Page MenuHomeFreeBSD

tuexen (Michael Tuexen)
User

Projects

User Details

User Since
Feb 4 2016, 4:45 PM (158 w, 6 d)

Recent Activity

Yesterday

tuexen added a comment to D19252: Unwire pages when cleaning up the kcov state.

Latest version tested again with running kcovtrace /usr/bin/uptime in a loop and in addition with running syzkaller. Both work fine.

Wed, Feb 20, 7:29 PM
tuexen added inline comments to D19254: Avoid dereferencing a NULL pointer when TCP_REASS_LOGGING is enabled.
Wed, Feb 20, 7:26 PM
tuexen updated the diff for D19254: Avoid dereferencing a NULL pointer when TCP_REASS_LOGGING is enabled.

Address bz@'s comment.

Wed, Feb 20, 7:23 PM
tuexen retitled D19254: Avoid dereferencing a NULL pointer when TCP_REASS_LOGGING is enabled from Avoid dereferencing a NULL point when TCP_REASS_LOGGING is enabled to Avoid dereferencing a NULL pointer when TCP_REASS_LOGGING is enabled.
Wed, Feb 20, 6:05 PM
tuexen closed D18941: RTO.Initial = 1 second for TCP.
Wed, Feb 20, 6:04 PM
tuexen committed rS344368: Reduce the TCP initial retransmission timeout from 3 seconds to.
Reduce the TCP initial retransmission timeout from 3 seconds to
Wed, Feb 20, 6:04 PM
tuexen committed rS344367: Use exponential backoff for retransmitting SYN segments as specified.
Use exponential backoff for retransmitting SYN segments as specified
Wed, Feb 20, 5:57 PM
tuexen closed D18974: Use exponential backoff also for SYN segments.
Wed, Feb 20, 5:57 PM
tuexen added a comment to D19252: Unwire pages when cleaning up the kcov state.

Using this version, the memory lockup observed does not occur anymore. Tested with running kcovtrace uptime in a loop.

Wed, Feb 20, 10:47 AM

Tue, Feb 19

tuexen created D19254: Avoid dereferencing a NULL pointer when TCP_REASS_LOGGING is enabled.
Tue, Feb 19, 10:14 PM
tuexen added inline comments to D19252: Unwire pages when cleaning up the kcov state.
Tue, Feb 19, 9:26 PM

Sun, Feb 17

tuexen requested changes to D19000: Avoid cwnd update for SYN sequence space.

Hi Richard,

Sun, Feb 17, 11:05 AM

Fri, Feb 15

tuexen committed rS344148: Fix a byte ordering issue for the advertised receiver window in ACK.
Fix a byte ordering issue for the advertised receiver window in ACK
Fri, Feb 15, 9:45 AM

Wed, Feb 13

tuexen created D19181: Improve input validation when when using IP_HDRINCL socket option on SOCK_RAW sockets.
Wed, Feb 13, 8:00 PM

Tue, Feb 12

tuexen committed rS344048: Improve input validation for raw IPv4 socket using the IP_HDRINCL.
Improve input validation for raw IPv4 socket using the IP_HDRINCL
Tue, Feb 12, 10:17 AM
tuexen closed D18834: Check IPv4 options when using IP_HDRINCL socket option on SOCK_RAW sockets.
Tue, Feb 12, 10:17 AM

Mon, Feb 11

tuexen committed rS344012: Fix flags used when compiling kern_kcov.c and subr_coverage.c..
Fix flags used when compiling kern_kcov.c and subr_coverage.c.
Mon, Feb 11, 3:38 PM
tuexen closed D19135: Fix compiler flags related to kernel coverage.
Mon, Feb 11, 3:38 PM
tuexen added inline comments to D19135: Fix compiler flags related to kernel coverage.
Mon, Feb 11, 2:28 PM
tuexen updated the diff for D19135: Fix compiler flags related to kernel coverage.

Follow Andrew's advice.

