From: Christian Ehrhardt <ehrhardt@linux.vnet.ibm.com> To: Andre Przywara <andre.przywara@amd.com> Cc: aliguori@us.ibm.com, Avi Kivity <avi@redhat.com>, qemu-devel@nongnu.org, kvm@vger.kernel.org Subject: Re: [PATCH] qemu: fix configuring kvm probe when using --kerneldir Date: Mon, 12 Jan 2009 07:42:34 +0100 [thread overview] Message-ID: <496AE65A.1000701@linux.vnet.ibm.com> (raw) In-Reply-To: <496760F3.7020902@amd.com> Andre Przywara wrote: > ehrhardt@linux.vnet.ibm.com wrote: >> From: Christian Ehrhardt <ehrhardt@linux.vnet.ibm.com> >> >> There is already a variable kvm_cflags which gets the path of the kernel >> includes when using --kerneldir. But eventually with newer kernels we >> all will >> need arch/$arch/include too (my case was a incldue of asm/kvm.h which >> was not >> found anymore). Headers in a full kernel source are not flattened to >> one arch like they are if e.g. installed kernel headers are used. > I also stumbled over this recently (in kvm-userspace.git), but I had > problems with the qemu part not including KVM support because in > qemu/configure the KVM build test failed due to the missing asm/kvm.h. > I saw that --kerneldir gets not propagated to qemu, but > libkvm_kerneldir instead, which is hardcoded to point to `pwd`/kernel. > Shouldn't that be fixed, too? > I use kvm-userspace.git and a not-installed kernel from kvm.git for > compiling, so I say "./configure --kerneldir=/src/kvm.git > --with-patched-kernel". I eventually hacked KVM's configure to > propagate --kerneldir to qemu and added arch/x86/include to the > include path in qemu/configure. This is of course a hack (that's why I > don't append it here), but it worked ;-) > If someone proposes a clean and easy way to solve this, I'd be happy > to write a patch. I know this issue and reported it ~a month ago. I also had issues compiling against a --kerneldir kernel because the libkvm_kerneldir was propagated. Eventually in the discussion it came up that we don't need to fix configure "technically", but maybe we should find a way to better inform users/developüers about this (I guess up to 99% that this works for you in kvm-userspace too): (in a clean kvm-userspace) cd kernel make sync LINUX=path/to/your/kerneldir cd .. ./configure opt=whateveryouwant This way your kerneldir is synced and flattened into kvm-userspace and propagating libkvm_kerneldir is fine since that are your kerneldir headers now. Maybe a "fix" would be that if --kerneldir is provided to configure it has to ensure that THIS kerneldir is synced in before continuing. You should be aware that the fix I sent on Friday was for plain qemu which doesn't have that kernel subdir indirection and therefore works a bit different. >> To fix that, the includes added to cflags depending on --kerneldir >> should also >> contian the arch includes. The patch adds a special check for x86 >> because its >> source layout recently changed, all others directly use >> arch/$cpu/include if >> existent. > This is one problem I also noticed. $cpu is not the same as the Linux' > arch name, is there a suitable variable or do we have to do a large > switch/case? I looked around and there was no real 1:1 matching variable. But fortunately $cpu is similar enough to simplify that swicth/case a lot like I did in my patch here. > Regards, > Andre. > > >> >> Signed-off-by: Christian Ehrhardt <ehrhardt@linux.vnet.ibm.com> >> --- >> >> [diffstat] >> configure | 6 ++++++ >> 1 file changed, 6 insertions(+) >> >> [diff] >> diff --git a/configure b/configure >> --- a/configure >> +++ b/configure >> @@ -963,6 +963,12 @@ EOF >> EOF >> if test "$kerneldir" != "" ; then >> kvm_cflags=-I"$kerneldir"/include >> + if test \( "$cpu" = "i386" -o "$cpu" = "x86_64" \) \ >> + -a -d "$kerneldir/arch/x86/include" ; then >> + kvm_cflags="$kvm_cflags -I$kerneldir/arch/x86/include" >> + elif test -d "$kerneldir/arch/$cpu/include" ; then >> + kvm_cflags="$kvm_cflags -I$kerneldir/arch/$cpu/include" >> + fi >> else >> kvm_cflags="" >> fi -- Grüsse / regards, Christian Ehrhardt IBM Linux Technology Center, Open Virtualization
WARNING: multiple messages have this Message-ID (diff)
