)]}' { "log": [ { "commit": "6038f373a3dc1f1c26496e60b6c40b164716f07e", "tree": "a0d3bbd026eea41b9fc36b8c722cbaf56cd9f825", "parents": [ "1ec5584e3edf9c4bf2c88c846534d19cf986ba11" ], "author": { "name": "Arnd Bergmann", "email": "arnd@arndb.de", "time": "Sun Aug 15 18:52:59 2010 +0200" }, "committer": { "name": "Arnd Bergmann", "email": "arnd@arndb.de", "time": "Fri Oct 15 15:53:27 2010 +0200" }, "message": "llseek: automatically add .llseek fop\n\nAll file_operations should get a .llseek operation so we can make\nnonseekable_open the default for future file operations without a\n.llseek pointer.\n\nThe three cases that we can automatically detect are no_llseek, seq_lseek\nand default_llseek. For cases where we can we can automatically prove that\nthe file offset is always ignored, we use noop_llseek, which maintains\nthe current behavior of not returning an error from a seek.\n\nNew drivers should normally not use noop_llseek but instead use no_llseek\nand call nonseekable_open at open time. Existing drivers can be converted\nto do the same when the maintainer knows for certain that no user code\nrelies on calling seek on the device file.\n\nThe generated code is often incorrectly indented and right now contains\ncomments that clarify for each added line why a specific variant was\nchosen. In the version that gets submitted upstream, the comments will\nbe gone and I will manually fix the indentation, because there does not\nseem to be a way to do that using coccinelle.\n\nSome amount of new code is currently sitting in linux-next that should get\nthe same modifications, which I will do at the end of the merge window.\n\nMany thanks to Julia Lawall for helping me learn to write a semantic\npatch that does all this.\n\n\u003d\u003d\u003d\u003d\u003d begin semantic patch \u003d\u003d\u003d\u003d\u003d\n// This adds an llseek\u003d method to all file operations,\n// as a preparation for making no_llseek the default.\n//\n// The rules are\n// - use no_llseek explicitly if we do nonseekable_open\n// - use seq_lseek for sequential files\n// - use default_llseek if we know we access f_pos\n// - use noop_llseek if we know we don\u0027t access f_pos,\n// but we still want to allow users to call lseek\n//\n@ open1 exists @\nidentifier nested_open;\n@@\nnested_open(...)\n{\n\u003c+...\nnonseekable_open(...)\n...+\u003e\n}\n\n@ open exists@\nidentifier open_f;\nidentifier i, f;\nidentifier open1.nested_open;\n@@\nint open_f(struct inode *i, struct file *f)\n{\n\u003c+...\n(\nnonseekable_open(...)\n|\nnested_open(...)\n)\n...+\u003e\n}\n\n@ read disable optional_qualifier exists @\nidentifier read_f;\nidentifier f, p, s, off;\ntype ssize_t, size_t, loff_t;\nexpression E;\nidentifier func;\n@@\nssize_t read_f(struct file *f, char *p, size_t s, loff_t *off)\n{\n\u003c+...\n(\n *off \u003d E\n|\n *off +\u003d E\n|\n func(..., off, ...)\n|\n E \u003d *off\n)\n...+\u003e\n}\n\n@ read_no_fpos disable optional_qualifier exists @\nidentifier read_f;\nidentifier f, p, s, off;\ntype ssize_t, size_t, loff_t;\n@@\nssize_t read_f(struct file *f, char *p, size_t s, loff_t *off)\n{\n... when !\u003d off\n}\n\n@ write @\nidentifier write_f;\nidentifier f, p, s, off;\ntype ssize_t, size_t, loff_t;\nexpression E;\nidentifier func;\n@@\nssize_t write_f(struct file *f, const char *p, size_t s, loff_t *off)\n{\n\u003c+...\n(\n *off \u003d E\n|\n *off +\u003d E\n|\n func(..., off, ...)\n|\n E \u003d *off\n)\n...