Mon, Feb 11, 2:27 PM
tuexen added inline comments to D19135: Fix compiler flags related to kernel coverage.
Mon, Feb 11, 1:40 PM

Sun, Feb 10

tuexen added inline comments to D19135: Fix compiler flags related to kernel coverage.
Sun, Feb 10, 4:29 PM
tuexen updated the diff for D19135: Fix compiler flags related to kernel coverage.

Use underscores instead of hyphens as suggested by Ed.

Sun, Feb 10, 4:27 PM
tuexen added a comment to D19135: Fix compiler flags related to kernel coverage.

I tested this patch on a kernel build with clang in base and with the amd64-xtoolchain-gcc cross toolchain.

Sun, Feb 10, 2:23 PM
tuexen committed rS343961: Fix a locking issue when reporing outbount messages..
Fix a locking issue when reporing outbount messages.
Sun, Feb 10, 2:02 PM
tuexen committed rS343960: Fix a locking issue in the IPPROTO_SCTP level SCTP_PEER_ADDR_THLDS socket.
Fix a locking issue in the IPPROTO_SCTP level SCTP_PEER_ADDR_THLDS socket
Sun, Feb 10, 1:55 PM
tuexen committed rS343954: Fix a locking bug in the IPPROTO_SCTP level SCTP_EVENT socket option..
Fix a locking bug in the IPPROTO_SCTP level SCTP_EVENT socket option.
Sun, Feb 10, 10:42 AM
tuexen committed rS343951: Fix locking for IPPROTO_SCTP level SCTP_DEFAULT_PRINFO socket option..
Fix locking for IPPROTO_SCTP level SCTP_DEFAULT_PRINFO socket option.
Sun, Feb 10, 8:29 AM

Sat, Feb 9

tuexen created D19135: Fix compiler flags related to kernel coverage.
Sat, Feb 9, 2:48 PM

Fri, Feb 8

tuexen closed D19071: Get TCP CDG CC working if net.inet.tcp.cc.cdg.smoothing_factor = 0 .
Fri, Feb 8, 8:43 PM
tuexen committed rS343920: Ensure that when using the TCP CDG congestion control and setting the.
Ensure that when using the TCP CDG congestion control and setting the
Fri, Feb 8, 8:43 PM

Tue, Feb 5

tuexen added a comment to D19071: Get TCP CDG CC working if net.inet.tcp.cc.cdg.smoothing_factor = 0 .

Suggest moving stailq flush loops out of cdg_cb_destroy() into an inline function, changing smoothing_factor sysctl to a SYSCTL_PROC with custom handler similar to the exp_backoff_scale sysctl, and calling flush function from both cdg_cb_destroy() and sysctl handler when smoothing_factor is set to zero

Tue, Feb 5, 2:39 PM
tuexen committed rS343770: Only reduce the PMTU after the send call. The only way to increase it, is.
Only reduce the PMTU after the send call. The only way to increase it, is
Tue, Feb 5, 10:29 AM
tuexen committed rS343769: Fix an off-by-one error in the input validation of the SCTP_RESET_STREAMS.
Fix an off-by-one error in the input validation of the SCTP_RESET_STREAMS
Tue, Feb 5, 10:14 AM

Mon, Feb 4

tuexen committed rS343742: Add missing SCTP_EOR entry..
Add missing SCTP_EOR entry.
Mon, Feb 4, 1:31 PM

Sun, Feb 3

tuexen created D19071: Get TCP CDG CC working if net.inet.tcp.cc.cdg.smoothing_factor = 0 .
Sun, Feb 3, 1:09 PM

Fri, Feb 1

tuexen closed D19033: Initialise snd_wnd on the server side always before processing the ack.
Fri, Feb 1, 12:33 PM
tuexen committed rS343661: When handling SYN-ACK segments in the SYN-RCVD state, set tp->snd_wnd.
When handling SYN-ACK segments in the SYN-RCVD state, set tp->snd_wnd
Fri, Feb 1, 12:33 PM
tuexen updated the diff for D19033: Initialise snd_wnd on the server side always before processing the ack.

