Message ID | 20250307215623.524987-1-pierrick.bouvier@linaro.org (mailing list archive) |
---|---|
Headers | show |
Series | hw/hyperv: remove duplication compilation units | expand |
On 7/3/25 22:56, Pierrick Bouvier wrote: > Work towards having a single binary, by removing duplicated object files. > > hw/hyperv/hyperv.c was excluded at this time, because it depends on target > dependent symbols: > - from system/kvm.h > - kvm_check_extension > - kvm_vm_ioctl Bug, these should be declared outside of COMPILING_PER_TARGET. > - from exec/cpu-all.h | memory_ldst_phys.h.inc > - ldq_phys Yeah, not an easy one.
Hi Maciej, On 7/3/25 22:56, Pierrick Bouvier wrote: > Work towards having a single binary, by removing duplicated object files. > Pierrick Bouvier (7): > hw/hyperv/hv-balloon-stub: common compilation unit > hw/hyperv/hyperv.h: header cleanup > hw/hyperv/vmbus: common compilation unit > hw/hyperv/hyperv-proto: move SYNDBG definition from target/i386 > hw/hyperv/syndbg: common compilation unit > hw/hyperv/balloon: common balloon compilation units > hw/hyperv/hyperv_testdev: common compilation unit If you are happy with this series and provide your Ack-by tag, I can take it in my next hw-misc pull request if that helps. Regards, Phil.
Hi Philippe, On 7.03.2025 23:25, Philippe Mathieu-Daudé wrote: > Hi Maciej, > > On 7/3/25 22:56, Pierrick Bouvier wrote: >> Work towards having a single binary, by removing duplicated object files. > >> Pierrick Bouvier (7): >> hw/hyperv/hv-balloon-stub: common compilation unit >> hw/hyperv/hyperv.h: header cleanup >> hw/hyperv/vmbus: common compilation unit >> hw/hyperv/hyperv-proto: move SYNDBG definition from target/i386 >> hw/hyperv/syndbg: common compilation unit >> hw/hyperv/balloon: common balloon compilation units >> hw/hyperv/hyperv_testdev: common compilation unit > > If you are happy with this series and provide your Ack-by tag, > I can take it in my next hw-misc pull request if that helps. There's nothing obviously wrong in the patch set, but if we can defer this to Monday then I could do a runtime check with a Windows VM too. > Regards, > > Phil. Thanks, Maciej
On 7/3/25 23:31, Maciej S. Szmigiero wrote: > Hi Philippe, > > On 7.03.2025 23:25, Philippe Mathieu-Daudé wrote: >> Hi Maciej, >> >> On 7/3/25 22:56, Pierrick Bouvier wrote: >>> Work towards having a single binary, by removing duplicated object >>> files. >> >>> Pierrick Bouvier (7): >>> hw/hyperv/hv-balloon-stub: common compilation unit >>> hw/hyperv/hyperv.h: header cleanup >>> hw/hyperv/vmbus: common compilation unit >>> hw/hyperv/hyperv-proto: move SYNDBG definition from target/i386 >>> hw/hyperv/syndbg: common compilation unit >>> hw/hyperv/balloon: common balloon compilation units >>> hw/hyperv/hyperv_testdev: common compilation unit >> >> If you are happy with this series and provide your Ack-by tag, >> I can take it in my next hw-misc pull request if that helps. > > There's nothing obviously wrong in the patch set, > but if we can defer this to Monday then I could do > a runtime check with a Windows VM too. Thanks, no rush on my side :)
On 3/7/25 14:06, Philippe Mathieu-Daudé wrote: > On 7/3/25 22:56, Pierrick Bouvier wrote: >> Work towards having a single binary, by removing duplicated object files. >> >> hw/hyperv/hyperv.c was excluded at this time, because it depends on target >> dependent symbols: >> - from system/kvm.h >> - kvm_check_extension >> - kvm_vm_ioctl > > Bug, these should be declared outside of COMPILING_PER_TARGET. > Yes, I noticed you solved it this way on one of your series. In the end, since QEMU does not have any specific boundaries between subsystems (i.e. no proper libraries with private/public symbols), we are mostly free to include any symbol we see in an header as long as it does not use a target dependent type which changes the signature. >> - from exec/cpu-all.h | memory_ldst_phys.h.inc >> - ldq_phys > > Yeah, not an easy one. >