)]}' { "commit": "f9a30554803a670f9b95a7794be00f03929d6ecd", "tree": "46da1dafce0c76ab5730540540aaf4093463f551", "parents": [ "291764a70e6d8b212680e311bfb0825abf2b9a2f" ], "author": { "name": "David Hendricks", "email": "dhendricks@fb.com", "time": "Sat May 23 20:30:30 2015 -0700" }, "committer": { "name": "David Hendricks", "email": "david.hendricks@gmail.com", "time": "Thu May 17 16:49:16 2018 +0000" }, "message": "linux_mtd: Import driver from ChromiumOS\n\nThis imports a series of patches from chromiumos for MTD support.\nThe patches are squashed to ease review and original Change-Ids have\nbeen removed to avoid confusing Gerrit.\n\nThere are a few changes to integrate the code:\n- Conflict resolution\n- Makefile changes\n- Remove file library usage from linux_mtd. We may revisit this and use\n it for other Linux interfaces later on.\n- Switch to using file stream functions for reads and writes.\n\nThis consolidated patch is\nSigned-off-by: David Hendricks \u003cdhendricks@fb.com\u003e\n\nThe first commit\u0027s message is:\nInitial MTD support\n\nThis adds MTD support to flashrom so that we can read, erase, and\nwrite content on a NOR flash chip via MTD.\n\nBUG\u003dchrome-os-partner:40208\nBRANCH\u003dnone\nTEST\u003dread, write, and erase works on Oak\n\nSigned-off-by: David Hendricks \u003cdhendrix@chromium.org\u003e\nReviewed-on: https://chromium-review.googlesource.com/272983\nReviewed-by: Shawn N \u003cshawnn@chromium.org\u003e\n\nThis is the 2nd commit message:\n\nlinux_mtd: Fix compilation errors\n\nThis fixes compilation errors from the initial import patch.\n\nSigned-off-by: David Hendricks \u003cdhendricks@fb.com\u003e\n\nThis is the 3rd commit message:\n\nlinux_mtd: Suppress message if NOR device not found\n\nThis just suppresses a message that might cause confusion for\nunsuspecting users.\n\nBUG\u003dnone\nBRANCH\u003dnone\nTEST\u003dran on veyron_mickey, \"NOR type device not found\" message\nno longer appears under normal circumstances.\nSigned-off-by: David Hendricks \u003cdhendrix@chromium.org\u003e\n\nReviewed-on: https://chromium-review.googlesource.com/302145\nCommit-Ready: David Hendricks \u003cdhendrix@chromium.org\u003e\nTested-by: David Hendricks \u003cdhendrix@chromium.org\u003e\nReviewed-by: Shawn N \u003cshawnn@chromium.org\u003e\n\nThis is the 4th commit message:\n\nlinux_mtd: Support for NO_ERASE type devices\n\nSome mtd devices have the MTD_NO_ERASE flag set. This means\nthese devices don\u0027t require an erase to write and might not have\nimplemented an erase function. We should be conservative and skip\nerasing altogether, falling back to performing writes over the whole\nflash.\n\nBUG\u003db:35104688\nTESTED\u003dZaius flash is now written correctly for the 0xff regions.\n\nSigned-off-by: William A. Kennington III \u003cwak@google.com\u003e\nReviewed-on: https://chromium-review.googlesource.com/472128\nCommit-Ready: William Kennington \u003cwak@google.com\u003e\nTested-by: William Kennington \u003cwak@google.com\u003e\nReviewed-by: Brian Norris \u003cbriannorris@chromium.org\u003e\n\nThis is the 5th commit message:\n\nlinux_mtd: do reads in eraseblock-sized chunks\n\nIt\u0027s probably not the best idea to try to do an 8MB read in one syscall.\nTheoretically, this should work; but MTD just relies on the SPI driver\nto deliver the whole read in one transfer, and many SPI drivers haven\u0027t\nbeen tested well with large transfer sizes.\n\nI\u0027d consider this a workaround, but it\u0027s still good to have IMO.\n\nBUG\u003dchrome-os-partner:53215\nTEST\u003dboot kevin; `flashrom --read ...`\nTEST\u003dcheck for performance regression on oak\nBRANCH\u003dnone\n\nSigned-off-by: Brian Norris \u003cbriannorris@chromium.org\u003e\nReviewed-on: https://chromium-review.googlesource.com/344006\nReviewed-by: David Hendricks \u003cdhendrix@chromium.org\u003e\n\nThis is the 6th commit message:\n\nlinux_mtd: make read/write loop chunks consistent, and documented\n\nTheoretically, there should be no maximum size for the read() and\nwrite() syscalls on an MTD (well, except for the size of the entire\ndevice). But practical concerns (i.e., bugs) have meant we don\u0027t quite\ndo this.