Re: try to map 64-bit mem space first

2018-04-09 Thread Mike Larkin
On Mon, Apr 09, 2018 at 08:27:49PM +0200, Mark Kettenis wrote: > > Date: Mon, 9 Apr 2018 18:31:49 +0200 > > From: Patrick Wildt > > > > Hi, > > > > some (probably newer) re(4) cards don't have the 32-bit memory BAR that > > we try to map first. Instead there's a 64-bit memory BAR in the follow-

Re: ksh: fix buffer overflow in u64ton

2018-04-09 Thread Theo Buehler
On Mon, Apr 09, 2018 at 08:56:28PM +0200, Tobias Stoeckmann wrote: > As tb@ pointed out, u64ton can overflow on ULONG_MAX. It could also > happen on systems with 64 bit int and INT_MIN, although we don't have > such a system supported by our code base. > > You can reach the u64ton function by prin

ksh: fix buffer overflow in u64ton

2018-04-09 Thread Tobias Stoeckmann
As tb@ pointed out, u64ton can overflow on ULONG_MAX. It could also happen on systems with 64 bit int and INT_MIN, although we don't have such a system supported by our code base. You can reach the u64ton function by printing the length of a string within a variable like this: $ a=string $ echo $

Re: try to map 64-bit mem space first

2018-04-09 Thread Mark Kettenis
> Date: Mon, 9 Apr 2018 18:31:49 +0200 > From: Patrick Wildt > > Hi, > > some (probably newer) re(4) cards don't have the 32-bit memory BAR that > we try to map first. Instead there's a 64-bit memory BAR in the follow- > ing BAR. On the MacchiatoBIN, where we currently do not support the IO >

more netinet bcopy -> mem{cpy|move}

2018-04-09 Thread David Hill
Hello - Two bcopy->memmove conversions where src and dst are in the same memory space. Perhaps not overlapping, but just being safe. Convert the bcopy's in tcp_respond to memcpy since the destination is a new mbuf header. OK? Index: netinet/ip_ah.c =

Re: subr_autoconf: allow _attach to fail? w/no void2int churn option

2018-04-09 Thread Theo de Raadt
I think this approach is wrong, insane, and fragile. DVF_ACTIVE doesn't work precisely that way. A better approach would be a whole-tree change such that attach functions are non-void. This has been proposed a few times, but noone ever sat down and did the whole-tree work. > could we allow conf

Re: video(1): don't use optarg as format string

2018-04-09 Thread Todd C. Miller
On Mon, 09 Apr 2018 11:56:46 -0500, Scott Cheloha wrote: > Always nifty: > > $ video -f '%d%d%d' > video: 31668978116843009-2139062144: No such file or directory > > So, use strlcpy and check for truncation. > > ok? OK millert@ - todd

video(1): don't use optarg as format string

2018-04-09 Thread Scott Cheloha
Always nifty: $ video -f '%d%d%d' video: 31668978116843009-2139062144: No such file or directory So, use strlcpy and check for truncation. ok? -- Scott Cheloha Index: app/video/video.c === RCS file: /cvs/xenocara/app/video/video.c

Re: C99 def for 'struct fileops'

2018-04-09 Thread Alexander Bluhm
On Mon, Apr 09, 2018 at 02:21:13PM +0200, Martin Pieuchot wrote: > Like we did for 'struct protosw', I'd like to convert these definitions > to help for grepping. > > Ok? OK bluhm@ > Index: kern/kern_event.c > === > RCS file: /cvs/s

re(4): try to map 64-bit mem space first

2018-04-09 Thread Patrick Wildt
Hi, some (probably newer) re(4) cards don't have the 32-bit memory BAR that we try to map first. Instead there's a 64-bit memory BAR in the follow- ing BAR. On the MacchiatoBIN, where we currently do not support the IO space, this means we need to attempt to map the 64-bit memory BAR. This dif

Re: [PATCH 7/7] em: Add magic delay for HP elitebook 820 G3

2018-04-09 Thread Jonathan Gray
On Sat, Apr 07, 2018 at 01:35:31PM +0200, Stefan Fritsch wrote: > On Fri, 6 Apr 2018, Jonathan Gray wrote: > > > On Thu, Apr 05, 2018 at 09:57:23PM +0200, Stefan Fritsch wrote: > > > Add another magic 1ms delay that seems to help with some remaining > > > issues on an HP elitebook 820 G3 with i219

