On 26/5/22 18:21, Patrick Venture wrote:
On Thu, May 26, 2022 at 8:54 AM Peter Maydell <peter.mayd...@linaro.org
<mailto:peter.mayd...@linaro.org>> wrote:
On Fri, 25 Feb 2022 at 17:45, Hao Wu <wuhao...@google.com
<mailto:wuhao...@google.com>> wrote:
>
> From: Shengtan Mao <st...@google.com <mailto:st...@google.com>>
>
> Reviewed-by: Hao Wu <wuhao...@google.com
<mailto:wuhao...@google.com>>
> Reviewed-by: Chris Rauer <cra...@google.com
<mailto:cra...@google.com>>
> Signed-off-by: Shengtan Mao <st...@google.com
<mailto:st...@google.com>>
> Signed-off-by: Patrick Venture <vent...@google.com
<mailto:vent...@google.com>>
Hi; John Snow tells me that this test fails in the tests/vm/netbsd
VM (you can test this with 'make vm-build-netbsd') because the
assert() on the ftruncate() call fails:
> + ret = ftruncate(fd, NPCM7XX_TEST_IMAGE_SIZE);
> + g_assert_cmpint(ret, ==, 0);
> +#define NPCM7XX_TEST_IMAGE_SIZE (1 << 30)
I haven't investigated the exact cause, but this is a
gigabyte, right? That's a pretty massive file for a test case to
create -- can we make the test use a more sensible size of
sd card image ?
It looks like the nuvoton part had an issue with a smaller image size,
but we can resurrect that thread and poke at it a bit and see what
shakes out.
Could you use the null-co block driver instead?
-blockdev driver=null-co,size=1G,read-zeroes=on,node-name=null0
-device sd-card,drive=null0
