- 31 Oct, 2005 40 commits
-
-
Clemens Ladisch authored
Fix a division by zero that happened when the HPET_INFO ioctl was called before a timer frequency had been set. Signed-off-by:
Clemens Ladisch <clemens@ladisch.de> Cc: Bob Picco <bob.picco@hp.com> Signed-off-by:
Andrew Morton <akpm@osdl.org> Signed-off-by:
Linus Torvalds <torvalds@osdl.org>
-
Clemens Ladisch authored
Fix a wrong memory access in hpet_ioctl_common(). It was not possible to use the HPET_INFO ioctl from kernel space because it always called copy_to_user(). Signed-off-by:
-
Clemens Ladisch authored
Reads from an HPET register require a round trip to the south bridge and are almost as slow as PCI reads. By caching the last value we've written to the comparator register, we can eliminate all HPET reads from the fast path in the emulated RTC interrupt handler. Signed-off-by:
Venkatesh Pallipadi <venkatesh.pallipadi@intel.com> Signed-off-by:
-
Clemens Ladisch authored
Make sure that the RTC timer is in non-periodic mode; some stupid BIOS might have initialized it to periodic mode. Furthermore, don't set the SETVAL bit in the config register. This wouldn't have any effect unless the timer was in period mode (which it isn't), and then the actual timer frequency would be half that of the desired one because incrementing the comparator in the interrupt handler would be done after the hardware has already incremented it itself. Signed-off-by:
-
Clemens Ladisch authored
When the emulated RTC interrupt is no longer needed, we better disable it; otherwise, we get a spurious interrupt whenever the timer has rolled over and reaches the same comparator value. Having a superfluous interrupt every five minutes doesn't hurt much, but it's bad style anyway. ;-) Signed-off-by:
-
Clemens Ladisch authored
This patch adds support for shared HPET interrupts. The driver previously acknowledged interrupts for both edge and level interrupts, but didn't actually allow a shared interrupt in the latter case. We use a new per-timer flag to save whether the timer's interrupt might be shared, and use it to do the processing required for level interrupts only if necessary. Signed-off-by:
Bob Picco <bob.picco@hp.com> Signed-off-by:
-
Clemens Ladisch authored
It was only the RTC hardware that restricted interrupt frequencies to a power of two. There is no reason to take over this restriction into the HPET driver, so remove the offending check. Signed-off-by:
-
Clemens Ladisch authored
This patch removes several reads of a timer's config register that serve no purpose whatsoever. Signed-off-by:
-
Clemens Ladisch authored
The variable hpet_ntimer is never read, so remove it. Signed-off-by:
-
Clemens Ladisch authored
On 32-bit architectures, the multiplication in the argument for hpet_time_div() often overflows. In the typical case of a 14.32 MHz timer, this happens when the desired frequency exceeds 61 Hz. To avoid this multiplication, we can precompute and store the hardware timer frequency, instead of the period, in the device structure, which leaves us with a simple division when computing the number of timer ticks. As a side effect, this also removes a theoretical bug where the timer interpolator's frequency would be computed as a 32-bit value even if the HPET frequency is greater than 2^32 Hz (the HPET spec allows up to 10 GHz). Signed-off-by:
-
Clemens Ladisch authored
Disallow setting an interrupt frequency of zero (which would result in a division by zero), and disallow enabling the interrupt when the frequency hasn't yet been set (which would use an interrupt period of zero). Signed-off-by:
-
Randy Dunlap authored
Convert most of the remaining "Using plain integer as NULL pointer" sparse warnings to use NULL. (Not duplicating patches that are already in -mm, -bird, or -kj.) Convert isdn driver struct initializer to use C99 syntax. Signed-off-by:
Randy Dunlap <rdunlap@xenotime.net> Signed-off-by:
-
Andrew Morton authored
- Various whitespace fixes - Use kzalloc() Acked-by:
Kylene Hall <kjhall@us.ibm.com> Signed-off-by:
-
Marcel Selhorst authored
Move the Infineon TPM driver off pci device and makes it a pure pnp-driver. It includes pnp-port validation and region requesting. Signed-off-by:
Marcel Selhorst <selhorst@crypto.rub.de> Signed-off-by:
-
Kylene Jo Hall authored
This patch changes the nsc driver from a pci driver to a platform driver. Signed-off-by:
-
Kylene Jo Hall authored
This patch changes the atmel driver from a pci driver to a platform driver. Signed-off-by:
-
Kylene Jo Hall authored
This patch is in support of moving away from the lpc bus pci_dev. The power management prototypes used by platform drivers is different but the functionality remains the same. Signed-off-by: Kylene Hall <kjhall@us.ibm.com Signed-off-by:
-
Kylene Jo Hall authored
Since the tpm does not have it's own pci id we have been consuming the lpc bus. This is not correct and causes problems to support non lpc bus chips. This patch removes the dependency on pci_dev from tpm.c The subsequent patches will stop the supported chips from registering as pci drivers. Signed-off-by:
-
Kylene Jo Hall authored
This patch is in preparation of supporting chips that are not necessarily on the lpc bus and thus are not accessed with inb's and outb's. The patch replaces the call to get the chip's status in the tpm.c file with a vendor specific status function. The patch also defines the function for each of the current supported devices. Signed-off-by:
-
Andi Kleen authored
It's nasty to set random drivers to default m because people who just press enter on make oldconfig get these. Remove the default m Signed-off-by:
Andi Kleen <ak@suse.de> Signed-off-by:
-
Paul Jackson authored
Simplify the UP (1 CPU) implementatin of set_cpus_allowed. The one CPU is hardcoded to be cpu 0 - so just test for that bit, and avoid having to pick up the cpu_online_map. Also, unexport cpu_online_map: it was only needed for set_cpus_allowed(). Signed-off-by:
Paul Jackson <pj@sgi.com> Acked-by:
Ingo Molnar <mingo@elte.hu> Signed-off-by:
-
Yuri Vasilevski authored
I made a patch that detects if libintl.h (needed for nls) is present on the host system and if it's not, it nls support is disabled by providing dummies for the used nls functions. This way if there is nls support on the host system the *config targets will build according to Arnaldo Carvalho de Melo's i18n modifications, else it just uses the original English messages. I have also made a bug report at kernel's bugzilla: http://bugzilla.kernel.org/show_bug.cgi?id=5501 And there is a discussion about this problem in Gentoo's bugzilla: http://bugs.gentoo.org/show_bug.cgi?id=99810 Cc: Sam Ravnborg <sam@ravnborg.org> Signed-off-by:
-
Horms authored
Signed-off-by:
-
Paul Jackson authored
A couple of (char *) casts removed in a previous cleanup patch in lib/string.c:memmove() were actually useful, as they suppressed a couple of warnings: assignment discards qualifiers from pointer target type Fix by declaring the local variable const in the first place, so casts aren't needed to strip the const qualifier. Signed-off-by:
-
Matt Mackall authored
From: Hareesh Nagarajan <hnagar2@gmail.com> Signed-off-by:
Hareesh Nagarajan <hnagar2@gmail.com> Acked-by:
Matt Mackall <mpm@selenic.com> Signed-off-by:
-
Andrew Morton authored
If a filesystem passes an idiotic blocksize into bread(), __getblk_slow() will warn and will return NULL. We have a report (from Hubert Tonneau <hubert.tonneau@fullpliant.org>) of isofs_fill_super() doing this (passing in a silly block size) against an unplugged CDROM drive. But a couple of __getblk_slow() callers forgot to check for the NULL bh, hence oops. Signed-off-by:
-
Arthur Othieno authored
__MUTEX_INITIALIZER() has no users, and equates to the more commonly used DECLARE_MUTEX(), thus making it pretty much redundant. Remove it for good. Signed-off-by:
Arthur Othieno <a.othieno@bluewin.ch> Signed-off-by:
-
Bjorn Helgaas authored
The driver had incorrectly wrapped module_init(rp_init) in #ifdef MODULE, so it worked only when compiled as a module. Tested by Wolfgang Denk with this device: 00:0e.0 Communication controller: Comtrol Corporation RocketPort 8 port w/RJ11 connectors (rev 04) Control: I/O+ Mem- BusMaster- SpecCycle- MemWINV- VGASnoop- ParErr- Stepping- SERR- FastB2B- Status: Cap- 66Mhz- UDF- FastB2B- ParErr- DEVSEL=slow >TAbort- <TAbort- <MAbort- >SERR- <PERR- Interrupt: pin A routed to IRQ 11 Region 0: I/O ports at 7000 [size=64] Signed-off-by:Bjorn Helgaas <bjorn.helgaas@hp.com> Signed-off-by:
-
Roland McGrath authored
This change corrects an omission in posix_cpu_timer_schedule, so that it correctly propagates the overrun calculation to where it will get reported to the user. Signed-off-by:
Roland McGrath <roland@redhat.com> Signed-off-by:
-
Paul E. McKenney authored
This patch is a rewrite of the one submitted on October 1st, using modules (http://marc.theaimsgroup.com/?l=linux-kernel&m=112819093522998&w=2). This rewrite adds a tristate CONFIG_RCU_TORTURE_TEST, which enables an intense torture test of the RCU infratructure. This is needed due to the continued changes to the RCU infrastructure to accommodate dynamic ticks, CPU hotplug, realtime, and so on. Most of the code is in a separate file that is compiled only if the CONFIG variable is set. Documentation on how to run the test and interpret the output is also included. This code has been tested on i386 and ppc64, and an earlier version of the code has received extensive testing on a number of architectures as part of the PREEMPT_RT patchset. Signed-off-by:
"Paul E. McKenney" <paulmck@us.ibm.com> Signed-off-by:
-
Alexey Dobriyan authored
Signed-off-by:
Alexey Dobriyan <adobriyan@gmail.com> Signed-off-by:
-
Nikita Danilov authored
Fix comment describing BUILD_BUG_ON: BUG_ON is not an assertion (unfortunately). Signed-off-by:
Nikita Danilov <nikita@clusterfs.com> Signed-off-by:
-
Laurent Riffard authored
This updates .owner field of struct pci_driver. This allows SYSFS to create the symlink from the driver to the module which provides it. Signed-off-by:
Laurent Riffard <laurent.riffard@free.fr> Acked-by:
Alan Cox <alan@redhat.com> Signed-off-by:
-
Laurent Riffard authored
This updates .owner field of struct pci_driver. This allows SYSFS to create the symlink from the driver to the module which provides it. Signed-off-by:
-
Laurent Riffard authored
This updates .owner field of struct pci_driver. This allows SYSFS to create the symlink from the driver to the module which provides it. Signed-off-by:
-
Pozsar Balazs authored
This patch fixes a long-standing vgacon bug: characters with the bright bit set were left on the screen and not blacked out. All I did was that I lookuped up some examples on the net about setting the vga palette, and added the call missing from the linux kernel, but included in all other ones. It works for me. You can test this by writing something with the bright set to the console, for example: echo -e "\e[1;31mhello there\e[0m" and then wait for the console to blank itself (by default, after 10 mins of inactivity), maybe making it faster using setterm -blank 1 so you only have to wait 1 minute. Signed-off-by:
Pozsar Balazs <pozsy@uhulinux.hu> Cc: James Simmons <jsimmons@infradead.org> Cc: "Antonino A. Daplas" <adaplas@pol.net> Signed-off-by:
-
Glauber de Oliveira Costa authored
This patch adds tests for the return value of sb_getblk() in the ext2/3 filesystems. In fs/buffer.c it is stated that the getblk() function never fails. However, it does can return NULL in some situations due to I/O errors, which may lead us to NULL pointer dereferences Signed-off-by:
Glauber de Oliveira Costa <glommer@br.ibm.com> Signed-off-by:
-
Andrea Arcangeli authored
list_move(&inode->i_list, &inode_in_use); } else { list_move(&inode->i_list, &inode_unused); + inodes_stat.nr_unused++; } } wake_up_inode(inode); Are you sure the above diff is correct? It was added somewhere between 2.6.5 and 2.6.8. I think it's wrong. The only way I can imagine the i_count to be zero in the above path, is that I_WILL_FREE is set. And if I_WILL_FREE is set, then we must not increase nr_unused. So I believe the above change is buggy and it will definitely overstate the number of unused inodes and it should be backed out. Note that __writeback_single_inode before calling __sync_single_inode, can drop the spinlock and we can have both the dirty and locked bitflags clear here: spin_unlock(&inode_lock); __wait_on_inode(inode); iput(inode); XXXXXXX spin_lock(&inode_lock); } use inode again here a construct like the above makes zero sense from a reference counting standpoint. Either we don't ever use the inode again after the iput, or the inode_lock should be taken _before_ executing the iput (i.e. a __iput would be required). Taking the inode_lock after iput means the iget was useless if we keep using the inode after the iput. So the only chance the 2.6 was safe to call __writeback_single_inode with the i_count == 0, is that I_WILL_FREE is set (I_WILL_FREE will prevent the VM to free the inode in XXXXX). Potentially calling the above iput with I_WILL_FREE was also wrong because it would recurse in iput_final (the second mainline bug). The below (untested) patch fixes the nr_unused accounting, avoids recursing in iput when I_WILL_FREE is set and makes sure (with the BUG_ON) that we don't corrupt memory and that all holders that don't set I_WILL_FREE, keeps a reference on the inode! Signed-off-by:Andrea Arcangeli <andrea@suse.de> Signed-off-by:
-
Pavel Roskin authored
sparse complains about every MODULE_PARM used in a module: warning: symbol '__parm_foo' was not declared. Should it be static? The fix is to split declaration and initialization. While MODULE_PARM is obsolete, it's not something sparse should report. Signed-off-by:
Pavel Roskin <proski@gnu.org> Signed-off-by:
-
Nicolas Pitre authored
They aren't used anywhere in that file. Signed-off-by:
Nicolas Pitre <nico@cam.org> Signed-off-by:
-
