<html>
<head>
<meta http-equiv="Content-Type" content="text/html; charset=us-ascii">
</head>
<body style="word-wrap: break-word; -webkit-nbsp-mode: space; line-break: after-white-space;" class="">
On Aug 19, 2022, at 13:58, John Bauer <<a href="mailto:bauerj@iodoctors.com" class="">bauerj@iodoctors.com</a>> wrote:<br class="">
<div>
<blockquote type="cite" class=""><br class="Apple-interchange-newline">
<div class="">
<div class="">
<p class="">Andreas,</p>
<p class="">As I mentioned in an earlier email, this had been working for a long time. I think that using an old header file is at the root of the issue. On my development platform, which doesn't have Lustre installed, nor did I have e2fsprogs installed,
I had simply copied the Lustre files I needed from the site I was working with. The fiemap.h file I was using , the top of which is shown below ( I see you mentioned ) has fe_device explicitly in the structure. Was it this way before the #define fe_device
was implemented?</p>
</div>
</div>
</blockquote>
<div>Yes, we used to patch the fiemap.h header to add fe_device ourselves, but changed to the #define mechanism to reduce the changes to the core kernel.</div>
<div><br class="">
</div>
<div>Cheers, Andreas</div>
<blockquote type="cite" class="">
<div class="">
<div class="">
<p class=""> The #define was using the fe_reserved[0], which always had a 0 value. What puzzles me is why this ever worked at all. That will have to wait for a rainy day to mess with. What started me down this path at this time was getting my lustre extents
plotting program working with PFL.<br class="">
</p>
<p class="">Again, thanks much for your excellent/quick assistance in tracking this down.
<br class="">
</p>
<p class="">John<br class="">
</p>
<pre class=""><font color="#3465A4" class="">/*</font>
<font color="#3465A4" class=""> * FS_IOC_FIEMAP ioctl infrastructure.</font>
<font color="#3465A4" class=""> *</font>
<font color="#3465A4" class=""> * Some portions copyright (C) 2007 Cluster File Systems, Inc</font>
<font color="#3465A4" class=""> *</font>
<font color="#3465A4" class=""> * Authors: Mark Fasheh <a class="moz-txt-link-rfc2396E" href="mailto:mfasheh@suse.com"><mfasheh@suse.com></a></font>
<font color="#3465A4" class=""> * Kalpak Shah <a class="moz-txt-link-rfc2396E" href="mailto:kalpak.shah@sun.com"><kalpak.shah@sun.com></a></font>
<font color="#3465A4" class=""> * Andreas Dilger <a class="moz-txt-link-rfc2396E" href="mailto:adilger@sun.com"><adilger@sun.com></a></font>
<font color="#3465A4" class=""> */</font>
<font color="#75507B" class="">#ifndef _LINUX_FIEMAP_H</font>
<font color="#75507B" class="">#define _LINUX_FIEMAP_H</font>
<font color="#4E9A06" class="">struct</font> fiemap_extent {
__u64 fe_logical; <font color="#3465A4" class="">/* logical offset in bytes for the start of</font>
<font color="#3465A4" class=""> * the extent from the beginning of the file */</font>
__u64 fe_physical; <font color="#3465A4" class="">/* physical offset in bytes for the start</font>
<font color="#3465A4" class=""> * of the extent from the beginning of the disk */</font>
__u64 fe_length; <font color="#3465A4" class="">/* length in bytes for this extent */</font>
__u64 fe_reserved64[<font color="#CC0000" class="">2</font>];
__u32 fe_flags; <font color="#3465A4" class="">/* FIEMAP_EXTENT_* flags for this extent */</font>
__u32 fe_device; <font color="#3465A4" class="">/* device number (fs-specific if FIEMAP_EXTENT_NET)*/</font>
__u32 fe_reserved[<font color="#CC0000" class="">2</font>];
};
