Juan Quintela [Wed, 7 Oct 2009 00:41:03 +0000 (02:41 +0200)]
Generate config-devices.h
Generate config-devices.h for each target and config-all-devices.h for
common library. We don't want to name both config-devices.h to avoid
path problems
Juan Quintela [Wed, 7 Oct 2009 00:40:59 +0000 (02:40 +0200)]
Remove useless check for config-host.mak
If config-host.mak dont' exist, we have exited in the check at
the beginning of the file.
Once here, move the bits to the else part of the test at the beginning of
the file.
Like qdev_init(), but terminate program via hw_error() instead of
returning an error value.
Use it instead of qdev_init() where terminating the program on failure
is okay, either because it's during machine construction, or because
we know that failure can't happen.
Because relying in the latter is somewhat unclean, and the former is
not always obvious, it would be nice to go back to qdev_init() in the
not-so-obvious cases, only with proper error handling. I'm leaving
that for another day, because it involves making sure that error
values are properly checked by all callers.
Commit 22f84e73 added a qdev_init() missing on the path through
usb_host_device_open(), but that broke the path through
usb_host_auto_scan(), which already had one. Remove that one.
Making pci device cleanup msix automatically makes pci.c depend on
msix.c, which is IMO messy. Since devices do msix_init it's easy and
natural for them to also do msix_uninit.
PCI load routine has to be called with size equal to 256 (otherwise it
will crash in weird ways). So assert this, making code clearer.
Also avoid dynamically sized array on stack - good for portability.
Mark McLoughlin [Tue, 6 Oct 2009 11:17:16 +0000 (12:17 +0100)]
Final net cleanup after conversion to QemuOpts
Now that net_client_init() has no users, kill it off and rename
net_client_init_from_opts().
There is no further need for the old code in net_client_parse() either.
We use qemu_opts_parse() 'firstname' facitity for that. Instead, move
the special handling of the 'vmchannel' type there.
Simplify the vl.c code into merely call net_client_parse() for each
-net command line option and then calling net_init_clients() later
to iterate over the options and create the clients.
Mark McLoughlin [Tue, 6 Oct 2009 11:17:13 +0000 (12:17 +0100)]
Port host_net_add monitor command to QemuOpts
Here is where we rely on qemu_opts_parse() to handle an empty string.
We could alternatively explicitly handle this here by using
qemu_opts_create() when we're not supplied any parameters, but its
cleaner this way.
Mark McLoughlin [Tue, 6 Oct 2009 11:17:08 +0000 (12:17 +0100)]
Port -net tap to QemuOpts
Some parameters are not valid with fd=. Rather than having a separate
parameter description table for validating fd=, it's easir to just
check for those invalid parameters later.
Note, the need to possible lookup a file descriptor name from the
monitor is the reason why all these init functions are passed a Monitor
pointer.
Mark McLoughlin [Tue, 6 Oct 2009 11:17:07 +0000 (12:17 +0100)]
Port -net user to QemuOpts
The handling of guestfwd and hostfwd requires the previous changes
to allow multiple values for each parameter. The only way to access
those multiple values is to use qemu_opt_foreach().
Mark McLoughlin [Tue, 6 Oct 2009 11:17:06 +0000 (12:17 +0100)]
Port -net none and -net nic to QemuOpts
We use a table of network types to look up the initialization function
and parameter descriptions in net_client_init().
For now, we use QemuOpts for the 'none' and 'nic' types. Subsequent
patches port the other types too and the special casing is removed.
We're not parsing the full -net option string here as the type has
been stripped from the string, so we do not use qemu_opts_parse()
'firstname' facility. This will also be rectified in subsequent
patches.
No functional changes are introduced by this patch.
Mark McLoughlin [Tue, 6 Oct 2009 11:17:05 +0000 (12:17 +0100)]
Add qemu_net_opts
The first step in porting -net to QemuOpts. We do not include parameter
descriptions in the QemuOptsList because we use the first parameter to
choose which descriptions validate against.
Mark McLoughlin [Tue, 6 Oct 2009 11:17:04 +0000 (12:17 +0100)]
Never overwrite a QemuOpt
Rather than overwriting a QemuOpt, just add a new one to the tail and
always do a reverse search for parameters to preserve the same
behaviour. We use this order so that foreach() iterates over the opts
in their original order.
