From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from eggs.gnu.org ([2001:4830:134:3::10]:45864) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1YT62V-0006lW-Rd for qemu-devel@nongnu.org; Wed, 04 Mar 2015 04:56:31 -0500 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1YT62S-0004DW-KO for qemu-devel@nongnu.org; Wed, 04 Mar 2015 04:56:27 -0500 Received: from e28smtp07.in.ibm.com ([122.248.162.7]:46754) by eggs.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1YT62S-0004Aq-1K for qemu-devel@nongnu.org; Wed, 04 Mar 2015 04:56:24 -0500 Received: from /spool/local by e28smtp07.in.ibm.com with IBM ESMTP SMTP Gateway: Authorized Use Only! Violators will be prosecuted for from ; Wed, 4 Mar 2015 15:26:20 +0530 From: Nikunj A Dadhania In-Reply-To: <20150304103918.230ad0b6@oc7435384737.ibm.com> References: <1425387274-11133-1-git-send-email-nikunj@linux.vnet.ibm.com> <87zj7t2pit.fsf@blackfin.pond.sub.org> <87vbih6u7a.fsf@linux.vnet.ibm.com> <20150304095914.1d2af305@oc7435384737.ibm.com> <87sidl6smc.fsf@linux.vnet.ibm.com> <20150304101910.5dc7022b@oc7435384737.ibm.com> <87pp8p6rh2.fsf@linux.vnet.ibm.com> <20150304103918.230ad0b6@oc7435384737.ibm.com> Date: Wed, 04 Mar 2015 15:26:16 +0530 Message-ID: <87mw3t6q7z.fsf@linux.vnet.ibm.com> MIME-Version: 1.0 Content-Type: text/plain Subject: Re: [Qemu-devel] [PATCH] Introduce machine specific default memory size List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: Thomas Huth Cc: qemu-ppc@nongnu.org, Markus Armbruster , qemu-devel@nongnu.org Thomas Huth writes: > On Wed, 04 Mar 2015 14:59:13 +0530 > Nikunj A Dadhania wrote: > >> Thomas Huth writes: >> >> > On Wed, 04 Mar 2015 14:34:27 +0530 >> > Nikunj A Dadhania wrote: >> > >> >> Thomas Huth writes: >> >> >> >> > On Wed, 04 Mar 2015 14:00:17 +0530 >> >> > Nikunj A Dadhania wrote: >> >> > >> >> >> Markus Armbruster writes: >> >> >> >> >> >> > Nikunj A Dadhania writes: >> >> >> > >> >> >> >> diff --git a/vl.c b/vl.c >> >> >> >> index eb89d62..dd56754 100644 >> >> >> >> --- a/vl.c >> >> >> >> +++ b/vl.c >> >> >> >> @@ -4053,6 +4053,18 @@ int main(int argc, char **argv, char **envp) >> >> >> >> exit(1); >> >> >> >> } >> >> >> >> >> >> >> >> + if (machine_class->default_ram_size && ram_size < machine_class->default_ram_size) { >> >> >> >> + fprintf(stderr, "qemu: %s guest ram size defaulting to %ld MB\n", >> >> >> >> + machine_class->name, >> >> >> >> + machine_class->default_ram_size / (1024 * 1024)); >> >> >> > >> >> >> > If the user explicitly asks for something, we either provide it >> >> >> > silently, or we error out. This does neither. Why? >> >> >> >> >> >> In case the user has provided memory not enough to boot the machine, I >> >> >> could error out. My idea was to have a sane default which is provided by >> >> >> the machine. >> >> >> >> >> >> Initially, I had just "ram_size == default_ram_size", but then it was >> >> >> allowing "-m 128M" to go through. And the VM would not boot. >> >> >> >> >> >> This can as well be converted to an error report and fail here to boot >> >> >> the VM. >> >> > >> >> > What does exactly fail with 128MB? Linux? >> >> >> >> Linux kernel, and not much info as well on the console. >> > >> > Ok, but then I think it should still be possible to specify -m 128M on >> > the command line - in case the user wants to run an older Linux which >> > still works fine with that amount of memory. >> >> But how do we distinguish whether its old/new kernel in the distro? >> >> And the older kernel will boot with more memory, while the reverse isnt >> true. > > True, and it's IMHO certainly ok to increase the default memory size - > I just wanted to say that you should not disallow "-m 128M" in case the > users know what they are doing. That becomes a little bit tricky, as 128MB can come from default_ram_size and from the user as well. So here I am only changing the machine class for sPAPR. I have tried few distro installations and none of them boot with qemu default memory(128MB). Moreover, the installer kernel does not tell what went wrong. It silently goes till a point and is stuck. Regards Nikunj