots of edits got cancelled at once!) -Ctrl+Y (same thing : edits got resumed) -Alt+arrow left/right (in eclipse : previous/next location) I hope this will help! Frederic. > > Frederic. > > > > > > > Frederic. > > > ___ > > > Spice-devel mailing list > > > Spice-devel@lists.freedesktop.org > > > http://lists.freedesktop.org/mailman/listinfo/spice-devel > > > ___ > Spice-devel mailing list > Spice-devel@lists.freedesktop.org > http://lists.freedesktop.org/mailman/listinfo/spice-devel > ___ Spice-devel mailing list Spice-devel@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/spice-devel
Re: [Spice-devel] Bug : KeyEvent up not sent?
> What kind of network? I mean, can you reproduce this with localhost? > (client == server) Or can you create > some synthetic environment to trigger it? Is it a specific key (ok, > I'm reaching)? Does it happen when the guest > is in X only, or also console? Network is a Wifi, I cannot test client==server (server is headless...), but I will test with an other client plugged through ethernet. It is not for a specific key, and I did not test it in the console. Frederic. > > > > Frederic. > > ___ > > Spice-devel mailing list > > Spice-devel@lists.freedesktop.org > > http://lists.freedesktop.org/mailman/listinfo/spice-devel > ___ Spice-devel mailing list Spice-devel@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/spice-devel
[Spice-devel] Bug : KeyEvent up not sent?
Hi all, I'm currently using Spice to develop on a remote computer (more powerfull than my laptop, which makes me spare quite a lot of time waiting for the builds!), and I come to a quite annoying bug when I use it : relatively often, some event seem not to be sent. The most noticeable is a keyevent up, since the character keeps printing, but I cannot say that other events are not dropped. I waited for few seconds and the event is definitely lost : the only way to stop the characters from printing (or removing if I pressed backspace!) is to press/release another key. I don't really know how to get more information about this bug : there is nothing in the logs (client, server, and tail -f /var/log/* on the guest shows nothing). My versions are as follow : -server : git, spice.kvm.v28, fedora 14, all updates -client : built from git, 0.7.3, Ubuntu 10.10, all updates -guest : fedora 14, all updates This is quite annoying, since it happens almost once every 20 key pressed I would say, which is quite a lot! It happens whether I am in fullscreen or not. Tell me if you need more information about this, I would be pleased to help! Frederic. ___ Spice-devel mailing list Spice-devel@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/spice-devel
[Spice-devel] Difficulty in building qemu with spice enabled
Hello everyone, Because of a bug I mentioned on the v...@lists.fedoraproject.org mailing list (available here : http://lists.fedoraproject.org/pipermail/virt/2011-February/002539.html ), I tried to manually build my own copy of qemu-kvm, with spice enabled. I'm not sure if that is a "./configure" pre-requisite problem, but I spent few hours before understanding the following message : ./configure --enable-spice ERROR ERROR: User requested feature spice ERROR: configure was not able to find it ERROR So, for information to anyone encountering the same problem : this does not mean that your source repository hasn't got the feature, but this means that pkg-config could not find the "spice-server" description file. In my case, I had to set "PKG_CONFIG_PATH=/usr/local/lib/pkg-config" before launching configure. Maybe this information can find a place on the wiki... I'll see if I have time to do it tomorrow. Frederic. PS : By the way, after building qemu by myself (tag spice.kvm.v28 from git), the above-mentioned bug was solved. ___ Spice-devel mailing list Spice-devel@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/spice-devel
Re: [Spice-devel] [Qemu-devel] paravirtual mouse/tablet
> So it'd end up being (x,y,pressure) N times (I think 16 is fine for > the foreseeable future). It's highly speculative, but Microsoft just released Surface 2, and it is able to handle up to 40 different pressure points... It is aimed at people in reunion meeting around a table and manipulation/exchanging documents with each other... If there is no hard/soft limit at 16, maybe it would be a good idea to allow some more from the beginning... Frederic. ___ Spice-devel mailing list Spice-devel@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/spice-devel
Re: [Spice-devel] RFC; usb redirection protocol
> > Do you know whenever certain low-level usb ops can work with this? > > > > I expect most usb devices to work with this, I don't know about > really weird ones. > > > Specifically iphone firmware flashing was mentioned on the list. > > > > I think that should work, but that is an interesting case about which > I don't know enough yet. If that can help you, I may try to capture the usb activities while doing the next firmware upgrade. Unfortunately, it'll be under Windows so I don't know what tool I could use for that. If I can confirm that the upgrade works with Oracle Virtualbox, I could also try that, and capture it at guest /and/ host level. I heard that next update should happen within a few week, so I'll keep you in touch about it. Frederic. ___ Spice-devel mailing list Spice-devel@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/spice-devel
Re: [Spice-devel] [Qemu-devel] RFC: usb redirection over the network, interesting outside of spice?
