From: "Andreas Färber" <afaerber@suse.de>
To: Markus Armbruster <armbru@redhat.com>, qemu-devel@nongnu.org
Cc: pbonzini@redhat.com, kraxel@redhat.com
Subject: Re: [Qemu-devel] [PATCH v2 3/7] qdev-monitor: Stop error avalance in qbus_find_recursive()
Date: Fri, 19 Jun 2015 14:47:02 +0200 [thread overview]
Message-ID: <55840F46.7000009@suse.de> (raw)
In-Reply-To: <1434194302-26589-4-git-send-email-armbru@redhat.com>
Did you mean avalanche?
Am 13.06.2015 um 13:18 schrieb Markus Armbruster:
> Reproducer:
>
> $ qemu-system-x86_64 -nodefaults -device virtio-rng-pci -device virtio-rng-pci -device virtio-rng-device,bus=virtio-bus
> qemu-system-x86_64: -device virtio-rng-device,bus=virtio-bus: Bus 'virtio-bus' is full
> qemu-system-x86_64: -device virtio-rng-device,bus=virtio-bus: Bus 'virtio-bus' is full
> qemu-system-x86_64: -device virtio-rng-device,bus=virtio-bus: Bus 'virtio-bus' not found
>
> qbus_find_recursive() reports the "is full" error itself, and leaves
> reporting "not found" to its caller. The result is confusion. Write
> it a function contract that permits leaving all error reporting to the
> caller, and implement it. Update callers to detect and report "is
> full".
>
> Screwed up when commit 1395af6 added the max_dev limit and the "is
> full" error condition to enforce it.
>
> Signed-off-by: Markus Armbruster <armbru@redhat.com>
> Reviewed-by: Eric Blake <eblake@redhat.com>
> ---
> qdev-monitor.c | 62 ++++++++++++++++++++++++++++++++++++----------------------
> 1 file changed, 39 insertions(+), 23 deletions(-)
>
> diff --git a/qdev-monitor.c b/qdev-monitor.c
> index 7dd62dd..1408c86 100644
> --- a/qdev-monitor.c
> +++ b/qdev-monitor.c
> @@ -364,43 +364,55 @@ static DeviceState *qbus_find_dev(BusState *bus, char *elem)
> return NULL;
> }
>
> +static inline bool qbus_is_full(BusState *bus)
> +{
> + BusClass *bus_class = BUS_GET_CLASS(bus);
> + return bus_class->max_dev && bus->max_index >= bus_class->max_dev;
> +}
> +
> +/**
This should probably just be "/*".
I'll trust you on getting the functional logic right, can't do a full
review right now.
Andreas
> + * Search the tree rooted at @bus for a bus.
> + * If @name, search for a bus with that name. Note that bus names
> + * need not be unique. Yes, that's screwed up.
> + * Else search for a bus that is a subtype of @bus_typename.
> + * If more than one exists, prefer one that can take another device.
> + * Return the bus if found, else %NULL.