+\u003e\n}\n\n@ write_no_fpos @\nidentifier write_f;\nidentifier f, p, s, off;\ntype ssize_t, size_t, loff_t;\n@@\nssize_t write_f(struct file *f, const char *p, size_t s, loff_t *off)\n{\n... when !\u003d off\n}\n\n@ fops0 @\nidentifier fops;\n@@\nstruct file_operations fops \u003d {\n ...\n};\n\n@ has_llseek depends on fops0 @\nidentifier fops0.fops;\nidentifier llseek_f;\n@@\nstruct file_operations fops \u003d {\n...\n .llseek \u003d llseek_f,\n...\n};\n\n@ has_read depends on fops0 @\nidentifier fops0.fops;\nidentifier read_f;\n@@\nstruct file_operations fops \u003d {\n...\n .read \u003d read_f,\n...\n};\n\n@ has_write depends on fops0 @\nidentifier fops0.fops;\nidentifier write_f;\n@@\nstruct file_operations fops \u003d {\n...\n .write \u003d write_f,\n...\n};\n\n@ has_open depends on fops0 @\nidentifier fops0.fops;\nidentifier open_f;\n@@\nstruct file_operations fops \u003d {\n...\n .open \u003d open_f,\n...\n};\n\n// use no_llseek if we call nonseekable_open\n////////////////////////////////////////////\n@ nonseekable1 depends on !has_llseek \u0026\u0026 has_open @\nidentifier fops0.fops;\nidentifier nso ~\u003d \"nonseekable_open\";\n@@\nstruct file_operations fops \u003d {\n... .open \u003d nso, ...\n+.llseek \u003d no_llseek, /* nonseekable */\n};\n\n@ nonseekable2 depends on !has_llseek @\nidentifier fops0.fops;\nidentifier open.open_f;\n@@\nstruct file_operations fops \u003d {\n... .open \u003d open_f, ...\n+.llseek \u003d no_llseek, /* open uses nonseekable */\n};\n\n// use seq_lseek for sequential files\n/////////////////////////////////////\n@ seq depends on !has_llseek @\nidentifier fops0.fops;\nidentifier sr ~\u003d \"seq_read\";\n@@\nstruct file_operations fops \u003d {\n... .read \u003d sr, ...\n+.llseek \u003d seq_lseek, /* we have seq_read */\n};\n\n// use default_llseek if there is a readdir\n///////////////////////////////////////////\n@ fops1 depends on !has_llseek \u0026\u0026 !nonseekable1 \u0026\u0026 !nonseekable2 \u0026\u0026 !seq @\nidentifier fops0.fops;\nidentifier readdir_e;\n@@\n// any other fop is used that changes pos\nstruct file_operations fops \u003d {\n... .readdir \u003d readdir_e, ...\n+.llseek \u003d default_llseek, /* readdir is present */\n};\n\n// use default_llseek if at least one of read/write touches f_pos\n/////////////////////////////////////////////////////////////////\n@ fops2 depends on !fops1 \u0026\u0026 !has_llseek \u0026\u0026 !nonseekable1 \u0026\u0026 !nonseekable2 \u0026\u0026 !seq @\nidentifier fops0.fops;\nidentifier read.read_f;\n@@\n// read fops use offset\nstruct file_operations fops \u003d {\n... .read \u003d read_f, ...\n+.llseek \u003d default_llseek, /* read accesses f_pos */\n};\n\n@ fops3 depends on !fops1 \u0026\u0026 !fops2 \u0026\u0026 !has_llseek \u0026\u0026 !nonseekable1 \u0026\u0026 !nonseekable2 \u0026\u0026 !seq @\nidentifier fops0.fops;\nidentifier write.write_f;\n@@\n// write fops use offset\nstruct file_operations fops \u003d {\n... .write \u003d write_f, ...\n+\t.llseek \u003d default_llseek, /* write accesses f_pos */\n};\n\n// Use noop_llseek if neither read nor write accesses f_pos\n///////////////////////////////////////////////////////////\n\n@ fops4 depends on !fops1 \u0026\u0026 !fops2 \u0026\u0026 !fops3 \u0026\u0026 !has_llseek \u0026\u0026 !nonseekable1 \u0026\u0026 !nonseekable2 \u0026\u0026 !seq @\nidentifier fops0.fops;\nidentifier read_no_fpos.read_f;\nidentifier write_no_fpos.write_f;\n@@\n// write fops use offset\nstruct file_operations fops \u003d {\n...