> > Not me at the moment, but unless you tunnel it inside another > > protocol, you'd really want to look at the existing USB-over-IP > > protocols instead of reinventing the wheel: > > http://usbip.sourceforge.net/ (some support in Linux already IIRC) > > (and there are others which I don't recall) > > Doesn't look very useful on a quick glance. > I'm note sure about what I will say, but will a kernel approach handle specific disconnection/reconnection of devices, that libusb cannot? I'm thinking in particular of what happens when one updates an iP* (*=hone, ad, odTouch, od) device, where it gets connected and disconnected several times during the process. During the updates, the device reboots, generating a disconnection event, and at reboot the guest OS (and iTunes) has to to catch the reconnection event directly (before iOS loads) to upload the new firmware. I'm not sure about it, but I think Virtualbox (non-ose) took care of this issue, and updates are now possible in a vm (TbC). I don't think that an approach at libusb level would work, while I could understand that a kernel module would be able to catch the "connection" event directly, and pipe it through the network to the guest. I don't know for sure if this use case affects only those devices, nor if it is an objective for spice to allow Apple device updates, but I think it has to be handled at pretty low level, so is worth mentioning before any implementation choices. Still, if such thing were to work, I think it would be a huge step for all the people in the iP* community that maintain a double boot only because they bought a closed phone once :-) However, the guest side could be handled at qemu-level, since it would certainly provide an OS-independent implementation without the need for linux-/windows-/other-specific driver. Frederic. ___ Spice-devel mailing list Spice-devel@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/spice-devel
Re: [Spice-devel] RFC: Updating spice documentation for 0.6 release
Hi, I don't have so much time to work on it, but I would be happy to occasionnaly help documenting spice. I'm using it regularly for a winXP guest on F14 host, and my skills are mainly on linux administration (network, virt...). I also think that the wiki is more accessible, especially for occasionnal contributers like me. Still, the pdf manual could stay there for "historical purposes", but it needs to be specified, so that people don't take those information for what they are not anymore If that helps, and since I've been following the list for quite a long time now, I think I could for example feed the wiki with the questions/answer that are found here : when the docs will have a defined structure they could each go in there respective categories, otherwise it could be some starting point for a FAQ. Regards, Frederic. -- frede...@grelot.net Tél : 05 61 62 42 14 / 06 82 23 46 17 - Original Message - > I suggest we go for keeping the wiki updated. It's easier accessible, > easier searchable, and > easier indexable by search engines than pdf files. > > I would use the wiki of Mikrotik > (http://wiki.mikrotik.com/wiki/Category:Manual) as an example of > structure. I find the information I need very easy in this wiki. > > Any other suggestions? > > > Rgds, > Siggi > > > > > On Fri, November 19, 2010 11:00, Hans de Goede wrote: > > Hi All, > > > > > > As Sigbjorn Lie pointed out in the > > "How can I move the mouse between spice client and local desktop > > without Shift+F12?" > > Thread our documentation is currently a bit lacking / outdated. > > > > > > Sigbjorn has indicated that he is willing to help out with > > updating the documentation. Looking at for example: > > http://spice-space.org/docs/spice_user_manual.pdf > > > > > > Quite a bit of updating is needed. So this brings me to the > > following > > questions: > > > > > > 1) What and where is the source of the pdf's we currently have? > > > > > > 2) In which format do we want to have our documentation? Currently: > > -We have some docs on the wiki > > -We have some pdf's > > Some info is available in both, ie building from source instructions > > can be found in both: > > http://spice-space.org/docs/spice_user_manual.pdf > > http://spice-space.org/page/Building_Instructions > > > > > > With the second one being more or less up2date and the pdf > > containing > > instructions