Dear all,
The below changeset is a from scratch & pledged C implementation of the
'ts' perl utility found in the moreutils collection by Joey Hess.
OK to add to /usr/bin?
Kind regards,
Job
Index: ts/Makefile
===
RCS file: ts/Makef
Hello,
I like the idea. I would just consider to make ip6_hbhchcheck()
to always free mp on error and set NULL as return value.
for example here in current ip6_input_if() we may leak memory
on error:
308 int
309 ip6_input_if(struct mbuf **mp, int *offp, int nxt, int af, struct ifnet
*ifp)
31
Hi,
Deep in ip6_hbhchcheck() may modify the mbuf pointer. Instead of
having dangling pointer that we must not use, pass pointer to mbuf
pointer and keep it consistent in the caller.
ok?
bluhm
Index: netinet6/ip6_input.c
===
RCS fi
On Tue, Jun 28 2022, Martin Pieuchot wrote:
> On 28/06/22(Tue) 18:17, Jeremie Courreges-Anglas wrote:
>>
>> Initially I just wandered in syscall_mi.h and found the locking scheme
>> super weird, even if technically correct. pledge_syscall() better be
>> safe to call without the kernel lock so I
On Tue, Jun 28, 2022 at 09:05:09PM +0300, Vitaliy Makkoveev wrote:
> The updated diff. It was triggered by Hrvoje Popovski, we could do
> direct (*if_qstart)() call in pipex(4) PPPOE input path, and this
> provides deadlock. The updated diff uses ip{,6}_send() instead of
> ipv{4,6}_input().
>
I th
On 28/06/22(Tue) 18:17, Jeremie Courreges-Anglas wrote:
>
> Initially I just wandered in syscall_mi.h and found the locking scheme
> super weird, even if technically correct. pledge_syscall() better be
> safe to call without the kernel lock so I don't understand why we're
> sometimes calling it w
The updated diff. It was triggered by Hrvoje Popovski, we could do
direct (*if_qstart)() call in pipex(4) PPPOE input path, and this
provides deadlock. The updated diff uses ip{,6}_send() instead of
ipv{4,6}_input().
r420-1# witness: acquiring duplicate lock of same type: "&session->pxs_mtx"
1st
Initially I just wandered in syscall_mi.h and found the locking scheme
super weird, even if technically correct. pledge_syscall() better be
safe to call without the kernel lock so I don't understand why we're
sometimes calling it with the kernel lock and sometimes not.
ps_pledge is 64 bits so i
On 27/06/22(Mon) 15:44, Martin Pieuchot wrote:
> Diff below contain 3 parts that can be committed independently. The 3
> of them are necessary to allow the pagedaemon to make progress in OOM
> situation and to satisfy all the allocations waiting for pages in
> specific ranges.
>
> * uvm/uvm_pager
This is bgplgd that can be used to provide a JSON REST API for bgpd.
It can be used to integrate bgpd into external looking glasses.
The bgplgd uses a modified version of slowcgi. It forks bgpctl per request
and returns the output via the slowcgi socket.
Securing the API needs to be done in the ht
We can't predict netlock state when we executing pipex(4) related
(*if_qstart)() handlers, so we can't use netlock for pipex(4)
protection.
We already use `pipex_list_mtx' for global pipex(4) data, so
Use per-session `pxs_mtx' mutex(9) for pipex session context protection.
We already use `pipex_l
On Sat, Jun 25, 2022 at 04:21:54AM -0700, Mike Larkin wrote:
> On Fri, Jun 24, 2022 at 07:19:47PM +1000, Jonathan Gray wrote:
> > +void
> > +amd_cpu_cacheinfo(struct cpu_info *ci)
> > +{
> > + u_int eax, ebx, ecx, edx;
> > +
> > + /* used by vmm */
> > +
>
> If this is the only user of these f
I'd like to abstract the use of PG_WANTED to start unifying & cleaning
the various cases where a code path is waiting for a busy page. Here's
the first step.
ok?
Index: uvm/uvm_amap.c
===
RCS file: /cvs/src/sys/uvm/uvm_amap.c,v
retr
The current handling of network interfaces in the installer is rather
confusing when somebody has an unsupported network device - it offers
to configure vlan0, but there is no interface on which a vlan can run
anyway.