Simplification of the fix within the RACK code.

Fri, Feb 1, 9:01 AM

Thu, Jan 31

tuexen added a comment to D19033: Initialise snd_wnd on the server side always before processing the ack.

RACK seems to have a similar code in rack_do_syn_recv.

Thu, Jan 31, 12:10 PM
tuexen updated the diff for D19033: Initialise snd_wnd on the server side always before processing the ack.

Also fix the issue in the RACK stack.

Thu, Jan 31, 12:09 PM

Wed, Jan 30

tuexen created D19033: Initialise snd_wnd on the server side always before processing the ack.
Wed, Jan 30, 4:02 PM

Mon, Jan 28

tuexen closed D18996: Disallow broken ECN TCP stacks from using ECN.
Mon, Jan 28, 12:46 PM
tuexen committed rS343525: Fix the detection of ECN-setup SYN-ACK packets..
Fix the detection of ECN-setup SYN-ACK packets.
Mon, Jan 28, 12:45 PM
tuexen requested changes to D18996: Disallow broken ECN TCP stacks from using ECN.

Please apply the same fix to tcp_stacks/rack.c.

Mon, Jan 28, 12:23 PM
tuexen accepted D18996: Disallow broken ECN TCP stacks from using ECN.
Mon, Jan 28, 12:04 PM

Sun, Jan 27

tuexen added inline comments to D18974: Use exponential backoff also for SYN segments.
Sun, Jan 27, 7:14 PM

Sat, Jan 26

tuexen added a reviewer for D18960: Patch to MFC TCP reassembly rewrite to stable/11: jtl.
Sat, Jan 26, 10:07 AM

Fri, Jan 25

tuexen added a comment to D18963: explicit NewReno module dependency.

I think this is not needed, since cc_newreno is statically compiled into the kernel and not compiled as a module.

Fri, Jan 25, 10:18 PM
tuexen added a comment to D18960: Patch to MFC TCP reassembly rewrite to stable/11.
In D18960#405325, @bz wrote:

Given the discussions in D18968 I think you'd be safe as I'd assume we don't otherwise have TCP_REASS_LOGGING in 11, do we?
I guess rrs may or may not give you a more technical feedback.

Fri, Jan 25, 9:26 PM
tuexen created D18974: Use exponential backoff also for SYN segments.
Fri, Jan 25, 9:04 PM
tuexen created D18960: Patch to MFC TCP reassembly rewrite to stable/11.
Fri, Jan 25, 5:41 PM
tuexen committed rS343439: Don't include two header files when not needed..
Don't include two header files when not needed.
Fri, Jan 25, 5:08 PM
tuexen committed rS343437: MFC r343089:.
MFC r343089:
Fri, Jan 25, 3:48 PM
tuexen committed rS343436: MFC r343089:.
MFC r343089:
Fri, Jan 25, 3:46 PM
tuexen committed rS343435: MFC r342879:.
MFC r342879:
Fri, Jan 25, 3:45 PM
tuexen committed rS343434: MFC r342857:.
MFC r342857:
Fri, Jan 25, 3:41 PM
tuexen committed rS343433: MFC r342857:.
MFC r342857:
Fri, Jan 25, 3:39 PM
tuexen committed rS343432: MFC r338138:.
MFC r338138:
Fri, Jan 25, 3:26 PM
tuexen committed rS343431: MFC r338137:.
MFC r338137:
Fri, Jan 25, 3:22 PM
tuexen added a comment to D18954: Implement Cubic-specific After-Idle reaction.

Looks good to me. Will let lstewart@ the final decision...

