Below the conversion to uiomove() for kern/spec_vnops.c. This diff
prevents truncation of uio_resid when passed to min().
Index: kern/spec_vnops.c
===
RCS file: /cvs/src/sys/kern/spec_vnops.c,v
retrieving revision 1.84
diff -u -p -u -
Below the uiomove() conversion for net/ppp_tty.c. M_TRAILINGSPACE()
returns int, but the result can't be negative, so using u_int for the
return value should be fine.
Index: net/ppp_tty.c
===
RCS file: /cvs/src/sys/net/ppp_tty.c,v
ret
Below the conversion from uiomovei() to uiomove() for kern/tty.c and
kern/tty_pty.c. 'cc' consistently contains small, non-negative integer
values, so leaving the type as int should be ok. It could as well be
changed to size_t, but I don't see a benefit in doing so for that
particular case, except
Stefan Kempf wrote:
> I'm a bit uneasy though with passing signed values as-is to uiomove().
> Can we somehow make it explicit that we know that the uiomove() argument is
> >= 0?
>
> Changing types all over the place would be too much churn though.
>
> I'm leaning towards an explicit size_t cast
I agree with Ingo, ls(1) shouldn't generate unsafe output. Regardless
of whether the output is to a terminal or some other file.
cheers,
natano
Hi,
below some comments regarding the vmctl disk image creation code.
Index: main.c
===
RCS file: /cvs/src/usr.sbin/vmctl/main.c,v
retrieving revision 1.13
diff -u -p -u -r1.13 main.c
--- main.c 5 Jan 2016 16:25:34 - 1
Below the conversion to uiomove() for isofs/udf/. Note that converting
size to size_t is not possible in udf_read(), as udf_readatoffset()
requires a pointer to an integer variable. Changing that would cause a
lot of code churn, so i chose to truncate uio_resid to INT_MAX instead.
udf_readatoffset(
Below the uiomove() conversion for msdosfs. This diff prevents
truncation of uio_resid in both msdosfs_read() and msdosfs_write().
Index: msdosfs/msdosfs_vnops.c
===
RCS file: /cvs/src/sys/msdosfs/msdosfs_vnops.c,v
retrieving revision
For what it's worth, your diff reads fine to me. It took me a while to
realize, that tlen in pppxwrite() and tun_dev_write() can't overflow
because of the range check beforehand, but on close look that code also
turns out to be correct.
cheers,
natano
On Fri, Jan 22, 2016 at 07:29:54PM +0100, St
Yes! This absolutely makes Youtube videos watchable for me (on a
Thinkpad T520). There still is occassional stuttering, but _far_ less
disruptive than before. Another usecase where I see improvements is
reloading a resource-heavy web page while switching tabs. Before
applying the patch, this caused
Below the conversion from uiomovei() to uiomove() for bktr. The patch
also replaces two occurrences of uio->uio_iov->iov_len with
uio->uio_resid. I don't see a reason why bktr should inspect iov_len
directly.
Index: dev/pci/bktr/bktr_core.c
=
On Wed, Jan 27, 2016 at 06:26:00AM +0100, Stefan Kempf wrote:
> Martin Natano wrote:
> > Below the conversion to uiomove() for isofs/udf/. Note that converting
> > size to size_t is not possible in udf_read(), as udf_readatoffset()
> > requires a pointer to an integer variable
In ufs, the calculation of i_modrev can produce signed overflow on 32
bit architectures (found on i386). The tv.tv_usec * 4294 calculation is
designed to move the microseconds part of a struct timeval to the upper
bits of an unsigned(!) 32 bit value to make room for simple i_modrev
increments, but
On Wed, Jan 27, 2016 at 12:27:46PM +0100, Mike Belopuhov wrote:
> On 27 January 2016 at 09:52, Martin Natano wrote:
> > In ufs, the calculation of i_modrev can produce signed overflow on 32
> > bit architectures (found on i386). The tv.tv_usec * 4294 calculation is
> >
Below the uiomove() conversion for nfs. I didn't change the type of 'n'
to be size_t, because it never exceeds the maximum rpc size (nm_rsize),
which is an integer too. (Also, to avoid unnecessary code churn.)
Index: nfs/nfs_bio.c
===
Below the conversion to uiomove() for ntfs. In the first three hunks the
size passed to uiomove(i) already was of type size_t. I also converted
the 'left' variable in ntfs_readattr() to size_t, because it tracks the
remainder of 'rsize', which also is size_t.
