Ex Parte Clark et alDownload PDFPatent Trial and Appeal BoardMar 26, 201311831985 (P.T.A.B. Mar. 26, 2013) Copy Citation UNITED STATES PATENT AND TRADEMARK OFFICE ________________ BEFORE THE PATENT TRIAL AND APPEAL BOARD ________________ Ex parte MICHAEL T. CLARK and JELENA ILIC ________________ Appeal 2010-011206 Application 11/831,985 Technology Center 2100 ________________ Before ERIC B. CHEN, TREVOR M. JEFFERSON, and JOHN G. NEW, Administrative Patent Judges. NEW, Administrative Patent Judge. DECISION ON APPEAL Appeal 2010-011206 Application 11/831,985 2 SUMMARY Appellants file this appeal under 35 U.S.C. § 134(a) from the Examiner’s Final Rejection of claims 1-22 as unpatentable under 35 U.S.C.§ 103(a) as being obvious over the combination of Gephardt (US 5,530,891, June 25, 1996) (“Gephardt”), Wilcox et al. (US 5,764,999, June 9, 1998) (“Wilcox”), and Deschepper et al. (US 6,199,134 B1, March 6, 2001) (“Deschepper”). We have jurisdiction under 35 U.S.C. § 6(b). We affirm. STATEMENT OF THE CASE Appellants’ invention is directed to a computer system, including a system memory, a plurality of processor cores, and an input/output (I/O) hub, that may communicate with each of the processor cores. In response to detecting an occurrence of an internal system management interrupt (SMI), each of the processor cores may save to a system management mode (SMM) save state in the system memory, information corresponding to a source of the internal SMI. In response to detecting the internal SMI, each processor core may further initiate a I/O cycle to a predetermined port address within the hub. The 110 hub may broadcast an SMI message to each of the processor cores in response to receiving the 110 cycle. Each of the processor cores may further save to the SMM save state in the system memory, respective internal SMI source information in response to receiving the broadcast SMI message. Abstract. Appeal 2010-011206 Application 11/831,985 3 GROUPING OF CLAIMS Because Appellants argue that the Examiner erred for substantially the same reasons with respect to claims 1, 9, and 17, we select claim 1 as representative of this group. App. Br. 5. Claim 1 recites: 1. A computer system comprising: a system memory; a plurality of processor cores coupled to the system memory, wherein in response to detecting an occurrence of an internal system management interrupt (SMI), each of the processor cores is configured to save to a system management mode (SMM) save state in the system memory, information corresponding to a source of the internal SMI; an input/output (I/O hub) configured to communicate with each of the processor cores; wherein in response to detecting the internal SMI, each processor core is further configured to initiate an I/O cycle to a predetermined port address within the I/O hub; wherein the I/O hub is configured to broadcast an SMI message to each of the plurality of processor cores in response to receiving the I/O cycle; wherein each of the processor cores is further configured to save to the SMM save state in the system memory, respective internal SMI source information in response to receiving the broadcast SMI message. App. Br. 12. Appeal 2010-011206 Application 11/831,985 4 Appellants argue that the Examiner erred for substantially the same reasons with respect to claims 2, 10, and 18 and that these claims are separately patentable. App. Br. 8. We therefore select claim 2 as representative of this group. Claim 2 recites: 2. The computer system as recited in claim 1, wherein a selected one of the plurality of processor cores is configured to read from the system memory, the SMM save state of all of the processor cores to determine within which processor core the internal SMI occurred. App. Br. 12. Appellants argue that the Examiner erred for substantially the same reasons with respect to claims 3, 11, and 19, and that these claims are separately patentable. App. Br. 9. We therefore select claim 3 as representative of this group. Claim 3 recites: 3. The computer system as recited in claim 2, wherein an SMI handler within the selected processor core is configured to service the internal SMI of the processor core within which the internal SMI occurred. App. Br. 12. Appellants argue that the Examiner erred for substantially the same reasons with respect to claims 5, 13, and 21 and that these claims are separately patentable. App. Br. 9. We therefore select claim 5 as representative of this group. Claim 5 recites: 5. The computer system as recited in claim 4, wherein the predetermined port address is programmed into a model specific register of each of the processor cores during the boot-up process by the BIOS. App. Br. 13. Appeal 2010-011206 Application 11/831,985 5 Appellants argue that the Examiner erred for substantially the same reasons with respect to claims 8, 14, and 22 and that these claims are separately patentable. App. Br. 10. We therefore select claim 8 as representative of this group. Claim 8 recites: 8. The computer system as recited in claim 1, wherein the information corresponding to a source