Re: [PATCH 4/7] em: Improve access logic for software flag

2018-04-09 Thread Jonathan Gray
On Thu, Apr 05, 2018 at 09:57:20PM +0200, Stefan Fritsch wrote: > Some em chips have a semaphore ("software flag") to synchronize access > to certain registers between OS and firmware (ME/AMT). > > Make the logic to get the flag match the logic in freebsd. This includes > higher timeouts and waiti

vput(9) for NFS

2018-04-09 Thread Martin Pieuchot
Diff below implements most of the missing locking goo for NFS without enabling it. It does the following: - Add a missing PDIRUNLOCK in nfs_lookup() - Change vrele(9) into vput(9) where necessary. nfs_nget() will in future return a locked NFSnode. Just like ffs_vget() returns a locked vnod

subr_autoconf: allow _attach to fail? w/no void2int churn option

2018-04-09 Thread Artturi Alm
Hi, could we allow config_attach() to 'fail' without leaving dead device behind? ie. with gpioow(4) it's easy to make a load of these while testing stuff, which does lead to inconsistent unit numbering unless you manually detach the broken ones before attaching another.. for devices which are fou

Re: SCHED_LOCK() vs kprintf_mutex

2018-04-09 Thread Mark Kettenis
> Date: Mon, 9 Apr 2018 13:51:34 +0200 > From: Martin Pieuchot > > Diff below avoids a lock ordering issue between the SCHED_LOCK() and > printf(9)'s mutex. This problem has been reported multiple times in > drm(4) code where debug printfs are placed between sleep_setup() and > sleep_finish(). >

C99 def for 'struct fileops'

2018-04-09 Thread Martin Pieuchot
Like we did for 'struct protosw', I'd like to convert these definitions to help for grepping. Ok? Index: kern/kern_event.c === RCS file: /cvs/src/sys/kern/kern_event.c,v retrieving revision 1.86 diff -u -p -r1.86 kern_event.c --- ker

SCHED_LOCK() vs kprintf_mutex

2018-04-09 Thread Martin Pieuchot
Diff below avoids a lock ordering issue between the SCHED_LOCK() and printf(9)'s mutex. This problem has been reported multiple times in drm(4) code where debug printfs are placed between sleep_setup() and sleep_finish(). The problem is easily fixed. logwakeup() must not be called while holding t

Re: httpd: listen on all IPv4 and IPv6 addresses for "listen on *"

2018-04-09 Thread Klemens Nanni
On Mon, Apr 09, 2018 at 10:58:43AM +0200, Florian Obser wrote: > This shuffles things around to make httpd listen on v4 and v6 for *. Thanks for beating me to it! This also fixes `listen on egress', as it would previously listen on my IPv6 link-local address only. With this diff httpd listens on a

Re: fstat -r flag to display rdomains on sockets

2018-04-09 Thread Henning Brauer
* Stuart Henderson [2018-04-07 12:51]: > How about skipping the -r flag, but printing it after the existing > "internet [...] addr:port", and only printing if it's a non-default > table? This should minimise breakage, and it's not like the format > is cast in stone (spliced sockets were added in t

httpd: listen on all IPv4 and IPv6 addresses for "listen on *"

2018-04-09 Thread Florian Obser
This shuffles things around to make httpd listen on v4 and v6 for *. OK? diff --git httpd.conf.5 httpd.conf.5 index afda0ac132b..3194a3400c2 100644 --- httpd.conf.5 +++ httpd.conf.5 @@ -52,12 +52,12 @@ addresses of the specified network interface. If .Sq * is given as an address, -it will be

Re: ldomctl.8: Do not discourage IPv6 usage

2018-04-09 Thread Stefan Sperling
On Mon, Apr 09, 2018 at 01:04:00AM +0200, Jeremie Courreges-Anglas wrote: > On Mon, Apr 09 2018, Klemens Nanni wrote: > > I was wondering for a second whether vnet(4) had IPv6 related bugs or so. > > Probably Mark's preferences, from when link-local IPv6 addresses were > created by default. > >

Re: Adding current bandwidth to ftp(1)

2018-04-09 Thread Matthias Schmidt
Hi Christian, On 08.04.2018 22:42, Christian Barthel wrote: Hello, i am often sitting behind a very slow internet link and thought, it might be useful to show the current bandwidth while downloading something with ftp(1). Since I sometimes work with crappy WLANs/connections I very much appre