Date: Tue, 22 Aug 2017 20:43:27 +0300 From: Konstantin Belousov <kostikbel@gmail.com> To: David Wolfskill <david@catwhisker.org>, current@freebsd.org Subject: Re: SIGSEGV in /bin/sh after r322740 -> r322776 update Message-ID: <20170822174327.GK1700@kib.kiev.ua> In-Reply-To: <20170822160703.GW1130@albert.catwhisker.org> References: <20170822114627.GC1130@albert.catwhisker.org> <20170822115923.GC1700@kib.kiev.ua> <20170822122836.GH1130@albert.catwhisker.org> <20170822123449.GD1700@kib.kiev.ua> <20170822124617.GN1130@albert.catwhisker.org> <20170822131958.GE1700@kib.kiev.ua> <20170822151738.GV1130@albert.catwhisker.org> <20170822153442.GG1700@kib.kiev.ua> <20170822160703.GW1130@albert.catwhisker.org>
next in thread | previous in thread | raw e-mail | index | archive | help
On Tue, Aug 22, 2017 at 09:07:03AM -0700, David Wolfskill wrote:
> On Tue, Aug 22, 2017 at 06:34:42PM +0300, Konstantin Belousov wrote:
> > ...
> > > Bisection time? Or if there's another approach (or even a suggestion
> > > for a revision to try first), I'm up for it. 9And yes, I'll just
> > > be rebuilding the kernel for the rest of this exercise, I think.
> > > That should speed things up significantly.)
> >
> > No need. It is clearly something with r322762 (more likely) or
> > r322763 (less likely).
>
> Ah.
>
> > Give me some time, I either fix it today or revert the commits.
>
> Cool. I'm at work now, but if there's anything I can do (e.g.,
> testing), I will do what I can: I have serial console access to the
> machine (for example).
Try this. The patch helped ae@, it seems.
I will commit it anyway in a hour, but more confirmations or nacks
would be good. This patch has some debugging bits which add noise on
console when a process traps. If this happens, please show me the lines.
Thank you for the patience.
diff --git a/sys/amd64/amd64/trap.c b/sys/amd64/amd64/trap.c
index e5a69d715a7..d1de62d89a9 100644
--- a/sys/amd64/amd64/trap.c
+++ b/sys/amd64/amd64/trap.c
@@ -147,7 +147,7 @@ static int prot_fault_translation;
SYSCTL_INT(_machdep, OID_AUTO, prot_fault_translation, CTLFLAG_RWTUN,
&prot_fault_translation, 0,
"Select signal to deliver on protection fault");
-static int uprintf_signal;
+static int uprintf_signal = 1;
SYSCTL_INT(_machdep, OID_AUTO, uprintf_signal, CTLFLAG_RWTUN,
&uprintf_signal, 0,
"Print debugging information on trap signal to ctty");
@@ -559,7 +559,7 @@ trap(struct trapframe *frame)
ksi.ksi_trapno = type;
ksi.ksi_addr = (void *)addr;
if (uprintf_signal) {
- uprintf("pid %d comm %s: signal %d err %lx code %d type %d "
+ printf("pid %d comm %s: signal %d err %lx code %d type %d "
"addr 0x%lx rsp 0x%lx rip 0x%lx "
"<%02x %02x %02x %02x %02x %02x %02x %02x>\n",
p->p_pid, p->p_comm, signo, frame->tf_err, ucode, type,
@@ -572,6 +572,8 @@ trap(struct trapframe *frame)
fubyte((void *)(frame->tf_rip + 5)),
fubyte((void *)(frame->tf_rip + 6)),
fubyte((void *)(frame->tf_rip + 7)));
+ printf("fsbase %#lx pcbfsbase %#lx flags %x\n", rdfsbase(),
+ td->td_pcb->pcb_fsbase, td->td_pcb->pcb_flags);
}
KASSERT((read_rflags() & PSL_I) != 0, ("interrupts disabled"));
trapsignal(td, &ksi);
diff --git a/sys/amd64/amd64/vm_machdep.c b/sys/amd64/amd64/vm_machdep.c
index db82da4c8fe..f71378b36f8 100644
--- a/sys/amd64/amd64/vm_machdep.c
+++ b/sys/amd64/amd64/vm_machdep.c
@@ -172,6 +172,7 @@ cpu_fork(struct thread *td1, struct proc *p2, struct thread *td2, int flags)
/* Ensure that td1's pcb is up to date. */
fpuexit(td1);
+ update_pcb_bases(td1->td_pcb);
/* Point the pcb to the top of the stack */
pcb2 = get_pcb_td(td2);
@@ -433,6 +434,7 @@ cpu_copy_thread(struct thread *td, struct thread *td0)
* Those not loaded individually below get their default
* values here.
*/
+ update_pcb_bases(td0->td_pcb);
bcopy(td0->td_pcb, pcb2, sizeof(*pcb2));
clear_pcb_flags(pcb2, PCB_FPUINITDONE | PCB_USERFPUINITDONE |
PCB_KERNFPU);
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20170822174327.GK1700>