of the internal SMI comprises a bit vector having a plurality of bits each corresponding to a respective source of an internal SMI. App. Br. 13. ISSUES AND ANALYSES We address each of Appellants’ arguments seriatim, as presented in Appellants’ Brief. A. Rejection of claims 1, 9, and 17 under 35 U.S.C. § 103(a). Issue 1 Appellants argue that the Examiner erred in finding that Gephardt teaches or suggests storing in the SMM save state in memory “information corresponding to a source” of the SMI. App. Br. 5-6. Appellants also argue that the Examiner erred in finding that saving SMI source information is inherent when a processor switches to SMM. App. Br. 6. We therefore address the issue of whether the Examiner so erred. Analysis Appellants argue that Gephardt discusses SMM and the handling of SMI interrupts using registers in a controller. App. Br. 6. Appellants assert that Gephardt does not teach or suggest saving the SMI source information Appeal 2010-011206 Application 11/831,985 6 within the SMM save state in the system memory. Id. Appellants contend that Gephardt uses registers within a central controller to communicate interrupt information. Id. (citing, e.g., Gephardt, Figs. 12, 13, and cols. 15- 16. Appellants argue further that processors use an SMI handler upon entering SMM, and because the SMI handler is a special handler that is executed upon entering SMM, the interrupt descriptor table (IDT) is not used to hold descriptors for SMI interrupts. App. Br. 6. Therefore, argue Appellants, although the IDT is stored in the SMM save state area in memory, SMI interrupt source information is not necessarily stored therein. Id. Consequently, Appellants contend, and contra the Examiner’s findings, the reason the processor state is saved is so that once the SMI has been serviced, the processor can restore the processor state as it was prior to the SMI, and NOT so that the processor can “actually servic[e] the interrupt.” Id. (quoting Final Rej., 2). The Examiner responds that, as an initial matter, Appellants have mischaracterized Examiner’s findings by asserting that the Examiner finds that it is inherent to the teachings of Gephardt to save the source information; rather, the Examiner finds, saving the state is inherent to Gephardt. Ans. 6. The Examiner finds further that Gephardt explicitly teaches that the Source ID of interrupts is stored in the processor registers, thus this is part of the state information taught by Gephardt. Ans. 6. The Examiner finds that Wilcox also teaches that “[f]or internally-signaled SMIs, the processor has available source and/or specific cause of the SMI” and that “[f]or externally-signaled SMIs, the chipset 49 includes SMI control registers that Appeal 2010-011206 Application 11/831,985 7 must be read by the processor to determine the source of an SMI.” Ans. 7 (citing Wilcox, col. 8, ll. 32-42). The Examiner finds that Deschepper also explicitly teaches identifying the source of the interrupt (column 12 lines 40-65). The Examiner concludes that, since it is inherent that the processor save state and the Source ID is part of Gephardt’s state, and that Wilcox expressly states the source of an SMI must be identified, it would have therefore been obvious to a person of ordinary skill in the contemporaneous art to save this part of the state information along with the rest of the state information, because saving state is required for interrupt handling. Ans. 6- 7. We are persuaded by the Examiner’s reasoning and adopt it as our own. We agree with the Examiner that Gephardt teaches that Source ID is saved in the registers. Ans. 7. Moreover, we are persuaded by the Examiner’s unrebutted finding that saving the state at the time of the SMI is inherent to the teachings of Gephardt and that furthermore state is saved to service an interrupt. Ans. 6, 7. Thus, since the source ID and is part of the state and since Wilcox teaches or suggests that the source of an SMI must be identified, we conclude that it would have been obvious to one of ordinary skill in the contemporaneous art to save this information when saving the state. We therefore conclude that the Examiner did not err in finding that the combination of Gephardt and Wilcox teaches or suggests storing in the SMM save state in memory “information corresponding to a source of the internal SMI,” as recited in claim 1. Appeal 2010-011206 Application 11/831,985 8 Issue 2 Appellants argue that the Examiner erred in finding that Deschepper teaches the limitations of claim 1 reciting “in response to detecting the internal SMI, each processor core is further configured to initiate an I/O cycle to a predetermined port address within the hub” and “the I/O hub broadcast[ing] an SMI message to each of the plurality of processor cores in response to receiving the I/O cycle.” App. Br. 8. We therefore address the issue of whether the Examiner so erred. Analysis Appellants argue that Deschepper discloses a “[s]outh bridge logic Device” includes “trap registers” that hold device “status information” for devices that “are powered down.” App. Br. 