Question about CVE-2022-43945
Subsystems (2)
Hi, Chuck Lever,
CVE-2022-43945(https://nvd.nist.gov/vuln/detail/CVE-2022-43945) describe that a normal request header ended with garbage data can trigger the nfsd overflow since nfsd share the request and response with the same pages array.
It seems that the patchset(https://lore.kernel.org/linux-nfs/166204973526.1435.6068003336048840051.stgit@manet.1015granger.net/T/#t) has solved NFSv2/NFSv3, but leave NFSv4 still vulnerably?
Another question, for stable branch like lts-5.10, since NFSv2/NFSv3 did not switch to xdr_stream, the nfs_request_too_big in nfsd_dispatch will reject the request like READ/READDIR with too large request. So it seems branch without that "switch" seems ok for NFSv2/NFSv3, but NFSv3 still vulnerably. right?
Looking forward to your reply!
Thanks,
Erkun Yang
Quoted 9 lines · yangerkun
Hi, Chuck Lever,
CVE-2022-43945(https://nvd.nist.gov/vuln/detail/CVE-2022-43945) describe that a normal request header ended with garbage data can trigger the nfsd overflow since nfsd share the request and response with the same pages array.
It seems that the patchset(https://lore.kernel.org/linux-nfs/166204973526.1435.6068003336048840051.stgit@manet.1015granger.net/T/#t) has solved NFSv2/NFSv3, but leave NFSv4 still vulnerably?
Another question, for stable branch like lts-5.10, since NFSv2/NFSv3 did not switch to
xdr_stream, thenfs_request_too_biginnfsd_dispatchwill reject the request like READ/READDIR with too large request. So it seems branch without that "switch" seems ok for NFSv2/NFSv3, but NFSv3 still vulnerably. right?Looking forward to your reply!
Sorry, notice that 76ce4dcec0dc"NFSD: Cap rsize_bop result based on send buffer size") fix same problem for NFSv4.
So, for the stable branch like lts-5.10 which NFSv2/NFSv3 do not switch to xdr_stream, it seems we only need 76ce4dcec0dc"NFSD: Cap rsize_bop result based on send buffer size"). Right?
Quoted 4 lines
Thanks,
Erkun Yang
.
Quoted 8 lines · yangerkun
Quoted 3 lines · yangerkun
Hi, Chuck Lever,
CVE-2022-43945(https://nvd.nist.gov/vuln/detail/CVE-2022-43945) describe that a normal request header ended with garbage data can trigger the nfsd overflow since nfsd share the request and response with the same pages array. It seems that the patchset(https://lore.kernel.org/linux-nfs/166204973526.1435.6068003336048840051.stgit@manet.1015granger.net/T/#t) has solved NFSv2/NFSv3, but leave NFSv4 still vulnerably?
I asked the folks who reported this issue to check NFSv4 as well. They were not able to exploit NFSv4 in the same way. For now we believe this vulnerability does not impact the NFSv4 code paths.
Quoted 4 lines
Quoted 2 lines
Another question, for stable branch like lts-5.10, since NFSv2/NFSv3 did not switch to
xdr_stream, thenfs_request_too_biginnfsd_dispatchwill reject the request like READ/READDIR with too large request. So it seems branch without that "switch" seems ok for NFSv2/NFSv3, but NFSv3 still vulnerably. right? Looking forward to your reply!Sorry, notice that
76ce4dcec0dc("NFSD: Caprsize_bopresult based on send buffer size") fix same problem for NFSv4.
76ce4dcec0dc is a defensive fix. But, as I stated above, we haven't yet found that NFSD's NFSv4 implementation is vulnerable to this issue.
Quoted 1 line
So, for the stable branch like lts-5.10 which NFSv2/NFSv3 do not switch to
xdr_stream, it seems we only need 76ce4dcec0dc"NFSD: Caprsize_bopresult based on send buffer size"). Right?
At this time we don't believe 76ce4dcec0dc is required. But if you want it applied to v5.10 (or any LTS kernel) please first test that it does not result in a regression, and then make a request to the usual stable maintainers.
--
Chuck Lever
Hi Chuck,
Quoted 24 lines · Chuck Lever III
Quoted 8 lines · yangerkun
Quoted 3 lines · yangerkun
Hi, Chuck Lever,
CVE-2022-43945(https://nvd.nist.gov/vuln/detail/CVE-2022-43945) describe that a normal request header ended with garbage data can trigger the nfsd overflow since nfsd share the request and response with the same pages array. It seems that the patchset(https://lore.kernel.org/linux-nfs/166204973526.1435.6068003336048840051.stgit@manet.1015granger.net/T/#t) has solved NFSv2/NFSv3, but leave NFSv4 still vulnerably?I asked the folks who reported this issue to check NFSv4 as well. They were not able to exploit NFSv4 in the same way. For now we believe this vulnerability does not impact the NFSv4 code paths.
Quoted 4 lines
Quoted 2 lines
Another question, for stable branch like lts-5.10, since NFSv2/NFSv3 did not switch to
xdr_stream, thenfs_request_too_biginnfsd_dispatchwill reject the request like READ/READDIR with too large request. So it seems branch without that "switch" seems ok for NFSv2/NFSv3, but NFSv3 still vulnerably. right? Looking forward to your reply!Sorry, notice that
76ce4dcec0dc("NFSD: Caprsize_bopresult based on send buffer size") fix same problem for NFSv4.