Index: ntfs/ntfs_subr.c
==
Hi,
The add_entropy_words() function performs a right shift by
(32 - entropy_input_rotate) bits, with entropy_input_rotate being an
integer between [0..31]. This can lead to a shift of 32 on a 32 bit
value, which is undefined behaviour in C. The standard says this: "If
the value of the right opera
On Tue, Feb 02, 2016 at 07:30:08PM +0100, Stefan Kempf wrote:
> Looks good. I agree with changing left to size_t. One small remark
> though: size_t is defined as unsigned long. Do the DPRINTFs that print
> the value of left have to be changed to use %zu in the format string?
Said DDPRINTFs are in f
> The following patch adds documentation for the APM_IOC_HIBERNATE ioctl
> to the manual pages.
There was a typo in the zaurus hunk. Updated diff below.
Index: share/man/man4/man4.amd64/apm.4
===
RCS file: /cvs/src/share/man/man4/ma
I think that 'cp += shellarglen + 1;' is a dead store. The 'cp' variable
is not used anymore after that line.
Either way your diff reads fine to me!
natano
> Index: sys/kern/exec_script.c
> ===
> RCS file: /cvs/src/sys/kern/exec_scr
Below the conversion to uiomovei() for ufs. While there I changed all
instances of 'blkoffset', 'size' and 'xfersize' that where declared as
long integers to be plain integers instead for consistency with the
surrounding code. These variables are limited by fs_bsize and e2fs_bsize
respectively, whi
> One unrelated thing noted while reviewing:
>
> ufs/ext2fs/ext2fs_readwrite.c:
> static int
> ext2_ind_read(struct vnode *vp, struct inode *ip, struct m_ext2fs *fs,
> struct uio *uio)
> {
> ...
>
> if (vp->v_type == VLNK) {
> if ((int)ext2fs_size(ip) < vp->v_mount->
On Wed, Feb 10, 2016 at 09:29:26AM -0700, Todd C. Miller wrote:
> On Wed, 10 Feb 2016 08:36:42 -0700, "Todd C. Miller" wrote:
>
> > Making those values unsigned int should be sufficient but it is a
> > fairly intrusive change.
>
> Actually, this is trickier than it appears since diff uses the sig
When diff encounters a line that consists of a single dot, it emits two
dots instead, stops the current command and emits a substitute command
to replace the double dot with a single one. Then it restarts the
(original) command if necessary and inserts further lines. This is done
because a single d
On Thu, Feb 11, 2016 at 03:50:24PM +0100, Alexander Schrijver wrote:
> I ran into this issue when setting up my public_html folder using this
> configuration.
>
> prefork 2
>
> server "default" {
>listen on * port 80
>
>location match "/~*" {
> root "/users"
>}
> }
>
> types
The fusefs_checkexp() function returns 0, indicating that export via NFS
is allowed, while in fact fusefs doesn't support NFS at all. (There is
a vfs_export() call missing in fusefs_mount() and maybe other stuff too.)
Furthermore, it does so without setting *extflagsp and *credanonp, which
results
The fusefs_fhtovp() function makes use of the ROOTINO ((ufsino_t)2)
define instead of using FUSE_ROOTINO ((ino_t)1), which is used
everywhere else in the fuse filesystem. This causes a file handle for
the filesystem root to be falsely rejected with ESTALE.
Comments?
Index: miscfs/fuse/fuse_vfsops
Casting the result of ext2fs_size() and DIP(ip, size) to int potentially
truncates the result. Issue found by Stefan Kempf, see
https://marc.info/?l=openbsd-tech&m=145495905416536 .
While there I also removed the cast in the ext2fs_chmod() call, because
the function expects a mode_t argument anywa
On Wed, Feb 17, 2016 at 11:27:29AM -0700, Todd C. Miller wrote:
> On Wed, 17 Feb 2016 10:22:04 +0100, Martin Natano wrote:
>
> > Casting the result of ext2fs_size() and DIP(ip, size) to int potentially
> > truncates the result. Issue found by Stefan Kempf, see
> > http
On Thu, Feb 18, 2016 at 09:11:18PM +0100, Stefan Kempf wrote:
>
> I think we don't mix declarations and code.
> Would this be an option?