</pre>
<div class="moz-cite-prefix">On 8/18/22 23:44, Andreas Dilger wrote:<br class="">
</div>
<blockquote type="cite" cite="mid:F46BFD2B-476D-4216-BD2B-5586EAEF5026@whamcloud.com" class="">
The "fe_device" field is actually Lustre-specific, so it is a macro that overlays on fe_reserved[0]:
<div class=""><br class="">
</div>
<div class=""> #define fe_device fe_reserved[0]</div>
<div class=""><br class="">
</div>
<div class="">but that shouldn't affect compiler alignment. On my system, "pahole lustre/llite/lustre.ko" reports:</div>
<div class=""><br class="">
</div>
<div class="">struct fiemap_extent {<br class="">
__u64 fe_logical; /* 0 8 */<br class="">
__u64 fe_physical; /* 8 8 */<br class="">
__u64 fe_length; /* 16 8 */<br class="">
__u64 fe_reserved64[2]; /* 24 16 */<br class="">
__u32 fe_flags; /* 40 4 */<br class="">
__u32 fe_reserved[3]; /* 44 12 */<br class="">
<br class="">
/* size: 56, cachelines: 1, members: 6 */<br class="">
/* last cacheline: 56 bytes */<br class="">
};</div>
<div class=""><br class="">
</div>
<div class="">So there is definitely something going wrong with the struct alignment for fe_reserved, even though there doesn't need to be (all of the fields have "natural" alignment on their 4/8-byte sizes.</div>
<div class=""><br class="">
</div>
<div class="">The other thing that is strange is that you show only 2 fe_reserved[] fields, when I have 3. Is there some other field added to your version of struct fiemap_extent after fe_flags? I don't see anything in the upstream kernel, nor in the Lustre
headers.</div>
<div class=""><br class="">
</div>
<div class="">You could try adding "__attribute__((packed))" at the end of the struct definition to see if that fixes the problem.<br class="">
<div class=""><br class="">
</div>
<div class="">Cheers, Andreas</div>
<div class=""><br class="">
</div>
<div class="">
<blockquote type="cite" class="">
<div class="">On Aug 18, 2022, at 21:54, John Bauer <<a href="mailto:bauerj@iodoctors.com" class="moz-txt-link-freetext" moz-do-not-send="true">bauerj@iodoctors.com</a>> wrote:</div>
<br class="Apple-interchange-newline">
<div class="">
<div class="">
<p class="">Andreas,</p>
<p class="">This is no longer Lustre related, but I hope you can shed some light on this. It appears that my compilier, gcc 8.5.0, which I upgraded to recently when I upgraded my build system to Centos 8, is not padding the struct fiemap_extent correctly.
I put the following prints in to see whats going on. The sizeof the structure is good at 56, but notice that both fe_device and fe_reserved[0] have an offset of 48 bytes into the structure. Odd that the sizeof fe_flags is 4, but fe_device is 8 bytes away
from it. I traced the compile to ensure that I am getting the lustre_include/ext2fs/fiemap.h and there is nothing odd in the fiemap.h ( it's the one I've been using for years ). Any thoughts on how to remedy this?</p>
<p class="">John<br class="">
</p>
<pre class="">fprintf(stderr,"%s() logical %d %ld\n", __func__, sizeof(fm_ext->fe_logical ), (char *)&fm_ext->fe_logical -(char *)fm_ext);
fprintf(stderr,"%s() physical %d %ld\n", __func__, sizeof(fm_ext->fe_physical), (char *)&fm_ext->fe_physical -(char *)fm_ext);
fprintf(stderr,"%s() length %d %ld\n", __func__, sizeof(fm_ext->fe_length ), (char *)&fm_ext->fe_length - (char *)fm_ext);
fprintf(stderr,"%s() res64[0] %d %ld\n", __func__, sizeof(fm_ext->fe_reserved64[0]), (char *)&fm_ext->fe_reserved64[0] - (char *)fm_ext);