\n .write \u003d write_f,\n .read \u003d read_f,\n...\n+.llseek \u003d noop_llseek, /* read and write both use no f_pos */\n};\n\n@ depends on has_write \u0026\u0026 !has_read \u0026\u0026 !fops1 \u0026\u0026 !fops2 \u0026\u0026 !has_llseek \u0026\u0026 !nonseekable1 \u0026\u0026 !nonseekable2 \u0026\u0026 !seq @\nidentifier fops0.fops;\nidentifier write_no_fpos.write_f;\n@@\nstruct file_operations fops \u003d {\n... .write \u003d write_f, ...\n+.llseek \u003d noop_llseek, /* write uses no f_pos */\n};\n\n@ depends on has_read \u0026\u0026 !has_write \u0026\u0026 !fops1 \u0026\u0026 !fops2 \u0026\u0026 !has_llseek \u0026\u0026 !nonseekable1 \u0026\u0026 !nonseekable2 \u0026\u0026 !seq @\nidentifier fops0.fops;\nidentifier read_no_fpos.read_f;\n@@\nstruct file_operations fops \u003d {\n... .read \u003d read_f, ...\n+.llseek \u003d noop_llseek, /* read uses no f_pos */\n};\n\n@ depends on !has_read \u0026\u0026 !has_write \u0026\u0026 !fops1 \u0026\u0026 !fops2 \u0026\u0026 !has_llseek \u0026\u0026 !nonseekable1 \u0026\u0026 !nonseekable2 \u0026\u0026 !seq @\nidentifier fops0.fops;\n@@\nstruct file_operations fops \u003d {\n...\n+.llseek \u003d noop_llseek, /* no read or write fn */\n};\n\u003d\u003d\u003d\u003d\u003d End semantic patch \u003d\u003d\u003d\u003d\u003d\n\nSigned-off-by: Arnd Bergmann \u003carnd@arndb.de\u003e\nCc: Julia Lawall \u003cjulia@diku.dk\u003e\nCc: Christoph Hellwig \u003chch@infradead.org\u003e\n" }, { "commit": "40565f1962c5be9b9e285e05af01ab7771534868", "tree": "ae84097778a8adfc5a9aad8a5428fe803d54346a", "parents": [ "d096f3e9898d469493fc0afe88d7285c4bdc3ce2" ], "author": { "name": "Jiri Slaby", "email": "jirislaby@gmail.com", "time": "Mon Feb 12 00:52:31 2007 -0800" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@woody.linux-foundation.org", "time": "Mon Feb 12 09:48:30 2007 -0800" }, "message": "[PATCH] Char: timers cleanup\n\n- Use timer macros to set function and data members and to modify\n expiration time.\n- Use DEFINE_TIMER for global timers and do not init them at run-time in\n these cases.\n- del_timer_sync is common in most cases -- we want to wait for timer\n function if it\u0027s still running.\n\nSigned-off-by: Jiri Slaby \u003cjirislaby@gmail.com\u003e\nCc: Dave Airlie \u003cairlied@linux.ie\u003e\nCc: David Woodhouse \u003cdwmw2@infradead.org\u003e\nCc: Dominik Brodowski \u003clinux@dominikbrodowski.net\u003e\nCc: Alessandro Zummo \u003ca.zummo@towertech.it\u003e\nCc: Paul Fulghum \u003cpaulkf@microgate.com\u003e\nCc: Kylene Jo Hall \u003ckjhall@us.ibm.com\u003e\nCc: Wim Van Sebroeck \u003cwim@iguana.be\u003e\nAcked-by: Dmitry Torokhov \u003cdtor@mail.ru\u003e\t(Input bits)\nSigned-off-by: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "7d12e780e003f93433d49ce78cfedf4b4c52adc5", "tree": "6748550400445c11a306b132009f3001e3525df8", "parents": [ "da482792a6d1a3fbaaa25fae867b343fb4db3246" ], "author": { "name": "David Howells", "email": "dhowells@redhat.com", "time": "Thu Oct 05 14:55:46 2006 +0100" }, "committer": { "name": "David Howells", "email": "dhowells@warthog.cambridge.redhat.com", "time": "Thu Oct 05 15:10:12 2006 +0100" }, "message": "IRQ: Maintain regs pointer globally rather than passing to IRQ handlers\n\nMaintain a per-CPU global \"struct pt_regs *\" variable which can be used instead\nof passing regs around manually through all ~1800 interrupt handlers in the\nLinux kernel.