>
> diff --git a/dev/rnd.c b/dev/rnd.c
> index 819ce0d..0f57b1b 100644
> --- a/dev/rnd.c
> +++ b/dev/rnd.c
> @@ -421,7 +421,7 @@ add_entropy_words(const u_int3
On Wed, Feb 17, 2016 at 11:27:29AM -0700, Todd C. Miller wrote:
> There is currently code that checks for mnt_maxsymlinklen <= 0.
> Removing the cast will cause other problems for ffs if the maxsymlinklen
> value is negative. I don't think it is safe to make this change
> unless mnt_maxsymlinklen
> Hmm, something still does not seem right. For example, try
> diff -e on the contents below (some other tests attached):
>
> A
> B
> C
>
> vs.
>
> W
> X
> .
> Y
> Z
>
> [...]
>
> About the changes below, is something like this the idea behind it?
> The first change command has already removed
When inserting multiple lines with the insert or change command the
lines get reordered when the file is empty before inserting. e.g.
following ed script produces incorrect output (apply to a file with
three lines in it).
1,3c
x
y
z
.
The expected order of lines would be, x, y, z but patch produ
patch(1) doesn't recognize ed-style diffs that start with an insert
command as such. Same holds true for substitution command, but with the
current state of support for s/, I don't think it makes sense to support
diffs starting with s/.//.
Thoughts?
natano
Index: pch.c
=
On Mon, Feb 22, 2016 at 06:31:00AM -0700, Todd C. Miller wrote:
> On Sun, 21 Feb 2016 16:33:27 +0100, Stefan Kempf wrote:
>
> > Should we really mount the FS in that case? If the FS was of the
> > "new" format, then short symlinks would store the destination path in the
> > inode directly. I think
For most architectures there still is an entry for the removed RAIDframe
device lurking in chrtoblktbl. While there I truncated the tables to the
minimum required size; chrtoblk() and blktochr() are designed to handle
a table shorter than cdevsw.
Ok?
natano
Index: arch/alpha/alpha/conf.c
==
.
natano
On Fri, Feb 26, 2016 at 11:32:08PM +0100, Martin Natano wrote:
> The concept of differentiating between "short" and "long" symlinks is
> specific to ufs/, but the relevant variable is contained in struct
> mount, which is used by all filesystems. I think it wou
Sorry, the second message was not supposed to go to the list...
The concept of differentiating between "short" and "long" symlinks is
specific to ufs/, but the relevant variable is contained in struct
mount, which is used by all filesystems. I think it would be a worthwile
effort to clear this up by moving maxsymlinklen to struct ufsmount
instead. Filesystem co
Last chance to voice objections. I plan to commit this in the next days.
natano
> Index: diffreg.c
> ===
> RCS file: /cvs/src/usr.bin/diff/diffreg.c,v
> retrieving revision 1.90
> diff -u -p -r1.90 diffreg.c
> --- diffreg.c 26 Oct 2
The ext2fs_read() and ffs_read() functions return EFBIG when uio_offset
if smaller than zero or larger than the maxium file size. However this
doesn't seem to be in accordance with the POSIX read(2) specification,
which requires EINVAL for an invalid offset (< 2) and a return of 0
(zero bytes trans
On Sat, Feb 27, 2016 at 05:44:49PM -0500, Michael McConville wrote:
> * There wasn't yet a list of possible errors for tmpfs in mount(2).
>That said, I question the value of maintaining these lists. It's
>almost impossible for them to be comprehensive - for example, some
>*_mount() fun
On Sun, Feb 28, 2016 at 04:40:21PM +0100, Stefan Kempf wrote:
> Stefan Kempf wrote:
> > Martin Pieuchot wrote:
> > > I'm also wondering when you say "an offset that's at or paste the
> > > EOF" does that include ``uio_resid''? I mean shouldn't you check
> > > for:
> > >
> > > if ((uio->uio_offs
On Sun, Feb 28, 2016 at 08:14:39PM +0100, Stefan Kempf wrote:
> This changes uiomovei calls to uiomove in usb. It fixes
> a few integer truncations due to use of min, and uses
> unsigned types for count variables where it makes sense.
> This also allows us to get rid of a couple of 'if (len < 0)' c
On Mon, Feb 29, 2016 at 07:56:22PM +0100, Jeremie Courreges-Anglas wrote:
>
> ok?
Looks good to me.
