diff for duplicates of <20200903225424.331f1d94@suse.de> diff --git a/a/1.txt b/N1/1.txt index 9305f69..28f7ee5 100644 --- a/a/1.txt +++ b/N1/1.txt @@ -32,7 +32,7 @@ On Thu, 3 Sep 2020 10:36:58 -0500, Michael Christie wrote: > > So is the answer to my question a maybe but it probably will never happen? -A src=dest XCOPY? I think it's just as likely as a cross device XCOPY. +A srcŢst XCOPY? I think it's just as likely as a cross device XCOPY. The UUID collision error is probably unlikely to be triggered because: - XCOPY is a pretty exotic SCSI command mostly used by ESXi - Users may already provide a vpd_unit_serial with enough unique @@ -60,7 +60,7 @@ reasons for RFC are: - I've tested this with libiscsi's iscsi-dd (-x) and test suite, but not against the ESXi initiator yet. -> For example, I create 2 tcmu devices. They both point to the same real device. Then export dev1 through target port1 and dev2 through target port2. Each tcmu device would then have it’s own data/cmd ring and locking, so you do not hit those perf issues. I do this for perf testing. I don’t think that type of thing is common or ever done, so I think the patch would be ok if that is a concern and it’s better than possible data corruption. +> For example, I create 2 tcmu devices. They both point to the same real device. Then export dev1 through target port1 and dev2 through target port2. Each tcmu device would then have it’s own data/cmd ring and locking, so you do not hit those perf issues. I do this for perf testing. I don’t think that type of thing is common or ever done, so I think the patch would be ok if that is a concern and it’s better than possible data corruption. > > Code wise it looks ok to me. diff --git a/a/content_digest b/N1/content_digest index 77f7fd1..fcaa266 100644 --- a/a/content_digest +++ b/N1/content_digest @@ -4,7 +4,7 @@ "ref\07F596A9A-2116-4BA8-8A32-E98EDE996D8C@oracle.com\0" "From\0David Disseldorp <ddiss@suse.de>\0" "Subject\0Re: [RFC PATCH] scsi: target: detect XCOPY NAA descriptor conflicts\0" - "Date\0Thu, 3 Sep 2020 22:54:24 +0200\0" + "Date\0Thu, 03 Sep 2020 20:54:24 +0000\0" "To\0Michael Christie <michael.christie@oracle.com>\0" "Cc\0target-devel@vger.kernel.org" " linux-scsi@vger.kernel.org\0" @@ -44,7 +44,7 @@ "> \n" "> So is the answer to my question a maybe but it probably will never happen?\n" "\n" - "A src=dest XCOPY? I think it's just as likely as a cross device XCOPY.\n" + "A src\305\242st XCOPY? I think it's just as likely as a cross device XCOPY.\n" "The UUID collision error is probably unlikely to be triggered because:\n" "- XCOPY is a pretty exotic SCSI command mostly used by ESXi\n" "- Users may already provide a vpd_unit_serial with enough unique\n" @@ -72,7 +72,7 @@ "- I've tested this with libiscsi's iscsi-dd (-x) and test suite, but not\n" " against the ESXi initiator yet.\n" "\n" - "> For example, I create 2 tcmu devices. They both point to the same real device. Then export dev1 through target port1 and dev2 through target port2. Each tcmu device would then have it\342\200\231s own data/cmd ring and locking, so you do not hit those perf issues. I do this for perf testing. I don\342\200\231t think that type of thing is common or ever done, so I think the patch would be ok if that is a concern and it\342\200\231s better than possible data corruption.\n" + "> For example, I create 2 tcmu devices. They both point to the same real device. Then export dev1 through target port1 and dev2 through target port2. Each tcmu device would then have it\303\242\342\202\254\342\204\242s own data/cmd ring and locking, so you do not hit those perf issues. I do this for perf testing. I don\303\242\342\202\254\342\204\242t think that type of thing is common or ever done, so I think the patch would be ok if that is a concern and it\303\242\342\202\254\342\204\242s better than possible data corruption.\n" "> \n" "> Code wise it looks ok to me.\n" "\n" @@ -80,4 +80,4 @@ "\n" Cheers, David -6dcac343bac1bcdd89a42d8ea93e51f0a5a7f01de031c199b391c04a022882c4 +f5cf55522a2821f10ad4f7da2434887a260e5ac57204a8d1bb061ef96165b103
This is an external index of several public inboxes, see mirroring instructions on how to clone and mirror all data and code used by this external index.