> + */
> static BusState *qbus_find_recursive(BusState *bus, const char *name,
> const char *bus_typename)
> {
> - BusClass *bus_class = BUS_GET_CLASS(bus);
> BusChild *kid;
> - BusState *child, *ret;
> - int match = 1;
> + BusState *pick, *child, *ret;
> + bool match;
>
> - if (name && (strcmp(bus->name, name) != 0)) {
> - match = 0;
> - } else if (bus_typename && !object_dynamic_cast(OBJECT(bus), bus_typename)) {
> - match = 0;
> - } else if ((bus_class->max_dev != 0) && (bus_class->max_dev <= bus->max_index)) {
> - if (name != NULL) {
> - /* bus was explicitly specified: return an error. */
> - qerror_report(ERROR_CLASS_GENERIC_ERROR, "Bus '%s' is full",
> - bus->name);
> - return NULL;
> - } else {
> - /* bus was not specified: try to find another one. */
> - match = 0;
> - }
> + assert(name || bus_typename);
> + if (name) {
> + match = !strcmp(bus->name, name);
> + } else {
> + match = !!object_dynamic_cast(OBJECT(bus), bus_typename);
> }
> - if (match) {
> - return bus;
> +
> + if (match && !qbus_is_full(bus)) {
> + return bus; /* root matches and isn't full */
> }
>
> + pick = match ? bus : NULL;
> +
> QTAILQ_FOREACH(kid, &bus->children, sibling) {
> DeviceState *dev = kid->child;
> QLIST_FOREACH(child, &dev->child_bus, sibling) {
> ret = qbus_find_recursive(child, name, bus_typename);
> - if (ret) {
> - return ret;
> + if (ret && !qbus_is_full(ret)) {
> + return ret; /* a descendant matches and isn't full */
> + }
> + if (ret && !pick) {
> + pick = ret;
> }
> }
> }
> - return NULL;
> +
> + /* root or a descendant matches, but is full */
> + return pick;
> }
>
> static BusState *qbus_find(const char *path)
> @@ -423,6 +435,10 @@ static BusState *qbus_find(const char *path)
> if (!bus) {
> qerror_report(QERR_BUS_NOT_FOUND, elem);
> return NULL;
> + } else if (qbus_is_full(bus)) {
> + qerror_report(ERROR_CLASS_GENERIC_ERROR, "Bus '%s' is full",
> + elem);
> + return NULL;
> }
> pos = len;
> }
> @@ -529,7 +545,7 @@ DeviceState *qdev_device_add(QemuOpts *opts)
> }
> } else if (dc->bus_type != NULL) {
> bus = qbus_find_recursive(sysbus_get_default(), NULL, dc->bus_type);
> - if (!bus) {
> + if (!bus || qbus_is_full(bus)) {
> qerror_report(ERROR_CLASS_GENERIC_ERROR,
> "No '%s' bus found for device '%s'",
> dc->bus_type, driver);
>
--
SUSE Linux GmbH, Maxfeldstr. 5, 90409 Nürnberg, Germany
GF: Felix Imendörffer, Jane Smithard, Dilip Upmanyu, Graham Norton; HRB
21284 (AG Nürnberg)
next prev parent reply other threads:[~2015-06-19 12:47 UTC|newest]
Thread overview: 17+ messages / expand[flat|nested] mbox.gz Atom feed top
2015-06-13 11:18 [Qemu-devel] [PATCH v2 0/7] qdev: Mostly wean off QError Markus Armbruster
2015-06-13 11:18 ` [Qemu-devel] [PATCH v2 1/7] qdev: Deprecated qdev_init() is finally unused, drop Markus Armbruster
2015-06-19 12:36 ` Andreas Färber
2015-06-19 13:17 ` Markus Armbruster
2015-06-13 11:18 ` [Qemu-devel] [PATCH v2 2/7] qdev: Un-deprecate qdev_init_nofail() Markus Armbruster
2015-06-19 12:42 ` Andreas Färber
2015-06-19 13:18 ` Markus Armbruster
2015-06-13 11:18 ` [Qemu-devel] [PATCH v2 3/7] qdev-monitor: Stop error avalance in qbus_find_recursive() Markus Armbruster
2015-06-19 12:47 ` Andreas Färber [this message]
2015-06-19 13:19 ` Markus Armbruster
2015-06-13 11:18 ` [Qemu-devel] [PATCH v2 4/7] qdev-monitor: Fix check for full bus Markus Armbruster
2015-06-13 11:18 ` [Qemu-devel] [PATCH v2 5/7] qdev-monitor: Convert qbus_find() to Error Markus Armbruster
2015-06-13 11:18 ` [Qemu-devel] [PATCH v2 6/7] qdev-monitor: Propagate errors through set_property() Markus Armbruster
2015-06-19 12:49 ` Andreas Färber
2015-06-13 11:18 ` [Qemu-devel] [PATCH v2 7/7] qdev-monitor: Propagate errors through qdev_device_add() Markus Armbruster
2015-06-19 13:08 ` Andreas Färber
2015-06-19 13:49 ` Markus Armbruster
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=55840F46.7000009@suse.de \
--to=afaerber@suse.de \
--cc=armbru@redhat.com \
--cc=kraxel@redhat.com \
--cc=pbonzini@redhat.com \
--cc=qemu-devel@nongnu.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).