8. Appellants assert that the bridge traps PCI accesses from a single CPU or other PCI device to I/O devices that are in a low power state. Id. Moreover, Appellants contend, the bridge device “ACPI/power management logic causes an SMI signal to be transmitted to the processor” in response to detecting “an access to the target that has been placed in the low power mode.” Id. (quoting Deschepper, col. 4, ll. 14-65). Appellants argue that the predetermined address in the I/O hub is used for just that purpose, and is not an address of some peripheral device that has been powered down and that, furthermore, the claimed I/O hub “broadcast[s] an SMI message” to all the other processor cores, and does not simply generate an SMI to the one CPU in the system.” App. Br. 8. The Examiner responds that obviousness rejection of claim 1 was over the combination of Gephardt, Wilcox, and Deschepper. Ans. 9. The Appeal 2010-011206 Application 11/831,985 9 Examiner finds that Wilcox teaches or suggests that an address is used to generate the SMI, and that “internally-signaled SMI's are all memory mapped or I/O traps,” and that, therefore, the internal SMI’s source is an address Id. (quoting Wilcox, col. 8 ll. 25-31). The Examiner finds that although Wilcox does not teach or suggest that such address trapping can be performed in a HUB, Deschepper teaches that an address can be used in a hub to generate the SMI’s. Ans. 9. The Examiner further finds that Gephardt teaches or suggests a “command” sent to the hub for the purpose of broadcasting an SMI to the other cores. Id. The Examiner concludes that it would therefore have been obvious to one of ordinary skill in the contemporary art to have the core provide an address signal to the HUB so as to have it generate the SMI that is broadcast, as opposed to the processor core generating the SMI to all the other cores. Id. An artisan of ordinary skill would have been motivated to so combine the teachings of the prior art references because this would allow for the HUB to handle the signaling to the other cores, leaving the processor to have to signal only one other device (i.e., the HUB), thus removing the additional signaling requirement from each of the processor cores and increasing system efficiency. Id. Appellants reply that the address of the SMI, as taught by Wilcox, would not be the address sent to the hub. Reply Br. 5. Appellants argue that Gephardt does send a command to the hub to tell the hub to broadcast an SMI, and that the Deschepper teaches that the hub places addresses in a table as peripheral devices are powered down, and simply traps on those addresses when some device tries to normally access the peripherals because those are considered illegal accesses and require an SMI to be Appeal 2010-011206 Application 11/831,985 10 initiated. Id. However, argue Appellants, there is no suggestion to send the command of Gephardt to a predetermined and dedicated port address in the hub. Id. We are persuaded by the Examiner’s reasoning and adopt it as our own. “One cannot show non-obviousness by attacking references individually where, as here, the rejections are based on combinations of references.” In re Keller, 642 F.2d 413, 426 (C.C.P.A. 1981). In this instance, we agree with the Examiner’s reasoning that the combination of Gephardt, Wilcox, and Deschepper teaches the disputed limitations. Specifically, we are persuaded that the teachings or suggestions of Deschepper, combined with those of Gephardt and Wilcox, teaches or suggests an address trap that causes the generation of an SMI, and that it would have been obvious to an artisan of ordinary skill in the contemporaneous art to use a port address in the I/O hub to generate the global SMI because this would have eliminated any need for an extra signal line to signal the I/O HUB for generating the broadcast SMI. Ans. 5. B. Rejection of claims 2, 10, and 18 under 35 U.S.C. § 103(a). Issue Appellants argue that the Examiner erred in finding that Gephardt teaches or suggests the limitation of claim 1 reciting added) “wherein a selected one of the plurality of processor cores is configured to read from the system memory, the SMM save state of all of the processor cores to determine within which processor core the internal SMI occurred.” App. Br. 8. We therefore address the issue of whether the Examiner so erred. Appeal 2010-011206 Application 11/831,985 11 Analysis Appellants argue that claim 2 requires “a selected” one of the processor cores that “read[s] from the system memory, the SMM save state of all of the processor cores to determine within which processor core the internal SMI occurred,” and assert that this limitation is neither taught nor suggested in any of the prior art references. App. Br. 8. Moreover, argue Appellants, Gephardt teaches that the originating processor ID (i.e., the processor) which caused the interprocessor interrupt (IPI), and not an SMI, is stored in a register. Id. (citing Gephardt, Table 8). Appellants maintain that Wilcox and Deschepper both teach CPU systems and so cannot possibly teach