| Message ID | 20190107163117.16269-25-peter.maydell@linaro.org |
|---|---|
| State | Not Applicable |
| Headers | show |
| Series | target-arm queue | expand |
On Mon, Jan 07, 2019 at 04:31:04PM +0000, Peter Maydell wrote: > From: Steffen Görtz <contrib@steffen-goertz.de> > > Adds a new qtest command "set_irq_in" which allows > to set qemu gpio lines to a given level. > > Based on https://lists.gnu.org/archive/html/qemu-devel/2012-12/msg02363.html > which never got merged. > > Signed-off-by: Steffen Görtz <contrib@steffen-goertz.de> > Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com> > Reviewed-by: Thomas Huth <thuth@redhat.com> > Reviewed-by: Laurent Vivier <lvivier@redhat.com> > Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com> > Message-id: 20190103091119.9367-2-stefanha@redhat.com > Originally-by: Matthew Ogilvie <mmogilvi_qemu@miniinfo.net> It is kind of interesting to see a part of my old thread(s) resurrected, even if it isn't the part I was actually interested in. Note that my miniinfo email account no longer exists, but I can still be reached at mmogilvi+qemu@zoho.com if desired. Regarding the patch, I haven't kept up with changes to qemu well enough to really have a useful opinion, although superficially it looks good. Perhaps it would make sense to include both my old and new email addresses in the sign-off section, to both tie it to the old threads and make it easier to reach me if someone wants to contact me? (But I'm not sure of a "clean" way to do so, and it is probably low priority. Perhaps this current email is good enough?) - Matthew Ogilvie P.S.: Explanation and keywords to find this message (and me) with search tools: With regards to qemu, my primary interest (from that thread and related 2012 threads) is/was being able to run "Microport UNIX System V/386 v2.1" (ca. 1987) under qemu, which especially required: - Fixing various inaccuracies in qemu's model of the i8259 interrupt controller. - Implementing very old graphics card model(s) that Microport UNIX actually knows how to control properly. Perhaps original IBM monochrome display adaptor (MDA), CGA, and/or hercules graphics card. Or else my nasty qemu hackish patches I wrote that allowed VGA to work in text mode despite Microport driving it like CGA in a way that didn't work with real VGA hardware.
On Wed, 9 Jan 2019 at 05:59, Matthew Ogilvie <mmogilvi+qemu@zoho.com> wrote: > > On Mon, Jan 07, 2019 at 04:31:04PM +0000, Peter Maydell wrote: > > From: Steffen Görtz <contrib@steffen-goertz.de> > > > > Adds a new qtest command "set_irq_in" which allows > > to set qemu gpio lines to a given level. > > > > Based on https://lists.gnu.org/archive/html/qemu-devel/2012-12/msg02363.html > > which never got merged. > > > > Signed-off-by: Steffen Görtz <contrib@steffen-goertz.de> > > Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com> > > Reviewed-by: Thomas Huth <thuth@redhat.com> > > Reviewed-by: Laurent Vivier <lvivier@redhat.com> > > Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com> > > Message-id: 20190103091119.9367-2-stefanha@redhat.com > > Originally-by: Matthew Ogilvie <mmogilvi_qemu@miniinfo.net> > > It is kind of interesting to see a part of my old thread(s) > resurrected, even if it isn't the part I was actually interested > in. > > Note that my miniinfo email account no longer exists, but I > can still be reached at mmogilvi+qemu@zoho.com if desired. > > Regarding the patch, I haven't kept up with changes to qemu > well enough to really have a useful opinion, although superficially > it looks good. Perhaps it would make sense to include both > my old and new email addresses in the sign-off section, to both > tie it to the old threads and make it easier to reach me > if someone wants to contact me? (But I'm not sure of a "clean" > way to do so, and it is probably low priority. Perhaps this > current email is good enough?) Thanks for the followup. Unfortunately the commit is already in master so we can't tweak the commit message at this point. That shouldn't matter much, people don't often fish emails out of old commit message logs to email anyway :-) thanks -- PMM