Available network interfaces are: vlan0.
Which network interface do you wish to
On Tue, Jun 28, 2022 at 12:44:30PM +0300, Vitaliy Makkoveev wrote:
> ether_input() called with shared netlock, but pppoe(4) wants it to be
> exclusive. Do the pppoe(4) input within netisr handler with exclusive
> netlok held, and remove kernel lock hack from ether_input().
>
> This is the step bac
On Tue, 2022-06-28 at 12:21 +0200, Martijn van Duren wrote:
> On Tue, 2022-06-28 at 10:21 +0200, Martijn van Duren wrote:
> > Back in 2020 florian@ added the filter-pf-addresses keyword.
> > Although useful, I always felt it was a bit too case-specific. The diff
> > below adds a new blocklist featu
On Tue, 2022-06-28 at 10:21 +0200, Martijn van Duren wrote:
> Back in 2020 florian@ added the filter-pf-addresses keyword.
> Although useful, I always felt it was a bit too case-specific. The diff
> below adds a new blocklist feature/backend, which takes hold of an
> entire subtree, effectively rem
On Tue, Jun 28, 2022 at 10:51:29AM +0200, Alexander Bluhm wrote:
> arpintr() looks MP safe and I cannot trigger a crash with this diff.
Running parallel forwarding and arp -d in a loop triggers route
NULL pointer dereference after a while. I have to figure out if
it is related to this diff.
bluh
As noticed by landry, going to the ports tree for dependencies can fail
hilariously in case of a PORTSDIR_PATH.
Most specifically, assuming you are building something like firefox-102
in one of your ports directory, when building the debug package,
dependency look-up in pkg_create is going to foll
ether_input() called with shared netlock, but pppoe(4) wants it to be
exclusive. Do the pppoe(4) input within netisr handler with exclusive
netlok held, and remove kernel lock hack from ether_input().
This is the step back, but it makes ether_input() path better then it
is now.
ok?
Index: sys/ne
ok kettenis@
> Op 28-06-2022 07:20 schreef Anton Lindqvist :
>
>
> Hi,
> No need to pass a copy of the bcmdmac_channel structure to predicate
> routines.
>
> Comments? OK?
>
> diff --git sys/dev/fdt/bcm2835_dmac.c sys/dev/fdt/bcm2835_dmac.c
> index 145810dd7af..e9b31af568c 100644
> --- sys/de
meh
not really worth fixing, but since you already wrote the diff
ok kettenis@
> Op 28-06-2022 07:23 schreef Anton Lindqvist :
>
>
> Hi,
> A com_acpi_softc pointer is used as the interrupt callback cookie which
> is later on interpreted as a com_softc pointer. This is not a problem in
> pract
When playing with the blocklist I noticed that the oid directive only
supports numeric OIDs. So if florian wants to use filter-pf-table in the
future he'd have to set:
blocklist 1.3.6.1.4.1.30155.1.9.129
which is pure insanity, if:
blocklist pfTblAddrTable
could be an option.
Diff below changes th
Hi,
arpintr() looks MP safe and I cannot trigger a crash with this diff.
Hrvoje, can you try it?
bluhm
Index: net/if.c
===
RCS file: /data/mirror/openbsd/cvs/src/sys/net/if.c,v
retrieving revision 1.654
diff -u -p -r1.654 if.c
---
Back in 2020 florian@ added the filter-pf-addresses keyword.
Although useful, I always felt it was a bit too case-specific. The diff
below adds a new blocklist feature/backend, which takes hold of an
entire subtree, effectively removing it from the tree.
With this I've deprecated the filter-pf-add
Stefan Hagen wrote:
> Stuart Henderson wrote (2022-06-28 08:13 CEST):
> > On 2022/06/27 17:12, Bryan Steele wrote:
> > >
> > > Shouldn't this also take into consideration hw.power as well? If it
> > > doesn't make sense for perfpolicy=high then it probably doesn't for
> > > perfpolicy=auto when
26 matches
Mail list logo