which are pretty much useless for the 0.6 series. > > > > > > All in all I think our documentation needs an overhaul, starting > > with > > defining a structure for the different documents and in which > > format(s) we want to make the various > > docs available. > > > > Note I'm not saying we should rewrite all the docs we currently > > have, > > but I think it would be a good idea to remove some things (like > > build instructions) from the pdf > > docs, and replace them with a pointer to the wiki, were they can be > > kept up2date much easier. > > > > I would very much appreciate input from others on this. > > > > > > Regards, > > > > > > Hans > > ___ > > Spice-devel mailing list > > Spice-devel@lists.freedesktop.org > > http://lists.freedesktop.org/mailman/listinfo/spice-devel > > > > > > > ___ > Spice-devel mailing list > Spice-devel@lists.freedesktop.org > http://lists.freedesktop.org/mailman/listinfo/spice-devel ___ Spice-devel mailing list Spice-devel@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/spice-devel
Re: [Spice-devel] RFC: Updating spice documentation for 0.6 release
Hi, I don't have so much time to work on it, but I would be happy to occasionnaly help documenting spice. I'm using it regularly for a winXP guest on F14 host, and my skills are mainly on linux administration (network, virt...). I also think that the wiki is more accessible, especially for occasionnal contributers like me. Still, the pdf manual could stay there for "historical purposes", but it needs to be specified, so that people don't take those information for what they are not anymore If that helps, and since I've been following the list for quite a long time now, I think I could for example feed the wiki with the questions/answer that are found here : when the docs will have a defined structure they could each go in there respective categories, otherwise it could be some starting point for a FAQ. Regards, Frederic. ___ Spice-devel mailing list Spice-devel@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/spice-devel
Re: [Spice-devel] How much CPU load will spice server take? Is it possible to take spice client to do more processing?
> > And if I'm understanding it right, in the futur features "Video > > acceleration" is expected to address this use case? > > > > Yes, note though that there is no ETA of this feature yet. Yep, I've noticed that (i've been following the spice-space website for more than a year now...). > > > By the way, let's imagine that I'm playing some flash video (youtube > > for example) in full-screen. Is the flash plugin actually using the > > "upscaling" 2D command, so that this is done client-side ? > > Yes, if you play normal resolution flash videos fullscreened on a > monitor with say 1920x1080, you will > see this will work fine. OTOH if you play hd resolution flashvideos, > you will see the same problem > as with playing local hd videos multiplied by the flash is a cpu hog > factor :( Hmmm... Unfortunately this is a well known flash issue They just started to provide versions that use Intel/AMD-ATI/nVidia acceleration, I'm not sure they are considering QXL yet! By the way for my personal culture, I'm not aware of qxl internals but do you already know how it should work? : provide OpenGL generic APIs to the applications, transmit them through the network layer and execute them on client's side, or provide specific wrapper that reflects the client's card driver? Frederic. > > Regards, > > Hans ___ Spice-devel mailing list Spice-devel@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/spice-devel
Re: [Spice-devel] How much CPU load will spice server take? Is it possible to take spice client to do more processing?
> > http://www.spice-space.org/features.html > > Graphic commands - processes and transmits 2D graphic commands > > Correct, and video decoding is not a 2d graphics command (otoh > upscaling is). > Understood. And if I'm understanding it right, in the futur features "Video acceleration" is expected to address this use case? By the way, let's imagine that I'm playing some flash video (youtube for example) in full-screen. Is the flash plugin actually using the "upscaling" 2D command, so that this is done client-side ? Frederic. ___ Spice-devel mailing list Spice-devel@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/spice-devel
[Spice-devel] QXL Driver working in X11?