fprintf(stderr,"%s() res64[1] %d %ld\n", __func__, sizeof(fm_ext->fe_reserved64[1]), (char *)&fm_ext->fe_reserved64[1] - (char *)fm_ext);
fprintf(stderr,"%s() flags %d %ld\n", __func__, sizeof(fm_ext->fe_flags ), (char *)&fm_ext->fe_flags -(char *)fm_ext);
fprintf(stderr,"%s() device %d %ld\n", __func__, sizeof(fm_ext->fe_device ), (char *)&fm_ext->fe_device -(char *)fm_ext);
fprintf(stderr,"%s() res32[0] %d %ld\n", __func__, sizeof(fm_ext->fe_reserved[0]), (char *)&fm_ext->fe_reserved[0] - (char *)fm_ext);
fprintf(stderr,"%s() res32[1] %d %ld\n", __func__, sizeof(fm_ext->fe_reserved[1]), (char *)&fm_ext->fe_reserved[1] - (char *)fm_ext);
</pre>
<pre class="">StripeChunks_get() fm_ext->fe_device=0 fe_logical=0 sizeof(struct fiemap_extent)56
StripeChunks_get() logical 8 0
StripeChunks_get() physical 8 8
StripeChunks_get() length 8 16
StripeChunks_get() res64[0] 8 24
StripeChunks_get() res64[1] 8 32
StripeChunks_get() flags 4 40
StripeChunks_get() device 4 48
StripeChunks_get() res32[0] 4 48
StripeChunks_get() res32[1] 4 52
</pre>
<div class="moz-cite-prefix">On 8/18/22 16:11, Andreas Dilger wrote:<br class="">
</div>
<blockquote type="cite" cite="mid:17DE058E-FEF9-43FD-B39B-3A8BE7583B8F@whamcloud.com" class="">
<div class="">
<div class="">On Aug 18, 2022, at 14:28, John Bauer <<a href="mailto:bauerj@iodoctors.com" class="moz-txt-link-freetext" moz-do-not-send="true">bauerj@iodoctors.com</a>> wrote:</div>
<blockquote type="cite" class=""><br class="Apple-interchange-newline">
<div class="">
<div class="">
<p class="">Andreas,</p>
<p class="">Thanks for the reply. I don't think I'm accessing the Lustre filefrag ( see below ). Where would I normally find that installed? I downloaded the lustre-release git repository and can't find filefrag stuff to build my own. Is that somewhere else?</p>
</div>
</div>
</blockquote>
filefrag is part of the e2fsprogs package ("rpm -qf $(which filefrag)"), so you need to download and install the Lustre-patched e2fsprogs from <u class=""><a href="https://downloads.whamcloud.com/public/e2fsprogs/latest/" class="moz-txt-link-freetext" moz-do-not-send="true">https://downloads.whamcloud.com/public/e2fsprogs/latest/</a></u><br class="">
<div class=""><br class="">
</div>
<blockquote type="cite" class="">
<div class="">
<div class="">
<p class="">More info: <br class="">
</p>
<pre class="">pfe27.jbauer2 334> cat /sys/fs/lustre/version
2.12.8_ddn12
</pre>
</div>
</div>
</blockquote>
<div class=""><br class="">
</div>
You should really use "lctl get_param version", since the Lustre /proc and /sys files move around on occasion.</div>
<div class=""><br class="">
</div>
<div class="">The PFL/FLR change for FIEMAP is not included in this version, but it _should_ be irrelevant because the file you are testing is using a plain layout, not PFL or FLR.<br class="">
<blockquote type="cite" class="">
<div class="">
<div class="">
<pre class="">pfe27.jbauer2 335> filefrag -v /nobackupp17/jbauer2/dd.dat
Filesystem type is: bd00bd0
File size of /nobackupp17/jbauer2/dd.dat is 104857600 (25600 blocks of 4096 bytes)
/nobackupp17/jbauer2/dd.dat: FIBMAP unsupported
pfe27.jbauer2 336> which filefrag
/usr/sbin/filefrag
</pre>
<div class="moz-cite-prefix">John<br class="">
</div>
<div class="moz-cite-prefix"><br class="">
</div>
<div class="moz-cite-prefix">On 8/18/22 14:57, Andreas Dilger wrote:<br class="">
</div>
<blockquote type="cite" cite="mid:E166F2A4-5F6B-43CC-8E8E-055B25B9A97C@whamcloud.com" class="">
What version of Lustre are you using? Does "filefrag -v" from a newer Lustre e2fsprogs (1.45.6.wc3+) work properly?