Fri, Jan 25, 2:40 PM
tuexen added inline comments to D18954: Implement Cubic-specific After-Idle reaction.
Fri, Jan 25, 2:34 PM
tuexen committed rS343428: Fix a bug in the restart window computation of TCP New Reno.
Fix a bug in the restart window computation of TCP New Reno
Fri, Jan 25, 1:57 PM
tuexen closed D18940: Consolidate Initial Window calculations.
Fri, Jan 25, 1:57 PM
tuexen accepted D18940: Consolidate Initial Window calculations.
Fri, Jan 25, 1:44 PM
tuexen added inline comments to D18954: Implement Cubic-specific After-Idle reaction.
Fri, Jan 25, 1:22 PM
tuexen requested changes to D18940: Consolidate Initial Window calculations.

Can you please make use of tcp_compute_initwnd() also in sys/netinet/tcp_stacks/rack.c to minimise code duplication?

Fri, Jan 25, 10:35 AM
tuexen added a reviewer for D18834: Check IPv4 options when using IP_HDRINCL socket option on SOCK_RAW sockets: lstewart.
Fri, Jan 25, 9:04 AM
tuexen updated the diff for D18834: Check IPv4 options when using IP_HDRINCL socket option on SOCK_RAW sockets.

Use explicit page number range instead of 15ff as suggested by lstewart@.

Fri, Jan 25, 9:04 AM

Thu, Jan 24

tuexen created D18941: RTO.Initial = 1 second for TCP.
Thu, Jan 24, 5:18 PM
tuexen committed rS343403: Get the arithmetic right....
Get the arithmetic right...
Thu, Jan 24, 4:47 PM
tuexen committed rS343402: Kill a trailing whitespace character....
Kill a trailing whitespace character...
Thu, Jan 24, 4:43 PM
tuexen committed rS343401: Update a comment to reflect the current reality..
Update a comment to reflect the current reality.
Thu, Jan 24, 4:40 PM
tuexen accepted D18940: Consolidate Initial Window calculations.
Thu, Jan 24, 10:58 AM
tuexen added a reviewer for D18940: Consolidate Initial Window calculations: tuexen.
Thu, Jan 24, 10:58 AM

Jan 16 2019

tuexen committed rS343089: Limit the user-controllable amount of memory the kernel allocates.
Limit the user-controllable amount of memory the kernel allocates
Jan 16 2019, 11:33 AM

Jan 13 2019

tuexen created D18834: Check IPv4 options when using IP_HDRINCL socket option on SOCK_RAW sockets.
Jan 13 2019, 3:34 PM

Jan 10 2019

tuexen added inline comments to D14599: Implement kernel code coverage (kcov).
Jan 10 2019, 1:03 PM

Jan 9 2019

tuexen added inline comments to D14599: Implement kernel code coverage (kcov).
Jan 9 2019, 3:01 PM
tuexen closed D18753: Fix getsockopt(..., IPPROTO_IP, IP_OPTIONS, ..., ...).
Jan 9 2019, 6:37 AM
tuexen committed rS342879: Fix getsockopt() for IP_OPTIONS/IP_RETOPTS..
Fix getsockopt() for IP_OPTIONS/IP_RETOPTS.
Jan 9 2019, 6:37 AM

Jan 8 2019

tuexen closed D18763: Avoid overflow in vtruncbuf().
Jan 8 2019, 9:04 AM
tuexen committed rS342857: Avoid overfow in vtruncbuf().
Avoid overfow in vtruncbuf()
Jan 8 2019, 9:04 AM

Jan 7 2019

tuexen added a comment to D14599: Implement kernel code coverage (kcov).

Also Andrew, I saw you mention in another thread that you had a man page started but not quite ready. If you'd like help reviewing or finishing it let me know, as I had started a basic one locally myself.

I think now I'll commit without the manpage. We can then work on one based on the final interface.

Jan 7 2019, 5:12 PM

Jan 6 2019

tuexen created D18763: Avoid overflow in vtruncbuf().
Jan 6 2019, 9:33 PM

Jan 5 2019

tuexen created D18753: Fix getsockopt(..., IPPROTO_IP, IP_OPTIONS, ..., ...).
Jan 5 2019, 12:44 PM

Jan 3 2019

tuexen accepted D18728: Support MSG_DONTWAIT in sendmsg().
Jan 3 2019, 11:48 PM

Dec 30 2018

tuexen added inline comments to D14599: Implement kernel code coverage (kcov).
Dec 30 2018, 7:32 PM
tuexen added a comment to D14599: Implement kernel code coverage (kcov).