76ce4dcec0dcis a defensive fix. But, as I stated above, we haven't yet found that NFSD's NFSv4 implementation is vulnerable to this issue.
Quoted 6 lines
Quoted 1 line
So, for the stable branch like lts-5.10 which NFSv2/NFSv3 do not switch to
xdr_stream, it seems we only need 76ce4dcec0dc"NFSD: Caprsize_bopresult based on send buffer size"). Right?At this time we don't believe
76ce4dcec0dcis required. But if you want it applied to v5.10 (or any LTS kernel) please first test that it does not result in a regression, and then make a request to the usual stable maintainers.
I was reviewing open CVEs for Debian, based on the 5.10.y stable series, and noticed CVE-2022-43945 is yet unfixed in 5.10.162. I see SUSE did some backporting, with Neil Brown, according to
https://bugzilla.suse.com/show_bug.cgi?id=1205128#c4 . From the set of fixes the first two of
Commit 90bfc37b5ab9 ("SUNRPC: Fix svcxdr_init_decode's end-of-buffer calculation") Commit 1242a87da0d8 ("SUNRPC: Fix svcxdr_init_encode's buflen calculation") Commit 00b4492686e0 ("NFSD: Protect against send buffer overflow in NFSv2 READDIR") Commit 640f87c190e0 ("NFSD: Protect against send buffer overflow in NFSv3 READDIR") Commit 401bc1f90874 ("NFSD: Protect against send buffer overflow in NFSv2 READ") Commit fa6be9cc6e80 ("NFSD: Protect against send buffer overflow in NFSv3 READ") Commit 76ce4dcec0dc ("NFSD: Cap rsize_bop result based on send buffer size")
would not be needed, but still the others, though won't apply cleanly as thei need substantial changes. Neil, would it be possible to have the fixes backported to the 5.10.y series as well?
Regards,
Salvatore
Quoted 48 lines · Salvatore Bonaccorso
Hi Chuck,
Quoted 24 lines · Chuck Lever III
Quoted 8 lines · yangerkun
Quoted 3 lines · yangerkun
Hi, Chuck Lever,
CVE-2022-43945(https://nvd.nist.gov/vuln/detail/CVE-2022-43945) describe that a normal request header ended with garbage data can trigger the nfsd overflow since nfsd share the request and response with the same pages array. It seems that the patchset(https://lore.kernel.org/linux-nfs/166204973526.1435.6068003336048840051.stgit@manet.1015granger.net/T/#t) has solved NFSv2/NFSv3, but leave NFSv4 still vulnerably?I asked the folks who reported this issue to check NFSv4 as well. They were not able to exploit NFSv4 in the same way. For now we believe this vulnerability does not impact the NFSv4 code paths.
Quoted 4 lines
Quoted 2 lines
Another question, for stable branch like lts-5.10, since NFSv2/NFSv3 did not switch to
xdr_stream, thenfs_request_too_biginnfsd_dispatchwill reject the request like READ/READDIR with too large request. So it seems branch without that "switch" seems ok for NFSv2/NFSv3, but NFSv3 still vulnerably. right? Looking forward to your reply!Sorry, notice that
76ce4dcec0dc("NFSD: Caprsize_bopresult based on send buffer size") fix same problem for NFSv4.
76ce4dcec0dcis a defensive fix. But, as I stated above, we haven't yet found that NFSD's NFSv4 implementation is vulnerable to this issue.Quoted 6 lines
Quoted 1 line
So, for the stable branch like lts-5.10 which NFSv2/NFSv3 do not switch to
xdr_stream, it seems we only need 76ce4dcec0dc"NFSD: Caprsize_bopresult based on send buffer size"). Right?At this time we don't believe
76ce4dcec0dcis required. But if you want it applied to v5.10 (or any LTS kernel) please first test that it does not result in a regression, and then make a request to the usual stable maintainers.I was reviewing open CVEs for Debian, based on the 5.10.y stable series, and noticed CVE-2022-43945 is yet unfixed in 5.10.162. I see SUSE did some backporting, with Neil Brown, according to
https://bugzilla.suse.com/show_bug.cgi?id=1205128#c4 . From the set of fixes the first two ofCommit
90bfc37b5ab9("SUNRPC: Fix svcxdr_init_decode's end-of-buffer calculation") Commit1242a87da0d8("SUNRPC: Fix svcxdr_init_encode's buflen calculation") Commit00b4492686e0("NFSD: Protect against send buffer overflow in NFSv2 READDIR") Commit640f87c190e0("NFSD: Protect against send buffer overflow in NFSv3 READDIR") Commit401bc1f90874("NFSD: Protect against send buffer overflow in NFSv2 READ") Commitfa6be9cc6e80("NFSD: Protect against send buffer overflow in NFSv3 READ") Commit76ce4dcec0dc("NFSD: Caprsize_bopresult based on send buffer size")
Quoted 3 lines
would not be needed, but still the others, though won't apply cleanly as thei need substantial changes. Neil, would it be possible to have the fixes backported to the 5.10.y series as well?
You can find my 5.3 patches at
https://github.com/SUSE/kernel-source/commit/e93318a3f8b5618afcda871b6d8201466af333a8
and earlier - follow the "parent" link.
Feel free to check they apply to 5.10.y, review them, and forward to [email protected].
NeilBrown
Quoted 4 lines
Regards,
Salvatore