<div class=""><br class="">
</div>
<div class="">There was a small change to the Lustre FIEMAP handling in order to handle overstriped files and PFL/FLR files with many stripes and multiple components, since the FIEMAP "restart" mechanism was broken for files that had multiple objects on the
same OST index. See LU-11484 for details. That change was included in the 2.14.0 release.</div>
<div class=""><br class="">
</div>
<div class="">In essence, the fe_device field now encodes the absolute file stripe number in the high 16 bits of fe_device, and the device number in the low 16 bits (as it did before). Since "filefrag -v" prints fe_device in hex and would show as "0x<stripe><device>"
instead of "0x0000<device>", this was considered an acceptable tradeoff compared to other "less compatible" changes that would have been needed to implement PFL/FLR handling.</div>
<div class=""><br class="">
</div>
<div class="">That said, I would have expected this change to result in your tool reporting very large values for fe_device (e.g. OST index + N * 65536), so returning all-zero values is somewhat unexpected. </div>
<div class=""><br class="">
</div>
<div class="">Cheers, Andreas</div>
<div class="">
<div class=""><br class="">
<blockquote type="cite" class="">
<div class="">On Aug 18, 2022, at 06:27, John Bauer <<a href="mailto:bauerj@iodoctors.com" class="moz-txt-link-freetext" moz-do-not-send="true">bauerj@iodoctors.com</a>> wrote:</div>
<br class="Apple-interchange-newline">
<div class="">
<div class="">Hi all,<br class="">
<br class="">
I am trying to get my llfie program (which uses fiemap) going again, but now the struct fiemap_extent structures I get back from the ioctl call, all have fe_device=0. The output from lfs getstripe indicates that the devices are not all 0. The sum of the fe_length
members adds up to the file size, so that is working. The fe_physical members look reasonable also. Has something changed? This used to work.<br class="">
<br class="">
Thanks, John<br class="">
<br class="">
pfe27.jbauer2 300> llfie /nobackupp17/jbauer2/dd.dat<br class="">
LustreStripeInfo_get() lum->lmm_magic=0xbd30bd0<br class="">
listExtents() fe_physical=30643484360704 fe_device=0 fe_length=16777216<br class="">
listExtents() fe_physical=30646084829184 fe_device=0 fe_length=2097152<br class="">
listExtents() fe_physical=5705226518528 fe_device=0 fe_length=16777216<br class="">
listExtents() fe_physical=5710209351680 fe_device=0 fe_length=2097152<br class="">
listExtents() fe_physical=30621271326720 fe_device=0 fe_length=16777216<br class="">
listExtents() fe_physical=31761568366592 fe_device=0 fe_length=16777216<br class="">
listExtents() fe_physical=24757567225856 fe_device=0 fe_length=16777216<br class="">
listExtents() fe_physical=14196460748800 fe_device=0 fe_length=16777216<br class="">
listExtents() nMapped=8 byteCount=104857600<br class="">
<br class="">
<br class="">
pfe27.jbauer2 301> lfs getstripe /nobackupp17/jbauer2/dd.dat<br class="">
/nobackupp17/jbauer2/dd.dat<br class="">
lmm_stripe_count: 6<br class="">
lmm_stripe_size: 2097152<br class="">
lmm_pattern: raid0<br class="">
lmm_layout_gen: 0<br class="">
lmm_stripe_offset: 126<br class="">
lmm_pool: ssd-pool<br class="">