>
> Index: libexec/ftpd/monitor_fdpass.c
> ===
> RCS file: /cvs/src/libexec/ftpd/monitor_fdpass.c,v
> retrieving revision 1.6
> dif
On Wed, Mar 02, 2016 at 05:07:21PM +0100, Michal Mazurek wrote:
> kern_sched.c:
> - remove unused functions
> - mark static functions as static
Functions shouldn't be static in the kernel. "In userland,
functions local to one source module should be declared ‘static’. This
should not be done in k
The VOP_UNLOCK() function has a flags parameter, which is documented to
should be zero in most cases. It turns out that the flags argument is
zero for all ~200 callers in the tree. On a closer look it is revealed
that VOP_UNLOCK() uses lockmgr() for all filesystems that support
locking. The only l
> Maybe it's just the alias being ignored by nice:
nice executes the utility with execvp(), so it doesn't have a way to
know about your alias, as the shell is not involved. The alias is also
not expanded before calling nice, because it is not the first word in
the line.
natano
On Sat, Mar 05, 2016 at 11:32:04AM -0800, Philip Guenther wrote:
> On Sat, Mar 5, 2016 at 8:10 AM, Michal Mazurek wrote:
> > sys/compat is gone. I don't know if PID_MAX can or should be increased,
> > but I don't think the comment is now useful.
>
> I think we can go a bit further now and include
On Sun, Mar 06, 2016 at 05:32:16AM +0100, Martijn Dekker wrote:
> The command 'set -o' shows the current shell options in an unspecified
> format. Less well-known is the variant 'set +o', which should output the
> current shell options "in a format that is suitable for reinput to the
> shell as com
> Index: bin/ksh/misc.c
> ===
> RCS file: /cvs/src/bin/ksh/misc.c,v
> retrieving revision 1.53
> diff -u -p -u -r1.53 misc.c
> --- bin/ksh/misc.c21 Dec 2015 04:57:50 - 1.53
> +++ bin/ksh/misc.c6 Mar 2016 04:27:20 -
There are bunch of ioctl definitions kdump(1) doesn't know about yet,
with the result, that kdump prints the numerical value of the ioctl and
not the name of the definition.
This patch adds support for ipmi, vscsi, pvbus, udl, fuse, trunk, pipex
and memrange.
OK?
natano
Index: Makefile
===
On Mon, Mar 14, 2016 at 10:57:36AM +0100, Marc Espie wrote:
>
> So, does it make sense to put COLUMNS and SIZE forward ?
> I think this is the first important question to ask...
>
> (I remember having COLUMNS and LINES hardcoded in my old, old .profile
> around SunOS4... we can probably assume
On Mon, Mar 14, 2016 at 02:02:47PM -0600, Abel Abraham Camarillo Ojeda wrote:
> On Mon, Mar 14, 2016 at 5:32 AM, Martin Natano wrote:
> > On Mon, Mar 14, 2016 at 10:57:36AM +0100, Marc Espie wrote:
> >>
> >> So, does it make sense to put COLUMNS and SIZE forward ?
>
I don't see how execute permissions on a file system image would be
useful.
Ok?
Index: newfs_ext2fs.c
===
RCS file: /cvs/src/sbin/newfs_ext2fs/newfs_ext2fs.c,v
retrieving revision 1.24
diff -u -p -r1.24 newfs_ext2fs.c
--- newfs_ext2
I'm currently working on a diff to make bpf a cloning device. Therefore
it is necessary to increase the number of clones possible of a cloning
device, as there are users with a need for more than 64 open bpf devices
at the same time. mikeb@ pointed me to this thread:
https://marc.info/?l=openbsd-te
Thank you all for the input. Allocatig the bitmap via malloc() really
seems like the way to go, so we don't waste space for non-cloning
devices. See updated patch below.
Would it make sense to move the (rdev == VCHR && ...) condition to a
macro in ? I figured it's only used twice, so I inlined
it.
On Thu, Mar 31, 2016 at 12:31:14PM +, Miod Vallat wrote:
>
> > Thank you all for the input. Allocatig the bitmap via malloc() really
> > seems like the way to go, so we don't waste space for non-cloning
> > devices. See updated patch below.
> >
> > Would it make sense to move the (rdev == VCHR
On Thu, Mar 31, 2016 at 06:02:07AM -0600, Todd C. Miller wrote:
> On Thu, 31 Mar 2016 09:34:32 +0200, Martin Natano wrote:
>
> > Thank you all for the input. Allocatig the bitmap via malloc() really
> > seems like the way to go, so we don't waste space for non-cloning
&
Below is the fixed v_specbitmap enlargements diff, including some tweaks
by mikeb@. I have tested this with fuse _and_ drm on amd64 and macppc. I
also tested on macppc with cloning bpf (not in the tree). Can anyone
come up with another interesting test case?