\n\nFor reads:\nBug https://b/35573113 shows that some SPI-based MTD drivers don\u0027t yet\nhandle very large transactions. So we artificially limit this to\nblock-sized chunks.\n\nFor writes:\nIt\u0027s not clear there is a hard limit. Some drivers will already split\nlarge writes into smaller chunks automatically. Others don\u0027t do any\nsplitting. At any rate, using *small* chunks can actually be a problem\nfor some devices (b:35104688), as they get worse performance (doing an\ninternal read/modify/write). This could be fixed in other ways by\nadvertizing their true \"write chunk size\" to user space somehow, but\nthis isn\u0027t so easy.\n\nAs a simpler fix, we can just increase the loop increment to match the\nread loop. Per David, the original implementation (looping over page\nchunks) was just being paranoid.\n\nSo this patch:\n * clarifies comments in linux_mtd_read(), to note that the chunking is\n somewhat of a hack that ideally can be fixed (with bug reference)\n * simplifies the linux_mtd_write() looping to match the structure in\n linux_mtd_read(), including dropping several unnecessary seeks, and\n correcting the error messages (they referred to \"reads\" and had the\n wrong parameters)\n * change linux_mtd_write() to align its chunks to eraseblocks, not page\n sizes\n\nNote that the \"-\u003epage_size\" parameter is still somewhat ill-defined, and\nonly set by the upper layers for \"opaque\" flash. And it\u0027s not actually\nused in this driver now. If we could figure out what we really want to\nuse it for, then we could try to set it appropriately.\n\nBRANCH\u003dnone\nBUG\u003db:35104688\nTEST\u003dvarious flashrom tests on Kevin\nTEST\u003dReading and writing to flash works on our zaius machines over mtd\n\nChange-Id: I3d6bb282863a5cf69909e28a1fc752b35f1b9599\nSigned-off-by: Brian Norris \u003cbriannorris@chromium.org\u003e\nReviewed-on: https://chromium-review.googlesource.com/505409\nReviewed-by: David Hendricks \u003cdhendrix@chromium.org\u003e\nReviewed-by: Martin Roth \u003cmartinroth@chromium.org\u003e\nReviewed-by: William Kennington \u003cwak@google.com\u003e\nReviewed-on: https://review.coreboot.org/25706\nTested-by: David Hendricks \u003cdavid.hendricks@gmail.com\u003e\nTested-by: build bot (Jenkins) \u003cno-reply@coreboot.org\u003e\nReviewed-by: Nico Huber \u003cnico.h@gmx.de\u003e\nReviewed-by: Philipp Deppenwiese \u003czaolin.daisuki@gmail.com\u003e\n", "tree_diff": [ { "type": "modify", "old_id": "d150dcb0872762c9fedd6e93cb51490984ef114c", "old_mode": 33188, "old_path": "Makefile", "new_id": "943d88d76f77c4aa97fd877a7081279bbed2a042", "new_mode": 33188, "new_path": "Makefile" }, { "type": "modify", "old_id": "72d345a4d877a88121c9a0c22b976209194cbdfd", "old_mode": 33188, "old_path": "flash.h", "new_id": "fca377506c091f67273b39a5a731bda34f1f086b", "new_mode": 33188, "new_path": "flash.h" }, { "type": "modify", "old_id": "16001137449a5557f8a457bbb2b247c8377385ea", "old_mode": 33188, "old_path": "flashrom.8.tmpl", "new_id": "e732bcf21403b99a8f93a86a4960071eb0dd0610", "new_mode": 33188, "new_path": "flashrom.8.tmpl" }, { "type": "modify", "old_id": "776ad70b7b34f64f259897f0fdfb4fbaae7fb83b", "old_mode": 33188, "old_path": "flashrom.c", "new_id": "f85dbb1f836c62f44f7fcaa0f363a15c780214e2", "new_mode": 33188, "new_path": "flashrom.c" }, { "type": "modify", "old_id": "9d2d8ad2f557c1d82fb2e79f8d0674d2d1a468c1", "old_mode": 33188, "old_path": "internal.c", "new_id": "1d6cff6b638915f05ce73a8c83d775adb6c228ed", "new_mode": 33188, "new_path": "internal.c" }, { "type": "add", "old_id": "0000000000000000000000000000000000000000", "old_mode": 0, "old_path": "/dev/null", "new_id": "e65f093c0c217a3f2cb18b1ae7c24abb7c9b19d3", "new_mode": 33188, "new_path": "linux_mtd.c" }, { "type": "modify", "old_id": "5d9610a465dc713644f8efacf559eb660cfeb9a9", "old_mode": 33188, "old_path": "programmer.h", "new_id": "e49f2621ea1e1bac21ff64cde0920a1f64d40af8", "new_mode": 33188, "new_path": "programmer.h" } ] }