the features of claim 2. App. Br. 8. The Examiner responds that all three references teach or suggest that the processor that services the SMI determines the source of an interrupt. Ans. 9. The Examiner finds that Wilcox teaches that “[f]or internally- signaled SMIs, the processor has available source and/or specific cause of the SMI” and that “[f]or externally-signaled SMIs, the chipset 49 includes SMI control registers that must be read by the processor to determine the source of an SMI.” Id. (citing Wilcox, col. 8, ll. 32-42). The Examiner further finds that Deschepper also teaches or suggests identifying the source of the interrupt. Ans. 9 (citing Deschepper, col., 12 ll. 40-65). Therefore, the Examiner finds a “selected one” (i.e., the processor that services the interrupt) of the plurality of processor cores is configured to read the source information. Ans. 9. The Examiner therefore finds that all three references teach or suggest that the processor that services the interrupt reads the source information from where it is saved. Id. Appeal 2010-011206 Application 11/831,985 12 Appellants reply that they disagree with the Examiner’s assertion that a “selected one” of the cores “read[s] from the system memory, the SMM save state of all of the processor cores to determine within which processor core the internal SMI occurred.” Reply Br. 6. Appellants argue that all three references simply disclose servicing an SMI and that, therefore, the combination of references does not teach or suggest the “selected one” core reading “the SMM save state of all of the processor cores.” Id. We are persuaded by the Examiner’s reasoning and adopt it as our own. As an initial matter, Appellants provide no specific definition of “selected” in their Specification. Consequently, the Examiner may employ the broadest reasonable meaning of the words in their ordinary usage as they would be understood by one of ordinary skill in the art. See In re Morris, 127 F.3d 1048, 1054 (Fed. Cir. 1997). In this instance “selected” may be reasonably defined as “chosen from a number or group by fitness or preference.” Merriam-Webster Online Dictionary, available at http://www.merriam-webster.com/dictionary/selected?show=1&t= 1363985686 (last visited March 22, 2013). Consequently, the Examiner’s definition of “selected” as “the processor that services the interrupt” is not unreasonable. See Ans. 9. Moreover, Appellants admit that all three prior art references teach or suggest that the references disclose identifying the source of the interrupt, but argue that Deschepper and Wilcox are irrelevant because they only teach single-processor systems. Reply Br. 7. As discussed supra, it is the combination of teachings that is relevant to an obviousness analysis. We agree with the Examiner that Gephardt teaches a plurality of processor cores, and, when combined with the teachings of Deschepper and Wilcox, Appeal 2010-011206 Application 11/831,985 13 teaches or suggests the limitation of claim 2 reciting “wherein a selected one of the plurality of processor cores is configured to read from the system memory, the SMM save state of all of the processor cores to determine within which processor core the internal SMI occurred.” Ans. 9. We consequently conclude that the Examiner did not err. C. Rejection of claims 3, 11, and 19 under 35 U.S.C. § 103(a). Issue Appellants argue that the Examiner erred in finding that Gephardt teaches or suggests the limitation of claim 3 reciting “wherein an SMI handler within the selected processor core is configured to service the internal SMI of the processor core within which the internal SMI occurred.” App. Br. 9. We therefore address the issue of whether the Examiner so erred. Analysis Appellants argue that that Gephardt does not teach or suggest internal SMI interrupts. App. Br. 9. Additionally, Appellants argue, Gephardt teaches servicing SMI interrupts from a power management unit. Id. (citing Gephardt, col. 21, ll. 35-37). Therefore, Appellants contend, Gephardt does not teach or suggest the disputed limitation. The Examiner responds that the obviousness rejection of claim 3 is not made in view of Gephardt alone, but is based upon a combination of the references. Ans. 10. The Examiner finds that Wilcox teaches internal interrupts and that Gephardt teaches the processor that receives the SMI is the master processor. Id. Since the interrupt is internal in Wilcox, the Appeal 2010-011206 Application 11/831,985 14 process in which the interrupt originates would be the processor to first receive the interrupt. Id. We agree with the Examiner. “One cannot show non-obviousness by attacking references individually where, as here, the rejections are based on combinations of references.” Keller, 642 F.2d at 426. It is not disputed that Wilcox teaches internal interrupts. See Wilcox, col. 8, ll. 27-30. We agree with the Examiner that the combination of the prior art references therefore teaches the disputed limitation. D. Rejection of claims 5, 13, and 21 under 35 U.S.C. § 103(a). Issue Appellants next argue that the