Comments? Ok?
natano
Index: sys/spec
On Sun, Apr 03, 2016 at 09:14:40AM +0200, Otto Moerbeek wrote:
> On Sat, Apr 02, 2016 at 09:13:20PM +, H??ctor Luis Gimbatti wrote:
>
> > Hi,
> > Apparently the error seems to be in /usr/src/usr.bin/grep/util.c at line
> > 400:
> >
> > if ((!(lflag || cflag)) && ((!(bol || eol)) &&
>
On Sun, Apr 03, 2016 at 08:20:57PM -0700, Philip Guenther wrote:
>
> libc has a number of macros for dealing with thread safety such that it
> can operate efficiently when single-threaded but Do The Right Thing when
> multi-threaded. In include/thread_private.h are two sets of macros that
> lo
On Mon, Apr 04, 2016 at 07:18:36PM -0700, Philip Guenther wrote:
> On Mon, 4 Apr 2016, Martin Natano wrote:
> > On Sun, Apr 03, 2016 at 08:20:57PM -0700, Philip Guenther wrote:
> ...
> > > So, diff below converts several uses of _THREAD_PRIVATE_MUTEX_*() to
> > > _
The following patch enables device cloning for bpf, which allows to have
just one bpf device node in /dev, that services all bpf consumers (up to
1024).
Cloning bpf offers some advantages to the current situation:
- Users with high bpf usage won't have to clutter /dev with device
nodes.
- A lot
On Wed, Apr 06, 2016 at 03:42:20PM +0200, Michal Mazurek wrote:
> MPSAFE is never used, and doesn't look like it's even supported (no
> matching SY_MPSAFE anywhere).
SY_MPSAFE seems to be unused since it's introduction in 2007 and it
doesn't have any effect; ok natano@, if no one comes up with a r
On Wed, Apr 06, 2016 at 09:47:35AM +0200, Michal Mazurek wrote:
> relebad used to have more body:
> relebad:
> PRELE(t);
> return (error);
> But then PRELE(t); was removed. This diff gets rid of what remains of
> relebad.
Looks good to me.
>
> Index: sys/kern/s
On Wed, Apr 06, 2016 at 02:45:46PM +0200, Michal Mazurek wrote:
> * document NOLOCK - copy and paste from the commit message.
That's useful.
> * mention INDIR and NOLOCK in 'Fields'.
How about something like "one of the types listed below, or one of the
compatibility options defined in syscalls
In rbootd a struct bpf_timeval (with 32bit tv_sec) is copied to a struct
timeval (with 64 tv_sec) via bcopy(). This most likely causes
connections to not time out correctly in rbootd. I don't have an HP
machine to test this with. Who owns such a machine and is willing to
test this?
natano
Index:
On Sat, Apr 16, 2016 at 11:10:14AM +0100, Stuart Henderson wrote:
> On 2016/04/15 23:07, Philip Guenther wrote:
> > On Thu, Apr 14, 2016 at 3:15 PM, Martin Natano wrote:
> > > In rbootd a struct bpf_timeval (with 32bit tv_sec) is copied to a struct
> > > timeval (with 6
On Mon, Apr 18, 2016 at 07:11:46PM -0400, Christian Heckendorf wrote:
> Hi,
> After reading the man page description of tr -s, I couldn't understand
> what it meant to "squeeze" characters. It could be that 'squeeze' is a
> mnemonic for '-s' but it's not helpful as a technical explanation.
Your di
There seem to be a number of issues with statfs related code in the
kernel. The first issue is inside of the copy_statfs_info() function
which is designed to be used by the filesystem's .vfs_statfs
implementations to copy data from mp->mnt_stat to the target stat
buffer. copy_statfs_info() always c
On Sun, Apr 24, 2016 at 01:48:04PM +0200, Stefan Kempf wrote:
>
> Diff reads good to me. Any reason why you changed setting f_mntfromname
> from "fusefs" to "fuse"?
No, it's a typo; updated diff below. Thanks!
natano
Index: isofs/cd9660/cd9660_vfsops.c
Following diff replaces /dev/bpf[0-9] with only /dev/bpf and /dev/bpf0.