\n\nThe regs pointer is used in few places, but it potentially costs both stack\nspace and code to pass it around. On the FRV arch, removing the regs parameter\nfrom all the genirq function results in a 20% speed up of the IRQ exit path\n(ie: from leaving timer_interrupt() to leaving do_IRQ()).\n\nWhere appropriate, an arch may override the generic storage facility and do\nsomething different with the variable. On FRV, for instance, the address is\nmaintained in GR28 at all times inside the kernel as part of general exception\nhandling.\n\nHaving looked over the code, it appears that the parameter may be handed down\nthrough up to twenty or so layers of functions. Consider a USB character\ndevice attached to a USB hub, attached to a USB controller that posts its\ninterrupts through a cascaded auxiliary interrupt controller. A character\ndevice driver may want to pass regs to the sysrq handler through the input\nlayer which adds another few layers of parameter passing.\n\nI\u0027ve build this code with allyesconfig for x86_64 and i386. I\u0027ve runtested the\nmain part of the code on FRV and i386, though I can\u0027t test most of the drivers.\nI\u0027ve also done partial conversion for powerpc and MIPS - these at least compile\nwith minimal configurations.\n\nThis will affect all archs. Mostly the changes should be relatively easy.\nTake do_IRQ(), store the regs pointer at the beginning, saving the old one:\n\n\tstruct pt_regs *old_regs \u003d set_irq_regs(regs);\n\nAnd put the old one back at the end:\n\n\tset_irq_regs(old_regs);\n\nDon\u0027t pass regs through to generic_handle_irq() or __do_IRQ().\n\nIn timer_interrupt(), this sort of change will be necessary:\n\n\t-\tupdate_process_times(user_mode(regs));\n\t-\tprofile_tick(CPU_PROFILING, regs);\n\t+\tupdate_process_times(user_mode(get_irq_regs()));\n\t+\tprofile_tick(CPU_PROFILING);\n\nI\u0027d like to move update_process_times()\u0027s use of get_irq_regs() into itself,\nexcept that i386, alone of the archs, uses something other than user_mode().\n\nSome notes on the interrupt handling in the drivers:\n\n (*) input_dev() is now gone entirely. The regs pointer is no longer stored in\n the input_dev struct.\n\n (*) finish_unlinks() in drivers/usb/host/ohci-q.c needs checking. It does\n something different depending on whether it\u0027s been supplied with a regs\n pointer or not.\n\n (*) Various IRQ handler function pointers have been moved to type\n irq_handler_t.\n\nSigned-Off-By: David Howells \u003cdhowells@redhat.com\u003e\n(cherry picked from 1b16e7ac850969f38b375e511e3fa2f474a33867 commit)\n" }, { "commit": "9ec52099e4b8678a60e9f93e41ad87885d64f3e6", "tree": "a68fe051b39f8f8e8be469cbd3c2f653b9b71a9d", "parents": [ "1a657f78dcc8ea7c53eaa1f2a45ea2315738c15f" ], "author": { "name": "Cedric Le Goater", "email": "clg@fr.ibm.com", "time": "Mon Oct 02 02:19:00 2006 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@g5.osdl.org", "time": "Mon Oct 02 07:57:25 2006 -0700" }, "message": "[PATCH] replace cad_pid by a struct pid\n\nThere are a few places in the kernel where the init task is signaled. The\nctrl+alt+del sequence is one them. It kills a task, usually init, using a\ncached pid (cad_pid).\n\nThis patch replaces the pid_t by a struct pid to avoid pid wrap around\nproblem. The struct pid is initialized at boot time in init() and can be\nmodified through systctl with\n\n\t/proc/sys/kernel/cad_pid\n\n[ I haven\u0027t found any distro using it ? ]\n\nIt also introduces a small helper routine kill_cad_pid() which is used\nwhere it seemed ok to use cad_pid instead of pid 1.