This will allow us handle options where multiple values for the same
parameter is allowed - e.g. -net user,hostfwd=
Mark McLoughlin [Tue, 6 Oct 2009 11:17:03 +0000 (12:17 +0100)]
Add qemu_opts_validate() for post parsing validation
Several qemu command line options have a parameter whose value affects
what other parameters are accepted for the option.
In these cases, we can have an empty description table in the
QemuOptsList and once the option has been parsed we can use a suitable
description table to validate the other parameters based on the value of
that parameter.
Mark McLoughlin [Tue, 6 Oct 2009 11:17:02 +0000 (12:17 +0100)]
Make qemu_opts_parse() handle empty strings
Rather than making callers explicitly handle empty strings by using
qemu_opts_create(), we can easily have qemu_opts_parse() handle
empty parameter strings.
Mark McLoughlin [Tue, 6 Oct 2009 11:16:59 +0000 (12:16 +0100)]
Remove bogus error message from qemu_opts_set()
The only way qemu_opts_create() can fail is if a QemuOpts with that id
already exists and fail_if_exists=1. In that case, we already print
an error which makes more sense than the one in qemu_opts_set().
Diagnostic output goes to stderr, except when we're in a monitor
command, when it goes to the monitor instead.
config_error() implements this with a monitor argument: if it's
non-null, report there, else to stderr. This obliges us to pass the
monitor down various call chains, to make it available to
config_error().
The recently created qemu_error() doesn't need a monitor argument to
route output. Use it.
There's one user-visible difference: config_error() prepended "qemu: "
to a message bound for stderr. qemu_error() doesn't, which means the
prefix goes away with this commit. If such a prefix is desired for
stderr, then I figure it should be slapped on all error messages, not
just the ones that used to go through config_error().
Propagating errors up the call chain is tedious. In startup code, we
can take a shortcut: terminate the program. This is wrong elsewhere,
the monitor in particular.
config_error() tries to cater for both customers: it terminates the
program unless its mon parameter tells it it's working for the
monitor.
Its users need to return status anyway (unless passing a null mon
argument, which none do), which their users need to check. So this
automatic exit buys us exactly nothing useful. Only the dangerous
delusion that we can get away without returning status. Some of its
users fell for that. Their callers continue executing after failure
when working for the monitor.
This bites monitor command host_net_add in two places:
* net_slirp_init() continues after slirp_hostfwd(), slirp_guestfwd(),
or slirp_smb() failed, and may end up reporting success. This
happens for "host_net_add user guestfwd=foo": it complains about the
invalid guest forwarding rule, then happily creates the user network
without guest forwarding.
* net_client_init() can't detect slirp_guestfwd() failure, and gets
fooled by net_slirp_init() lying about success. Suppresses its
"Could not initialize device" message.
Add the missing error reporting, make sure errors are checked, and
drop the exit() from config_error().
Make net_client_init() consume slirp_configs even on error
net_slirp_init() walks slirp_configs, and stops when it encounters one
that doesn't work. Instead of consuming slirp_configs members there,
consume them in the sole caller. This makes sure all are consumed.
Before, the tail starting with the non-working one was left in place,
where it made the next net_slirp_init() fail again.
Adrian Bridgett [Thu, 1 Oct 2009 10:56:10 +0000 (11:56 +0100)]
add host_device format description to qemu-img manpage
Converting files using "qemu-img convert" onto logical volumes (or any
block device) you need to use the currently undocumented "host_device"
format. This patch adds the required documentation to the manpage.
Gerd Hoffmann [Thu, 1 Oct 2009 14:42:33 +0000 (16:42 +0200)]
Reorganize option rom (+linux kernel) loading.
This patch adds infrastructure to maintain memory regions which must be
restored on reset. That includes roms (vga bios and option roms on pc),
but is also used when loading linux kernels directly. Features:
- loading files is supported.
- passing blobs is supported.
- target address range is supported (for optionrom area).
- fixed target memory address is supported (linux kernel).
New in v2:
- writes to ROM are done only at initial boot.
- also handle aout and uimage loaders.
- drop unused fread_targphys() function.
The final memory layout is created once all memory regions are
registered. The option roms get addresses assigned and the
registered regions are checked against overlaps. Finally all data
is copyed to the guest memory.