The /dev/bpf node is unused for now, but I plan to convert all programs
in base to use it in a future diff. /dev/bpf0 is for compatibility with
existing binaries and is to be removed after a transition period.
install.sub cont
This is the first diff in a series to move base to /dev/bpf. I think
tcpdump is a good starting point, because it is easy to test.
Ok?
natano
Index: privsep_pcap.c
===
RCS file: /cvs/src/usr.sbin/tcpdump/privsep_pcap.c,v
retrieving
Diff below simplifies the device open path and removes an explanation
about bpf device nodes from the manpage.
Ok?
natano
Index: bpf.c
===
RCS file: /cvs/src/sbin/dhclient/bpf.c,v
retrieving revision 1.38
diff -u -p -r1.38 bpf.c
--
And again: /dev/bpf instead of /dev/bpf*. Ok?
natano
Index: pcap-bpf.c
===
RCS file: /cvs/src/lib/libpcap/pcap-bpf.c,v
retrieving revision 1.32
diff -u -p -r1.32 pcap-bpf.c
--- pcap-bpf.c 22 Dec 2015 19:51:04 - 1.32
+++ pc
On Tue, May 03, 2016 at 05:09:47AM -0600, Theo de Raadt wrote:
> > On 2016-04-27, Martin Natano wrote:
> >
> > > /dev/bpf0 is for compatibility with existing binaries and is to
> > > be removed after a transition period.
> >
> > What's the impac
On Thu, May 05, 2016 at 09:30:43PM +0100, Stuart Henderson wrote:
> On 2016/05/02 09:34, Martin Natano wrote:
> > Diff below simplifies the device open path and removes an explanation
> > about bpf device nodes from the manpage.
>
> There's a problem with this. If someo
On Fri, May 06, 2016 at 07:17:50PM +0100, Stuart Henderson wrote:
>
> Personally, my preferred course of action would be to switch to
> looking for /dev/bpf0 for now (bpf0 only: *not* revert to the old
> loop, that's not necessary), and then switch to /dev/bpf
> immediately after 6.0 is released.
Despite being redundant, the LK_CANRECURSE flag isn't checked anywhere.
Ok?
natano
Index: sys/kern/kern_lock.c
===
RCS file: /cvs/src/sys/kern/kern_lock.c,v
retrieving revision 1.46
diff -u -p -r1.46 kern_lock.c
--- sys/kern/kern_l
On Tue, Jun 13, 2017 at 08:02:13PM +0200, Klemens Nanni wrote:
> Unify option checking and simply logic.
>
> F_HDRINCL and F_ROUTE are mutually exclusive, thus check the latter only
> if the former one is not set.
>
> Index: ping.c
> ===
ping?
On Sun, May 21, 2017 at 09:46:45AM +0200, Martin Natano wrote:
> While switching init and reboot to CPU_LIDACTION, I forgot about the
> #ifdef's. Ok?
>
> natano
>
>
> Index: init/init.c
> ===
&
On Fri, Jun 30, 2017 at 09:19:31AM -0600, Anthony J. Bentley wrote:
>
> Here's a diff that updates the STANDARDS section in the manual.
> I think it will be enough to talk about our POSIX compliance. POSIX
> discusses its differences from historical ex/vi in great detail,
> including an explicit r
Go ahead replacing machdep.lidsuspend with cpu.lidaction. OpenBSD 6.1
contains both MIBs, so there is a clear path for migration.
ok?
natano
Index: arch/amd64/amd64/machdep.c
===
RCS file: /cvs/src/sys/arch/amd64/amd64/machdep.c,v
On Tue, Jul 11, 2017 at 02:40:16PM -0700, Mike Larkin wrote:
> On Tue, Jul 11, 2017 at 11:19:17PM +0200, Martin Natano wrote:
> > Go ahead replacing machdep.lidsuspend with cpu.lidaction. OpenBSD 6.1
>
> cpu.lidaction or machdep.lidaction?
machdep.lidaction; I got confused by the #define name.
Words are more descriptive than numbers. Let's use them!
This diff removes
sysctl machdep.lidaction=0
sysctl machdep.lidaction=1
sysctl machdep.lidaction=2
in favor of the more descriptive
sysctl machdep.lidaction=none
sysctl machdep.lidaction=suspend
On Sat, Jul 29, 2017 at 02:03:22PM +0200, Mark Kettenis wrote:
>
> I don't think we want to add string parsing like this in the kernel.