<span class="Apple-tab-span"></span>obdidx<span class="Apple-tab-span"> </span><span class="Apple-tab-span"></span>objid<span class="Apple-tab-span">
</span><span class="Apple-tab-span"></span>objid<span class="Apple-tab-span"> </span>
<span class="Apple-tab-span"></span>group<br class="">
<span class="Apple-tab-span"></span> 126<span class="Apple-tab-span"> </span> 13930025<span class="Apple-tab-span">
</span> 0xd48e29<span class="Apple-tab-span"> </span> 0<br class="">
<span class="Apple-tab-span"></span> 113<span class="Apple-tab-span"> </span> 13115889<span class="Apple-tab-span">
</span> 0xc821f1<span class="Apple-tab-span"> </span> 0<br class="">
<span class="Apple-tab-span"></span> 120<span class="Apple-tab-span"> </span> 14003176<span class="Apple-tab-span">
</span> 0xd5abe8<span class="Apple-tab-span"> </span> 0<br class="">
<span class="Apple-tab-span"></span> 109<span class="Apple-tab-span"> </span> 12785483<span class="Apple-tab-span">
</span> 0xc3174b<span class="Apple-tab-span"> </span> 0<br class="">
<span class="Apple-tab-span"></span> 102<span class="Apple-tab-span"> </span> 13811117<span class="Apple-tab-span">
</span> 0xd2bdad<span class="Apple-tab-span"> </span> 0<br class="">
<span class="Apple-tab-span"></span> 116<span class="Apple-tab-span"> </span> 13377285<span class="Apple-tab-span">
</span> 0xcc1f05<span class="Apple-tab-span"> </span> 0<br class="">
<br class="">
_______________________________________________<br class="">
lustre-discuss mailing list<br class="">
<a href="mailto:lustre-discuss@lists.lustre.org" class="moz-txt-link-freetext" moz-do-not-send="true">lustre-discuss@lists.lustre.org</a><br class="">
<a class="moz-txt-link-freetext" href="http://lists.lustre.org/listinfo.cgi/lustre-discuss-lustre.org" moz-do-not-send="true">http://lists.lustre.org/listinfo.cgi/lustre-discuss-lustre.org</a><br class="">
</div>
</div>
</blockquote>
</div>
<br class="">
<div class="">
<div dir="auto" class="">
<div dir="auto" class="">
<div dir="auto" class="">
<div dir="auto" class="">
<div dir="auto" class="">
<div dir="auto" class="">
<div class="">Cheers, Andreas</div>
<div class="">--</div>
<div class="">Andreas Dilger</div>
<div class="">Lustre Principal Architect</div>
<div class="">Whamcloud</div>
<div class=""><br class="">
</div>
<div class=""><br class="">
</div>
<div class=""><br class="">
</div>
</div>
</div>
</div>
</div>
</div>
<br class="Apple-interchange-newline">
</div>
<br class="Apple-interchange-newline">
<br class="Apple-interchange-newline">
</div>
<br class="">
</div>
</blockquote>
</div>
</div>
</blockquote>
</div>
<br class="">
<div class="">
<div dir="auto" class="">
<div dir="auto" class="">
<div dir="auto" class="">
<div dir="auto" class="">
<div dir="auto" class="">
<div dir="auto" class="">
<div class="">Cheers, Andreas</div>
<div class="">--</div>
<div class="">Andreas Dilger</div>
<div class="">Lustre Principal Architect</div>
<div class="">Whamcloud</div>
<div class=""><br class="">
</div>
<div class=""><br class="">
</div>
<div class=""><br class="">
</div>
</div>
</div>
</div>
</div>
</div>
<br class="Apple-interchange-newline">
</div>
<br class="Apple-interchange-newline">
<br class="Apple-interchange-newline">