Hi, I just read the post of Hans about vdagent support for linux, which reminded me that the last time I tried, I couldn't manage to make the QXL driver work with Fedora : did anything evolve about this? I've been using spice a lot with windows XP, but would prefer to use it with my F13/14 virtual machines... If it is not ready yet, are there any plan to make it working before F14 release? Thanks for your answers. Frederic. ___ Spice-devel mailing list Spice-devel@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/spice-devel
Re: [Spice-devel] Available resolutions
Hi, Thanks four your fast answer, I'll study that today. By the way, is there any reason why the resolution should be limited to a fixed set? I think there is some option in VMWare that allow the user to resize the guest's window, and the resolution "just follows" the movement. This must work together with the agent though, since it is responsible for sending the correct resolution to the guest's os as the user resizes the window, but I found it really comfortable when I've had to use it... Thus, if you have a display running with, say, 1024*768, you could enlarge the guest's window up to 1020*720 (roughly) because of the borders&title/bottom bar , but would still have your full desktop (in particular, the bottom bar with the start button). And if, at this point, you hit shift+F11, the window becomes full screen, and the guest's resolution adapts itself to fit your 1024*768. I'm not sure about the implications of this feature, but I think it could be interesting to have it. At least, I'll see qxl_modes and make it fit my small screen! Thanks again, Frederic. - Mail original - > On 07/08/2010 01:22 AM, Frédéric Grelot wrote: > > Hi all, > > > > I've set up a spice server (unstable version), and few clients : as > > you may remember from my previous mails, it works quite well in > > windows, but still not with fedora (i gave up with this, after > > trying all the drivers I found on the net...). > > However, I still have one question : is it possible to change the > > available resolutions (for the windows driver in particular) ? > > I've installed the client on an eee-pc (1024x600), and would like to > > use it to access my spice desktop, but this resolution does not > > appear in the "display properties" page of the guest. > > I'm eager to use it, concerning the tests I already performed by > > opening an "HD" movie on youtube (C) in full screen : locally I > > get<1 frame/s, but with spice (through wifi and set up as 1024x768) > > it is really smooth! > > > > Thanks again for your help (and, if there is more info about the X11 > > driver, or you need more from my tests, please say so/ask!) > > > > Frederic. > > ___ > > Spice-devel mailing list > > Spice-devel@lists.freedesktop.org > > http://lists.freedesktop.org/mailman/listinfo/spice-devel > > Hi, > You can add the resolution to the struct qxl_modes in qxl.c under > qemu/hw. But, sometimes, tough a resolution is supported, Windows menu > does not display it. If for this resolution this will be the case, use > the spice client option --full-screen=auto-conf ==> this will force > the > client resolution on the guest (after you add it to the supported > ones). > Then, it will also appear on the windows menu. > > We are currently considering the option of Spice automatically > supporting at least all the resolutions the client support. > > Yonit. ___ Spice-devel mailing list Spice-devel@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/spice-devel
[Spice-devel] Available resolutions
Hi all, I've set up a spice server (unstable version), and few clients : as you may remember from my previous mails, it works quite well in windows, but still not with fedora (i gave up with this, after trying all the drivers I found on the net...). However, I still have one question : is it possible to change the available resolutions (for the windows driver in particular) ? I've installed the client on an eee-pc (1024x600), and would like to use it to access my spice desktop, but this resolution does not appear in the "display properties" page of the guest. I'm eager to use it, concerning the tests I already performed by opening an "HD" movie on youtube (C) in full screen : locally I get <1 frame/s, but with spice (through wifi and set up as 1024x768) it is really smooth! Thanks again for your help (and, if there is more info about the X11 driver, or you need more from my tests, please say so/ask!) Frederic. ___ Spice-devel mailing list Spice-devel@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/spice-devel
