similar to: Re: supermin init segfaults when kernel has large modules

On Fri, 27 May 2016 11:10:33 +0100 "Richard W.M. Jones" <rjones@redhat.com> wrote: > Continuing the discussion from: > https://bugzilla.redhat.com/show_bug.cgi?id=1339691#c17 > > [Luiz: There's no need to subscribe to the mailing list, once I've > moderated your first message the others will go through.] > > > IMO, supermin should use the kernel

The driver was renamed from ibmvscsic to ibmvscsi on kernel 3.7. See https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?id=9965c2f19be470c452357ae4f6304467cdeada55 The old name is kept so supermin works with recent kernels as well as kernel version older than 3.7. --- src/format_ext2_initrd.ml | 1 + 1 file changed, 1 insertion(+) diff --git a/src/format_ext2_initrd.ml

If the modules are unstripped and/or especially large, then the stack can overflow. --- init/init.c | 8 +++++++- 1 file changed, 7 insertions(+), 1 deletion(-) diff --git a/init/init.c b/init/init.c index 106be02..733d66e 100644 --- a/init/init.c +++ b/init/init.c @@ -314,7 +314,11 @@ insmod (const char *filename) exit (EXIT_FAILURE); } size = st.st_size; - char buf[size]; + char

On 03/17/2016 12:07, Richard W.M. Jones wrote: > Also drop the sr_mod module used by the deprecated guestfs_add_cdrom > interface. > --- > src/ext2_initrd.ml | 5 ----- > 1 file changed, 5 deletions(-) > > diff --git a/src/ext2_initrd.ml b/src/ext2_initrd.ml > index 730f806..5e841ea 100644 > --- a/src/ext2_initrd.ml > +++ b/src/ext2_initrd.ml > @@ -40,12 +40,7 @@

Also drop the sr_mod module used by the deprecated guestfs_add_cdrom interface. --- src/ext2_initrd.ml | 5 ----- 1 file changed, 5 deletions(-) diff --git a/src/ext2_initrd.ml b/src/ext2_initrd.ml index 730f806..5e841ea 100644 --- a/src/ext2_initrd.ml +++ b/src/ext2_initrd.ml @@ -40,12 +40,7 @@ let kmods = [ "virtio*.ko*"; "libata*.ko*"; "piix*.ko*"; -

On Sun, Mar 20, 2016 at 01:30:17PM +0200, Török Edwin wrote: > On 03/17/2016 12:07, Richard W.M. Jones wrote: > > Also drop the sr_mod module used by the deprecated guestfs_add_cdrom > > interface. > > --- > > src/ext2_initrd.ml | 5 ----- > > 1 file changed, 5 deletions(-) > > > > diff --git a/src/ext2_initrd.ml b/src/ext2_initrd.ml > > index

Dearl list, can anyone point me to a function or library that can create a graph similar to the one in the following powerpoint presentation? http://bmi.osu.edu/~khuang/IBGP705/BMI705-Lecture7.ppt (pages 36-37) In order to try to explain the graph, the way I see it in R terms is something like this: the "p-q" axis is a vector of positions (for example, seq(0,5000000,1)) the

When building the mini initrd, previously we copied the modules into the initrd as-is, so for example if the module was xz-compressed, we copied the foo.ko.xz file to the initrd. This requires that the mini init binary is linked to zlib & lzma, so that it knows how to uncompress these modules when insmoding them at boot time. Also since the init is statically linked, it required _static_

This patch series fixes the way supermin sorts the list of installed packages when resolving a name, picking the right package for the host architecture. Pino Toscano (5): rpm: do not unpack parameters rpm: fix version comparison rpm: query the RPM architecture rpm: fix package sorting (RHBZ#1696822) utils: remove unused 'compare_architecture' function src/librpm-c.c | 10

On Wednesday 17 February 2016 14:22:34 Richard W.M. Jones wrote: > --- > init/init.c | 9 +++++++++ > 1 file changed, 9 insertions(+) > > diff --git a/init/init.c b/init/init.c > index 25d6bc6..38392c4 100644 > --- a/init/init.c > +++ b/init/init.c > @@ -94,6 +94,15 @@ main () > print_uptime (); > fprintf (stderr, "supermin: ext2 mini initrd starting

No functional change, just change the style to the one used by libguestfs. --- init/init.c | 41 +++++++++++++++++++++++++---------------- 1 file changed, 25 insertions(+), 16 deletions(-) diff --git a/init/init.c b/init/init.c index e3d1107..473a5c5 100644 --- a/init/init.c +++ b/init/init.c @@ -96,6 +96,18 @@ static char line[1024]; int main () { + FILE *fp; + size_t n; + char *root,

Don't know why it works, but it works ... Rich.

Doing this breaks bash in virt-rescue with the error: bash: cannot set terminal process group (-1): Inappropriate ioctl for device bash: no job control in this shell Also ^C etc does not work. Removing this ioctl call fixes this. I noticed the problem because supermin's init compiled with dietlibc worked, since the dietlibc header files don't define TIOCSCTTY and hence the ioctl

--- init/init.c | 9 +++++++++ 1 file changed, 9 insertions(+) diff --git a/init/init.c b/init/init.c index 25d6bc6..38392c4 100644 --- a/init/init.c +++ b/init/init.c @@ -94,6 +94,15 @@ main () print_uptime (); fprintf (stderr, "supermin: ext2 mini initrd starting up: " PACKAGE_VERSION +#if defined(__dietlibc__) + " dietlibc" +#elif

We want to exclude virtio-gpu since it's irrelevant, large, and pulls in other unwanted dependencies (modeswitching, drm). Add a second list of kmods which is a blacklist, applied after the first. This reduces the libguestfs initrd size from 39M down to 17M, with concomitant small reductions in boot time. --- src/ext2_initrd.ml | 10 +++++++++- 1 file changed, 9 insertions(+), 1 deletion(-)

--- init/init.c | 19 ++++++++++++------- 1 file changed, 12 insertions(+), 7 deletions(-) diff --git a/init/init.c b/init/init.c index 473a5c5..a6279b8 100644 --- a/init/init.c +++ b/init/init.c @@ -101,13 +101,19 @@ main () char *root, *path; size_t len; int dax = 0; - uint64_t delay_ns = 250000; + uint64_t delay_ns; int virtio_message = 0; - struct timespec t; int major,

On Wednesday 17 February 2016 14:22:30 Richard W.M. Jones wrote: > v1 -> v2: > > - If we split out the init program into a separate init/ directory, > that makes it much easier to build against an alternate libc. > > I tried to build against uClibc, but uClibc requires an entire build > chain, which looked like it was going to be a massive ballache. Patches #1/2/3

On Wed, Apr 19, 2017 at 10:57:28AM +0100, Richard W.M. Jones wrote: > + struct timespec t; > +#define NANOSLEEP(ns) do { \ > + t.tv_sec = delay_ns / 1000000000; \ > + t.tv_nsec = delay_ns % 1000000000; \ > + nanosleep (&t, NULL); \ > + } while(0) > + Defining 't' outside its context is a bit

After supermin has finished running, the initramfs files sit around occupying swappable memory but serving no further purpose. This saves a little memory, at the cost of about 1ms of extra boot time. --- init/init.c | 50 ++++++++++++++++++++++++++++++++++++++++++++++++-- 1 file changed, 48 insertions(+), 2 deletions(-) diff --git a/init/init.c b/init/init.c index 733d66e..5ac53e9 100644 ---

v1 -> v2: - I fixed the end condition of the loop in parse_root_uuid. - Retested. Didn't change the busy wait loop into a function or macro, as per discussion on IRC. Rich.