</div>
<br class="">
</blockquote>
</div>
</div>
</blockquote>
</div>
<br class="">
<div class="">
<div dir="auto" class="">
<div dir="auto" class="">
<div dir="auto" class="">
<div dir="auto" class="">
<div dir="auto" class="">
<div dir="auto" class="">
<div class="">Cheers, Andreas</div>
<div class="">--</div>
<div class="">Andreas Dilger</div>
<div class="">Lustre Principal Architect</div>
<div class="">Whamcloud</div>
<div class=""><br class="">
</div>
<div class=""><br class="">
</div>
<div class=""><br class="">
</div>
</div>
</div>
</div>
</div>
</div>
<br class="Apple-interchange-newline">
</div>
<br class="Apple-interchange-newline">
<br class="Apple-interchange-newline">
</div>
<br class="">
</div>
</blockquote>
</div>
</div>
</blockquote>
</div>
<br class="">
<div class="">
<div dir="auto" style="caret-color: rgb(0, 0, 0); color: rgb(0, 0, 0); letter-spacing: normal; text-align: start; text-indent: 0px; text-transform: none; white-space: normal; word-spacing: 0px; -webkit-text-stroke-width: 0px; text-decoration: none; word-wrap: break-word; -webkit-nbsp-mode: space; line-break: after-white-space;" class="">
<div dir="auto" style="caret-color: rgb(0, 0, 0); color: rgb(0, 0, 0); letter-spacing: normal; text-align: start; text-indent: 0px; text-transform: none; white-space: normal; word-spacing: 0px; -webkit-text-stroke-width: 0px; text-decoration: none; word-wrap: break-word; -webkit-nbsp-mode: space; line-break: after-white-space;" class="">
<div dir="auto" style="caret-color: rgb(0, 0, 0); color: rgb(0, 0, 0); letter-spacing: normal; text-align: start; text-indent: 0px; text-transform: none; white-space: normal; word-spacing: 0px; -webkit-text-stroke-width: 0px; text-decoration: none; word-wrap: break-word; -webkit-nbsp-mode: space; line-break: after-white-space;" class="">
<div dir="auto" style="caret-color: rgb(0, 0, 0); color: rgb(0, 0, 0); letter-spacing: normal; text-align: start; text-indent: 0px; text-transform: none; white-space: normal; word-spacing: 0px; -webkit-text-stroke-width: 0px; text-decoration: none; word-wrap: break-word; -webkit-nbsp-mode: space; line-break: after-white-space;" class="">
<div dir="auto" style="caret-color: rgb(0, 0, 0); color: rgb(0, 0, 0); letter-spacing: normal; text-align: start; text-indent: 0px; text-transform: none; white-space: normal; word-spacing: 0px; -webkit-text-stroke-width: 0px; text-decoration: none; word-wrap: break-word; -webkit-nbsp-mode: space; line-break: after-white-space;" class="">
<div dir="auto" style="caret-color: rgb(0, 0, 0); color: rgb(0, 0, 0); letter-spacing: normal; text-align: start; text-indent: 0px; text-transform: none; white-space: normal; word-spacing: 0px; -webkit-text-stroke-width: 0px; text-decoration: none; word-wrap: break-word; -webkit-nbsp-mode: space; line-break: after-white-space;" class="">
<div>Cheers, Andreas</div>
<div>--</div>
<div>Andreas Dilger</div>
<div>Lustre Principal Architect</div>
<div>Whamcloud</div>
<div><br class="">
</div>
<div><br class="">
</div>
<div><br class="">
</div>
</div>
</div>
</div>
</div>
</div>
<br class="Apple-interchange-newline">
</div>
<br class="Apple-interchange-newline">
<br class="Apple-interchange-newline">
</div>
<br class="">
</body>
</html>