Re: [Spice-devel] cannot boot from virtio / no -qxl option
Well... I'll wait :-) I think your indications may already help me setup it with my f13 guest, that'll be a first step. (I was sure that you transformed the options into something like "-device ..."!) Frederic. - Mail original ----- > - "Frédéric Grelot" wrote: > > > Hi List & Gerd, > > > [snip] > > By the way, I made winxp and win7 work with spice without much > > trouble, but is it normal that there is no "special" device anymore? > > I remember installing two specific drivers "miniport" and > > "vdi_port", then vdagent... with the version I run now, there is no > > unknown devices (in config panel), and vdagent doesn't work... > > > > [AL] I can answer that - the vdiport is in the process of being > replaced, the replacement is not upstream yet.. so actually you will > probably have > to wait a bit. If you really really need this I could give you the > vdagent, and you need to supply the switch "-device virtio-serial > -device spicevmc" > to get the device, and you'll need a driver for windows (linux has one > already upstream) that I'll need to package for you as well. So, more > work for me, and you get almost no extra functionality at this point. > So if you > really really need this :) > > > Thanks in advance, > > > > Frederic. > > > > - Mail original - > > > Hi, > > > > > > > About the qxl driver in F13, it definitely won't build... > > > > > > It builds fine, needs some xorg devel packages though, try 'yum > > > install xorg-x11-server-sdk'. > > > > > > /me tries to roll up a rpm with it ... > > > > > > cheers, > > > Gerd > > ___ Spice-devel mailing > > list Spice-devel@lists.freedesktop.org > > http://lists.freedesktop.org/mailman/listinfo/spice-devel ___ Spice-devel mailing list Spice-devel@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/spice-devel
Re: [Spice-devel] cannot boot from virtio / no -qxl option
Hi List & Gerd, Your driver (the rpm that appeared on the repository) doesn't want to work... One first thing : the underscore of its name makes it be refused by xorg (due to its way of finding drivers .so file from name, it tries to load "qxlunstable_drv.so" instead of "qxl_unstable_drv.so" Still, I'm trying to build it myself, and noticed that the difference with the rpm was the file /usr/share/hwdata/videoaliases/qxl_unstable.xinf I modified it to "qxl", rebuilt/reinstalled the driver from the url you gave me few days ago, but the driver doesn't load... When I don't force it with system-config-display (and the appropriate created xorg.conf), the default VESA loads... When I force it, I get in Xorg.0.log : [ 1502.332] (II) LoadModule: "qxl" [ 1502.333] (II) Loading /usr/lib64/xorg/modules/drivers/qxl_drv.so [ 1502.333] (II) Module qxl: vendor="X.Org Foundation" [ 1502.333]compiled for 1.8.0, module version = 0.0.0 [ 1502.333]Module class: X.Org Video Driver [ 1502.333]ABI class: X.Org Video Driver, version 7.0 [ 1502.333] (II) qxl: Driver for QXL virtual graphics: QXL 1 [ 1502.333] (--) using VT number 8 [ 1502.389] (II) Primary Device is: PCI 0...@00:02:0 [ 1502.389] (EE) No devices detected. I even tried to force the BusID in xorg.conf, to no success... Here is my qemu command-line, if it may help (generated by libvirt...) : /usr/bin/qemu-spice -spice port=5930,disable-ticketing -S -M pc-0.13 -enable-kvm -m 4096 -smp 1,sockets=1,cores=1,threads=1 -name spice-test-f13 -uuid 7a56343e-f36a-a729-da43-d087f6189bee -nodefaults -chardev socket,id=monitor,path=/var/lib/libvirt/qemu/spice-test-f13.monitor,server,nowait -mon chardev=monitor,mode=readline -rtc base=utc -boot c -drive file=/raid/iso_images/Fedora-13-Beta-x86_64-DVD.iso,if=none,media=cdrom,id=drive-ide0-1-0,readonly=on,format=raw -device ide-drive,bus=ide.1,unit=0,drive=drive-ide0-1-0,id=ide0-1-0 -drive file=/sync/vms/spice-f13-boot.qcow2.img,if=none,id=drive-ide0-0-0,format=qcow2 -device ide-drive,bus=ide.0,unit=0,drive=drive-ide0-0-0,id=ide0-0-0 -drive file=/sync/vms/spice-test-f13.qcow2.img,if=none,id=drive-virtio-disk0,format=qcow2 -device virtio-blk-pci,bus=pci.0,addr=0x4,drive=drive-virtio-disk0,id=virtio-disk0 -device virtio-net-pci,vlan=0,id=net0,mac=52:54:00:00:11:72,bus=pci.0,addr=0x5 -net tap,fd=50,vlan=0,name=hostnet0 -chardev pty ,id=serial0 -device isa-serial,chardev=serial0 -usb -device usb-tablet,id=input0 -vnc 127.0.0.1:3 -vga qxl -device virtio-balloon-pci,id=balloon0,bus=pci.0,addr=0x3 By the way, I made winxp and win7 work with spice without much trouble, but is it normal that there is no "special" device anymore? I remember installing two specific drivers "miniport" and "vdi_port", then vdagent... with the version I run now, there is no unknown devices (in config panel), and vdagent doesn't work... Thanks in advance, Frederic. - Mail original - > Hi, > > > About the qxl driver in F13, it definitely won't build... > > It builds fine, needs some xorg devel packages though, try 'yum > install xorg-x11-server-sdk'. > > /me tries to roll up a rpm with it ... > > cheers, > Gerd ___ Spice-devel mailing list Spice-devel@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/spice-devel