I fixed a some other small issue in syzkaller and now got it in a state, where the coverage information is collected initially, but then stops working. The reason, I think, is based on the following observation, how the state machines work.
When a thread exits, the following happens:

Dec 30 2018, 6:17 PM

Dec 29 2018

tuexen added inline comments to D14599: Implement kernel code coverage (kcov).
Dec 29 2018, 12:32 PM

Dec 28 2018

tuexen added a comment to D14599: Implement kernel code coverage (kcov).

I was trying to use this patch with syzkaller, but couldn't get it running. To figure out what is going on, I created a small test program,

, which results in a reboot of the system. When changing the code such that the ioctl(fd, KIODISABLE, mode) command is executed, to reboot happens. I think the system should not reboot...

Can you try with the new patch. There were some issues I found with the tests that should be fixed now.

Dec 28 2018, 3:43 PM

Dec 27 2018

tuexen added a comment to D14599: Implement kernel code coverage (kcov).

I was trying to use this patch with syzkaller, but couldn't get it running. To figure out what is going on, I created a small test program,

, which results in a reboot of the system. When changing the code such that the ioctl(fd, KIODISABLE, mode) command is executed, to reboot happens. I think the system should not reboot...

Can you try with the new patch. There were some issues I found with the tests that should be fixed now.

Dec 27 2018, 4:04 PM

Dec 26 2018

tuexen added a comment to D14599: Implement kernel code coverage (kcov).

I was trying to use this patch with syzkaller, but couldn't get it running. To figure out what is going on, I created a small test program,

, which results in a reboot of the system. When changing the code such that the ioctl(fd, KIODISABLE, mode) command is executed, to reboot happens. I think the system should not reboot...

Dec 26 2018, 1:26 PM

Dec 23 2018

tuexen committed rS342378: MFC r342280:.
MFC r342280:
Dec 23 2018, 9:48 AM

Dec 20 2018

tuexen added a comment to D18580: Accept always TCP segment with SEG.SEQ==RCV.NXT.
In D18580#397052, @jhb wrote:

In terms of an EN, the only thing you probably need to do is reply to the commit with the reply sent to re@ explaining this should be an EN and which existing releases it applies to.

Dec 20 2018, 5:22 PM
tuexen closed D18580: Accept always TCP segment with SEG.SEQ==RCV.NXT.
Dec 20 2018, 4:05 PM
tuexen committed rS342280: Fix a regression in the TCP handling of received segments..
Fix a regression in the TCP handling of received segments.
Dec 20 2018, 4:05 PM
tuexen added a comment to D18580: Accept always TCP segment with SEG.SEQ==RCV.NXT.

So it seems that there are more issues than the one I mentioned above.

So is it the case that this patch is needed to fix one issue, but another remains?

Dec 20 2018, 8:57 AM
tuexen added a comment to D18580: Accept always TCP segment with SEG.SEQ==RCV.NXT.
In D18580#396953, @jhb wrote:

So it seems that there are more issues than the one I mentioned above.

So is it the case that this patch is needed to fix one issue, but another remains?

No, Michael uploaded a newer diff after that comment that fixed the git clone fully. I think it fixes two places to allow retransmits to always make forward progress. Not sure why it hasn't been committed yet?

I'm just trying to figure out if I need to do anything special due to this being a candidate for an EN (haven't done that before).

Dec 20 2018, 8:54 AM

Dec 17 2018

tuexen updated the diff for D18580: Accept always TCP segment with SEG.SEQ==RCV.NXT.

Also add an exception to the mbuf count limit rule. This is the one being triggered by the git clone https://github.com/freebsd/freebsd.git command. With this fix, it runs to completion without a problem.

Dec 17 2018, 11:04 PM