\n\n[akpm@osdl.org: cleanups, build fix]\nSigned-off-by: Cedric Le Goater \u003cclg@fr.ibm.com\u003e\nCc: Eric W. Biederman \u003cebiederm@xmission.com\u003e\nCc: Martin Schwidefsky \u003cschwidefsky@de.ibm.com\u003e\nCc: Paul Mackerras \u003cpaulus@samba.org\u003e\nSigned-off-by: Andrew Morton \u003cakpm@osdl.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@osdl.org\u003e\n" }, { "commit": "62322d2554d2f9680c8ace7bbf1f97d8fa84ad1a", "tree": "904f8b8354a4079cc1daaa355fe30200dcae9dc9", "parents": [ "d459ec0baa5d22e111dfb139c46d6d788a9eac20" ], "author": { "name": "Arjan van de Ven", "email": "arjan@infradead.org", "time": "Mon Jul 03 00:24:21 2006 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@g5.osdl.org", "time": "Mon Jul 03 15:26:59 2006 -0700" }, "message": "[PATCH] make more file_operation structs static\n\nMark the static struct file_operations in drivers/char as const. Making\nthem const prevents accidental bugs, and moves them to the .rodata section\nso that they no longer do any false sharing; in addition with the proper\ndebug option they are then protected against corruption..\n\n[akpm@osdl.org: build fix]\nSigned-off-by: Arjan van de Ven \u003carjan@linux.intel.com\u003e\nSigned-off-by: Andrew Morton \u003cakpm@osdl.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@osdl.org\u003e\n" }, { "commit": "0f2ed4c6bae23d2b7ef0ea2d272377e3de700c0c", "tree": "94a24c9209eb97027e7035b2790a1592d4b2c40c", "parents": [ "69ab3912d1b4dbf27ea1a383cb5731251fc0e109" ], "author": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Sat Jul 01 19:29:33 2006 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@g5.osdl.org", "time": "Sun Jul 02 13:58:49 2006 -0700" }, "message": "[PATCH] irq-flags: drivers/char: Use the new IRQF_ constants\n\nSigned-off-by: Thomas Gleixner \u003ctglx@linutronix.de\u003e\nCc: Ingo Molnar \u003cmingo@elte.hu\u003e\nCc: \"David S. Miller\" \u003cdavem@davemloft.net\u003e\nCc: Benjamin Herrenschmidt \u003cbenh@kernel.crashing.org\u003e\nSigned-off-by: Andrew Morton \u003cakpm@osdl.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@osdl.org\u003e\n" }, { "commit": "6ab3d5624e172c553004ecc862bfeac16d9d68b7", "tree": "6d98881fe91fd9583c109208d5c27131b93fa248", "parents": [ "e02169b682bc448ccdc819dc8639ed34a23cedd8" ], "author": { "name": "Jörn Engel", "email": "joern@wohnheim.fh-wedel.de", "time": "Fri Jun 30 19:25:36 2006 +0200" }, "committer": { "name": "Adrian Bunk", "email": "bunk@stusta.de", "time": "Fri Jun 30 19:25:36 2006 +0200" }, "message": "Remove obsolete #include \u003clinux/config.h\u003e\n\nSigned-off-by: Jörn Engel \u003cjoern@wohnheim.fh-wedel.de\u003e\nSigned-off-by: Adrian Bunk \u003cbunk@stusta.de\u003e\n" }, { "commit": "1da177e4c3f41524e886b7f1b8a0c1fc7321cac2", "tree": "0bba044c4ce775e45a88a51686b5d9f90697ea9d", "parents": [], "author": { "name": "Linus Torvalds", "email": "torvalds@ppc970.osdl.org", "time": "Sat Apr 16 15:20:36 2005 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@ppc970.osdl.org", "time": "Sat Apr 16 15:20:36 2005 -0700" }, "message": "Linux-2.6.12-rc2\n\nInitial git repository build. I\u0027m not bothering with the full history,\neven though we have it. We can create a separate \"historical\" git\narchive of that later if we want to, and in the meantime it\u0027s about\n3.2GB when imported into git - space that would just make the early\ngit days unnecessarily complicated, when we don\u0027t have a lot of good\ninfrastructure for it.\n\nLet it rip!\n" } ] }