On 1/10/19 9:53 AM, Peter Maydell wrote: >> Note that my miniinfo email account no longer exists, but I >> can still be reached at mmogilvi+qemu@zoho.com if desired. >> >> Regarding the patch, I haven't kept up with changes to qemu >> well enough to really have a useful opinion, although superficially >> it looks good. Perhaps it would make sense to include both >> my old and new email addresses in the sign-off section, to both >> tie it to the old threads and make it easier to reach me >> if someone wants to contact me? (But I'm not sure of a "clean" >> way to do so, and it is probably low priority. Perhaps this >> current email is good enough?) > > Thanks for the followup. Unfortunately the commit is already > in master so we can't tweak the commit message at this point. > That shouldn't matter much, people don't often fish emails out > of old commit message logs to email anyway :-) We can't change the commit message body, but if you are worried about people still being able to reach you based about a patch they found while browsing git history, we DO have .mailmap where you can list a new preferred email address to use in place of any commits you made under an old address. Patching .mailmap is a nice (but optional) way to let people know how to reach you, if you still expect to be reached. -- Eric Blake, Principal Software Engineer Red Hat, Inc. +1-919-301-3226 Virtualization: qemu.org | libvirt.org
diff --git a/tests/libqtest.h b/tests/libqtest.h index 9758c51be61..7ea94139b0c 100644 --- a/tests/libqtest.h +++ b/tests/libqtest.h @@ -230,6 +230,19 @@ void qtest_irq_intercept_in(QTestState *s, const char *string); */ void qtest_irq_intercept_out(QTestState *s, const char *string); +/** + * qtest_set_irq_in: + * @s: QTestState instance to operate on. + * @string: QOM path of a device + * @name: IRQ name + * @irq: IRQ number + * @level: IRQ level + * + * Force given device/irq GPIO-in pin to the given level. + */ +void qtest_set_irq_in(QTestState *s, const char *string, const char *name, + int irq, int level); + /** * qtest_outb: * @s: #QTestState instance to operate on. diff --git a/qtest.c b/qtest.c index 69b9e9962b5..451696b5dae 100644 --- a/qtest.c +++ b/qtest.c @@ -164,6 +164,17 @@ static bool qtest_opened; * where NUM is an IRQ number. For the PC, interrupts can be intercepted * simply with "irq_intercept_in ioapic" (note that IRQ0 comes out with * NUM=0 even though it is remapped to GSI 2). + * + * Setting interrupt level: + * + * > set_irq_in QOM-PATH NAME NUM LEVEL + * < OK + * + * where NAME is the name of the irq/gpio list, NUM is an IRQ number and + * LEVEL is an signed integer IRQ level. + * + * Forcibly set the given interrupt pin to the given level. + * */ static int hex2nib(char ch) @@ -326,7 +337,39 @@ static void qtest_process_command(CharBackend *chr, gchar **words) irq_intercept_dev = dev; qtest_send_prefix(chr); qtest_send(chr, "OK\n"); + } else if (strcmp(words[0], "set_irq_in") == 0) { + DeviceState *dev; + qemu_irq irq; + char *name; + int ret; + int num; + int level; + g_assert(words[1] && words[2] && words[3] && words[4]); + + dev = DEVICE(object_resolve_path(words[1], NULL)); + if (!dev) { + qtest_send_prefix(chr); + qtest_send(chr, "FAIL Unknown device\n"); + return; + } + + if (strcmp(words[2], "unnamed-gpio-in") == 0) { + name = NULL; + } else { + name = words[2]; + } + + ret = qemu_strtoi(words[3], NULL, 0, &num); + g_assert(!ret); + ret = qemu_strtoi(words[4], NULL, 0, &level); + g_assert(!ret); + + irq = qdev_get_gpio_in_named(dev, name, num); + + qemu_set_irq(irq, level); + qtest_send_prefix(chr); + qtest_send(chr, "OK\n"); } else if (strcmp(words[0], "outb") == 0 || strcmp(words[0], "outw") == 0 || strcmp(words[0], "outl") == 0) { diff --git a/tests/libqtest.c b/tests/libqtest.c index 1d75d3c9363..55750dd68de 100644 --- a/tests/libqtest.c +++ b/tests/libqtest.c @@ -753,6 +753,16 @@ void qtest_irq_intercept_in(QTestState *s, const char *qom_path) qtest_rsp(s, 0); } +void qtest_set_irq_in(QTestState *s, const char *qom_path, const char *name, + int num, int level) +{ + if (!name) { + name = "unnamed-gpio-in"; + } + qtest_sendf(s, "set_irq_in %s %s %d %d\n", qom_path, name, num, level); + qtest_rsp(s, 0); +} + static void qtest_out(QTestState *s, const char *cmd, uint16_t addr, uint32_t value) { qtest_sendf(s, "%s 0x%x 0x%x\n", cmd, addr, value);