Advantages:
(1) Filling memory on initial boot and on reset takes the same
code path, making reset more robust.
(2) The need to keep track of the option rom load address is gone.
(3) Due to (2) option roms can be loaded outside pc_init(). This
allows to move the pxe rom loading into the nic drivers for
example.
Additional bonus: There is a 'info roms' monitor command now.
The patch also switches over pc.c and removes the
option_rom_setup_reset() and load_option_rom() functions.
By making the error reporting include strerror(errno), it gives the user
a bit more indication as to why qemu failed. This is particularly
important for people running qemu as a non root user.
Monitor command "pci_add ADDR nic model=MODEL" uses pci_nic_init() to
create the NIC. When MODEL is unknown or "?", this prints to stderr
and terminates the program.
Change pci_nic_init() not to treat "?" specially, and to return NULL
on failure. Switch uses during startup to new convenience wrapper
pci_nic_init_nofail(), which behaves just like pci_nic_init() used to
do.
Bonus bug fix: we now check for qdev_init() failing there.
Before this patch, pci_nic_init() returns NULL when it can't find the
model in pci_nic_models[]. Except this can't happen, because
qemu_check_nic_model_list() just searched for model in
pci_nic_models[], and terminated the program on failure.
Repeating the search here is pointless. Instead, change
qemu_check_nic_model_list() to return the model's array index.
Fix pci_add storage not to exit on bad first argument
Monitor command "pci_add ADDR storage ..." does its work in
qemu_pci_hot_add_nic(). It called pci_create(..., ADDR) to create the
device. That's wrong, because pci_create() terminates the program
when ADDR is invalid.
Use pci_get_bus_devfn() and pci_create_noinit() instead.
Commit a414c306 converted all VGA devices to qdev. It used
pci_create_simple() for all devices, except for this one it used
pci_create(). That's wrong, because it uses PCI bus#0 regardless of
the bus argument. Fix by switching to pci_create_noinit().
Thomas Monjalon [Fri, 11 Sep 2009 16:45:40 +0000 (18:45 +0200)]
restore CFLAGS check for conflict and fix recursive CFLAGS issue
cc-option uses more make-syntax to replace the shell "if/else".
Issue with recursive += is fixed by doing the first assignment
simply-expanded, as explained in
http://www.gnu.org/software/make/manual/html_node/Appending.html
Amit Shah [Tue, 29 Sep 2009 10:21:04 +0000 (15:51 +0530)]
virtio-pci: return error if virtio_console_init fails
Currently only one virtio_console device is supported. Trying to add
multiple devices fails and such failure should be reported back to the
qdev init functions.
Stefan Weil [Sat, 19 Sep 2009 19:00:09 +0000 (21:00 +0200)]
vnc: Set invalid buffer pointers to NULL
After qemu_free, the pointers for input and output
buffers are no longer valid, so set them to NULL
(most other calls of qemu_free in vnc.c use this
pattern, too).
Kevin Wolf [Wed, 23 Sep 2009 15:42:42 +0000 (17:42 +0200)]
eepro100: Don't allow guests to fail assertions
The idea of using assert() for input validation is rather questionable.
Let's remove it from eepro100, so that guests need to find more interesting
ways if they want to crash qemu.
This patch replaces asserts that are directly dependent on guest-accessible
data by other means of error handling.
Kevin Wolf [Tue, 15 Sep 2009 10:30:43 +0000 (12:30 +0200)]
qcow2: Increase maximum cluster size to 2 MB
This patch increases the maximum qcow2 cluster size to 2 MB. Starting with 128k
clusters, L2 tables span 2 GB or more of virtual disk space, causing 32 bit
truncation and wraparound of signed integers. Therefore some variables need to
use a larger data type.
While being at reviewing data types, change some integers that are used for
array indices to unsigned. In some places they were checked against some upper
limit but not for negative values. This could avoid potential segfaults with
corrupted qcow2 images.
Since cpu_register_phys_memory does not require size to be a multiple of
target page size, simply make msix page size 0x1000. Do this in msix,
reverting part of 5e520a7d500ec2569d22d80f9ef4272a34cb3c80, as we no
longer have to pass target page around.
projects / qemu.git / log