[Spice-devel] cannot boot from virtio / no -qxl option
Hi all, I'm trying to set up Spice on a Fedora 13 (still beta) server. I'm using libvirt to set up my guests, qemu-spice from kraxel repo, and a small wrapper script of mine to make libvirt work (basically, it changes the version line when I run "qemu-spice -help" to make it understandable by libvirt, adds the "-spice" option, and changes "-vga cirrus" by "-vga qxl") I have come to 2 problems : -when the disk is set to "virtio", I can install F13, but boot fails (at bios stage). When I set it to IDE then, it starts OK -when I use "-vga qxl", f13 loads the non-graphical part (blue progress bar), but X doesn't start. Instead, I just get a blinking cursor at the top left corner... I also noticed that the "-qxl 1" option does not work anymore, is that normal? Thanks for your help. Frederic. -- frede...@grelot.net Tél : 05 61 27 81 79 / 06 82 23 46 17 ___ Spice-devel mailing list Spice-devel@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/spice-devel
Re: [Spice-devel] cannot boot from virtio / no -qxl option
> > For the virtio-boot problem, does that mean that I can't use virtio > > disk with windows?? > > Never tried, reportly for xp guests there isn't a big difference > between ide and virtio due to the way the windows xp virtio guest > drivers are > written. Well, it didn't seem to boot though... However, when I installed it, the disk was ide, I then installed the drivers (by adding a second drive in virtio mode), switched the main drive to virtio and restarted. I already did that a few time ago and it worked... I'm currently trying to install it from scratch in virtio mode, by using the latest drivers from kvm website : this latest version works with windows installer (with F6 at startup to give specific drivers) so it may be better. About the qxl driver in F13, it definitely won't build... Tanks anyway for all your work, it becomes easier and easier to test spice! (and I can't imagine what it will be when spice gets into upstream qemu!). Frederic. ___ Spice-devel mailing list Spice-devel@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/spice-devel
Re: [Spice-devel] cannot boot from virtio / no -qxl option
> >> Not sure whenever some qxl x11 driver for unstable exists > >> somewhere? > > > > Isn't qxl part of the "spice project"? I don't really understand why > > xorg-x11-drv-qxl can be found in fedora repos while spice is not > > "officially" merged? > > Both spice and xorg live @ freedesktop.org > > qxl driver is here: > http://cgit.freedesktop.org/xorg/driver/xf86-video-qxl/ and corresponds to the one that get shipped with fedora 13? > > experimental bits are here: > http://cgit.freedesktop.org/~sandmann/xf86-video-qxl/log/?h=surface0 and *should* work with qemu-unstable-spice from kraxel's repo? > > Didn't test whenever they are complete and actually work. hence should > Right now it is v5, will jump to v9 soon. Very good! Frederic. ___ Spice-devel mailing list Spice-devel@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/spice-devel
Re: [Spice-devel] cannot boot from virtio / no -qxl option
> Well, there is upstream qemu and there is qemu-kvm which carries > additional patches (including extboot which handles booting from > virtio). fedora ships qemu-kvm ... I can confirm that my other vms run with qemu-kvm. Does this difference also explains that "-smp 2" does not work with spice? > Oh, right, this is unstable. I never actually tried X11 with this, > only windows guests. The qxl device is *not* compatible with older > spice qxl > devices, which is likely the reason why it fails. We plan to fix this > (so the new qxl device can handle guests with old qxl drivers just > fine), but it doesn't work yet. OK, I'll wait for these, and am ready to test compatibility if you want me to. > > Not sure whenever some qxl x11 driver for unstable exists somewhere? Isn't qxl part of the "spice project"? I don't really understand why xorg-x11-drv-qxl can be found in fedora repos while spice is not "officially" merged? > Well, almost. Secondary qxl devices are added via "-device qxl". OK, thanks for this precision (I also saw this in your last patch set for v8). By the way, could you tell me from which branch the "qemu-unstable-spice" of kraxel's repo comes from? Is it always updated to the latest? Frederic. ___ Spice-devel mailing list Spice-devel@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/spice-devel