Examiner erred in finding that Wilcox neither teaches nor suggests the limitation of claim 5 reciting “wherein the predetermined port address is programmed into a model specific register of each of the processor cores during the boot-up process by the BIOS.” We therefore address whether the Examiner so erred. Analysis Appellants argue that Wilcox teaches or suggests I/O traps that generate internal SMIs in the processor. App. Br. 9. Appellants assert that this is different from generating a global SMI in response to an I/O cycle to a predetermined port address of an I/O hub. Id. Appellants contend further that such I/O traps are a VSA extension for virtualization of peripheral software functions. Id. (citing Wilcox at col. 9, ll. 6-20). Appellants argue that there would therefore be no reason to store in BIOS I/O trap addresses for a software system that may change from version to version. App. Br. 9. Appeal 2010-011206 Application 11/831,985 15 The Examiner responds that Wilcox teaches or suggests I/O traps and BIOS, that Gephardt teaches or suggests a global SMI, and that Deschepper teaches or suggests that the trap can be in a hub. Ans. 10. Consequently, the Examiner finds, the combination of the three references teaches or suggests generating a global SMI trap in the hub. Id. Appellants reply that the Examiner’s assertion that Wilcox teaches or suggests I/O traps and BIOS does not reveal why would someone of skill in the art be motivated to have the BIOS program the predetermined address into a model specific register, as required by claim 5. Reply Br. 7. We are not persuaded by Appellants’ arguments. Motivation to combine has been discussed supra. We agree with the Examiner’s finding that Wilcox teaches a “trapped” address for an I/O cycle but is silent upon how the initial determination that the address should be “trapped.” Ans. 5. Moreover we agree with the Examiner’s finding that Wilcox also teaches BIOS but does not give any description of what is provided by the BIOS. Id. However, we also agree with the Examiner’s conclusion that it would have been obvious to one of ordinary skill in the contemporaneous art to store the “Trap” addressees in the BIOS because this would have thereby provided an initial trap value. Id. E. Rejection of claims 8, 14, and 22 under 35 U.S.C. § 103(a). Issue Appellants argue that the Examiner erred in finding that Gephardt does not teach or suggest the limitation of claim 8 reciting “wherein the information corresponding to a source of the internal SMI comprises a bit vector having a plurality of bits each corresponding to a respective source Appeal 2010-011206 Application 11/831,985 16 of an internal SMI.” App. Br. 10. We therefore address the issue of whether the Examiner so erred. Analysis Appellants argue that the entry in Table 8 of Gephardt refers to the originating processor ID of an interprocessor interrupt (IPI), which is the processor ID of the processor that generated or has scheduled an IPI, and not the source of an internal SMI within a processor. App. Br. 10. Furthermore, argue Appellants, the processor ID of Gephardt does not constitute a “bit vector” in which each bit corresponds “to a respective source of an internal SMI.” We refer Appellants to our discussion supra with respect to why we affirm the Examiner’s finding that the combination of the prior art references teaches “information corresponding to a source of the internal SMI.” Moreover, Appellants’ Specification states that [i]n response to detecting an occurrence of an internal system management interrupt (SMI), each of the processor cores may save to a system management mode (SMM) save state in the system memory, information such as a bit vector, for example, corresponding to a source of the internal SMI,” but provides no explicit definition of “bit vector” We therefore adopt the broadest reasonable definition of the phrase “source of the internal SMI comprises a bit vector having a plurality of bits each corresponding to a respective source of an internal SMI” as corresponding to the location address taught by Deschepper as the source of an interrupt. See Deschepper, col. 12, ll. 45-49 (“The master controller may be programmed in the fully nested mode to permit the slave controller to send the correct interrupt vector back to the Appeal 2010-011206 Application 11/831,985 17 CPU identifying the source of the interrupt”); see also Ans. 7. We consequently conclude that the Examiner did not err in finding that the combination of prior art references teaches or suggests the limitation of claim 8 reciting “wherein the information corresponding to a source of the internal SMI comprises a bit vector having a plurality of bits each corresponding to a respective source of an internal SMI.” DECISION The Examiner’s rejection of claims 1-22 as unpatentable under 35 U.S.C. 103(a) is affirmed. TIME PERIOD FOR RESPONSE No time period for taking any subsequent action in connection with this appeal may be extended under 37 C.F.R. § 1.136(a) (1)(iv) . AFFIRMED ke Copy with citationCopy as parenthetical citation