• Sanitize element status bits with ses_status_to_ctrl

Ok, I'll commit for now. However, this vendor is sometimes responsive. They might be willing to choose a better error type if I ask. Do you have any particular SCSI sense code in mind?

In D39141#894562, @mav wrote:

In D39141#894560, @asomers wrote:

@mav do you have any of those dual-actuator HDDs? If so, would you be willing to test a patch that adds their LUN numbers to the physical path string? I can send that separately.

I have couple of those in my office desk, but tomorrow morning I am leaving for AsiaBSDCon and will be back only after April 10.

@mav do you have any of those dual-actuator HDDs? If so, would you be willing to test a patch that adds their LUN numbers to the physical path string? I can send that separately.

@mav given my comments above, are you ok with merging this? Do you have any better ideas for how to fix the problem?

ping. Does this look good, @mav ?

Yep, I see that you're correct. This is a classic case of two branches solving the same problem in different ways (if you haven't already figured out, I've been trying to reconcile some of Spectra's old ses changes). I'll abandon this revision.

In D39141#891271, @mav wrote:

The original physical path format came from Solaris/Illumos. While I agree that it is ugly and not very documented, I am not a big fan of incompatible changes. I'd think twice.

• Respond to grahamperrin's comments

@scottl are you happy with the latest changes?

In D39017#888701, @mav wrote:

It may be OK, but I am not sure I like this approach. It creates additional kernel requests, it is potentially racy. Doesn't the enclosure return some reasonable status that kernel could convert into EACCESS or something more reasonable than EINVAL? At least it would be good to do both.

LGTM. BTW, if you want this to run in FreeBSD's own CI, you should add gtar to https://github.com/freebsd/freebsd-ci/blob/1b56d3fb6cbec4ac4864e42cad056ecf0d0bf474/scripts/build/build-test_image-head.sh#L63 .

Duplicate of D38719

• style
• Fully initialize elmstat during do_setelmstat_cleanup

LGTM. But is there any file system that you expect to support cross-mountpoint copy_file_range ?

• Correctly zero-initialize MPI2_IOC_FACTS_REQUEST
• Check for the size of the MPI2_IOC_FACTS_REPLY at runtime

In D38739#882170, @scottl wrote:

So yeah, just ask for 4 bytes, get the MsgLength, then ask for the rest based on that. No more magic numbers needed when you do that.

@scottl as discussed in bugzilla, the actual msgver is 2.5, at least in my case. So the original commit wrongly guessed which versions require the larger struct. Also, what's so bad about unconditionally allocating an extra 4 bytes? It's cheaper than checking the sysctl.
@imp what determines the message format version number? Is it the driver? Will an mps card ever use msgversion >= 2.5? Will an mpr card ever use msgversion < 2.5 ?

Hopefully this kind of thing won't happen any more once we have a better ping test suite.

Good find, @imp . I think this means we need to split mps_cmd into separate source files for mps and mpr.

• Fix size comparisons during FUSE_WRITE
• Add another size check in bmap.cc

Superseded by D38619

Is the purpose of this change to accept either a "/dev/foo/bar" or a "foo/bar"? Also, in your commit message could you check the word choice in "know the path passed in" ?

I suppose this is ok, not that I know how to test it or anything. But do fix the typo in the commit message, and convert spaces to tabs in the code.

Is this redundant with efidp -e? I'm not sure how to use efidp, though. Also, don't forget to update the man page's .Dd date.

I see nothing wrong with this, though I'm not very familiar with this library, and I don't know how to test it. However, there are two typos in your commit message, and the first line should read "libefivar" instead of "efivars".

Works for me now.

This fixes the path returned in abspath. However, it doesn't fix what's returned in dev. You can demonstrate by doing something like efivar -Op 8be4df61-93ca-11d2-aa0d-00e098032b8c-Boot0014. That will print the partition name instead of the label name.

https://reviews.freebsd.org/D38614 won't solve the original problem that Corvin described in PR 269564. The find_geom_efimedia function doesn't even get executed by "efibootmgr -pE". Are you planning to fix that?

We previously discussed publishing a join table to associate targets with luns. But I realized that I actually don't need that for our purposes, so I didn't implement it.

Closing in favor of D38500

I like it. And yes I think you should add stddev to match the regular program output at termination.

Philosophically, SIGQUIT is different than SIGINT. It's more like "I the programmer have determined that this program is behaving badly, and would like a core dump". But I could go either way, if people disagree with me here.

I sort of forgot about this review. I've committed it as-is, and next I'll make the memory allocations more efficient, as @mav suggested.

I don't understand why you need this revision. For what will it serve as a reference implementation?