Re: [Spice-devel] cannot boot from virtio / no -qxl option
- Mail original - > Hi, > > > -when the disk is set to "virtio", I can install F13, but boot fails > > (at bios stage). When I set it to IDE then, it starts OK > > Upstream qemu simply doesn't support that (yet). I'm quite surprised, I have others vms (running with "normal" qemu from virt-preview repo for F13), that boot without problem, and I think their main disk is virtio... I'm not sure however and can't check this right now... (and it's more likely that you are right...) Anyway the easiest solution would be to have a 500Mb/1Gb disk for /boot (+grub), IDE, and everything else on a virtio disk I imagine. > > There have been seabios patches floating around which should fix this > though, so this should improve soon. OK, good > > -when I use "-vga qxl", f13 loads the non-graphical part (blue > > progress bar), but X doesn't start. Instead, I just get a blinking > > cursor at the top left corner... > > Try installing xorg-x11-drv-qxl. I did'nt mentionned that I did install it, and, instead of being still, the cursor was just blinking... What I need to do now is set up the network services, connect by ssh to the guest when it is locked at this stage, and have a look at its X's log files. However, my last "yum update" failed (network got broken in the middle, and I was doing it remotely, which was not a good idea), so I'm afraid I'll have to set up the whole guest again. I'll do that with the 1Gb IDE boot disk and tell you the results... > > > I also noticed that the "-qxl 1" option does not work anymore, is > > that normal? > > It has been replaced by -vga qxl Ok, and I think that, if I want multiple qxl devices, I just have to use it several times, just like any other video device! Thanks for your help, and for your progess in making it more compatible with upstream qemu : being able to manage my spice instance with libvirt spares me lot of pain! If anybody's interested with the mentionned wrapper script, I could post it here. Frederic. ___ Spice-devel mailing list Spice-devel@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/spice-devel
[Spice-devel] cannot boot from virtio / no -qxl option
Hi all, I'm trying to set up Spice on a Fedora 13 (still beta) server. I'm using libvirt to set up my guests, qemu-spice from kraxel repo, and a small wrapper script of mine to make libvirt work (basically, it changes the version line when I run "qemu-spice -help" to make it understandable by libvirt, adds the "-spice" option, and changes "-vga cirrus" by "-vga qxl") I have come to 2 problems : -when the disk is set to "virtio", I can install F13, but boot fails (at bios stage). When I set it to IDE then, it starts OK -when I use "-vga qxl", f13 loads the non-graphical part (blue progress bar), but X doesn't start. Instead, I just get a blinking cursor at the top left corner... I also noticed that the "-qxl 1" option does not work anymore, is that normal? Thanks for your help. Frederic. ___ Spice-devel mailing list Spice-devel@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/spice-devel
Re: [Spice-devel] Building QXL
Have a look at that discussion, I once managed to build it on winXP without too much harm : http://www.mail-archive.com/spice-space-de...@lists.sourceforge.net/msg00077.html Frédéric - "Izik Eidus" a écrit : > On Wed, 17 Feb 2010 09:46:58 -0600 (CST) > David Martin wrote: > > > I'd like to test the latest patches that were committed yesterday. > Any tips on how to build the QXL driver? > > > I will search the documents for that, but maybe it would be enougth if > I just > give you compiled driver? > > (I plan to upload it to the spice-space site) > > > > > David Martin > > ___ > > Spice-devel mailing list > > Spice-devel@lists.freedesktop.org > > http://lists.freedesktop.org/mailman/listinfo/spice-devel > > ___ > Spice-devel mailing list > Spice-devel@lists.freedesktop.org > http://lists.freedesktop.org/mailman/listinfo/spice-devel ___ Spice-devel mailing list Spice-devel@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/spice-devel
