On Wed, 01/15 14:40, Peter Maydell wrote:
> On 15 January 2014 14:36, Paolo Bonzini wrote:
> > Il 15/01/2014 15:20, Peter Maydell ha scritto:
> >>> >
> >>> > On RHEL6 I tried "qemu-system-x86_64 -cdrom
> >>> > http://boot.ipxe.org/ipxe.iso"; and it worked, but it failed on Fedora.
> >>> > I don't
Il 15/01/2014 16:40, Peter Maydell ha scritto:
> Further hacking suggests this is probably because we're not
> implementing the curl timeout callback, which curl uses to say
> "call me back in X milliseconds, even if there's no filedescriptor
> action". It looks like the first time or two around cu
On 15 January 2014 15:29, Peter Maydell wrote:
> I threw some debug printouts in, and it looks like libcurl
> is just never calling the curl_sock_cb, so we never register
> the fd with QEMU's event loop, and so nothing happens once
> the first read AIO has commenced.
Further hacking suggests this
On 15 January 2014 14:51, Paolo Bonzini wrote:
> Il 15/01/2014 15:40, Peter Maydell ha scritto:
(a non module build works, or at least sits there without giving an
error message; I didn't feel like sitting there without a progress bar
while it downloaded a whole ISO image, so I ^C'
Il 15/01/2014 15:40, Peter Maydell ha scritto:
> Module loading should IMHO work with similar semantics
> to the data dir we use for BIOS file loading: one of the places checked
> is relative to the executable path.
Makes sense.
>>> (a non module build works, or at least sits there without giving
On 15 January 2014 14:36, Paolo Bonzini wrote:
> Il 15/01/2014 15:20, Peter Maydell ha scritto:
>>> >
>>> > On RHEL6 I tried "qemu-system-x86_64 -cdrom
>>> > http://boot.ipxe.org/ipxe.iso"; and it worked, but it failed on Fedora.
>>> > I don't know if it's a QEMU or curl bug.
>> Doesn't work on Ma
Il 15/01/2014 15:20, Peter Maydell ha scritto:
>> >
>> > On RHEL6 I tried "qemu-system-x86_64 -cdrom
>> > http://boot.ipxe.org/ipxe.iso"; and it worked, but it failed on Fedora.
>> > I don't know if it's a QEMU or curl bug.
> Doesn't work on MacOSX either: it says
> qemu-system-x86_64: -cdrom http:
Il 15/01/2014 14:11, Peter Maydell ha scritto:
> On 15 January 2014 08:48, Fam Zheng wrote:
>> A few changes on Paolo's v15, to fix MacOSX build (in fact fix Linux as well)
>> and get rid of $RANDOM:
>
> Thanks. I applied a local patch to use shasum, and it seems
> to build OK with modules enable
On 15 January 2014 14:11, Paolo Bonzini wrote:
> Il 15/01/2014 14:11, Peter Maydell ha scritto:
>> On 15 January 2014 08:48, Fam Zheng wrote:
>>> A few changes on Paolo's v15, to fix MacOSX build (in fact fix Linux as
>>> well)
>>> and get rid of $RANDOM:
>>
>> Thanks. I applied a local patch to
On 15 January 2014 08:48, Fam Zheng wrote:
> A few changes on Paolo's v15, to fix MacOSX build (in fact fix Linux as well)
> and get rid of $RANDOM:
Thanks. I applied a local patch to use shasum, and it seems
to build OK with modules enabled. Is there a simple test I can
do that will exercise a c
A few changes on Paolo's v15, to fix MacOSX build (in fact fix Linux as well)
and get rid of $RANDOM:
[05/09] build-sys: introduce common-obj-m and block-obj-m for DSO
Add " -undefined dynamic_lookup" to Darwin LDFLAGS. Otherwise the
linker complains about undefined sy
11 matches
Mail list logo
