| Message ID | 1559132524-228613-1-git-send-email-andrey.shinkevich@virtuozzo.com |
|---|---|
| State | New |
| Headers | show |
| Series | hw/block/fdc: floppy command FIFO memory initialization | expand |
On 5/29/19 8:22 AM, Andrey Shinkevich wrote: > The uninitialized memory allocated for the command FIFO of the > floppy controller during the VM hardware initialization incurs > many unwanted reports by Valgrind when VM state is being saved. > That verbosity hardens a search for the real memory issues when > the iotests run. Particularly, the patch eliminates 20 unnecessary > reports of the Valgrind tool in the iotest #169. > > Signed-off-by: Andrey Shinkevich <andrey.shinkevich@virtuozzo.com> > --- > hw/block/fdc.c | 4 ++++ > 1 file changed, 4 insertions(+) > > diff --git a/hw/block/fdc.c b/hw/block/fdc.c > index 6f19f12..54e470c 100644 > --- a/hw/block/fdc.c > +++ b/hw/block/fdc.c > @@ -2647,6 +2647,10 @@ static void fdctrl_realize_common(DeviceState *dev, FDCtrl *fdctrl, > > FLOPPY_DPRINTF("init controller\n"); > fdctrl->fifo = qemu_memalign(512, FD_SECTOR_LEN); > + if (fdctrl->fifo) { > + /* To avoid using the uninitialized memory while saving VM state */ > + memset(fdctrl->fifo, 0, FD_SECTOR_LEN); > + } qemu_memalign doesn't look like it can fail (looking at util/oslib-posix); is this conditional necessary? I think you could just: fdctrl->fifo = qemu_memalign(512, FD_SECTOR_LEN); memset(fdctrl->fifo, 0, FD_SECTOR_LEN); > fdctrl->fifo_size = 512; > fdctrl->result_timer = timer_new_ns(QEMU_CLOCK_VIRTUAL, > fdctrl_result_timer, fdctrl); >
On 29/05/2019 16:40, John Snow wrote: > > > On 5/29/19 8:22 AM, Andrey Shinkevich wrote: >> The uninitialized memory allocated for the command FIFO of the >> floppy controller during the VM hardware initialization incurs >> many unwanted reports by Valgrind when VM state is being saved. >> That verbosity hardens a search for the real memory issues when >> the iotests run. Particularly, the patch eliminates 20 unnecessary >> reports of the Valgrind tool in the iotest #169. >> >> Signed-off-by: Andrey Shinkevich <andrey.shinkevich@virtuozzo.com> >> --- >> hw/block/fdc.c | 4 ++++ >> 1 file changed, 4 insertions(+) >> >> diff --git a/hw/block/fdc.c b/hw/block/fdc.c >> index 6f19f12..54e470c 100644 >> --- a/hw/block/fdc.c >> +++ b/hw/block/fdc.c >> @@ -2647,6 +2647,10 @@ static void fdctrl_realize_common(DeviceState *dev, FDCtrl *fdctrl, >> >> FLOPPY_DPRINTF("init controller\n"); >> fdctrl->fifo = qemu_memalign(512, FD_SECTOR_LEN); >> + if (fdctrl->fifo) { >> + /* To avoid using the uninitialized memory while saving VM state */ >> + memset(fdctrl->fifo, 0, FD_SECTOR_LEN); >> + } > > qemu_memalign doesn't look like it can fail (looking at > util/oslib-posix); is this conditional necessary? > > I think you could just: > > fdctrl->fifo = qemu_memalign(512, FD_SECTOR_LEN); > memset(fdctrl->fifo, 0, FD_SECTOR_LEN); > >> fdctrl->fifo_size = 512; >> fdctrl->result_timer = timer_new_ns(QEMU_CLOCK_VIRTUAL, >> fdctrl_result_timer, fdctrl); >> Yes, that's right. Thank you, John. Andrey
On 5/29/19 9:56 AM, Andrey Shinkevich wrote: > > > On 29/05/2019 16:40, John Snow wrote: >> >> >> On 5/29/19 8:22 AM, Andrey Shinkevich wrote: >>> The uninitialized memory allocated for the command FIFO of the >>> floppy controller during the VM hardware initialization incurs >>> many unwanted reports by Valgrind when VM state is being saved. >>> That verbosity hardens a search for the real memory issues when >>> the iotests run. Particularly, the patch eliminates 20 unnecessary >>> reports of the Valgrind tool in the iotest #169. >>> >>> Signed-off-by: Andrey Shinkevich <andrey.shinkevich@virtuozzo.com> >>> --- >>> hw/block/fdc.c | 4 ++++ >>> 1 file changed, 4 insertions(+) >>> >>> diff --git a/hw/block/fdc.c b/hw/block/fdc.c >>> index 6f19f12..54e470c 100644 >>> --- a/hw/block/fdc.c >>> +++ b/hw/block/fdc.c >>> @@ -2647,6 +2647,10 @@ static void fdctrl_realize_common(DeviceState *dev, FDCtrl *fdctrl, >>> >>> FLOPPY_DPRINTF("init controller\n"); >>> fdctrl->fifo = qemu_memalign(512, FD_SECTOR_LEN); >>> + if (fdctrl->fifo) { >>> + /* To avoid using the uninitialized memory while saving VM state */ >>> + memset(fdctrl->fifo, 0, FD_SECTOR_LEN); >>> + } >> >> qemu_memalign doesn't look like it can fail (looking at >> util/oslib-posix); is this conditional necessary? >> >> I think you could just: >> >> fdctrl->fifo = qemu_memalign(512, FD_SECTOR_LEN); >> memset(fdctrl->fifo, 0, FD_SECTOR_LEN); >> >>> fdctrl->fifo_size = 512; >>> fdctrl->result_timer = timer_new_ns(QEMU_CLOCK_VIRTUAL, >>> fdctrl_result_timer, fdctrl); >>> > > Yes, that's right. > Thank you, John. > > Andrey > Thanks for valgrinding QEMU :) --js
diff --git a/hw/block/fdc.c b/hw/block/fdc.c index 6f19f12..54e470c 100644 --- a/hw/block/fdc.c +++ b/hw/block/fdc.c @@ -2647,6 +2647,10 @@ static void fdctrl_realize_common(DeviceState *dev, FDCtrl *fdctrl, FLOPPY_DPRINTF("init controller\n"); fdctrl->fifo = qemu_memalign(512, FD_SECTOR_LEN); + if (fdctrl->fifo) { + /* To avoid using the uninitialized memory while saving VM state */ + memset(fdctrl->fifo, 0, FD_SECTOR_LEN); + } fdctrl->fifo_size = 512; fdctrl->result_timer = timer_new_ns(QEMU_CLOCK_VIRTUAL, fdctrl_result_timer, fdctrl);
The uninitialized memory allocated for the command FIFO of the floppy controller during the VM hardware initialization incurs many unwanted reports by Valgrind when VM state is being saved. That verbosity hardens a search for the real memory issues when the iotests run. Particularly, the patch eliminates 20 unnecessary reports of the Valgrind tool in the iotest #169. Signed-off-by: Andrey Shinkevich <andrey.shinkevich@virtuozzo.com> --- hw/block/fdc.c | 4 ++++ 1 file changed, 4 insertions(+)