From: Christian Ehrhardt <ehrhardt@linux.vnet.ibm.com> To: Andre Przywara <andre.przywara@amd.com> Cc: aliguori@us.ibm.com, Avi Kivity <avi@redhat.com>, kvm@vger.kernel.org, qemu-devel@nongnu.org Subject: [Qemu-devel] Re: [PATCH] qemu: fix configuring kvm probe when using --kerneldir Date: Mon, 12 Jan 2009 07:42:34 +0100 [thread overview] Message-ID: <496AE65A.1000701@linux.vnet.ibm.com> (raw) In-Reply-To: <496760F3.7020902@amd.com> Andre Przywara wrote: > ehrhardt@linux.vnet.ibm.com wrote: >> From: Christian Ehrhardt <ehrhardt@linux.vnet.ibm.com> >> >> There is already a variable kvm_cflags which gets the path of the kernel >> includes when using --kerneldir. But eventually with newer kernels we >> all will >> need arch/$arch/include too (my case was a incldue of asm/kvm.h which >> was not >> found anymore). Headers in a full kernel source are not flattened to >> one arch like they are if e.g. installed kernel headers are used. > I also stumbled over this recently (in kvm-userspace.git), but I had > problems with the qemu part not including KVM support because in > qemu/configure the KVM build test failed due to the missing asm/kvm.h. > I saw that --kerneldir gets not propagated to qemu, but > libkvm_kerneldir instead, which is hardcoded to point to `pwd`/kernel. > Shouldn't that be fixed, too? > I use kvm-userspace.git and a not-installed kernel from kvm.git for > compiling, so I say "./configure --kerneldir=/src/kvm.git > --with-patched-kernel". I eventually hacked KVM's configure to > propagate --kerneldir to qemu and added arch/x86/include to the > include path in qemu/configure. This is of course a hack (that's why I > don't append it here), but it worked ;-) > If someone proposes a clean and easy way to solve this, I'd be happy > to write a patch. I know this issue and reported it ~a month ago. I also had issues compiling against a --kerneldir kernel because the libkvm_kerneldir was propagated. Eventually in the discussion it came up that we don't need to fix configure "technically", but maybe we should find a way to better inform users/developüers about this (I guess up to 99% that this works for you in kvm-userspace too): (in a clean kvm-userspace) cd kernel make sync LINUX=path/to/your/kerneldir cd .. ./configure opt=whateveryouwant This way your kerneldir is synced and flattened into kvm-userspace and propagating libkvm_kerneldir is fine since that are your kerneldir headers now. Maybe a "fix" would be that if --kerneldir is provided to configure it has to ensure that THIS kerneldir is synced in before continuing. You should be aware that the fix I sent on Friday was for plain qemu which doesn't have that kernel subdir indirection and therefore works a bit different. >> To fix that, the includes added to cflags depending on --kerneldir >> should also >> contian the arch includes. The patch adds a special check for x86 >> because its >> source layout recently changed, all others directly use >> arch/$cpu/include if >> existent. > This is one problem I also noticed. $cpu is not the same as the Linux' > arch name, is there a suitable variable or do we have to do a large > switch/case? I looked around and there was no real 1:1 matching variable. But fortunately $cpu is similar enough to simplify that swicth/case a lot like I did in my patch here. > Regards, > Andre. > > >> >> Signed-off-by: Christian Ehrhardt <ehrhardt@linux.vnet.ibm.com> >> --- >> >> [diffstat] >> configure | 6 ++++++ >> 1 file changed, 6 insertions(+) >> >> [diff] >> diff --git a/configure b/configure >> --- a/configure >> +++ b/configure >> @@ -963,6 +963,12 @@ EOF >> EOF >> if test "$kerneldir" != "" ; then >> kvm_cflags=-I"$kerneldir"/include >> + if test \( "$cpu" = "i386" -o "$cpu" = "x86_64" \) \ >> + -a -d "$kerneldir/arch/x86/include" ; then >> + kvm_cflags="$kvm_cflags -I$kerneldir/arch/x86/include" >> + elif test -d "$kerneldir/arch/$cpu/include" ; then >> + kvm_cflags="$kvm_cflags -I$kerneldir/arch/$cpu/include" >> + fi >> else >> kvm_cflags="" >> fi -- Grüsse / regards, Christian Ehrhardt IBM Linux Technology Center, Open Virtualization
next prev parent reply other threads:[~2009-01-12 6:42 UTC|newest] Thread overview: 8+ messages / expand[flat|nested] mbox.gz Atom feed top 2009-01-09 12:10 [PATCH] qemu: fix configuring kvm probe when using --kerneldir ehrhardt 2009-01-09 12:10 ` [Qemu-devel] " ehrhardt 2009-01-09 14:36 ` Andre Przywara 2009-01-09 14:36 ` [Qemu-devel] " Andre Przywara 2009-01-12 6:42 ` Christian Ehrhardt [this message] 2009-01-12 6:42 ` Christian Ehrhardt 2009-01-09 20:05 ` Anthony Liguori 2009-01-09 20:05 ` [Qemu-devel] " Anthony Liguori
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=496AE65A.1000701@linux.vnet.ibm.com \ --to=ehrhardt@linux.vnet.ibm.com \ --cc=aliguori@us.ibm.com \ --cc=andre.przywara@amd.com \ --cc=avi@redhat.com \ --cc=kvm@vger.kernel.org \ --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: linkBe sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes, see mirroring instructions on how to clone and mirror all data and code used by this external index.