> Maybe the sysctl(8) frontend should do the mapping from strings to
> numbers?
Ok, I'll try to come up with an alternative diff that does the parsing
in sysctl(
On Sat, Jul 29, 2017 at 02:19:50PM +0200, Martin Natano wrote:
> On Sat, Jul 29, 2017 at 02:03:22PM +0200, Mark Kettenis wrote:
> >
> > I don't think we want to add string parsing like this in the kernel.
> > Maybe the sysctl(8) frontend should do the mapping from strin
On Thu, May 19, 2016 at 02:03:43AM +0200, Alexander Bluhm wrote:
> Hi,
>
> All msg buf counters are long, so lmin(9) should be used here.
>
> ok?
Looks good to me.
>
> bluhm
>
> Index: kern/subr_log.c
> ===
> RCS file: /cvs/src/
VOP_ALLOCBLKS() and related code is unused since the removal of
cluster_write(). This diff even shrinks /bsd by 8kb. Ok?
natano
Index: sbin/sysctl/sysctl.8
===
RCS file: /cvs/src/sbin/sysctl/sysctl.8,v
retrieving revision 1.194
diff
The doforce variable isn't modified anywhere. Also, the only filesystem
left using it is fuse. It has been removed from all other filesystems.
Ok to remove?
natano
Index: kern/vfs_subr.c
===
RCS file: /cvs/src/sys/kern/vfs_subr.c,v
I think it's time to get rid of all the bpf open() loops in base.
dhclient and libpcap do a plain open("/dev/bpf0", ...) since a couple of
weeks now and the upgrade issue (/dev/bpf vs. /dev/bpf0) has been fixed.
I didn't hear any other complaints in the meantime.
Ok? Too soon?
natano
Index: usr
See one comment inline.
On Thu, May 26, 2016 at 09:59:24AM -0600, Todd C. Miller wrote:
> Since /usr/bin/cpp still specified -traditional, use ${CC} -E
> instead. I've had this rotting in my tree for years now.
>
> - todd
>
> Index: lib/libcurses/Makefile
>
On Thu, May 26, 2016 at 02:14:29PM -0600, Todd C. Miller wrote:
> On Thu, 26 May 2016 22:07:14 +0200, Martin Natano wrote:
>
> > > diff -u -p -u -r1.11 Makefile
> > > --- share/locale/ctype/Makefile 20 Mar 2016 15:45:40 - 1.11
> > > +++ share/locale/c
On Fri, May 27, 2016 at 10:03:37PM +0200, Jeremie Courreges-Anglas wrote:
> Martin Natano writes:
>
> > I think it's time to get rid of all the bpf open() loops in base.
> > dhclient and libpcap do a plain open("/dev/bpf0", ...) since a couple of
> > weeks
On Fri, May 27, 2016 at 10:15:39PM +0200, Jeremie Courreges-Anglas wrote:
> Mathieu - writes:
>
> > Hello everyone,
> >
> > While playing a bit with ptrace to do some debugging I stumbled upon
> > something that looks like a bug.
> > While trying to write to the ptrace'd process using PT_IO in co
On Sat, May 28, 2016 at 07:55:00PM -0700, Philip Guenther wrote:
>
> About the only place userland code should use PF_* socket constants is
> with sysctl(3)'s CTL_NET hierarchy. All the standardized functions are
> defined as taking AF_* values. Let's use the preferred names in the
> getaddri
On Sat, May 28, 2016 at 07:47:50PM -0700, Philip Guenther wrote:
>
> Overriding the hash algorithm used by the Berkeley DB bits isn't support
> (it would break getpw* if nothing else) and hasn't been possible since the
> symbol hiding effort last fall. So eliminate the redirection through a
>
It is time for the lockmgr() api to die. The api is only used by
filesystems, where it is a trivial change to use rrw locks instead. All
it needs is LK_* defines for the RW_* flags. (See the sys/lock.h hunk in
the diff below.)
The ffs regress tests display the same number of fail/ok results before
On Wed, Jun 01, 2016 at 01:43:06AM -0400, Ted Unangst wrote:
> Let's make the defaults be the defaults.
Reads fine. ok natano@
>
> Index: Makefile
> ===
> RCS file: /cvs/src/usr.bin/newsyslog/Makefile,v
> retrieving revision 1.6
>
1 - 100 of 265 matches
Mail list logo
